The Modal System of Earlier Egyptian Complement Clauses
Probleme der Ägyptologie Herausgegeben von
Wolfgang Schenkel und Antonio Loprieno
26. BAND
The Modal System of Earlier Egyptian Complement Clauses A Study in Pragmatics in a Dead Language
By
Sami Uljas
LEIDEN • BOSTON 2007
This book is printed on acid-free paper. Library of Congress Cataloging-in-Publication data is available on https://catalog.loc.gov.
ISSN 0169-9601 ISBN 978 90 04 15831 3 © Copyright 2007 by Koninklijke Brill NV, Leiden, The Netherlands. Koninklijke Brill NV incorporates the imprints Brill, Hotei Publishing, IDC Publishers, Martinus Nijhoff Publishers and VSP. All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher. Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill NV provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA. Fees are subject to change. printed in the netherlands
table of contents
v
TABLE OF CONTENTS Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ix
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.1 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.1.1 Orientation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.1.2 Assertion and Non-assertion . . . . . . . . . . . . . . 0.1.3 The System in Earlier Egyptian and the Structure of the Present Work . . . . . . . . . . . . . 0.2 Methodology, Terminology, Morphology and Sources
1 1 1 13 23 26
PART ONE MODALITY IN AFFIRMATIVE COMPLEMENT CLAUSES AFTER GOVERNING VERBS 1.
Introduction to Part One . . . . . . . . . . . . . . . . . . . . . . . . . .
39
2.
Affirmative Object Complementation after Notionally Assertive Verbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 2.1 Locution: The Verb Dd ‘Say’ and Indirect Speech . . 50 2.1.1 Core . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 2.1.2 An Example Text . . . . . . . . . . . . . . . . . . . . . . . 66 2.2 Verbs of Cognition and Perception . . . . . . . . . . . . . . 73 2.2.1 Seeing is Believing and Other Truisms: Realis, Irrealis, Knowledge and Perception . . . . . . . . . 73 2.2.2 The ‘Evaluative’ Use of Irrealis . . . . . . . . . . . . 84 2.3 Interim Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 2.4 The Paradigm after ntt/wnt: Form and Function Revisited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.
Affirmative Object Complementation after Notionally Non-assertive Verbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 3.1 wD and dbH: Between ‘Ontological’ and ‘Attitudinal’ Modality . . . . . . . . . . . . . . . . . . . . . . . . . 126 3.2 Verbs of Preventing and rdi . . . . . . . . . . . . . . . . . . . . 140
vi
table of contents 3.3 3.4 3.5
Pragmatics of Desire: the Verb mri . . . . . . . . . . . . . . 153 Some Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 Interim Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 PART TWO
MODALITY IN OTHER TYPES OF EARLIER EGYPTIAN COMPLEMENT CLAUSES 4.
Modality in Affirmative Subject Complement Clauses . . . . 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Asserted versus Non-asserted Subject Complements . 4.3 The Verb xpr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
173 173 181 194
5.
Modality in Negative Complement Clauses after Governing Predicates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Assertion: Complement Clauses with iwt and ntt n 5.3 Non-assertion: tm and nfr-n . . . . . . . . . . . . . . . . . . . .
203 203 206 210
6.
7.
8.
Modality in Complement Clauses after Prepositions . . . . . 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 The Typology of Irrealis in Complement Clauses of Preposition: the ‘Obvious’ Instances . . . . . . . . . . . 6.3 ‘Indicative Non-indicatives?’ The Broader System of Preposition Complement Modality in Earlier Egyptian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4 Clauses with m-xt . . . . . . . . . . . . . . . . . . . . . . . . . . . .
221 221
238 260
Earlier Egyptian Supplementary Patterns of Complementation after Verbs and Prepositions . . . . . . . . . 7.1 The sDm.n=f . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2 The Particle is . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3 The Element r-Dd . . . . . . . . . . . . . . . . . . . . . . . . . . . .
264 264 278 284
225
Predicate Complement Clauses . . . . . . . . . . . . . . . . . . . . . . 287 8.1 The Syntax of Type B Nominal Sentences with Clausal Predicates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
table of contents 8.2
8.3 9.
vii
Assertion and Non-assertion in Bipartite Nominal Sentences with Clausal Predicates and the Limits of ‘Nominal’ Predication . . . . . . . . . . . . . . . . . . . . . . . . . 296 Interim Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
At the Crossroads of Tempus and Modus: the AspectualModal Correspondence and the Conceptual Foundations of Irrealis Modality in Earlier Egyptian Complementation 9.1 The Interrelation between Temporal and Modal Functions of the Active Suffix-Conjugation Forms in Complementation and Beyond. . . . . . . . . . . . . . . . 9.2 Too Far or Too Close? The Modal-Aspectual Continuum of the Earlier Egyptian Irrealis-Marked sDm=f Forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3 Interim Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . .
321 329
10. Conclusion: Retrospect and Prospect . . . . . . . . . . . . . . . . . . 10.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2 The Diachronic Status of the System . . . . . . . . . . . . . 10.3 Extending the Hypothesis . . . . . . . . . . . . . . . . . . . . . . 10.4 Final Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
335 335 341 347 360
308
309
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363 Index of Cited Passages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379 General Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419 Index of Languages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427 Key to Abbreviations and Symbols Used . . . . . . . . . . . . . . . . . . 429
viii
table of contents
table of contents
ix
ACKNOWLEDGEMENTS This book represents a revised and extended version of my 2005 University of Liverpool doctoral thesis. The original work underwent a series of transformations before reaching its current form. At one point it was turned into a shorter study on complements of verbs that in 2004 gained me a research fellowship at Trinity College, University of Cambridge. Certain parts of it also integrate results from work carried out in the framework of the project ‘Basel Diachronic Grammar’, which began in 2005. I would like to express my gratitude to all those who have assisted me and in various ways taken part in the preparation of this book: Dr Mark Collier for his supervision of and comments on the work on which this volume is based as well as to my examiners Dr Christopher Eyre and Dr Jean Winand for their critical remarks. Prof. Antonio Loprieno and Prof. Wolfgang Schenkel not only deemed my work worthy of being published in the series Probleme der Ägyptologie, but also offered invaluable criticism and suggested many improvements to the text. Prof. James P. Allen similarly drew my attention to certain issues that I had overlooked as well as kindly provided me with information and material not available in a published form. Mr Michiel Swormink of Brill Publishing offered assistance of a more technical kind, besides displaying considerable patience and flexibility as regards matters such as dates of submission of the manuscript etc. A special word of gratitude I would like to express to my friend and colleague Dr Matthias Müller with whom I discussed a great many topics treated below and who, as the first critical reader of the manuscript, saved me from various embarrassing slips and inaccuracies. Of those that still remain, I bear the sole responsibility. Also my own and my wife’s family have acted as an ever-reliable supporting team at every stage of my research. Finally, my most heartfelt thanks I wish to tender to my wife Tuuli, to whom I am happy to dedicate this work. Cambridge, December 2006 Sami Uljas
x
table of contents
introduction
1
INTRODUCTION 0.1 Preliminaries 0.1.1. Orientation “With the means at our disposal it is not possible to distinguish different moods in Egyptian, if such existed”. Thus Sir Alan Gardiner in his Egyptian Grammar, still justly the most admired and widely used work on the Ancient Egyptian language in general and Middle Egyptian in particular.1 After nearly eighty years since Gardiner first made his brief remarks, the role of modality—a cover term for a wide array of subjective speaker ‘attitudes and opinions’—and its different manifestations in the grammar of Earlier (Old and Middle) Egyptian remain relatively little explored.2 Indeed, gleaned through the prism of modern Egyptology, the early language of the pharaohs, the medium of some of the earliest and most cultivated literary and philosophical works in world history, appears as somewhat of a linguistic anomaly. Of the three verbal ‘meaning domains’ of tense, aspect and modality, (henceforth tam) the first two seem to have had a notable role in the grammatical organisation of the language, whereas the third is at best marginally represented. Earlier Egyptian appears to have been manifestly a ‘talanguage’ largely bereft of the sophistication of expression attained by means of grammatical mood and modality in almost all other linguistic systems, both ancient and modern. The present work seeks to elevate the role of modality in Egyptological linguistics by proposing a novel, semantic-pragmatically based analysis of the grammar of Earlier Egyptian complementation, with a focus on fully clausal complements.3 Traditionally, complement clauses have been viewed as subordinate construals whose basic character is analogous to 1
Gardiner 1957 (henceforth GEG) § 294. It is not the purpose of the present work to provide a historical overview of modality in Egyptological linguistics; the most relevant works pertaining thereto will be noted and discussed as occasion arises. ‘Earlier Egyptian’ is defined here as in Loprieno 1995, 5–6, but excluding the so-called Égyptien de tradition. 3 With the infinitive, Earlier Egyptian governing verbs occasionally develop meanings different from their lexical semantics, some of which are clearly modal (see Uljas 2003, 393–95). However, the infinitive itself has no TAM-profile and is not marked for 2
2
introduction
substantives. Rather akin to nouns, in English a clause may function e.g. as an object complement of a transitive verb as in (i) below, or as a subject complement of e.g. a passive predicate as in (ii):4 (i) (ii)
Jill knew [that Jack had left] [That Jack had left] was known (by Jill).
Besides entering the structure of the main clauses, the subordinate propositions in (i)–(ii) also flesh out the semantics of the superordinate verb. Characteristically, the former constitutes a compulsory, albeit not a very prototypical, semantic argument of a main (matrix) verb.5 Yet, this is only insofar as the governing verb is a truly lexical predicate. In Egyptian as elsewhere, auxiliary verbs also occur with syntactic complements, but the argument-projecting facilities of the former are no longer functional.6 Further, the complement clause need not be associated with a verb at all, but may also appear e.g. as a predicate of a nominal sentence, as in (iii) below, or as a complement of a preposition (iv):7 (iii) (iv)
Jill’s opinion was [that Jack had left] It was a question of [whether Jack had left].
In Earlier Egyptian, clausal object- and subject complements of verbs display two main types of construal.8 In the affirmative, the more
person/number/gender. Consequently, infinitival complement clauses fall beyond the scope of the present study. 4 However, particularly in object complementation this provides merely the most basic blueprint for the structure, which is much less static than might seem. See chapter 1 for a brief discussion. 5 Notably, a clausal complement can never serve e.g. as an affected patient, which is the most typical semantic role of nominal objects of active and subjects of passive verbs. 6 As a signal of this, the complements themselves are almost invariably infinitival (see Uljas 2003, passim for Egyptian). Auxiliation is a diachronic process and auxiliaries are derived from earlier fully lexical verbs. One case of such a development in Earlier Egyptian is discussed in depth in 4.3 below. 7 Cf. Ransom 1986, 29 n.2. This shows e.g. that complements are not necessarily ‘entailed’ by the semantics of some other state of affairs (contra Cristofaro 2003, 38, 95). 8 In addition to the construction types illustrated below, clausal complements may also be used as direct or indirect genitives in Earlier Egyptian (see GEG §§ 191–92; Gunn 1949). Discussion of these must be left to a further occasion.
introduction
3
prevalent of these is directly embedded bare, or un-introduced, suffix-conjugation forms, typically a sDm=f of some sort.9 In example (1) below, the clause qnn=f functions syntactically as the object of the governing matrix predicate mAA, ‘see’, as does Hs Tn nTr=tn in relation to the verb mri ‘want’ in (2) and sipw sw wrty-HqAw in relation to wD ‘order’ in (3). In example (4), mss=s serves as a clausal subject of the adjective verb qsn ‘be difficult’: (1)
Amenemheb testifies to the valour of his royal master on battlefield:
iw wHm.n=i {n} mAA qnn=f iw=i m Smswt=f Again I saw how brave he was, when I was in his following. (Urk IV 892, 6–7)
(2)
From an Appeal to the Living:
[i anxw tpw] tA iw.t(y)=s[n] r [b]iA pn mr=Tn Hs Tn nTr=tn Dd=Tn [xA] m t Hnqt O the living upon earth who will come to this mining-region; as you will want your god to favour you, may you say: “A (Sinai 167, 1–3) thousand bread and beer…” (3)
The deceased is stated to be under Osiris’ command:
iw wD.n wsir sipw sw wrty-HqAw Osiris has ordered the two Great-of-Magic to examine him. (CT I 131a-b/L2Li)
(4)
The story of the birth of three kings describes their mother’s travails:
wa m nn hrw xpr wn.in rwd-Ddt Hr Snt=s qsn mss=s One of these days, Redjedet was suffering because her labour (pWestcar 9, 21–22) was difficult. As can be seen, the un-introduced sDm=f of mutable roots occurs in various different forms and sometimes shows characteristic endings -w and -y. In example (1) the ultimae infirmae root qni shows gemination (qnn), whereas in (2) the root Hsi from the same morphological class does not (Hs). In example (3) the caus. 2rad root sip shows an ending -w.
9
Discussion of complements with bare sDm.n=f will be postponed to 7.1 below.
4
introduction
These morphological peculiarities have, as shall be discussed shortly, aroused much speculation among Egyptologists. The second principal type of affirmative complements after verbs is clauses introduced by the element ntt/wnt:10 (5)
The king ends his letter announcing his accession to his Nubian viceroy:
hAb pw r rdit rx=k st ntt pr-nsw aD wDA This is a correspondence to let you know it, and that the royal (Urk IV 81, 2–3) house is sound and prosperous. In Earlier Egyptian clauses may also appear in various other complement environments, most notably as objects of prepositions used as conjuncts. In the affirmative, suffix-conjugation forms, particularly sDm=f, may again appear directly embedded, as in examples (6)–(8) below, or the complement may be introduced by ntt (rarely, wnt) as in example (9): (6)
Sinuhe expresses his opinion to his host on the motives of an adversary:
rqt-ib [pw] Hr mAA=f wi Hr irt wp[wt]=k It is but envy because he sees me carrying out your business. (Sin R 141–42)
(7)
Hapdjefa gives instructions on the use of a taper by an official performing his mortuary-cult:
Aw=f s(y) n Hm-kA=i r-sA iri=f irt=f im=s m Hwt-nTr He is to pass it on to my kA-priest after he has done what he is (Siut I 297–98) to do with it in the temple. (8)
From a medical instruction on treating a bleeding swelling:
dd=k Hr=s Ast nt ds dd=k st r hAw snf Whenever you put a splinter of firestone on it, apply it in (pEbers 88, 19) such a way that blood can flow down.
10
Less common methods of introducing complements will be discussed in 7.2 and 7.3 below.
introduction (9)
5
Amenemhat I, assassinated whilst taking repose, advices his son from beyond the grave:
sDr=k sAw n=k ib=k Ds=k Hr ntt nn wn{n}11 mr n s hrw n qsnt When asleep, keep yourself alert, because no man has a servant (Amenemhat IIIa–b) on a difficult moment. Fully clausal complements occur also as predicates in bipartite nominal sentences with the element pw. However, in this class of complement structures there exist no ntt/wnt-introduced variant to the type of sentence exemplified in (10) below with the bare geminating sDm=f. The construal in example (11) is modally equivalent to the former, but it is affiliated to a wholly different set of construals than complement clauses and has its own semantic-pragmatic characteristics, which explain its peculiar form: (10)
A medical gloss explaining a term:
ir nnw mwyt=f hAA mwyt pw m Hnn=f n sA.n=s n=f As for ‘his urine is lazy’; this means urine issues from his (pEdwin Smith 10, 21–22) penis without control. (11)
Queen Hatshepsut tells the background for her decision to erect two large obelisks for Amun:
ink pw snDm.n=i m aH sxA.n=i qmA wi I relaxed in the palace, and I took thought of the one who (Urk IV 364, 16–17) created me. In negative complements, the same overall formal division appears. Again, after verbs and prepositions the complement may be unintroduced, in which case the negative verb tm or the negative construction nfr-n appears. In introduced complements, the element iwt (or, alternatively, ntt + n-negation) is used: (12)
Hapdjefa describes his impeccable conduct in life:
ink dr bXbX m qA-sA sgr qA-xrw r tm=f mdw I was one who removed pride from the arrogant and silenced the loud-mouthed one so that he would not speak. (Siut I 229) 11
So pMillingen 1, 5; other variants have correctly wn.
6 (13)
introduction The king notes to his envoy concerning a dwarf that the latter is bringing:
Dd.n=k xr Hm(=i) iwt zp in.t(i) mit(y)=f in ky nb You have said to my majesty that never has his like been brought (Urk I 129, 2–3) by anyone else. In negative nominal sentence predicate complements only un-introduced clauses with tm are attested: (14)
A medical gloss explaining a term:
ir r=f mr... tm=f wn r=f pw mdw=f As for ‘his (the patient’s) mouth is tied’… this means that he does not/cannot open his mouth so that he might speak. (pEdwin Smith 4, 2–3)
The current opinio communis among students of Egyptian is that the guiding parameters of this grammatical and morphological organisation are primarily syntactic in character. The roots of this conception lie to a notable degree in wissenschaftsgeschichtliche trends peculiar to Egyptological linguistics. Early discussions of Earlier Egyptian complementation were already characterised by a rather scant concern with the particulars of meaning of clausal complements, most scholars contenting themselves with descriptions of their syntax and typology.12 However, the structural orientation in the analysis of the grammar of these constructions was, it seems, perpetuated by the so-called Standard Theory (henceforth ST) of Egyptian based on the work of H.J. Polotsky.13 Although since the early 1990s many of the basic tenets of the ST have been challenged, the effects of the ‘Polotskyan revolution’ are still widely felt in Egyptological linguistics and, arguably, nowhere more so than in the manner which the grammar of complement clauses is viewed. Prior to the advent of the ST, the use of the different complement patterns had been described in terms of dialectics between form and function. 12 See e.g. Sethe 1899, §§ 142, 150–01, 366; Erman 1928, §§ 356, 523–25a, 531– 32d; GEG §§ 154–57, 183–92, 223; Lefebvre 1955 (henceforth LGEC) §§ 612–16, 687–711; DeBuck 1952 §§ 119, 126; Edel 1955–64 (henceforth EAG) §§ 1014–27; Westendorf 1962 (henceforth WGMT) § 436. 13 For bibliographies and summaries of the rise and key hypotheses of the ST, see Depuydt 1983; 1995; Junge 1989, 11–18; Schenkel 1990, 145–58.
introduction
7
For example, according to Gardiner, in un-introduced complements the various verb-forms employed as subjects and objects of verbs and after prepositions had these syntactic roles only ‘virtually’ without being specialised for such uses.14 In the ST, however, form and function were largely synonymous; the geminating sDm=f and its non-geminating counterpart, analysed as a separate (or various separate) ‘prospective’ form(s), were interpreted as nominal forms or nominal ‘transpositions’. Their ability to function directly as complements, as well as their negation therein by tm, seen as the negator of all ‘nominal forms of the verb’, was cited as particularly transparent testimony of this inherent syntactic nature.15 One of the most persuasive arguments seemingly favouring this interpretation is the conspicuous absence of the bare geminating sDm=f/tm from the paradigm following the elements ntt/wnt as well as iwt.16 Accordingly, these elements were seen to allow forms and constructions somehow syntactically inappropriate for ‘nominal’ use to function thus.17 The ST analysis of the assumed inherent syntactic properties of the forms and elements employed in complementation as the primus motor of their grammar not only shared some resemblance with structuralist theories aiming at mathematising language and greatly in vogue between 1950s and 1970s. It also seemed, for the first time, to provide a model capable of accounting for all complement patterns by reference to the shared criterion of ‘nominality’. For example, before the ST there had been no explanation to the use of ntt, which e.g. Gardiner describes merely
14 GEG § 182; this is a clear functionalist formulation of grammatical use; see 0.2 below. 15 Polotsky 1944, §§ 25, 28, 30–31; 1964, 276–77; 1969, 470; 1976, 2.7.2, 2.7.5, 2.7.6; 1987, 19; De Cenival 1972, 41–42; Frandsen 1975, 18–20, 56–57, 69; Schenkel 1975, 41; Silverman 1985, 281; Depuydt 1983, 29–30; 1993, 19 among others. tm is characterised as ‘nominal’ already in Sethe 1899 § 994. Further proof of this was also seen in the obvious morpho-syntactic similarities between the ‘nominal’ and relative forms, the former of which were, from early on, seen as actual ‘abstract relative forms’ (Polotsky 1944, § 25 and passim; most recently Allen 2000, chapter 25 and Schenkel 2006). This topic will not be discussed in the present work, but certain remarks pertaining thereto will be found in the conclusion (section 10.3). 16 Cf. Polotsky 1944 §§ 25, 28. 17 Gilula 1970, 213; 1971, 16; Junge 1979, 84; Satzinger 1986, 307; 1989, 216; Allen 1986b, 25–27, 33; Silverman 1985, 272; 1986a, 38; cf. also Doret 1986, 34 n.264. Already Gunn (1924, 176) and Erman (1928, § 531; 1933 § 723) refer to ntt/wnt as ‘nominalisers’. Hypotheses on the precise syntactic character of the forms/constructions after ntt/wnt/iwt are discussed in 2.4 below; for the ‘prospective’, see 0.2 and 2.4 below.
8
introduction
as an element “occasionally used for ‘that’” after selected verbs and prepositions, without proposing an explanation as to what might condition its use.18 However, the ST ‘nominal hypothesis’ entailed an assumption of the grammar of complement clauses (and Egyptian in general) as almost wholly driven by syntactic rules. In many discussions within this school of thought one discerns a tendency to regard the ‘transpositions’ system an sich as a sufficient explanation for grammatical phenomena. Polotsky himself reserved some of his most ‘revolutionary’ fervour against the earlier views of Gardiner et al when discussing just the complement-uses of his ‘nominal forms’, at times even arguing that they represented purely syntactic entities wholly beyond the system of tam.19 Junge’s dismissal of the complement form-variation as mere hesitation of ‘speakers’ intuition’ or as nebulous irregularity in the ‘surface structure’, expressive of tam ‘Zusatzinformation’ represents the most explicit formulation of this theoretical vantage point.20 Yet, most ST treatises did usually not share such extreme views. Indeed, the Polotskyan School never sought a complete breakaway from earlier views that had stressed the role of tempus, particularly verbal aspect, in the organisation of the different complement patterns.21 Also Polotsky’s own later work is characterised by increasing stress laid on the importance of semantics (again mostly aspect) in the verbal system of Earlier Egyptian, particularly in case of the geminating sDm=f.22 Nevertheless, the ST analysis of complementation as based on incorporation of nominal verb-forms and nominalised clauses into their predetermined syntactic slots has proved to be lasting. At pace with the gradual rise of the ST as the most widely accepted paradigm for analysing Egyptian, also in the domain of complementation the number of approaches deviating from its precepts was reduced to naught.23 Among the advocates of the ST or more broadly in the 18 GEG §§ 187, 237; regarding prepositions + ntt, Gardiner notes that these complexes primarily express cause, (§ 223) but not in all instances. See 6.3 below. 19 Particularly the geminating sDm=f; Polotsky 1964, 281. 20 Junge 1978a, 98, 104, 109. 21 See Schenkel 1990, 151 for this point. 22 See Polotsky 1976, 2.3.1 and chapter 9.2 below. 23 Even the few approaches seeking an alternative to the ST make concessions to its tenets in case of complement clauses. For example, Thacker (1954, 331) claims that by Middle Kingdom the geminating sDm=f had come to express ‘energic’ sense—except in complementation, for which he accepts Polotsky’s analysis of the forms used as relative
introduction
9
intellectual climate permeated by it, there appears to have been little enthusiasm for specifically studying the semantic or pragmatic characteristics of the ‘nominal forms’ in complementation. Practically all post-1950s and pre-1990s discussions of the sDm=f-forms and tm in these construals assume their putative ‘nominality’ therein. Also in the current predominantly ‘post-Polotskyan’ landscape of Egyptological linguistics, views on the grammar of complement clauses continue to reflect the deeply entrenched conception of its syntactic underpinning. The assumption of the ‘nominal’ or ‘nominalised’ character of the bare complement sDm=f forms and tm/nfr-n-clauses, as well as the role of ntt/wnt as ‘nominalisers’ or nominal converters of various ‘nonnominal’ patterns not ‘specialised’ for such use still serves as the point of departure for well-nigh all discussions, differences in theoretical meta-language notwithstanding.24 The hypotheses on tempus/aspect and on the other semanticpragmatic characteristics of Earlier Egyptian complementation will be discussed later in the present work. Yet, due to its fundamental status in all research since Polotsky’s seminal work, the assumption of the role of syntax as the driving force behind the grammatical organisation of complement constructions must be addressed presently. In spite of the admirable consensus prevailing among researchers, at closer inspection the current ‘nominal hypothesis’ of Earlier Egyptian complementation turns out to be bedevilled by various sorts of problems. The most immediately apparent difficulties pertain to the mutual distribution of the bare ‘nominal forms’ and ntt/wnt. The ‘nominal hypothesis’ suggests that complements with and without the latter are in principle mutually substitutable. The tacit assumption seems to be that, as syntactic ‘converters’, ntt/wnt themselves are essentially meaningless. However, their actual distribution does not support this view. For example, ntt and wnt are widely used to introduce object complements forms of some sort (209). Even Gardiner, in his otherwise highly critical review of Polotsky 1944, felt obliged to state that it contained “no more penetrating pages” than those devoted to tm (1947, 99 n.4). Borghouts (1985, 36) argues that the use of the geminating form is motivated solely by temporal factors, but notes also that in complementation the ‘nominal’ character of the verb is undeniable. Yet, since all suffix-conjugation forms (allegedly) originate in nouns, here this inherent character merely “shimmers through” (for an outright rejection of this argument, see Satzinger 1993a, 205). 24 See e.g. Collier 1990a, 83–84; 1991a, passim; 1999, 57; Grandet & Mathieu 1997, 406, 444; Malaise & Winand 1999 §§ 574, 601–03, 611, 612–18, 895–923; Allen 2000, 137. This claim has been made by the present author as well (Uljas 2000, 130).
10
introduction
of verbs of locution, cognition and perception, but they never occur e.g. after verbs of volition, ordering or preventing:25 †
iw mr.n=f ntt ink Dd=i st He wanted that it would be me who will say it †
iw dbH.n=f ntt Dd=s n=f st He asked that she would say it to him †
iw wD.n=f ntt=f r Dd st He ordered that he say it
†
iw xsf.n=f ntt Dd=f st He prevented him from saying it
Similarly, there are only certain prepositions that combine with ntt/ wnt-clauses, whereas many combinations such as e.g. †m ntt, †n-mrwt ntt or †n-aAt-n ntt are never found.26 This clearly has nothing to do with the syntactic nature of the construals following ntt/wnt, and it remains unclear why such restrictions should apply if the role of these elements is simply to ‘nominalise’ the following construal and if the clauses thus introduced represent mere “analytic counterparts” or substitutes of the ‘nominal forms’.27 There are also certain gaps in the ‘nominalising’ function of the ‘converters’. For example, ntt/wnt cannot ‘nominalise’ such clearly ‘non-nominal’ expressions as e.g. adverbial phrases:28 † †
iw mA.n=f ntt m pr/aA He saw that in the house/here
The same holds also for fully clausal adjuncts introduced by prepositions: 25
Cf. Uljas 2000, 126; 2003, 388. However, at this juncture it should be noted that in many analyses the elements ntt/wnt after prepositions are not interpreted as introducing complements at all, but are seen to form non-isomorphic ‘adjunct connectors’ of the type ‘preposition-ntt’ with the preceding prepositions. For discussion of this topic, see 6.3 below. 27 Polotsky 1976, 2.3.1; Satzinger 1986, 306; cf. also Meltzer 1991, 227. 28 Cf. Collier 1990a, 78. 26
introduction † †
11
iw mA.n=f ntt xft irr=f st He saw that when he does it
Worse still, as will be discussed later on, there are main verbs that accept some or all of the so-called ‘non-nominal’ forms and constructions as their objects and subjects without a ‘converter’, and a similar anomaly is found in nominal sentences, where un-introduced ‘non-nominals’ nevertheless appear as predicate complements.29 Yet, the syntagmatic properties of the alleged ‘converters’ have aroused remarkably little interest. As before the ST halcyon days, comments to this effect have largely amounted to mentions that they occur only after ‘certain verbs’ and prepositions.30 Most researchers have focussed on their paradigmatic character and the nature of the forms and constructions after ntt/ wnt rather than on the clauses as a whole. These are, naturally, of fundamental importance.31 However, the restrictions in distribution of the ‘converters’ noted show that it is at least equally important to understand where and when ntt/wnt can or cannot occur and why. Moreover, there is a more general issue involved here. If ntt and wnt were truly required to introduce forms and constructions syntactically unsuited for being embedded as complements, one may justly ask why did they need to be so embedded at all if the language allegedly possessed also a set of ‘nominal forms’ as the ‘analytic counterparts’ of many of them? Why were the syntactically ‘unsuitable’ patterns not simply replaced by the ‘nominal forms’ consistently? Why was the clause ntt pr-nsw aD wDA in example (5) above not replaced by aD wDA pr-nsw, i.e. ‘nominal sDm=f’s’ of some sort? What motivated the use of ntt + the ‘unsuitable’ stative here but not e.g. in example (1), where the complement is similarly an intransitive state, but now appears as qnn=f, a bare geminating sDm=f? In a similar vein, if prepositions are assumed to govern ntt-complements, it may be asked what exactly prompted the use of ntt + nn wn in example (9) above, when a bare preposition Hr + the ‘nominal’ negation tm wn mr n s hrw n qsnt could, in principle, have been ‘substituted’ in its stead? Issues and questions such as these make it abundantly clear that it simply 29
See 2.4; 4.3 and chapter 8 below for all these construction types. Gilula 1970, 213; 1971, 16. Cf. also e.g. Allen 1986a, 11; 1986b, 25; Doret 1986, 28 n.146; 34 n.263; 48; Grandet & Mathieu 1997, 444, 446. 31 See 2.4 below for a full discussion. 30
12
introduction
does not suffice to insist that ntt/wnt are required for purely syntactic reasons when they obviously are not so used universally in what is, after all, constantly the same syntactic environment. Consequently, in order to understand what determines their use, clauses with ntt/wnt should be treated first as units. After the parameters conditioning their occurrence have been clarified, one may turn into the question whether or not the forms and patterns they contain have some shared feature(s).32 Just as analysing ntt/wnt as mere ‘nominalisers’ or ‘converters’ seems thus to be an oversimplification, the alleged ‘nominal’ character of the bare suffix-conjugation forms used in complementation is similarly disputable. In terms of finiteness, valency, and the manner in which the bare complement forms project arguments, they clearly behave as verbs, not nouns. Alternatively, the label ‘nominal form’ should perhaps not be understood simply as denoting a ‘noun equivalent’, but rather to mean something like ‘nominalised verb’. However, when assessed against properties whose clustering is usually seen as prototypical indicators of such a status of verbs, the bare suffix-conjugation forms used in complementation fare rather poorly.33 In these environments, they can indeed be seen as behaving as the syntactic ‘head noun’, but they do not e.g. acquire nominal morphology,34 do not mark their subjects or objects as genitives, do not lose their ability to express tam, nor accept nominal determiners such as articles, demonstratives, adjectives, numerals or relative clauses. Yet e.g. with the infinitive, a truly ‘nominal’ verb-form, most of these criteria are met.35 Thus, since they are clearly not substantives, but neither fit the profile of verbal nominalisations, it is wholly unclear exactly what sort of ‘nominals’ or ‘syncategorisations’ the bare forms are supposed to be. The ‘nominal’
32
This latter issue will be the topic of section 2.4. Sasse 2001; Givón 2001, vol.2, 25, 35; for Egyptian in particular, see Claudi & Mendel 1991. 34 Cf. Polotsky 1944 § 26.1–2; Gardiner 1947, 98; Reintges 1997, 115 n.10. The gemination displayed by the ultimae infirmae roots is surely not to be seen as such (cf. Eyre 1986, 134) particularly as this would not apply to the ‘nominal’ ‘prospective’ and sDm. n=f form(s). 35 Thus pA prt ‘the going’, prt iri.n=f ‘the going which he did”, prt 3 ‘three goings’ etc. The only exception here is the feature of nominalising morphology, which is universally absent from infinitives. 33
introduction
13
character of the geminating sDm=f, the sDm.n=f 36, the alleged ‘prospective’ and the negation tm is also seriously in doubt generally, and not just in complementation. All the counter-arguments above against the ‘nominal’ nature of these forms in the latter environment apply similarly to their other ‘nominal’ uses. Further, the assumption of the ‘nominal’ character of the ‘nominal forms’ in second tenses, the most widely attested use of the geminating sDm=f, rests on a dubious syntactic analogy (not similarity, as in complementation) with adverbial sentences plus a highly suspect theory on ‘predication’.37 It is also disputable whether any ‘nominal sDm.n=f’ exists at all, 38 and insofar as there are one or more ‘prospective’ sDm=f forms, these also occur in patently non-nominal environments such as in final ‘so that’-clauses, which even the most ingenious ST postulates have failed to analyse as nominal in character.39 The last property pertains also to tm, whose capability to negate adjunct clauses is in fact not restricted to final clauses,40 and which is also the negation of sequential sentences with the elements ix, xr and kA.41 When it is added that there are good reasons to doubt the entire division of the sDm=f forms in Earlier Egyptian,42 the assumption of ‘nominality’ as the founding principle in the grammar of complement clauses in this language seems unlikely to be correct and should be seriously reconsidered. 0.1.2 Assertion and Non-assertion It is argued in the present work that instead of syntactic characteristics of the forms and constructions employed, the grammatical organisation 36 The issue whether the sDm.n=f in the ‘nominal’ uses differs morphologically from that employed in ‘non-nominal’ environments will be discussed in 7.1 below. 37 See Collier 1990a, 80–82; 1992 and 10.3 below. 38 See 7.1 below for full discussion. 39 Attempts to divide the ‘prospective’ sDm=f into ‘adverbial’ and ‘nominal’ forms (Allen 1982, 25; Depuydt 1993a) as well as hypotheses of the possibility of using ‘nominal’ forms ‘adverbially’ (Allen 1984 § 290; Doret 1986, 43) have been rife. The initial uses have been argued to represent complements of some zero ‘ø main predicate’ (Schenkel 1975, 41; 1978, 113–14; Junge 1978a, 122; Depuydt 1983, 46; Allen 1984 § 255; 1991, 7–8; Kammerzell 1988, 41; Grandet & Mathieu 1997, 253. Jansen-Winkeln 1995 extends this to even final uses; for counter-arguments, see Uljas 2000, 126– 27) or ‘single noun clauses’ (Polotsky 1964, 271). 40 See 5.3 below for a brief discussion. 41 There has been no shortage of ST attempts to interpret constructions with ix/xr/ kA as ‘nominal’, see Satzinger 1968, §§ 73, 78 for one. 42 See 0.2 below.
14
introduction
of Earlier Egyptian complement clauses is based on modality and differentiation between asserted and non-asserted complements. Since these concepts belie great complexity, a discussion and illustration of their meaning and nature is indispensable. The term assertion refers to a particular kind of speech act, the performing of which is regulated by the co(n)text of the communication and its intent. Asserting requires the following preparatory conditions to be fulfilled:43
The S(peaker) has a reason (evidence etc.) for believing in the veracity of the p(roposition); S is committed to p; p is not obvious to both S and the H(earer).
In addition, and following from the last criterion above, the speaker’s aim must be that of informing H of p and of his commitment thereto, typically in view of convincing H of p. This requirement for assertion is known as the speaker’s illocutionary intention:
S assumes that saying p is relevant to, and in the interest of H; S wishes H to believe p and that S is committed to p.
Thus, p is asserted if S indicates some degree of positive commitment towards the information it conveys and the latter is not something that S knows or assumes to be already known by H or lacking in relevance—i.e. asserted propositions have a proportionally high information value. Conversely, and abstracting away from the more precise taxonomy of speech acts, propositions in which one or more of these criteria are not fulfilled represent non-assertions. This provides a notional definition for assertion and non-assertion as particular kinds of illocutionary forces. However, the grammatical coding of the notional concepts is an altogether different issue, and here assertion and nonassertion become a matter of modal organisation. In general terms, modal systems and grammatical coding for modality are activated by the status of propositions as assertions or non-assertions. In particular, the use of grammatical indicative or realis corresponds to expression 43 See Searle 1969, 67; Bach & Harnish 1979, 42; Edmondson 1981, 145; Levinson 1983, 277; Allan 1986, 193; Wierzbicka 1987, 321; cf. also Allan 1998, 925–26; Langacker 1991, 496.
introduction
15
of asserted propositions; in case of non-assertion, languages resort to non-indicative or irrealis patterns. The formal method of coding these two categories displays notable variation. Realis may be overtly indicated as such, but is usually unmarked,44 whereas expression of the often more marked irrealis ranges from particles and modal auxiliaries to verbal inflection, primus inter pares the subjunctive mood of numerous (particularly Romance) ‘Standard Average European’ languages.45 For example, in Spanish the use of the subjunctive is directly linked to non-assertion; a proposition may appear in the subjunctive if46 The speaker has doubts of its veracity (v below) It describes an unrealised situation (vi) The information conveyed is presupposed (vii) (v)
Dudo que sea (sub) buena idea “I doubt that’s a good idea”
(vi)
Necesito que me devuelvas (sub) ese libro “I need you to return that book to me”
(vii)
Me alegra que sepas (sub) la verdad “I’m glad that you know the truth”
Particularly the first and the last of these properties fail the preparatory condition for assertion: doubt equals lack of commitment and presupposition represents unchallenged information which the speaker assumes to be already obvious to, and accepted by the audience.47 44
Cf. Givón 1982, 155; 2001 vol.1, 330–31. Many languages are said to have a specific ‘irrealis’ rather than subjunctive mood, but this is largely a (Whorfian) research tradition: many modern linguists seem to have opted to employ the label ‘irrealis’ in their descriptions of various ‘exotic’ languages, perhaps to stress their uniqueness (cf. Palmer 2000, 185). Yet, both subjunctives and the ‘irrealis’ are grammaticalisations of the same notional category of non-assertion. The traditional view of the subjunctive, apparent in the etymology of the word itself, (Italian subiung , ‘subjoin’) is that it is the mood of subordination. This old assumption has been rightly discarded in linguistics; its inadequacy is clear from the data below. 46 Data from Lunn 1995, 430; see also the fundamental study by Hooper & Terrell 1974. 47 It is to be noted that presupposition is not tantamount to logical ‘necessarily true’. Presupposition as a linguistic concept refers solely to assumption of shared knowledge between speakers and hearers; it is perfectly possible to assert propositions whose logical 45
16
introduction
Presupposed information is part of the discourse background and hence lacks relevance: communicating what is already shared cannot be the speaker’s illocutionary intention and remains, by definition, unasserted. Notably, the above examples involve complement clauses, and there are no syntactic reasons barring the use of the indicative. Instead, the subjunctive is chosen for each proposition because “potentially assertable information must have two qualities: it must be both reliable… and informative as to news value. Information that is lacking in either of these qualities… is unlikely to be asserted”.48 However, the words ‘potentially’ and ‘unlikely’ in the last quote already raise various issues concerning this seemingly simple set of principles that are decisive for the understanding of Earlier Egyptian complementation. Assertion and non-assertion are prototype-based, rather than ‘classical’ Aristotelian categories whose all members are equally representative of the category as a whole: both can be viewed as encompassing certain core semantic notions surrounded by a peripheral domain of other more or less ‘good examples’ of (non-)assertion.49 As a result, languages vary considerably in what they classify as modally realis and irrealis within and between themselves, and to what extent. Particularly in irrealis, devices for expressing gradations of non-assertion are commonplace. Thus the grammatical treatment of e.g. epistemic expressions, i.e. speaker attitude towards the authenticity-status of a proposition, as either realis or irrealis universally reflects the degree of tentativeness expressed. Cf. the following Spanish and English expressions:50 (viii)
Tal vez me-estaba (ind)/me-estuviera (sub) esperando “Perhaps s/he was/might have been waiting for me”
truth-value is co(n)textually uncancellable, provided that the speaker does not assume this to be common ground with the hearer; cf. Hooper & Terrell 1974, 485; Lambrecht 1994, 61–63; Palmer 2000, 3–4; Cristofaro 2003, 30. This issue is also relevant with socalled ‘factive’ and ‘implicative’ verbs (see chapter 1 and sections 2.2.2 and 3.2 below). 48 Lunn 1989b, 691, emphasis in the original. 49 See Lunn 1989b and Bybee 1998 in particular. Thus e.g. when referring to irrealis-use, many authors prefer to speak of ‘reduced assertion’ (Bell 1980) or use expressions such as ‘suspension of affirmation’ (cf. Lunn 1989b, 687). 50 Spanish data from Givón 2001 vol.1, 314–15; cf. 1994, 301.
introduction (ix)
17
He’s a spy; I know it He must be a spy: look at him He may be a spy, who knows”
Much of deontic modality, i.e. expression of will, is grammatically irrealis across languages. But deontic is also subject to gradations of politeness, the assessed potential for control and manipulation over the ‘targets’ of the attitude, and thus in the likelihood of the state of affairs referred to.51 Consequently, the strongly demanding imperative often varies with less direct modal auxiliaries or subjunctives/jussives as in the English and Biblical Hebrew examples (x)–(xi) below. In Caddo, (N. Iroquoian) affirmative imperatives are realis-marked whereas negative imperatives and ‘obligative’ expressions (‘should’) are irrealis, reflecting the degree of expectation of compliance and realisation (xii):52 (x)
Go!
(xi) (xii)
You must/should/may/might go
( שמרimp)
( תשמרjuss)
“Guard!”
“May you guard”
(real) (irr) (irr)
“Look at it” “Don’t look at it” “He should sing”
This also shows an important faculty of ‘non-realisation’ vis-à-vis irrealis. There is no requirement that situations treated as irrealis be ‘unreal’. It is not ontology or, again, ‘truth values’ that determine the modal status of propositions. These concepts are logico-philosophical ‘absolutes’, whereas irrealis relates to speaker assessments of low information value, as in case of most certainly ‘real’ presupposed and their “judgments concerning the degree to which their ideas accord with what they believe to be objective reality”.53 Hence languages often divide the expression of e.g. futurity into ‘more’ and ‘less’ ‘objective’ types. The following examples from Central Pomo (N. California) illustrate this well:54 51
Cf. Givón 1975a; 1994, 298–300; Fleischman 1989, 8–12; Mithun 1995, 377. Hebrew and Caddo data from Greenberg 1965, 51 and Chafe 1995, 356-58 respectively. For further comments on ‘obligatives’, see chapter 6 n.23 below. 53 Chafe 1995, 364; emphasis by SU; cf. also Givón 2001 vol.1, 302 and Bendix 1998, 253–54. 54 Data from Mithun 1995, 370, 378–79. 52
18
introduction
(xiii)
•• (real) •= “We’ll go around eating” (certain) •=• (irr) • == “I’ll go to town and bring it back” (less certain)
Another more familiar example of the same phenomenon is the variation in Earlier Egyptian between the bare initial ‘prospective’ sDm=f and the ‘pseudo-verbal’ iw=f r sDm, recognised by most commentators as that between ‘subjective’ and ‘objective’ future respectively.55 The latter tends to occur when there are sufficient grounds for the speaker to be committed to the eventual realisation of the situation, which makes it more assertable. It is no accident, then, that e.g. first person intentions are mostly expressed by iw=i r sDm, as people tend to be more committed to and in control of their own actions than of those of others, or that the same pattern is the favourite also when the situation is somehow expectable through ‘experience’, ‘norm’, or e.g. divine prognostication.56 The use of such variants is not based on any more ‘objective’ criteria than approximation of likelihood and possibility of the future state of affairs; the final say on what to consider realis or irrealis belongs to the speaker.57 This is also apparent in the way in which speakers may withhold from asserting if e.g. they do not accept the proposition, as in the Albanian example (xiv) below, or because they do not wish to imply that what they say represents their own commitment, as e.g. in German ‘journalistic’ style (xv):58
55
See e.g. Hannig 1982, 47; Vernus 1990, 24–27 and numerous others. Thus for example, Peas B1, 42 mk wi r nHm aA=k sxty “I am going to take your donkey, peasant” does not portray the speaker as “compelled to seize the donkey as a logical punishment of the transgression” (Vernus 1990, 14) but expresses a firm conviction to carry out the said act. However, these words do not protest the speaker’s willingness to act, unlike e.g. the long string of first person bare sDm=f’s in Sh.S. 139–46: i.e. “I will (=am willing to) spread your fame at home” etc. Similarly, the standard expression iry=i r Hst=k “I will do as you ask” (EAG § 472) expresses readiness and willingness to obey, but no personal motivation as such; the variant iw=i r irt r Hst=k is notably rare. 57 A fine empirical study to this effect in Spanish is Lavandera 1983. 58 The Albanian ‘admirative’ usually expresses surprise, but as here, it can also indicate mockery and non-acceptance, both of which represent reduced assertion. See Haiman 1995, 333–34 for this and for similar examples from other languages. For the German data, see Palmer 2000, 42; in French, the conditional is similarly used, see Fleischman 1995, 533. 56
introduction
19
(xiv)
E na i dashka (adm) bullgaret.. ai e.. e pse keshtu u pritka nje Bullgar? “And he ‘likes’ Bulgarians. Him? Ha! After all, is that how you treat a Bulgarian?”
(xv)
Er sei (sub) mit S in Streit geraten und habe (sub) sich von diesem bedroht gefühlt “He had allegedly become involved in a quarrel with S and had felt threatened by him”
Also the requirement of ‘informativeness’ and relevance from asserted propositions is a very flexible concept.59 For example, in Spanish pairs such as the following are not uncommon:60 (xvi)
Aunque es (ind)/sea (sub) mi hija, la encuentro muy guapa “Although she is my daughter, I find her very pretty”
Here the indicative signals that the information about ‘she’ being ‘my daughter’ is communicated both as actual and relevant, whereas the subjunctive codes it as concessive and presupposed. Variation like this reflects speaker evaluation of the proposition information value and ‘background-ness’ and does not follow from any mechanical rule. Understanding it here requires particular attention to the co(n)text, as speakers’ motives for treating propositions as more or less informative can only be recovered by examining the discourse as a whole rather than isolated clauses or sentences. Further, there is no hard and fast limit to what speakers may and may not evaluate as ‘informative’. This depends on the quality of their performance in a given linguistic and even ‘extra-linguistic’ context. For example, in Spanish the subjunctive can signal that the audience is assumed as already ‘reasonably’ acquainted with the information, and the extent to which this can be taken is quite remarkable. In journalistic Spanish a prior mention of some proposition in the headline may suffice to prompt the use of the subjunctive when it is resumed in the main text:61 59 A superb discussion of the concept of ‘relevance’ and communication in general is Sperber & Wilson 1995. 60 Data from Lunn 1989b, 697–98; cf. Klein 1975, 364 n.8. 61 Data from Lunn 1995, 433.
20 (xvii)
introduction Headline: La bandera que besó es la que, en su día, también besó (ind) el Rey don J, y bordó (ind) su tatarabuela la Reina doña M. “The flag that he kissed is the one that one day king J also kissed and his great-great-grandmother queen M embroidered” Text: al final, besó la bandera roja y gualda que hace treinta años besara (sub) su padre el Rey y que un día bordara (sub) su tatarabuela la Reina doña M. “At the end, he kissed the red and gold flag that his father the king had also kissed thirty years ago, and that his greatgreat-grandmother queen M had once embroidered”
The next example derives from a gossip-magazine. Its author assumes the information in the relative clause to be ‘old news’ to faithful readership and indicates this with non-assertion:62 (xviii)
La pareja, que se hiciera (sub) famosa por interpretar el papel de marido y mujer en El pájaro espino, es (ind) en la vida real un matrimonio feliz “The couple, who became famous for their role as husband and wife in ‘The Thorn Birds’, is happily married in real life”
Here the principal information with the highest ‘news value’ is in the main clause, where the indicative appears. However, e.g. Russian and Polish dispense with the notion of ‘relevance’ from their grammaticalisation of modality altogether and code even presupposed information as indicative.63 This emphasises that the grammaticalisation of meanings as realis and irrealis is also a crosslinguistic variable. There are certain senses such as e.g. counterfactual,
62 Data from Lunn 1989a, 254; 1989b, 693; 1995, 433; cf. also Lavandera 1983, passim. Interestingly enough, since the antecedent of the relative clause is definite, the use of subjunctive should be ungrammatical. Yet, this rule is clearly ignored herein in favour of greater expressiveness—see also n.69 below. 63 Noonan 1985, 99; Wierzbicka 1988, 152.
introduction
21
whose non-assertivity is somehow signalled in almost every language.64 Yet the treatment of e.g. interrogatives, negatives and conditionals, all notional non-assertions, varies greatly. For instance, simple negated propositions in Alamblak (Papuan) are irrealis-marked, but realis in Central Pomo:65 (xix)
fiñji noh-r-fë-r (irr) “He did not die”
(xx)
=í í (real) “Because they didn’t keep up the rancheria...”
This is because negation can be viewed either as expression of total lack of commitment (negated assertion) or of commitment to falsity (negative assertion)—or kept outside the system of modality altogether, as in most European languages. Finally, assertion and non-assertion overlap conceptually with semantic categories other than modality, and this has grammatical consequences. For example, in some languages certain types of verbal aspect prompt irrealis-marking due to perceived similarity with ‘unreality’. Thus e.g. in Caddo, infrequent (xxi) and in Bargam (Papuan) past habitual propositions (xxii) are irrealis:66 (xxi)
(irr) “I seldom see it”
(xxii)
miles-eq leh-id (irr) “when (the pig) used to return…”
These meanings involve characteristics that set them metaphorically ‘close’ to irrealis. Something taking place seldom may be subjectively ‘as good as nothing’, and in Caddo this conceptual ‘family resemblance’ results in grammatical marking of such situations as irrealis. Habitual refers to a mass-like series of events rather than any one particular 64 For representative samples of languages, see D. James 1982, 378–84; Roberts 1990, 392. For counterfactuals in Earlier Egyptian, see 6.2, end and 7.1 below. 65 Data from Roberts 1990, 390 and Mithun 1995, 382 respectively. 66 Data from Chafe 1995, 357 and Roberts 1990, 384.
22
introduction
situation occupying a specific locus in time. The degree of individuation of the ‘sub-events’ is low and the proposition may even refer to a mere tendency.67 Thus habitual events may be viewed with precisely that “lack of belief in or lack of commitment to… the reality, realization, or referentiality of an event or sequence of events”68 characteristic to irrealis.69 This apparent lack of ‘coherence’ of realis and irrealis has led to occasional doubts of their validity as grammatical concepts.70 However, it could be argued that it rather shows their fundamentally scalar nature. The basic classes of non-assertion and assertion receive almost as many expressions in modal systems as there are languages and they can be further modulated into a notable degree. Yet, there is remarkable agreement across languages on e.g. what sorts of nonassertions constitute the ‘nucleus’ of irrealis and qualify for modal coding as such. Overall, however, realis and irrealis form a continuum of linguistic meaning without clear-cut lines of demarcation mutually or with other categories of grammar. In addition, and as a corollary of the previous point, irrealis and realis meaning is associated with linguistic expressions (verb-forms, particles etc.) in a non-arbitrary manner. In certain languages, the metaphorical linkage between e.g. habitual aspect and irrealis results in a linguistic mapping of both onto the same or co-occurring grammatical expressions. The task of the linguist is to define and explain this sort of procedures in the grammar of the specific language under study.
67 Cf. Givón 1994, 270–71; 2001 vol.1, 305; Fleischman 1995, 537–39; Palmer 2000, 179, 190–01. This is reflected in various ways. For instance, nouns under the scope of the habitual ‘hybrid modality’ are non-referring: in “he used to buy a newspaper every day”, ‘newspaper’ is non-referential and lacks a specific counterpart in ‘reality’. Note also the (particularly American) English use of ‘would’ instead of ‘used to’ for past habitual. 68 Fleischman 1995, 522, 537. 69 Irrealis and realis marking of verbs is sometimes affected even by case-marking of associate nouns and their definiteness; see e.g. Martin 1998, 199. In Romance, attributive relative clauses of indefinite nouns commonly employ the subjunctive; cf. n.62 above.. 70 See e.g. Bybee 1998. In Bybee et al 1994, 236–39 it is maintained that irrealis and realis are not appropriate concepts for assertion and non-assertion, which they see as the decisive factors in modal organisation. See Givón 1994, 320–28 and Palmer 2001, 188–91 for counterarguments.
introduction
23
0.1.3. The System in Earlier Egyptian and the Structure of the Present Work In Egyptology attitudes towards even the general status of modality in Earlier Egyptian have tended to be somewhat reserved.71 Polotsky’s dismissal of anything like an ‘echter modus’ in Egyptian reverberated through the writings of his followers,72 and the ST preoccupation with syntax sidelined modality in favour of speculations on structure and tense/aspect.73 In complementation, modality has hardly been discussed at all, although some researchers have e.g. proposed geminating sDm=f complements to portray the situation as ‘objective fact’ or ‘indicative’ and those with a bare sDm=f without gemination to entail “an element of possibility or doubt”.74 Most have had little to say about the pragmatic profile of ntt/wnt/iwt-clauses,75 and comments on modality and tm/nfr-n-complements are extremely scarce.76 The term ‘assertion’ has been applied to the illocutionary force labelled thus—particularly in connection with the auxiliary iw, which has been characterised as an ‘assertion-particle’, but assertion has mostly not 71 However, Gardiner’s stance in this matter, quoted in 0.1.1 above, did not amount to denial of modality in Egyptian a priori, but merely recognised the then current lack of suitable methodological tools for exploring this phenomenon. These are apparent e.g. in Sander-Hansen’s (1941) inspired, but perhaps premature attempt to distinguish moods in Egyptian. In Late Egyptian, however, there has recently been a surge of extremely interesting research, such as Winand’s fascinating study on the conjunctive, (2001) which substantially extends the pioneering work by Borghouts (1979) and demonstrates the irrealis character of this construction beyond doubt. 72 Polotsky 1964, 272; one may note e.g. the marginal remark by Depuydt 1993a, 13 n.7 or Callender’s (1986, 11) claim that on ‘internal grounds’ the notion of mood is redundant in Egyptian, justly disputed by Junge (1989, 39 n.62). 73 An example of the results of this is the synopsis of modality in Earlier Egyptian future expressions by Vernus, (1990) which is praiseworthy for its pioneering approach but which is unfortunately marred by a too great focus on tempus and adherence to ST dogmas. This problem is only partially remedied by reviewers (Ritter 1992; Schenkel 1992). In some treatises modality and pragmatics generally are seen to have given way diachronically to syntax as the driving force in grammar, including that of ‘originally’ modal patterns such as the geminating sDm=f or tm in the mature ‘syntactic system’ of Classical Egyptian (Loprieno 1991a, 215, 223; 1995, 82). 74 Doret 1986, 23, 39, 49; see also Gilula 1971, 16; Allen 1984, §§ 365, 703, 706; Malaise & Winand 1999, § 908. However, there has been some divergence in views as to how many different ‘prospective forms’ the latter disguises (see 0.2 below). 75 Notable exceptions here are Satzinger 1968 § 100, Allen 1986a, and Loprieno 1991a, 214. See 2.1.1 and 5.1 below for discussion. 76 But see Loprieno 1986b, 280–81; 1991b, 231–35 and 5.3 below. Silverman (1985, 281) and Vernus (1990, 119) note tm to be the negation of both geminating and non-geminating sDm=f-complements, which may indicate belief that it inherits the properties of both.
24
introduction
been treated as a modal concept.77 ‘Irrealis’ has been evoked but seldom, and always as a synonym for ‘unreal’.78 It is argued here that when understood as modal categories, realis and irrealis are analysable as the determining concepts behind the grammatical organisation of Earlier Egyptian complementation. The system can be summarised as follows: 1. The basic division of complement clauses into introduced and un-introduced types corresponds to differentiation between assertion and non-assertion. When the prerequisites for assertion are satisfied, the subordinate clause is specifically marked as asserted by the modal operators ntt/wnt or the negative iwt. Such complements are modally realis. 2. If the complement is not asserted, ntt/wnt or iwt are not used, and the clause remains un-introduced. In the affirmative, bare suffix-conjugation forms are employed. The patterns mostly used are various types of sDm=f, but also the bare sDm.n=f has its role to play in the expression of non-assertion. In the negative tm and nfr-n appear. Although less marked than ntt/wnt-clauses, complements of this sort are modally irrealis. 3. Unlike in many modern languages with a singular conjugated irrealis, e.g. a subjunctive mood, in Earlier Egyptian unintroduced sDm=f forms divide up the domain of irrealis and non-assertion in a principled and meaningful way. Their mutual variation in complementation serves to differentiate between hierarchically ‘stronger’ and ‘weaker’ non-assertion. However, in negated clauses, the modal (and formal) distinctions between these two degrees of non-assertion are neutralised and they share the common irrealis negations tm and nfr-n. 77 See Assmann 1974, 65; Junge 1989, 105; Hannig 1982, 43; Loprieno 1995, 169; Ritter 1995, 99; Ray 2000, 225 among others. Only Eyre (1991, 113) and Winand (2006, 166) treat iw as modal in its ‘assertive’ role. Zonhoven (1997b, 387) contrasts assertion and presupposition, Assmann (ibid) treats the former as equivalent to ‘predication’ and views adjuncts as ‘non-asserted’. Callender (1983, 93) correctly calls Hm a particle of “enhanced veracity and… relevance” and of ‘assertion’. In Late Egyptian the issues of assertion and illocution have received more attention; the most recent overall study to this effect is Sweeney 2001. 78 Early uses of the term in Egyptology are Erman 1933 § 536 and Till 1933. See also Vernus 1990, 16; Reintges 1997, 56; Hannig 2003, 1 for Earlier Egyptian.
introduction
25
The remainder of this work demonstrates how this system functions in practise and what are the parameters for coding a complement as an assertion or non-assertion in Earlier Egyptian. The point of departure for the discussion are affirmative object complements of main clause predicates, after which both asserted and un-asserted clauses are attested and which best illustrate the key principles of their variation. This is followed by a discussion of transitive verbs that, for reasons explained, systematically disallow realis ntt/wntintroduced object clauses, but which are shown still to form part of the same overall system of assertion and non-assertion covering all complementation in Earlier Egyptian. Subsequent chapters elaborate the image of this organisation by focussing on the little-studied affirmative clausal subjects and negative complement clauses after verbs. The discussion is then extended to preposition complements, for which a modal reanalysis is proposed along the lines established for complementation after governing predicates. Finally, after a discussion of certain less central patterns of complement clauses and predicate complements of nominal sentences, the ‘cross-categorical’ and more abstract conceptual foundations of the grammatical ‘mapping’ of Earlier Egyptian irrealis modality as revealed in complementation are surveyed. The use of the bare complement sDm=f forms is found to share notable parallelism with their temporal properties, particularly aspect. These two apparently disparate domains of use and meaning are shown to be inherently related and many previous theories focussing on the aspectual characteristics of the forms in question are recast into a more global analysis merging their tam- properties into one semantic-pragmatic profile. The analysis presented here is also intended to provide an initiative and methodology for tackling many further problem-areas of Egyptian grammar from a modal perspective. It is suggested at various points, and in the conclusion, that modal irrealis and realis play a role in the grammar of Earlier Egyptian far beyond the relatively restricted domain of complementation. Nevertheless, complement clauses offer an ideal basis for investigating modality at this stage of the linguistic history of Egyptian. These constructions provide sufficient data (largely) free from the burden of various complicating factors pertaining to initial clauses, most notably the character of auxiliaries, the issue of predication and the question of theme-rheme/
26
introduction
topic-comment-oppositions.79 They reveal a system for expressing modal meaning whose basic characteristics are recognisable as language universals, whose expressive potential is remarkable, and whose functional flexibility is deeply impressive. 0.2 Methodology, Terminology, Morphology and Sources Before embarking on the discussion of the system outlined above, certain remarks on methodological and related issues must be made. The approach adopted in the present work has, to an extent, a comparative element. Studying Ancient Egyptian is archaeology of a dead language in which cross-linguistic comparisons provide the only support available for hypotheses on semantic-pragmatics. The imperfectly preserved evidence cannot be understood or analysed in isolation. In most instances, the comparisons will be drawn from languages with no relationship to Ancient Egyptian. Genetic matches demonstrate whether a certain feature exists or does not exist in some language group. However, parallelisms with unrelated languages have the advantage of showing that if a given meaning suspected as grammatically realised in the language under study is actually so treated in languages more far a field, this is likely to reflect some underlying cognitive aspect of human communication in general. Further, the present study entails a conscious rejection of Bloomfieldian views according to which researchers of language have no direct access to ‘meaning’ but only form. It is argued here that systematic application of comparisons from other languages and close attention to the co(n)text in which the expressions studied occur allow substantiated (and not necessarily purely hermeneutic) judgements to be made of even such subtleties as the expression of attitudinal information in extinct languages. As a concrete illustration of this, one may consider the Biblical Hebrew examples in (xi) above. The discussion relies extensively on linguistic data. The reason for this is not only the need to avoid idiosyncrasy in analysis; employing linguistic principles and methodology in the study of Egyptian also renders the language more accessible to non-Egyptologists. 79
All these issues will be touched upon several times in the course of the present work and in some more depth in the conclusion (section 10. 3).
introduction
27
Ancient Egyptian is but seldom quoted in linguistic literature,80 and a significant contributing factor to this state of affairs is arguably the combined effect of the syntactic models with which the language has been described and the sense of overwhelming ‘otherness’ entailed. Postulates such as the various ‘nominal transpositions’ and ‘conversions’ create a false impression of it as somehow fundamentally alien, incomprehensible and best left to Egyptologists. But, if and when it is shown that Earlier Egyptian possessed clear and definable means of expressing modality in complementation that share resemblance with those found in other languages, this should attract the interest of students of language more widely. The same holds also for the special features of Earlier Egyptian modality—in particular the internal division of irrealis, the use of irrealis negations, and the generally ‘upside-down’ character of the system where irrealis is less marked than realis. The relationship with linguistics can perhaps finally be made more reciprocal. Just as methods and data from the latter can be appropriated for the benefit of Egyptology, a solid description in similar terms of Egyptian, the language with the longest documented history, can contribute significantly to the general scientific discussion on language and communication. The analysis presented is not couched in the vocabulary of any particular theory, but shares affinities with Cognitive Grammar, Functional Grammar and the so-called (Radical) Construction Grammar. 81 From the first-mentioned approach is adopted the view of the ideational basis of modality as ‘force dynamics’, i.e. as metaphorical abstraction of physical forces and barriers into the domains of compulsion and obligation in deontic, as well as reasoning in epistemic modality.82 A modal ‘force’ may be someone’s authority 80 And, alas, when this happens, the evidence is often misinterpreted. Two examples will suffice: Sasse, in his discussion of the thetic/categorical divide, refers to the Earlier Egyptian element iw as an ‘existential marker’ used in ‘thetic’ statements (1987, 576 n.19). However, this analysis is clearly misguided. Cristofaro claims that Ancient Egyptian uses only direct speech to convey reported statements “and possibly also reported believes and commands” (2003, 46). As will be seen, this too is incorrect. 81 A succinct summary of the basic tenets of the first two closely related approaches is Horie & Comrie 2000; for an outline of Construction Grammar and its developed form Radical Construction Grammar, see Croft 2001, 14–29. 82 Talmy 1988 is the basic theoretical text, Sweetser 1990, chapter 3 an illuminating practical discussion. The most recent overall treatment is Talmy 2000, 409–549, particularly 440–54.
28
introduction
and status in deontic and some indubitable set of premises and evidence in epistemic modality, which function metaphorically as ‘real’ physical forces affecting material beings. This approach explains well e.g. the intuition that ‘epistemic necessity’ (must) quite literally presupposes ‘compelling evidence’ which ‘forces’ one to conclude something. The ‘forces’ involved are subjected to equally abstract ‘barriers’: e.g. in “he may not go” an authority in deontic, or some ‘mental block’ in the epistemic reading prevents the action or conclusion. From functionalist approaches to language derive the view of structures as non-arbitrary, of meaning as context-dependent and non-atomic, of categories as non-discrete and of grammar in general as non-static and emergent. From there is also adopted the focus on explaining the communicative function of grammatical structure as well as the demarcation between form and function, which is similarly one of the central tenets of Construction Grammar. From the latter derives the view of constructions as the central unit of syntactic description and their decisive role in determining and assigning semantic-pragmatic function to their constituent parts.83 These issues pertain particularly to the role active sDm=f forms in Earlier Egyptian complementation. Indeed, in addition to analysing the grammar of these constructions, the present work seeks to contribute generally to the ongoing discussion of the ‘sDm=f theory’ and propagate a particular sort of ‘minimalist’ approach to formal issues.84 The expression of non-assertion in affirmative complement clauses is primarily based on the use of un-introduced forms of the sDm=f formation. As seen, in such instances the sDm=f of verbs with ult(imae) inf(irmae) roots appear both with and without gemination and some also with additional consonants. Thus e.g. the 3ae inf. verbs iri ‘do, make’ and hAi ‘descend’ are found written respectively as ir, irr, iry, irw, and hA, hAA, hAy and hAw. Doubling verbs such as the 2ae gem. mAA ‘see’ occur both with and without doubling of the last consonant. Anomalous roots display further peculiarities—e.g. iwi ‘go’ occurs both as iw and iwt—whereas immutable ones show no variations.
83
Croft 2001, particularly 4–5, 45–46. For the issue of form and the sDm.n=f, see 7.1 below. The past passive sDm=f will be treated as a independent form due to its combinability (at least in Classical Egyptian) with noun subjects only, which indicates that it no longer forms part of the suffixconjugation. 84
introduction
29
The most widely shared assumption is that, of the various sDm=f forms postulated, employed herein are the geminating ‘nominal sDm=f’ and one or several ‘prospective forms’. The latter do not geminate and appear variously with and without the endings -w and -y in the ult. inf. and caus(ative) classes, show both doubling and no doubling in doubling classes plus diagnostic spellings such as iwt in anomalous roots.85 Opinions diverge on how many ‘prospective’ forms there are as well as whether they exist in Earlier Egyptian throughout or whether some are restricted to the earliest stratum of the language. A widely shared assumption is that originally there were (at least) two ‘prospective’ sDm=f’s, which either largely merged by Middle Egyptian or one suppressed the other;86 total separation after this is suggested but seldom.87 There is further debate on the more general question concerning the extent to which particularly the assumed prospective form(s) represent full paradigms. According to the predominant view, formal distinctions such as the endings -w and -y observable in singular roots can be generalised, and ‘forms’ defined as paradigms of writings of all roots sharing the same syntactic position with the ‘diagnostic’ writings.88 However, this old hypothesis, which underpins most current divisions of sDm=f ‘forms’ in general, is suspicious. It seems fallacious to argue that, say, a writing such as hA=f of the ult. inf. root hAi represents some ‘prospective sDm=f’ only because it occurs e.g. in final clauses with future-modal meaning. This is particularly so if it is simultaneously maintained that hA=f is ‘also’ the writing of some other—in this case ‘circumstantial’—sDm=f form of the same verb hAi in gnomic main clauses after iw or in adjuncts with relative present tense. What can actually be observed in writing suggests that hA=f is the same form in each case, but its semantic-pragmatic and syntactic function is not. Of course, theoretically it is possible that as in Semitic generally, the inflection of the Early Egyptian verb was not based only on consonantal, 85 Depuydt 1993a provides a comprehensive summary of all the views on ‘prospective forms’ in Egyptological language studies. See also Schenkel 2000b on the endings -w and -y in Coffin Texts. 86 Allen 1984 § 266; Schenkel 1981; Loprieno 1991a, 210–17; 1995, 81; Grandet & Mathieu 1997, 206; Malaise & Winand 1999 §§ 445–46; Schenkel 2005, 199, among others. 87 See n.39 above. 88 This principle goes back to Erman 1889a, 11 n.2 and was embraced by ST (Polotsky 1969, 466–67; 1990, 769 and especially 1965, 9).
30
introduction
but also vocalic variation. Native speakers confronted with an unfamiliar text might have utilised syntactic and syntagmatic context to separate forms differing only in vowels—i.e. having recognised the function of some string of written consonants, proceeded to inflect it into a fully vocalised form. However, it is not known whether or to what extent this strategy was actually employed, and its application is in any case forever lost to modern Egyptologists, who have access to Earlier Egyptian only as it survives in the vowel-less writing.89 When all is said and done, the unpleasant reality is that ultimately grammars of this language describe merely the behaviour of the bare consonantal skeletons of words. The extent to which this corresponds to the living, spoken mode of communication is mostly unknown, although here one should be reminded of Junge’s admirably common-sense argument that if a feature is of any grammatical significance, it will not be invisible in writing.90 In what survives of Earlier Egyptian, then, form must equal written morphology, and, as a consequence, researchers are bound primarily to write grammar of function rather than of form. The two cannot be equated, as this will result in postulating ‘forms’ with
89 There are, of course, not a few reconstructed vocalisations of the alleged forms. The sole sufficiently extensive piece of evidence here derives from the Coptic t-causatives, (tako, tanxo etc.) which since Erman 1884 have been used as proof for a ‘subjunctive’ sDm=f in hieroglyphic Egyptian (see Elanskaya 1981, 80–81 for a research history on this topic). In Coptic, the Formenbildung of the t-causatives is regular, which may be taken as indicative of an Earlier Egyptian unitary paradigm of sDm=f in the combination rdi + sDm=f, from which the Coptic lexemes derive (see 3.2 below). In this paradigm the same roots occur written with and without the ending -y but not with the ending -w. From this it may be deduced that forms with -y and -w do not belong to the same formal paradigm and that the y-form may sometimes be written without the ending. In other complement environments, -y and -w-forms occur together along with forms with gemination or no ending, i.e. the paradigms are clearly mixed. Now, one may assume that the last-mentioned are still ‘defectively’ written ‘subjunctives’, as they might be after rdi, but there is no definite proof of this, nor any certainty that the -w cannot also be subject to such omission. If it can, beyond rdi it is impossible to say which form is in question. In any case, there is a deeper problem with all this reasoning. Even if it were the case that after rdi a single paradigm of sDm=f was used in Earlier Egyptian, from this it does not follow that in other complement environments the choice must be only between the ‘subjunctive’ and the -w-form whenever there is no gemination. That is, the supposed fact that after rdi only one kind of sDm=f is used does not indicate that elsewhere in complementation the number of non-geminating sDm=f’s of weak roots cannot be higher than two. This assumption is, once again, based on theory-internal expectations of inherent syntactic properties of the forms involved rather than on phonological or diachronic evidence (cf. e.g. Schenkel 2000b, 71). 90 Junge 1989, 16.
introduction
31
no morphological identity whatsoever. For example, the paradigm of the so-called ‘circumstantial sDm=f’ contains not a single writing that does not also serve some other function than adjunct use. When ‘forms’ become defined as nothing but function, one runs a serious risk of a categorical error. Also the proliferation of ‘identically written forms’, which is an inevitable outcome of such an approach, amounts to a flagrant aberration of Occam’s Razor—one of the most sensible principles in philosophy of science. An alternative approach to the question of form and function has been advocated by Allen, who views the occurrence of differently written sDm=f’s in a given syntactic position as variation of functional counterparts.91 This interpretation accepts that e.g. such writings as hAA=f of the ult. inf. root hAi and spr=f of the immutable root spr are not the same form, but may share the same function. In complement clauses, Allen distinguishes three inflected bare sDm=f forms: a ‘nominal aorist’ active sDm=f signalled by gemination in root-classes showing this property, a ‘prospective’ sDm=f with the ending -w, and a ‘subjunctive’, morphologically distinct only in case of the verbs iwi and ini, which are written iwt/int.92 Other sDm=f’s without these morphological features are considered as functional counterparts of these three types of forms. The division also touches upon modality: the ‘aorist’ describes a “simple occurrence of an action” whereas the ‘subjunctive’ “carries a sense of contingency or necessity that the prospective seems to lack”.93 The present work seeks to build on Allen’s original insight, but treats form as strictly isomorphic of semantic-pragmatic and syntactic function, which relate to use instead.94 The geminating and ‘prospective’ forms of the sDm=f are not understood to exist as paradigms of the Earlier Egyptian verb as a whole.95 What will, for reasons explained in time, be termed proximal irrealis represents a modal function of the sDm=f that is a particular sort of non-assertion and contrasts with another irrealis function of the sDm=f, termed distal. Verbs with immutable 91
Allen 1984 § 212; Allen’s object of research are the Pyramid Texts. Allen 1984 §§ 258–64. 93 Allen 1984 §§ 258, 267; cf. also § 364 and Allen 1982, 22 as well as Malaise and Winand 1999 §§ 593–94. Rather differently in Loprieno 1991a, 215–17; 1995, 81. 94 Comparable views have been expressed by Eyre (1994, 118–19). 95 The same holds also with the ‘circumstantial’ sDm=f, which ought to disappear altogether. 92
32
introduction
roots are interpreted as allocating both these functions to the one and only form of sDm=f in which they occur. Verbs of ult. inf. class possess a specific form for the proximal irrealis, namely the geminating hAA=f/ nTrr=f, but assign the distal function mostly to the non-geminating sDm=f. However, many roots in this as well as in the causative class also possess forms for the latter function: hAw=f, sAxw=f, iry/mry/Hsy=f etc., which do not occur in other functions96 and are not seen as ‘full writings’ of something else any more than hA=f, sAx=f, iri=f etc. are understood to represent ‘Defektivschreibung’. These distal irrealis forms are but rarely used and apparently under diachronic pressure of elimination and morphological collapse, with their function(s) increasingly assigned to their counterparts the non-geminating sDm=f.97 The doubled form of doubling roots such as mAA is associated with proximal function(s), whereas in the classical language the form mA=f of the verb mAA at least seems to have been established as a specific distal irrealis form after the demise of the independent ‘indicative’, which survives only in the bound negative n sDm=f.98 The apparent discrepancy in the form-function divide between ult. inf. (geminating = specific proximal, non-geminating = non-specific distal) and doubling roots (doubling = non-specific proximal, no doubling = specific distal) is in no way surprising. Doubling and gemination are distinct phenomena in that the former is a matter of lexical root morphology whereas the latter is a method of creating semantic-pragmatic oppositions by mutating the root. As a result, the distribution of modal functions in and between the doubling and ult. inf. classes is different. The anomalous roots show very disparate properties, which shows that they do not form any unitary morphological group. The roots rdi ‘give/cause’ and ini ‘bring’ possess separate proximal forms dd=f and inn=f as does the 96 In the Pyramid- and the Coffin Texts, sDm=f written thus occur with future passive sense, but these most likely represent the same form with no fixed voice as yet; cf. chapter 5 n.32 below for a similar situation with the negatival complement. 97 Rather similarly Schenkel 1975, 62; 1981, 517; Loprieno 1986a, 32 n.20, n.21, 38 n.39, end. 98 Although the paradigm of the ‘indicative’ in the different roots is largely identical to that of the -w/-y-less non-geminating sDm=f forms used in complementation, in the historical language the last remaining independent uses of the indicative can be interpreted as involving a separate ‘form’ that is only formed of transitive verbs, accepts only noun subjects and occurs only in absolute initial position; see Doret 1986, 24–27 and 6.3 below. The status of the form mAn=f (of the root mAA) found occasionally in the same environments as mA=f is obscure.
introduction
33
auxiliary wnn, (wnn=f )99 but in complementation rdi and wnn allocate distal functions to the forms rdi/di and wn.100 The root iwi has no separate proximal or distal forms but rather assigns these functions respectively to the forms iw=f 101and iwt=f, the latter of which appears to be not a sDm=f, but a sDmt=f form in a special use.102 The verb ini does the same with int=f, but it also uses the form in=f for distal. The verb ii appears to have a separate proximal form iy=f,103 but allocates distal functions to i(i)=f and iit=f, the latter of which is, just like iwt=f and int=f, a specialised use of the sDmt=f.104 The geminating sDm=f of ult. inf. roots and forms with the endings -y and -w are thus formally distinct moods; other (bare) forms of active sDm=f have, in complementation, irrealis function only. The former two may be grouped as modally marked non-indicative forms; the profile of the latter forms will be defined later.
99 There is no reason why wnn should be left out of consideration here or given some specific treatment just because it functions as an auxiliary. In this its role, wnn is specialised for expressing TAM—and little else—which in fact makes it an indispensable part of a discussion on modality. wnn can hardly be classified as an ordinary 2ae gem. root seeing that the form used after iw is wn, not wnn, which compares ill with e.g. iw(=f) mAA=f. However, wnn is clearly not a ‘geminating’ form either, insofar as this term is understood technically as referring to the reduplication of a medial consonant. Nevertheless, wnn=f is functionally equivalent to the ‘geminating’ form of ult. inf. roots, but in contrast to e.g. the long form of doubling roots, it has this one function in all instances and its writing does not conceal ‘frozen’ ‘prospectives’ (Allen 1982, 25) etc. See 9.1 for discussion. Similarly, wn is simply the ‘non-geminating’ form of this verb, not several identically written ‘converters’ nor an idiosyncratic spelling for wn.n=f (see 9 n.35). 100 The root rdi also has the rare forms rdiw/rdy/dy but these are not attested in complementation. The forms di and rdi cannot be divided functionally. Examples of both are found in most syntactic or semantic-pragmatic functions argued to be a domain of only one or the other, including future cleft sentences, (PT 990c: di=f; 1093b/ M: di=sn) after the verb rdi, (PT 583c/P2: rdi N) in the past negation n, (MMA 35.7.55, 11 n di=i), in final- and adjunct clauses as well as second tenses (cf. EAG § 459). 101 Polotsky (1965, 5) argues that iw=f should be seen as a ‘circumstantial’ form of this root whereas according to Gilula (1985) it should be analysed as the ‘emphatic’ (‘geminating’) form of this root. Seeing that iw=f occurs after ntt (see 2 n.159) but also demonstrably in second tenses, it seems that iw=f covers both these functions. There exists also the rare form iww=f; (Gilula ibid. 137 n.3) the form iwy=f is rare. 102 See 9.3 below for discussion of this point. 103 Polotsky (1965, 5) views iy=f as ‘circumstantial’, Gilula (1985) as ‘emphatic’. Except for Sin R 15, (ti sw iy=f “Now he was returning”) the uses of iy=f rather support Gilula’s view (see 2 n.88, 2 n.106, 4 n.30 and example (196) below). 104 However, a singular instance of a complement form iw of the root ii will be found in example (168) below.
34
introduction
The Earlier Egyptian formal-functional system of expressing modality with un-introduced sDm=f (plus the three sDmt=f forms iwt=f/int=f/iit=f) complement clauses may thus be tabulated as follows (x = non-existent): proximal Root Class Immutable Mutable Doubling Ult. Inf. Anomalous
rdi iwi wnn ini ii
distal
Specific Form x
Counterpart sDm=f
Specific Form x
Counterpart sDm=f
x
mAA=f x x
mA=f hAw/hAy=f x x x x
x
hAA=f dd=f x wnn=f inn=f iy=f
iw=f x x x
iw=f
hA=f rdi/di=f iwt=f wn=f in/int=f i(i)/iit=f
In sum, some roots possess special forms of sDm=f for the proximal and distal irrealis functions, some do not, and instead employ counterpart forms that may also function differently, depending on where they are used. By necessity, the discussion in the present work will mostly focus on mutable roots, which show clear formal differentiation between the distal and proximal irrealis functions. It should be stressed that this organisation is based on the evidence from complementation only. Whether or not the analysis of the forms grouped as marked or unmarked for some modal value is applicable generally to all their uses is another issue. However, as will be seen in the course of the present study and particularly in chapter 9 below, there are good reasons to believe that this is indeed the case. Yet, it goes without saying that verifying this requires an extensive study of the other uses of all the forms listed above. Finally, the present work has a strongly philological and example-based focus. Particularly when dealing with dead languages, elusive modal nuances can only be extracted from a large amount of data rather
introduction
35
than selected examples. The sources utilised span the period between and inclusive of the late V dynasty and the reign of Amenhotep III. Pre-XVIII dynasty examples derive from texts representing all textual genres and registers, whereas later material is largely restricted to royal and private monumental inscriptions.105 Such a wide diachronic scope is required by the relatively infrequent occurrence of complement clauses with revealing morphology and is perhaps less than ideal. As a precautionary measure, therefore, and notwithstanding the unavoidable degree of arbitrariness associated, the Pyramid Texts have been excluded from consideration.106 Yet, the realis-irrealis divide appears to be discernible in these texts as well and Pyramid Texts passages will also be noted at several occasions for comparison against the data studied.107 On the whole, the modal organisation discussed in the present work remains intact throughout the material studied: signs of its collapse appear late and are at first limited in number.108 In view of the many quite thoroughgoing diachronic developments elsewhere in the language between late Old- and early New Kingdom, the modal system of complementation represents one of the most persistent characteristics of the Ancient Egyptian language.
105
Exceptions are the early New Kingdom medical papyri pEdwin Smith, pEbers and pHearst, as well as the Book of the Dead. Examples from the latter derive only from XVIII dynasty manuscripts, and specific preference is given to the versions in the fine papyri of Nu (pBM EA 10477) and Nebseni, (pBM EA 9900) particularly when the latter differ from the version(s) in Naville’s edition. New Kingdom royal Unterweltsbücher are not utilised. 106 Some Pyramid Texts spells are clearly contemporary Old Egyptian, some use language that seems exceedingly archaic. Although the Coffin Texts incorporate excerpts from the Pyramid Texts, they are nevertheless included in the present corpus, seeing that independently their idiom seems closer to ‘standard’ Middle Egyptian than the language of the Pyramid Texts (cf. Polotsky 1969, 466). 107 For a brief discussion of the system outlined and the PT, see 10.2. 108 See 7.3 and 10.2.
36
introduction
introduction to part one
37
PART ONE
MODALITY IN AFFIRMATIVE COMPLEMENT CLAUSES AFTER GOVERNING VERBS
38
chapter one
introduction to part one
39
CHAPTER ONE
INTRODUCTION TO PART ONE Of the two types of complementation after governing verbs, object complement clauses are most widely attested in all languages. Also in Earlier Egyptian, where survival and evenness of the evidence are always an issue, a wide variety of matrix verbs with object complements other than the bare infinitive or the sDm=f of immutable roots abound in the textual corpus. Regardless of the particular language in question, in object complements one notices a certain dichotomy between often relatively clear syntax and highly complex semantic-pragmatics. It has been noted that the degree to which the situations described in the main and subordinate clauses are integrated is reflected iconically in their grammar. The complement displays signs of increasing morphosyntactic dependence from the main clause the more the situations they describe are semantically assimilated.1 Yet, the more generally ‘asymmetric’ hypotactic relationship of the clauses—i.e. one between a dependent subordinate- and a dominant main clause—remains the same.2 However, there also exists a related strategy of clausal combination known as direct speech, (oratio recta) which belongs to the same scale of inter-clausal cohesion between transitive verbs and following clauses as complementation, but which is best considered apart from the latter.3 In direct speech, or ‘quoting’, the relationship between the combined clauses is ‘symmetric’ or paratactic: they have equal status and maintain their grammatical independence.4 For instance, in examples (15) and (16) below the direct quotes after the verbs say and answer are self-contained sentences:
1 See e.g. Givón 1980; 1994, 278–79; 2001 vol. 2, chapter 12; Hopper & Traugott 1993, 171; Cristofaro 2003, 117–22. For Egyptian, see Uljas 2003. 2 Halliday 1985, 218; this functional formulation does not presuppose that the complement clause be morphologically marked for subordination or even treated as syntactically embedded, even if this is usually the case (but see Cristofaro 2003, 96–98 for synchronic and diachronic exceptions). 3 Cf. Li 1986, 36–37; Givón 2001 vol. 1, 155–56; Cristofaro 2003, 47, 99 n.4. 4 Halliday 1985, 250–52.
40
chapter one
(15)
The king has been told that a magician has arrived:
Dd.in Hm=f is in n=i sw Then his majesty said: “Go, bring him to me”. (pWestcar 8, 8–9)
(16)
The king has just implied that Sinuhe need not fear, then:
wSb.n=i st m wSb snDw ptr Ddt n=i nb=i I answered to this with an answer of a frightened man: (Sin B 260–61) “What is my lord saying to me?” ‘Direct speech’ need not involve a verb of locution. Humans use language also for thinking, and just as it is possible to quote words spoken, it is also perfectly feasible to quote thoughts: a sentence such as “I thought: ‘I’ll quickly see if he’s in’” is a near-paraphrase for “I said to myself: ‘…’”5 This is possible also in Egyptian: (17)
The sailor is about to encounter the giant snake for the first time:
aHa.n sDm.n=i xrw qri ib.kw wAw pw n wAD-wr Then I heard a tumultuous noise, but only thought: “It is (Sh.S. 56–59) (just) a wave of the sea”. Here the speaker’s thoughts are given a wording, and as in locution proper, this is signalled by the grammatically independent nominal sentence after ‘think’.6 A quote never represents a semantic argument of the governing predicate, but purports to be an authentic reproduction of the words (sounds) or thoughts quoted.7 Also when quoting others, as e.g. in (15) above, the real speaker (here the narrator) lends his voice to the original speaker, (the king) but the quote displays no signs
5
Halliday 1985, 255–56. Thus wAw pw is not an instance of a nominal sentence direct complement of ib, as analysed e.g. in GEG § 186.3 n.7 and Callender 1975, 74. Quoting gives here a more fragmented, dynamic shape to the situation-description, which is well in accord with the texture of the narrative: I only thought: “…”—i.e. in fleeting and without great concern at first. The text does not lose any of its elegance; indeed, the more one is able to recognise different rhetoric devices employed by the ancient authors, the lesser the danger of underestimating their expressive skills. However, in Egyptian parataxis is often evoked simply to explain away complementation violating the ST principles of the use of ‘nominal forms’. See 2 n.167 below. 7 See Haiman 1985, 222–28 for a fine discussion on this issue. 6
introduction to part one
41
of “removal from the reality of the original utterance”.8 In object complementation, by contrast, the deictic elements of the subordinate clause are modified to conform to the perspective of the real speaker— his identity and spatio-temporal locus—regardless of whether or not these are shared with the original speaker, ‘thinker’ etc. For example, in a sentence such as: (xxiii)
Jack said that Jill was going there
the complement stands for the presumed original “Jill is coming here” uttered by the subject ‘Jack’, who is not the real, but the original speaker.9 This is an exemplar of prototypical reported ‘indirect speech’ (oratio obliqua) where a verbum dicendi is followed by a complement clause that does not simply reproduce the original words. For the study of complement modality, direct speech is of limited interest, seeing that, as in example (15) with the imperative, it expresses also these features in exactly the same manner as independent main clauses.10 In object complementation proper, however, assigning a modal status to the subordinate clause is far more complex a matter. Here the semantic-pragmatic dependency of the complement from its governing predicate, the often extreme sensitivity of this to co(n)textual factors, and the role of the real speaker play a decisive role. The effects of these factors must now be briefly illustrated. Object complement modality is inherently linked to the type of semantic-pragmatic dependency of the complement from the main verb. A complement may be ‘meaning-dependent’ of a governing verb with respect to its11 time propositional attitude information value 8
Fleischman 1995, 532; see also Coulmas 1986, 2; Li 1986, 38. The changes in temporal deixis in particular have often been assigned to a set of mechanical rules of ‘sequence of tenses’ (systematised by Comrie 1985, 104–17). But see Palmer 1986, 166–67; Langacker 1991, 253–60 and 9.2 below for the role of tensechanges in modality. 10 But see Palmer 2000, 197–98 and Martin 1998, 205 for some exceptions. 11 Adapted from Noonan 1985, 92. 9
42
chapter one
Complement non-assertion typically results when the time of the subordinate clause is determined by the main verb, the situation it describes is made subject to negative propositional attitude, or presupposed—all characteristic motivations for the use of irrealis/ subjunctive.12 However, this is not merely a matter of main verb taxonomy. In most languages finite complements whose time-reference is predetermined as unrealised vis-à-vis the governing verb are treated as irrealis, and in no language are they asserted.13 Such verbs are e.g. those of commanding, requesting, intending, desiring, preventing etc. which are notionally non-assertive in that their complements are never presented as assertions.14 This is because for the original speaker the complement situation is unrealised and impossible to predict for certain. However, in reports there is potential for the real speaker’s perspective to be taken into account as well.15 In present and future reports, this does not substantially differ from that of the original speaker. For example, in “They order that he go”, it is quite obvious from the sentence that no real situation is referred to in the subordinate clause and whether or not the ordering is or will be successful cannot be known. On the other hand, when it is said “They ordered that he go”, it may actually be the case that at the time of speaking—i.e. in the real rather than the original speech context—the complement situation has been realised. The sentence may e.g. answer the question “Why did Jack go away?” Most languages, including English, Spanish and French, do not take the real speaker’s perspective into account in the modal marking of the object complement here at all. Instead, they assign the modality of complements of notionally non-assertive verbs such as order from the perspective of the original speaker and speech context. Thus, in these languages irrealis is used in the complements of non-assertive verbs also when the former represent past reports:16 (xxiv)
12
Spanish:
Le mandaron que se callara (SUB) “They ordered that he keep quiet”
See 0.1.2 above. In e.g. Bulgarian, this is argued to be the only factor conditioning complement modality (Noonan 1985, 93–94). 14 Cf. Givón 1994, 272; 2001 vol. 1, 309. 15 See Cristofaro 2003, 111–17 for a detailed discussion. 16 Data from Givón 1994, 281 and Judge & Healey 1983, 138 respectively. 13
introduction to part one (xxv)
French:
43
J’ai ordonné qu’il parte (SUB) “I ordered that he go”
However, as can be seen from these examples, the deictic shifts take place as normal. This shows that the overall perspective to the situation is still that of the real speaker, and some languages pay further attention to this vantage point. For example, in Bemba, a Bantu language, in past reported orders different complement patterns are used to indicate whether the situation ordered took place at some point between the time of ordering and the time of speaking:17 (xxvi)
Jack a-à-koonkomeshya Jill a-à-boombele (FIN) “Jack ordered Jill to work” (and she did) Jack a-à-koonkomeshya Jill a-boombe (SUB) “Jack ordered Jill to work” (but maybe she didn’t)
The different complement construals after the same main verb indicate respectively that the ‘working’ took place or that this was not necessarily so. This information is broadly attitudinal seeing that it reflects the subjective relevance attached by the speaker to whether or not the situation was realised. The ‘realisation not indicated’-alternative in (xxvi) is irrealis just as in the more familiar languages. The realised situation is not, but it is not realis either.18 The differentiation is, accordingly, part of the modal system. It does not pertain to the subject of the governing verb, but is added by the real speaker and represents a clear ‘intrusion’ of the latter to the complement situation frame. With verbs that express propositional attitude towards the complement situation/information, or content of communication, the role of the co(n)text and of the real versus original speaker is highly significant across languages. Many such verbs are notionally assertive in that they describe committed attitude and their complements have independent time-reference.19 Verbs of locution, knowledge and perception are obvious representatives of this class. What is said is of course potentially asserted, 17
Adapted from Givòn 1972, 150–51. In other words, the generalisation that the complements of notionally nonassertive verbs are never asserted holds. But see 3.1 for an alleged exception. 19 Hooper 1975, 95. 18
44
chapter one
and knowledge of something equals strongest attainable certainty and commitment to facilitate assertion. Similarly, vision is most objective kind of support for certainty concerning situations in the world, followed closely by auditory evidence. 20 In contrast, some verbs such as doubt describe negative attitude. Expressions of positive and negative propositional attitude are typically associated with asserted and nonasserted complements respectively, but here again various subjective and contextual factors make their effects felt. Although complements of notionally assertive verbs may be assertions, they do not always assert. For example, verbs of belief express commitment and are generally followed by the indicative in Spanish, but the subjunctive may appear if the matter ‘believed’ is still viewed somewhat tentatively:21 (xxvii)
Sospecho que es (IND)/sea (SUB) mentira “I suspect it is/might be a lie”
Similarly, the verb think is followed by the indicative in Spanish and many other languages if it is affirmative, but by the subjunctive, if negated:22 (xxviii) Creo que habla (IND) inglés “I think she speaks English” No creo que hable (SUB) inglés “I don’t think she speaks English”
20
Cf. Sweetser 1990, 37–40; Uljas 2003, 393. Here a note should be made on the issue of ‘factivity’ (Kiparsky & Kiparsky 1971). This term refers to verbs such as ‘see’ and ‘know’ that render their complements ‘necessarily true’. Factivity has been seen as synonymous with presupposition, but this assumption entails confusion between ‘semantic’ and ‘pragmatic’ presupposition (see introduction n.47 above as well as Hooper & Terrell 1974, 485; Hooper 1975, 116–17 and Levinson 1983, 179–81). ‘Factive’ verbs presuppose the logical truth of their complements, but do not in any sense indicate that the information therein is assumed as shared by the speaker and hearer. Nothing prevents from asserting something known and seen, as long as the speaker does not take it for granted that this is already ‘old news’ to the audience. However, there are nevertheless certain instances in which factivity may preclude the latter; see 2.2.2 below for an example of this in Earlier Egyptian. 21 Butt & Benjamin 2000, 258. 22 See Butt & Benjamin 2000, 241; Hooper & Terrell 1974, 486; Hooper 1975, 121 n.21. Also e.g. French, Italian, Portuguese, Russian and Polish behave similarly.
introduction to part one
45
This is because ‘not-think’ is tantamount to doubt, which cancels the assertion in the complement.23 As in main clauses, complement assertion and non-assertion thus often depend of the semantics of the overall co(n)text and of whole sentences rather than individual verbs. The speaker attitude affects the asserted/non-asserted alternation of the complements of ‘assertive’ verbs also when the speaker and the subject of the main clause are, unlike above, not the same. Such sentences potentially contain two assertions; verbs of speaking are the most obvious examples in this respect: (xxix)
Oedipus said that his mother was beautiful → Assertion 1: Oedipus said X Assertion 2: his mother was beautiful
The assertion in the first clause is assigned to the real speaker, whereas the complement represents something said by the referentially different main clause subject. But the deixis in the latter shows that it is anything but faithful to the ‘original’. The real speaker is free to modify and even add information to the complement not conveyed by the original speaker. In (xxix) it may be that Oedipus actually said e.g. “my wife is beautiful”, and it is the real speaker who provides the information that the wife of the original speaker is also his mother.24 In that case the complement represents even less what the original speaker said; it provides a report of this, but it is certainly not his assertion. The real speaker’s ability, willingness etc. to assert or non-assert the complement, and so comment on its content, is decisive to modal status. This is evident e.g. in German reported speech:25 (xxx)
Er sagt, er müsse (SUB) nach Hause/dass er nach Hause muss (IND)
“He says he must go home” The subjunctive in the complement indicates that the speaker does
23 Conversely, the verb doubt is notionally an explicit signal of inability to assert, and e.g. in Portuguese takes subjunctive complements when affirmative, but indicative if negated (Willis 1974, 294). 24 Coulmas 1986, 3–4; this is known as the de dicto/de re -ambiguity. 25 Palmer 2000, 114.
46
chapter one
not necessarily believe or accept the reported proposition, whereas the indicative shows that he does.26 Variation of this sort has little to do directly with the governing verb. The speaker’s role is also revealed in instances where the ‘assertivity’ of say is manipulated, e.g. again through negation. For example, the following German and English utterances are more typical reports of questionings in judiciary systems where suspects are presumed innocent unless proven otherwise: (xxxi)
Er gibt zu, in der Nähe gewesen zu sein, aber er sagt nicht, er wäre (SUB) der Einbrecher “He admits to having been nearby, but he does not say that he would be the burglar”
The non-indicative subjunctive and if-clauses communicate that the speakers are not aware of the true state of affairs described in the complement and cannot assert it. But e.g. in communist East Germany, an interrogator might have reported a failure to extract a confession from an accused whose guilt had already been decided beforehand, as follows: (xxxii)
Er gibt zu, Feindsender empfangen zu haben, aber er sagt nicht, dass er ein Klassenfeind ist (IND) “He does admit to having watched western TV-programs, but he does not say that he is a class-enemy”
Here it is explicitly told that the subject never asserted the complement, but the real speaker is committed to ‘his’ being a ‘class-enemy’ and frames it as an assertion.27 Verbs of cognition and perception abide
26 Cf. Coulmas 1986, 16; Wierzbicka 1988, 150; Palmer 2000, 198–99. The same principle lies behind the main clause use of subjunctive or conditional in French and German ‘journalistic style’ (see 0.1.2 above) where evading responsibility for the ‘truth’ of reported statements is of fundamental importance for reasons of libel. 27 This shows clearly that it is incorrect to argue that the complement assertion is simply assigned to the subject and that the “S(peaker) is not himself asserting the existence of T” (the situation; Langacker 1991, 255; cf. Cristofaro 2003, 36–37). The indicative pertains solely to the real speaker and its presence e.g. here is inexplicable by reference to the original speaker only. Cristofaro (ibid, 37 n.7) argues that the complement clause may be ‘dominant’ in the discourse context, but is never asserted. However, this leaves the variation between indicative and subjunctive mood in coding of reported speech completely unexplained. Cf. also Butt & Benjamin 2000, 254 for subjunctive/
introduction to part one
47
to the same principles based on the knowledge and acceptance of the real speaker. In reported speech proper a reference to the complement as a ‘subject assertion’ may be appropriate when this participant is positively reported to have carried out a speech act. However, otherwise reference to ‘assertion’ is meaningful only with respect to the real speaker. If the complement e.g. merely informs what someone knows or perceives, it need not report anything beyond what the sentence subject can be imagined as having said. The real speaker provides the wording, and again is responsible of the complement modality. For example, in normal circumstances the following English and German sentences are ungrammatical: (xxxiii) ?I don’t see/know that he is here ?Ich sehe/weiss nicht, dass er hier ist (IND) If the speaker says that he does not know something, he cannot phrase the latter as an assertion, which signals that he does know it.28 Yet the English and Spanish sentences below are perfectly acceptable:29 (xxxiv) He does not see/know that/whether he is here Yo no sabía que él estuviera (IND)/estuviese (SUB) ahí “I didn’t know that/whether (s)he was there” In the first English sentence it is innocuous what the subject of the sentence sees or knows. The that-clause shows that the real speaker is in a position to assert the complement because he sees/knows that the situation it describes obtains.30 Whether indicates that this is not so, or indicative expressing the same contrast in Spanish. E.g. Swedish uses conditional vs. indicative and Finnish interrogative vs. indicative to identical effect. 28 Hooper 1975, 119; Wierzbicka 1988, 146. Cf. also Kiparsky & Kiparsky 1971, 349, who, however, explain this phenomenon by reference to ‘factivity’. 29 Cf. Butt & Benjamin 2000, 254. 30 However, unlike with speaking, the real speaker is (almost) bound to belief if reporting someone else’s visual confirmation or knowledge of something; hence the oddity of ?“Jack knows/sees that Jill is there but I don’t believe it”. Nevertheless, “Jack hears that Jill is there, but I don’t believe it” is more acceptable because auditory evidence is less reliable and easier for the real speaker to overrule. This scalar organisation shows that the issue is not that of ‘factivity’. Seeing and hearing are equally ‘factive’, but visual confirmation is more ‘believable’; cf. the different weight assigned to eye-witness testimonies and “hearsay” in court-proceedings, or the ranking of visual evidence higher than auditory in evidential modal systems (see Willett 1988).
48
chapter one
that the speaker does not wish to confirm this. The speaker may withhold the assertion also if it is not his illocutionary intention to provide his audience with the information of whether ‘he’ actually is/was ‘there’, regardless of whether or not he is presently aware of this. In the past tense Spanish sentences in (xxxiv), indicative is acceptable provided the speaker is now in a position to assert—i.e. capable and willing in the current speech-context to commit himself to the ‘reality’ of ‘him’ having been ‘there’. The non-indicative versions again signal a contrary situation: either the speaker is still unaware of the ‘real’ state of affairs, or simply deems it unnecessary to inform/confirm this. In other words, the assertion/non-assertion variation reflects the character of the overall discourse. Finally, there are verbs that render their complements ‘informatively’ dependent by indicating that the latter are presupposed background; a verb of this sort is e.g. regret.31 Not all languages code background information as irrealis. Thus, as noted, e.g. in Russian complements of presupposing verbs such as regret are indicative.32 Further, the number of genuinely presupposing verbs taking clausal object complements is small, and none are attested in Earlier Egyptian. However, speakers may also evaluate when a complement is new information and when not, as in (xxxiv) above. There the use of whether/subjunctive may have been motivated by a perceived lack of relevance of the information about ‘somebody’s being there’. That is, the utterance “He does not see whether he is there” may reflect the speaker’s assumption that the audience is already aware of the true state of affairs or that the complement information is immaterial in the current speech context, perhaps because the most salient issue is that the subject does not see something rather than what he might not see in the present context. Accordingly, whether or not an object complement is an assertion or non-assertion depends on a range of issues including the notional properties of the main verb, the discourse context and the perspective of the real speaker. In every instance the use of realis signals the complement to be believed, accepted and seen as communicatively 31 This is seen e.g. from the survival of the presupposition under negation: I regret/ do not regret that I left » I left. Presupposition defeasibility, as in “Jack won’t have to regret doing a PhD because he failed it” does not affect the choice of mood. 32 Noonan 1985, 99.
introduction to part one
49
relevant by the real speaker to whom it is assertable. Irrealis could be said to signal the speaker’s attitude and ‘epistemic perspective’ to be “I cannot/don’t want to/don’t need to affirm this”.33 In Earlier Egyptian affirmative object complements after governing predicates, ntt/wnt-introduced realis- and bare sDm=f irrealis clauses are used under the same conditions. When asserting the complement is ruled out, or the speaker opts to indicate that for whatever reason he is not fully committed to the proposition, does not accept it or intend to present it as optimally relevant information, ntt/wnt are not used. Rather, a bare sDm=f, which in certain verb-classes can be formally divided into two types of irrealis, appears instead.34 Bare sDm=f irrealis-, and ntt/wnt-introduced realis object complement clauses indicate what is the real speaker’s illocutionary intention and ‘position’ in relation to the subordinate proposition. These subjective stances are not grammatical or lexically determined constants in Earlier Egyptian any more than in other languages.
33 34
low.
Cf. Wierzbicka 1988, 159. See 0.2 above. One single lexical verb stands outside this organisation; see 2.4 be-
50
chapter two
CHAPTER TWO
AFFIRMATIVE OBJECT COMPLEMENTATION AFTER NOTIONALLY ASSERTIVE VERBS For a modal analysis of Earlier Egyptian affirmative object complement clauses and complementation more generally, the notionally assertive verbs of locution, perception and cognition are of decisive importance. Of all governing predicates, these verbs occur with the widest range of object-clause types, including ones introduced by ntt/wnt as well as bare active sDm=f forms of all sorts.1 Combining with any one of these construals is not a prerogative of the said verbs. Yet, examples of the latter are sufficiently common to allow the modal character and parameters of use to be established completely for clauses with ntt/wnt, and initially for those with bare geminating and non-geminating sDm=f and their functional counterparts. Further, with these verbs a general differentiating between ntt/wnt-introduced assertions and un-introduced non-assertions is possible even when the latter are represented by the morphologically unrevealing sDm=f of immutable roots, whose value as evidence with verbs not attested with ntt/wnt is limited. 2.1 Locution: The Verb Dd ‘Say’ and Indirect Speech 2.1.1 Core According to Gardiner, “the highly developed indirect speech found in Latin, where all the pronouns after ‘he said’ or the like are reduced to 3rd pers., hardly exists in Egyptian”.2 This is largely true for Middle Egyptian, but in Old Egyptian, complement clauses after Dd are characteristically introduced by the element wnt and assume the guise of ‘indirect speech’ of traditional Western grammars. Nevertheless, Egyptian evidently gradually gave up this sort of hypotactic linkage of 1 There are also examples of the bare sDm.n=f after these verbs; see 7.1 below for discussion. 2 GEG § 224.
affirmative object complementation
51
clauses and receded into a paratactic system of quoted direct speech after ‘say’. Indirect speech introduced by specific elements largely disappeared by the advent of the Middle Kingdom, occurring only in consciously archaising sources and the Coffin Texts, but not e.g. in the classical literature of the era, where it is replaced by direct quotes. However, as will be seen, the method of expressing indirect speech without an introducing element was not subjected to similar diachronic elimination. Yet, indirect speech complements of this variety differ from those ushered in by additional elements not only in construal, but also in pragmatic function. Proper ‘indirect speech’ was retained in the Earlier Egyptian grammatical oeuvre for a specific modal purpose only. In Old Egyptian wnt alone is used after Dd in the elite tombautobiographies and secular texts generally, ntt being reserved for other verbs attested with introduced object complements. This raises a question of the possible differences between these two elements.3 Notably, wnt appears after verbs other than Dd already in the Old Kingdom and, conversely, examples of Dd + ntt are attested in the PT.4 In the CT and later, ntt seems wholly interchangeable with wnt regardless of the governing predicate. Both ntt and wnt share the same negative equivalent iwt, and although there are no Old Egyptian examples of this element after verbs other than Dd, in the Coffin Texts iwt occurs both after Dd and predicates earlier more closely associated with ntt.5 This progressively increasing overlap of the syntagmatic properties of ntt and wnt from early on is also paralleled by their closely similar etymology.6 wnt appears to be originally a perfective feminine/neuter singular participle from the verb wnn, ‘exist’, whereas ntt is clearly a similarly feminine/neuter form of the relative adjective nty ‘what (is)’.7 Nevertheless, it is first and foremost their identical pragmatic function in complementation that suggests that ntt and wnt represent two near-synonymous elements with essentially the same role in the grammatical system whose differentiation appears to have been slight throughout Earlier Egyptian. 3
See Polotsky 1969, 480–81 and n.146 below. See e.g. PT 1102a; Nt 40–41. 5 See 5.2 below for examples and discussion. 6 For further comments on this issue, see 2.3 below. 7 See Gardiner 1920, 53; GEG §§ 233, 237; LGEC § 701; EAG § 1019; Allen 1986b, 25. Cf. also 10.3 below. 4
52
chapter two
Of all the verba dicendi of Earlier Egyptian, Dd is the only one certainly attested with ntt/wnt and indirect speech.8 Perhaps as a reflection of its diachronic and typological restrictions, the grammatical phenomenon of indirect speech in Earlier Egyptian has received relatively little special attention among Egyptologists.9 Nonetheless, indirect speech with dD + ntt/wnt is rather widely encountered; the following example could be said to represent a locus classicus: (18)
The young king Pepy I writes to his emissary concerning previous correspondence:
Dd.n=k [r] mDAt=k tn wnt in.n=k inw nb aA nfr According to this letter of yours, you have said that you have brought all (sorts of) fine and abundant tribute. (Urk I 128, 10–11)
Here the subject pronoun in the clause after wnt is altered from “I have (or “yours truly has”) brought”, what the recipient probably said ‘originally’, to conform to the real speaker’s perspective. Syntactically, wnt marks the boundary of the subordinate clause and in translation corresponds closely e.g. to English that, German dass, French que, Latin quod or at(t) of Scandinavian languages. However, this simple analysis has not been universally accepted. For example, according to Loprieno, wnt and the main verb here actually form a single ‘thematic N(ominal) P(hrase)’, with “an indefinite object (of the main verb—SU) ‘contained’ in wn.t” and followed by a ‘rhematic NP’ as a “nominalization of a complex adverbial sentence or of a jw-sentence by means of wn” (from wnt):10
8 The verbs smi ‘report’, wHm ‘repeat, relate’ mdw ‘speak’, wSb ‘reply’, Sni ‘curse’, Sdi ‘recite’ and the defective verbs i and xr are attested with direct speech only; cf. Jenni 2002, 239 n.3. 9 A partial exception to this is a study of Old Egyptian ‘quotations’ by Goedicke (1955), where it is argued that in wnt-clauses the object is actually wnt ‘what is real’, followed by an appositional proposition. This interpretation fails to take into account the thoroughly grammaticalised status of wnt; (and ntt) see 2.3 below and cf. Doret 1986, 14 n.13. Significant efforts have, in contrast, been made in study of indirect speech in Late Egyptian (see Peust 1996, with a historical summary on pp. 41–48; cf. also Junge 2001, 220–01; Peust 2005) and in Coptic (e.g. Quecke 1990). Kammerzell & Peust 2002 provides a concise diachronic survey of indirect speech from Old to Late Egyptian. 10 Loprieno 1988a, 69–70. Loprieno’s analysis shares similarities with those by Goedicke 1955, Gunn 1924, 176–77 and Lefebvre (LGEC § 701); see 2.3 below.
affirmative object complementation NP theme
S[[Dd.n=k st r mDAt=k tn]
53
NP rheme
S[wn[iw in.n=k inw nb]]
That is, the entire complex constitutes one large ‘balanced sentence’construction. Yet, there are problems with this analysis. It is not the case that wnt or any part thereof is the ‘object’ of Dd—this is clearly the content of the entire clause, which the element introduces.11 Thus, example (18) is rather an archetypal instance of reported indirect speech. Similar examples with wnt from Old Egyptian and later with ntt are:12 (19)
The king refers to his addressee’s ground plan for a ceremonial court:
sk Tw Dd=k xr Hm(=i) wnt iri.n=k sw r [Aw] mH 1000 [wsx] mH 440[...] xft wDDt n=k m stp-sA Now, you tell my majesty that you have made it to (be) 1000 cubits [in length] and 440 [+x] cubits in [width] according to (Urk I 63, 2–3) what you were ordained in the court. (20)
The king recalls what his addressee said concerning a buildingproject:
[sk Tw Dd=k]13 xr Hm(=i) wnt=k r irt S xft Ddd[t] m stp-sA [...] m xmt=k sk Tw Dd=k xr Hm(=i) wnt=k r [...] [You have said] to my majesty that you are going to construct a garden (?) in accordance with what was said in the court […] in your absence, but you tell my majesty that (Urk I 62, 1–3) you will […]
11 Loprieno’s analysis is prompted by problems encountered by Doret (1986, 81) with this particular example in explaining how the complement clause, which he assumes to be a vedette, can function as such after an ‘emphatic’ Dd.n=k. If there is anything ‘emphasised’ in the sentence, this is surely r mDAt=k tn. The king is not stressing what his addressee has said, but the fact that the latter has put his claims in writing delivered to his majesty (the translation of (18) above follows this interpretation). 12 For the reconstruction of examples (19) and (20), see Brovarski 2001, text figure 2. 13 Edel’s (EAG § 1025) restoration is undoubtedly valid, seeing what follows.
54 (21)
chapter two The young king Pepy I writes to his emissary concerning previous correspondence:
Dd.n=k r mDAt=k tn wnt in.n=k dng ibAw nTr m tA Axtw According to this letter of yours, you have said that you have brought a pygmy of divine dances from the land of the Horizon(Urk I 128, 14–16) dwellers. (22)
The god Orion says to the deceased concerning certain divine beings:
ii.n=k is Ax.ti apr.t(i) rx.n=k rnw=sn Dd.n=k n=sn ntt am.n=k dSrt Verily you have come transfigured and equipped, having learnt their names and told them that you have swallowed the Red (CT V 397m–o) Crown. (23)
The deceased asks a divinity to say to Osiris:
sDm=k mdw nTr pw StA Hr xnty s(my)t iAbtt Ddw wnt=f Hr Ts=i May you hear the word of this god whose face is hidden, one foremost of the eastern desert, who says that he is knitting me (CT VI 328f–g) together. (24)
The author tells about some information conveyed to him by a third party:
[...] Dd.n=f n(=i) wnt rdi.n=k sT.t(i) mm HqAt 10 m nxn He said to me that you have caused ten heqats of emmer seed (pTurin 54002, 8–9) to be sown in Hieraconpolis. In all these instances the original and the real speaker are different individuals. Only one example of reported speech with co-referential main clause subject and speaker is attested, in an explicitly ‘actionexplaining’ context: (25)
King Intef II sings praise to goddess Hathor:
dy(=i) rx=s Dd(=i) r-gs=s ntt w(i) Ha.kw m mAA=s I let her know, I say by her side that I am rejoicing at seeing (MMA 13.182.3, vertical 3–4) her! However, just as the shifts in pronouns show the real speaker’s orientation to be that employed in the deictic profiling of the complement
affirmative object complementation
55
situation-description, it is also the attitudes and opinions of the latter that determine its modal character. It is relatively obvious that the primary locus of information and the main assertion in the sentences above is actually the complement rather than the main clause.14 In most instances, the subject of the sentence is also the addressee, and it seems unlikely that the real speaker’s illocutionary intention is to particularly ‘inform’ the addressee that the latter has said something.15 That is, the discourse-function of these sentences is rather to focus on what was said than that it was said—perhaps in view of further commenting upon them, as is frequently the case in letters whence most of the above instances derive. However, in this capacity they represent expressions of the real speaker’s own degree of commitment to the reported statements and of his high regard of their informative value, i.e. his/her assertions. This is obvious in instances where the original speaker cannot have asserted anything, as e.g. when the governing verb refers to futurity: (26)
Coffin Texts Spell 228 begins with a declaration by the deceased:
i wr aq Dd n iab sS iry-aA n wsir ntt wi ii.kw wr.ki Ax.ki wsr.ki O Great Enterer; tell him who unites writings, the doorkeeper of Osiris, that I have come, being great, transfigured and (CT III 268/69a–270/71b) mighty.16 (27)
King Sebekhotep I says to a divine being:
i.Dd=k n Hr wnt wi Ha.kw m mAa-xrw=f m-bAH psDt17 May you tell Horus that I was delighted of his vindication before (Louvre C10, x + 9–11) the Ennead. 14 See Hooper 1975, 94–95. In fact, if some or all of the initial un-introduced sDm. n=f’s in the above examples are to be interpreted as second tenses, it may be that the governing clauses in fact constitute non-assertions. This depends ultimately on the modal analysis of second tenses; see 10.3 for some preliminary discussion. 15 Examples such as Urk I 104, 12 Dd.t(i) wnt btkw nxt m xAstw pn “It was told that there were strong troublemakers among these hill-dwellers” could be analysed as being construed with an impersonal .t(i) subject and an object wnt-clause. However, particularly in such early examples, the complement should perhaps be rather seen as the subject of a synthetic passive Dd.t(i). See 4.1 below for discussion of this problem. 16 The variant B5C has n ntt. A further similar example is the rather obscure CT VI 340j. In CT V 49c/B4C the ntt-clause also contains the particle is, (see 7.2 and example (353) below) other variants omitting ntt. 17 So Pierret (1874–78, vol. 2, 34) and all subsequent authors citing this example except Helck (1975, no. 10) who has i.Dd=k n Hr Hwntw Ha.kw m mAa-xrw=f m-bAH psDt.
56
chapter two
These examples are, of course, not instances of reported speech; instead, the addressees are told to say something that they have not (yet) said. In the next example the speaker expresses a wish that some wholly unspecified participant would say the complement proposition:18 (28)
A myth narrates the laments of Horus when subjected to attacks by Seth:
Ad=f wi pH=f wi i.in Hr iw Hw A Dd ø n mwt(=i) tw ntt wi snd.ki wrt iw.ki m pf gs ntt wDa Ad=f wi “He rages at me and attacks me”—so said Horus, stranded. “If only someone would tell that mother of mine that I am very afraid and stranded on the yonder side, and that the Outcast (CT VI 408m–q) rages at me”. Whatever the identity of the main clause subject, the complements in these last three examples are patently unasserted by this participant. Yet, the same does not hold for the real speaker who remains wholly committed to the subordinate situations. It matters but little that the governing clause does not confirm an act of speaking or assertion having ‘originally’ occurred at all. After the verb Dd, the elements ntt/wnt can thus be analysed to mark object complement clauses as modally realis speaker assertions. Various alternative but similarly semantic-pragmatically (rather than syntactically) based analyses of the use of ntt/wnt in object complementation have been proposed earlier. Loprieno has interpreted ntt separately from wnt as ‘actualiser’ whose function is to indicate that the following proposition is ‘thetic’ in character.19 ‘Theticity’ refers to a particular strategy of information structuring whereby the described situation and its associated participants are viewed as having equal communicative salience. Thetic sentences do not serve to predicate attributes of a topical subject, but instead ‘present’ entire situations and function as ‘sentence focus-’ rather than ‘argument-’ or 18 Given the future time reference, it seems that the governing clause contains an impersonal subject omitted under relevance (cf. Collier 1990c, 88) rather than that the subordinate clause is a subject complement of a passive sDm=f with future reference. 19 Loprieno 1991a, 214. However, in Loprieno 1995 (100, 109, 200) this and the earlier arguments on wnt (see above) seem to have been abandoned and both are simply referred to as ‘conjunctions’.
affirmative object complementation
57
‘predicate focus’ propositions.20 For example, a proposition “John is singing” may represent a categorical (non-thetic) assertion if it answers the question “What is John doing?” In such a case the reply is about ‘John’, which serves as a topic, and the information about what he is doing figures more salient. If, however, the same sentence answers to “What’s that noise?”, then the utterance ‘presents’ the situation as one un-analysable whole “(It is) [John singing]”, where neither the event or its argument(s) figure larger in the information-chunk as a whole. Thus, e.g. expressions of participant-focus are diagonally opposite to thetic.21 However, ntt/wnt-introduced cleft sentence (subject focus) object clauses occur freely and show that there is no such constraint with these elements.22 It may also be asked what precisely is there in verbs of locution, cognition and perception that allow ‘thetic’ (ntt/wnt-) complements when most other matrix verbs do not. Indeed, thetic or non-thetic status of a given proposition is not determined by lexical or other grammatical factors, but arises from speakers’ shaping of their contribution in response to what sort of information they believe their audience to be seeking. Instead of indicating ‘theticity’ of the complement, Allen has suggested ntt/wnt to signal that clauses introduced by these elements express absolute tense.23 In an example such as (18) above, wnt would show that the tense of the subordinate verb is viewed from the real speaker’s temporal perspective (i.e. the time of the actual speech act) and as “relevant to his own present situation”. This proposal captures superbly the importance of the real speaker’s position, but appears to be somewhat too tempus-oriented.24 Many patterns following ntt/wnt are non-verbal, and with them the question of absolute versus relative timing does not arise.25 For Allen, the lack of an introductory element 20 For what follows, see Sasse 1987; Lambrecht 1994, 137–46. For further discussion, see 8.2 below. 21 Sasse 1987, 572–73. 22 See examples (74) and (76) below. Also ntt-introduced ‘emphatic’ (adjunct focus) second tenses seem to appear after verbs (see example (103) below) and are certainly found after prepositions (section 6.3). 23 Allen 1986a, 11–12. 24 Meltzer (1987, 148–50) raises the same objection against Allen’s thesis, particularly as regards the latter’s discussion of iw, but his criticism is based largely on the assumption of correctness of the ST system and the morphological deductions inherent thereof. 25 See below for examples with nominal and adverbial sentences. Perhaps noticing this problem, Allen suggests that with nominal sentences ntt/wnt may be merely
58
chapter two
signals the tense of the subordinate clause to be relative (taxical) to that of the governing predicate.26 However, a closer scrutiny of unintroduced ntt/wnt-less complements after Dd actually provides further support for the analysis of these elements as marking the real speaker’s stance towards the subordinate proposition. As seen, e.g. in certain negative contexts the real speaker may be denied the capacity to assert the complement, and this affects the grammatical form of the complement.27 In Earlier Egyptian there are no examples of reported speech with a negative Dd + complement clause.28 Fortunately, negation is not the only operation that may alter the speaker’s capability to assert the complement: interrogative has often closely similar effects. If the main verb of locution is questioned, in German the subjunctive is characteristically used in the complement and a whether/if-clause in English:29 (xxxv)
Sagt er, er wäre (SUB) müde? “Does he say if he is tired (or not)?”
Yet, this does not result automatically from the interrogative, but concerns its scope. The if/subjunctive shows that the complement is under the scope of the interrogation as well, i.e. the speaker is not only (or even primarily) interested in whether someone said something, but also in finding out whether that someone is ‘tired’ or not. Here again apply the usual caveats pertaining to different contexts: “Does he say that he is tired?” is quite acceptable if the ‘tiredness’ is known to be the case and all that is enquired is the saying.30 The use of a non-indicative in (xxxv) is thus a signal that the real speaker is actually ‘appositional’, and that with adverbial constructions these elements “could be considered markers of syntactic function only” (1986b, 33). 26 Allen 1986a, 7–8; this is seen to apply generally and not only to complementation. 27 Chapter 1 above. 28 There is one example with a negative hortative forbidding an addressee from saying a certain proposition, and in this unique instance the complement clause is actually not construed with a bare sDm=f, but displays a cleft sentence subordinated by the particle is; see 7.2 below for discussion. Sin B 7 is hardly to be read n Dd=i anx(=i) r-sA=f “I did not say I would live after it” or even as direct speech. anx may here be an infinitive. 29 Palmer 2000, 118, 121; see also Hooper 1975, 98–99. 30 No comparable examples without other irrealis-inducing factors exist in Egyptian with Dd, but they are to be found after rx; see 2.2.1 below.
affirmative object complementation
59
unaware of the status of the complement clause situation as well, and asks for its confirmation or denial. In instances of a similar sort, ntt/wnt is not employed after Dd in Earlier Egyptian: (29)
The ferryman of the dead asks the deceased:
in Dd=k DA=k ir gs iAbti n pt Do you say you will ferry over to the eastern side of the (CT V 103e/T1C) heaven?31 (30)
Part of a dialogue between the deceased and gods:
hA i.n=sn i.n nTrw r N pn in Dd=Tn hAw N pn r wab r-gs iSnw pi nb Htm “Go down!” So they say, namely the gods, to this N (=me). Do you say N (I) should go down to bathe beside this iSnw?32 (CT VI 316p–r)
The non-occurrence of ntt/wnt here is not due to the relative future tense of the complement situation. As is clear from examples (20) above and (51) below, future complements after Dd or in general are not restricted to bare sDm=f forms.33 Similarly, it is neither the case that ‘theticity’ is cancelled in these examples. Instead, more than enquiring whether something is/has been said, the real speakers ask whether the complement situations will occur; i.e. “will you ferry?” and “should I go down?”, with the interrogative scope extended also over the subordinate clause.34 This is particularly clear in example (30) where the matter primarily enquired is surely not whether the addressees say something, seeing that this was just heard by the speaker. Rather similar effects may also occur with conditionals. In the following example the governing predicate and its complement are both part of a conditional protasis describing a mere open possibility; again no ntt/wnt is used:35 31
Similarly in CT V 80b–c/T1C; see n.38 below. Similarly CT VI 317b–c. 33 The inverse holds as well; see below for examples of bare sDm=f with a nonfuture meaning. 34 See (41) and (73) below for instances where this is not the case. 35 A further example is CT V 30e ir Dd=k wnm=i nn “If you say that I must eat this…” However, the interrelation of conditionals and irrealis modality is not straightforward. There is a clear scale between counterfactual conditionals such as “If you 32
60 (31)
chapter two Words of a magician to a serpent:
ir Dd.n=k wd=k a r rxyt axm irt Hr hhyt gr rmT36 If you have said that you thrust (your) hand against mankind, the flaming eye of Horus will go forth, and men will fall silent. (pTurin 54003, rto. 11–12)
In all these instances the real speaker remains in various ways unaware of the actuality or eventual actuality of the complement situations and hence incapable of committing himself to them. Thus the subordinate clauses are not introduced by ntt/wnt, but are coded as irrealis with a bare sDm=f. Rather than the geminating form, non-geminating sDm=f-forms appear in the morphologically revealing examples, and include those with the ending -w.37 That is, lack of gemination seems to correlate with propositions such as these, where assertion is for some reason impossible to the speaker. These considerations allow also a better insight into the following example, which is a textual variant to (29) above: (32)
The yon ferryman says to the deceased:
Dd=k DA=k ir gs (p)f iAbti n pt You should say if you will ferry over to that eastern side of the (CT V 103e/T1Be) heaven. The editor of the CT variant T1Be seems to have understood the ferryman’s words not as a question, but as a hortative. 38
knew/had known X…” that present the protasis situation as definitely false, and sentences such as “If you know that 2+2=4, just say it”, whose ‘conditionality’ actually depends on the context in which the sentence is uttered and which are more likely to read as concessive. Moreover, the speaker’s attitude and the view of the protasis situation as more or less acceptable seem to play its role here as well. For example, “if you know all about this, why don’t you say it?” has the appearance of a ‘concessive’ conditional, but actually signals total disbelief. See Akatsuka 1986 and Sweetser 1990, chapter 5 for discussion of conditionals vis-à-vis discourse and modality. 36 The reading of the lines in the papyrus follows Vernus 1990, 32. 37 The verb DAi shows DA also in the instance quoted in n.31 above. Example (31) writes wd of the verb wdi, which apparently also has a geminating form wdd (Allen 1984 §§ 165, 170). See also example (44) below for the verb iTi ‘take’ and examples (44)–(48) and (50) for the verb ini, ‘bring’. 38 The same variation is observable in CT V 80b–c, where the version T1C has in Dd=k, whereas all the other variants have either Dd=k or i.Dd=k. The latter version appears to be the original in this instance.
affirmative object complementation
61
What differentiates this instance from the kind exemplified by (26)–(27) above is the same difference as that between English “you should say if you will ferry over” and “you should say that you will ferry over”. The first one asks for clarification as to what the addressee is planning to do, whereas the second is a request for the latter to say something, but with the speaker already aware of his addressee’s intentions to carry out the action referred to in the complement. Thus, the absence of ntt/wnt again signals the lack of ‘assertability’ of the complement proposition to the speaker. The examples (29)–(32) do not manifest his knowledge of and commitment to the complement state of affairs, but rather the ‘negative epistemic stances’ of unawareness, doubt and ignorance. Yet, there are a formidable number of examples of affirmative object complements after Dd with seemingly nothing in the grammatical context that would render the subordinate proposition beyond the ‘mental scope’ of the speaker or affect his knowledge and beliefs concerning it, but where no ntt/wnt is employed regardless. Instead, they appear to be straightforward examples of reported speech with all the characteristic modifications in deixis; for example: (33)
The deceased says to a malevolent spirit:
ii.n=i xr=k swA=i arw=k nSnS=i mDAwt=k Hr nn n mrmr Dd.n=k iry=k r=i Hr iSt=i I have come to you that I may break your pens and tear up your books because of this double ill, which you said that you would do to me for the sake of my property. (CT V 66e–h/B2Lb)
(34)
The deceased expresses his faith in the protection by Atum:
nHm=f w(i) m-a Sat Dwt Ddw rmT nTrw Axw mwt iri=sn sw r bA=i He will save me from the evil slaughter, which men, gods, spirits and the dead say that they will do against my soul.39 (CT VI 93d–e)
39 The lack of agreement between Sat and Ddw apparently results from mechanical copying from the identical CT VI 93a where the antecedent is sdb ‘impediment’.
62 (35)
chapter two Sinuhe characterises the ‘Goliath of Retenu’ who challenged him to a duel:
pry pw nn sn-nw=f dr.n=f s(y) r-Dr=f Dd.n=f aHA=f Hna=i He was a champion without equal: he had come to dominate it (his tribe) entirely, and he said he would fight with me. (Sin B 110–11)
Yet, all these examples share one crucial characteristic in common. Whether the question is of some unspecified malice, a murderous act, or an unprovoked fight, the situations described in the subordinate clauses are all most unwelcome to the speaker. Although the complements are reported as certainly having been said and asserted by the main clause subjects, they constitute scenarios to which the real speaker certainly does not want to commit himself. In other words, the issue is not that of speaker incapability, but unwillingness to express commitment and accept an unpalatable complement situation. This sophisticated conveying of information about the commitment and acceptance in Earlier Egyptian is paralleled by the use of mood in modern languages; for example, in Italian and German:40 (xxxvi) I maligni dicono che il nonno N riscuotesse (SUB) fino alla sua morte… una gabella ‘sul coito’ “Wicked people say (i.e. would have us believe) that grandfather N until his death levied a tax ‘on copulation’” (xxxvii) Sie sagen, dass mein Freund ein schrecklicher Mensch sei (SUB)
“They say that my friend is horrible person” Sie sagen, dass mein Freund ein sehr netter Mensch ist (IND)
“They say that my friend a really nice person” The speaker may decide to withhold from asserting the complement if it does not seem acceptable as a proposition, as is most clearly betrayed by 40 The Italian data from Maiden & Robustelli 2000, 326. In German the subjunctive is increasingly common in colloquial language in both sentences and the difference is consequently lost.
affirmative object complementation
63
the pair of German sentences in (xxxvii) above. Exactly the same holds for the Earlier Egyptian examples above. Non-geminating sDm=f’s, including special forms in -y, (judging from the writings iry=k/iri=k in examples (33)–(34)) are again used in these irrealis expressions of strong lack of speaker acceptance of what is said. Non-assertion occurs again, although the original speakers are indeed reported to have said the propositions. Further examples of similar sort after Dd and with the same motivation for the absence of ntt/wnt are the following:41 (36)
A mocking comment made of a defeated serpent-demon:
ia[nw] n=k imy-nhd=f... mk hwt pr m pt m-Xnw TpHt sbi Dd.n=f sbi=f r ra iri=f awA r=f sA r=f in snS wr Woe unto you, imy-nhd=f-serpent… See, burning has gone forth from the sky into the cavern of the rebel, although he had said that he would rebel against Ra and commit robbery against him; now his mouth is guarded by the Great Annihilator. (CT V 244b–245a)
(37)
An Execration Text heaps curses upon enemies specified in the following terms:
sbi.t(y)=sn wA.t(y)=sn aHA.t(y)=sn Ddw aHA=sn Ddw sbi=sn m tA pn r-Dr=f Those who will rebel, those who will conspire, those who will fight, those who say that they will fight, and those who say that (Ächt d 1–6) they will rebel—in all of this land.42 (38)
A boast uttered by a worker in a reaping-scene:
iSst pw TA Dd i.iri=f m tr(=i) Who is the man who can say that he can do (reaping) at my (Ti pl. 123, middle register) pace?43 41 See also CT IV 385e; CT V 324j; CT VI 318g (all iry=k); Tb 90/Nu pl. 22, 4 (iri=k). CT V 247b–c has sbi=f… iri=f awA, CT VI 93a iri=sn and CT VII 111f awAw=sn. CT II 51b has iri=sn but the context is obscure; Ptahh 111 appears to be corrupt. 42 Similarly Ächt h 4–5, l 4–5 and n 4–6. 43 In the bottom register, one has similarly zy pw Dd i.iri=f m tr(=i), “Who is one who can say…” etc. Here the occurrence of the i-prefix is notable (see EAG §§ 449–54; Allen 1984 §§ 11, 141). Primarily a feature of Pyramid Texts, the function of this additional inflectional feature is not altogether clear. However, for the present argument it is
64 (39)
chapter two The author of a letter denies the accusation that he is disregarding his duties because of a certain perceived injustice:
ir pA Dd wrS=i Hr rm(t) Hr=s... Now, as for the saying that I (allegedly) spend my time crying (pUC 32200, 17) about it… (40)
The deceased states that he will not partake in consumption of faeces:
bwt=i pw Hs n swri=i snt=f wsSt nTr nb Dd wnm=i bwt=i wnm=f Hna=i Excrement is my detestation and I will not drink its ‘sister’ urine. Every god who says/has said that I must eat what I detest, he (CT VI 198n–p) will (have to) eat with me. In the first of these examples the motivation for the lack of commitment is apparent, particularly as it is clear from the context that by the time of speaking the malevolent intentions of the original speaker had failed to materialise. The purpose of the text in example (37) is to annul all such intentions described, including their very likelihood, rather than to express them as accepted. In (38) the speaker is so certain of his own excellence in the task at hand that he expects no answer whatsoever, and in (39) the tone of denial is apparent and quite akin to that seen in the Italian example (xxxvi) cited above. In these last two examples the complement clauses do not refer to future, but merely describe general situations towards which little ‘faith’ is expressed. Finally, in the rather unpleasant example (40) the reasons for the lack of speaker commitment hardly need to be spelt out. Also in the following instance the non-acceptance of the complement situation is most apparent, seeing that the latter is immediately denied in the following sentence:
significant that i never co-occurs with gemination in ultimae infirmae class sDm=f, but, on the other hand, does co-occur with non-geminating forms showing the endings -y and -w (EAG § 450; see also 10.2 below). A further similar (i-less) example is Tomb of Merefnebef, pl. 21/66, right, with an exchange between two reapers A i TA Dd iri=f m tr(=i) and B ink pw nt(y) Hna(=i), “Hey, you man who says he can do (reaping) at my pace!”; “That’s me, mate”.
affirmative object complementation (41)
65
A question addressed to the deceased concerning Seth:
in smA.n=f Tw Dd.n ib=f mt=k n mt=k mk Tw irf xpr.t(i) r=f m mnw n smA Did he slay you or did his heart say that you have died? You have not died. See, you have been transformed into an enduring (CT VII 34a–e/T9C) bull against him! The denial also shows that unlike in examples (29)–(30) above, the interrogative scope does not extend over the complement mt=k; i.e. the speaker is most certainly not asking, “Have you died?” Noteworthy is also the seemingly past sense of mt. Although futurity is not wholly excluded either, past interpretation is favoured by the final sentence.44 In the next example, a yet another variation of the ‘ferryman’-theme, the speaker’s less-than-full conviction to the complement proposition is similarly suggested by what follows:45 (42)
The yon ferryman questions the deceased prior to accepting him on board:
nn i.Ddw=k DA=k ir gs pf iAbti n pt in i(w)=k rx.t(i) wAt i.Smt=k Hr=s As for this your saying that you will ferry to that eastern side of the sky; do you (even) know the road on which you will go? (CT V 104j–05b/Sq2Sq)
Finally, in the following famous passage, the speaker’s tone and rhetorical intent is close to sarcastic: (43)
The sailor has just promised to send the snake myrrh and incense:
aHa.n sbt.n=f im=i m nn Dd.n=i m nf m ib=f Dd=f n=i n wr n=k antw xprt nb sntr ink is HqA pwnt antw n=i-im(y) sw Hknw pf Dd.n=k in.t(w)=f bw pw wr n iw pn Then he laughed at me and at these things I had said, which were foolish in his mind, saying to me: 44 Of course, given that the example derives from the CT, n mt=k may also be read as future. The variant T1C inserts n rdi(=i) mt=k n mt after the initial question. This may similarly be interpreted either as a future “I will not let you die” or as past “I did not allow you to die” (cf. n.47 below). The third variant B10C has an additional n mt=k in the same position. 45 Cf. also the rather similar CT V 115a–d.
66
chapter two “You don’t have much myrrh or any sort of incense, whereas I am the ruler of Punt, and myrrh is mine. And that incense which you said would be brought;46 it is the mainstay of this (Sh.S. 149–52) island”.
In every case the speaker does not accept the complement proposition, and no ntt/wnt appears.47 2.1.2 An Example Text The most extensive source of examples displaying the full richness of motivations for lack of speaker commitment after Dd, and also one that contains an example with a bare non-geminating sDm=f in close proximity to a wnt-clause, are Coffin Texts Spells 38–40.48 The text is of considerable complexity and furnishes a case for testing the hypothesis presented above in at least equal measure as it provides further examples. The spells present a remarkable account of a court-case between a deceased father and his son before a tribunal of divinities. The son is about to enter the yon realm himself and eager to usurp his father’s position there. He employs refined rhetoric to persuade the judges of the righteousness of his intentions. The father’s words betray clear reluctance to bow to the pressure and hand over his status. The tribunal plays a passive role throughout and makes no contribution, although the contestants address it as much as they do each other, and arguments, protests and accusations are hurled to and fro. The debate opens with the son addressing the tribunal and exhorting it to confirm that the transition of rights has their blessing and that the gods have specifically summoned him: 46 Here the alternative would be to read int=f as an infinitive “of whose bringing you spoke”, but then one should perhaps expect the preposition r or Hr to appear before int=f. 47 Also the PT provide the following particularly fine examples: PT 944a–b/MN ii.n=f ir=k Dd.n=f smA=f Tw n smA=f Tw Twt smA=k sw “Although he has come against you and said that he would slay you, he will/has not slay/slain you; it is you who will slay him”. PT 1477a–b/P has in ir.n=Tn ir=f Dd.n=Tn mt=f n mt=f “Have you acted against him and said that he has/will die(d)? He has/will not die(d)”. As usual, the timereference of the negations is not entirely clear. PT 481a–b is an earlier version of example (41) above. 48 See Faulkner 1962; Grieshammer 1970; 1975/76 and De Jong 1994 for previous studies on one or all of these spells. The versions primarily followed here are the variants B4L and B12Cb.
affirmative object complementation (44)
67
mTn it=i pf aHaw=i pf mHy=i pf HAy=i pf hA.n=i n=f pf imy imnt imy Xrt-nTr siwy n=f wi m DADAt Dd n=f int wi tp-r=Tn sar.tw hrw=i m tA pn n anxw nt(y) wi im=f Dd n=f iT=i st=f nHm=i saH=f m tA pf Dsr nt(y)=f im=f See that father of mine; that attendant, guardian and champion of mine for whom I have descended; him of the West and of the necropolis! Announce me to him in the tribunal. Tell him that your mandate49 has brought me (now that) my days in the land of the living where I was are over. Tell him that I may take his seat and assume his status in this holy land where he (CT I 158a–159c) is.50
By not phrasing the complements int wi and iT=i as assertions, the son cleverly avoids presenting their propositional content as something to which he himself is committed or wishes to present as a fact.51 This is because he is seeking the tribunal to accept culpability of the looming usurpation by announcing this to the father, as well as responsibility of his being brought to present his claims, which undoubtedly is not the case.52 ntt/wnt appears in neither instance and int seems to have past reference. The son then continues by asking the tribunal whether they have been told why he has been brought before them:53
49 tp-r, ‘Ausspruch’ (Erman & Grapow 1926–31 (henceforth Wb) V, 287; Hannig 2003, 1423) clearly refers here to an authorising pronouncement. 50 Faulkner, (1962, 36; 1973–78 vol. 1, 30; 1979, 2) Grieshammer (1970, 117) and Silverman (1980, 9) understand both instances of Dd as sDm.n=f’s reporting the father having said that his son may usurp his status. This misses the point of the subsequent exchange of arguments quite fundamentally. 51 I.e. the motivation for irrealis is the same as in German or French ‘journalistic style’ (see chapter 1 n.26 above). 52 Compare the latter instance with example (228) below where the complement situation is similarly ‘false’ but the clause is introduced by iwt, the negative equivalent of ntt/wnt. The difference seems to be based precisely on what the speaker wishes to present as his own committed stance. 53 However, the construction in this sentence in fact belongs elsewhere in the discussion. The subordinate clause is not an object- but a subject complement of the passive Dd; see chapter 4.
68 (45)
chapter two in Dd n=Tn in.t(w)=i n=Tn r sAr n=i it=i pf r iwa=i st=f r nHm=i saH=f in Dd=Tn in.tw=i n=Tn (n)54 tp-r=f Have you been told that I am brought to you so that my father might be dispossessed55 in my favour and that I may inherit his seat and assume his status? Will you (rather) say that I am brought to you because of his pronouncement? (CT I 159d–h)
The son’s point is clearly that if the tribunal has been told that his intentions are so negative, this is not to be believed; his words show a denial of that suggestion. The intent of the son’s second question is less clear. He seems to suggest that the tribunal may consider announcing the transfer of rights in the name of the father—without hearing the latter.56 At this juncture the father puts forth a rebuttal whose central claim is rather obscure. Next the son addresses his father and attempts to persuade him to testify that it is his will that the son succeed his father. His intentions are the same as when addressing the tribunal at the start of the debate, and again he uses non-assertion: (46)
in Dd=k in.tw=i r tA pw Dsr {rdi.n.t(w)=f}57 nt(y)=k im=f r st=k imt Xrt-nTr...r pXr rf n=i Axw=k r nHm=i n=i iAwt=k r Dd(=i) rf Axw(=i) mitw=k r=k Will you say that I be/am brought to this holy land where you are to your position in the necropolis…so that58 your might
54 The variants B12Cb and B16C have no n before tp-r=f. However, the damaged B13Cb omits the dative n=Tn but has […] in.t(w)=i n tp-r=f, which must mean “I am brought because of his pronouncement”. It seems thus that in the two first-mentioned versions the n is lost in a haplography with the final n of n=Tn. 55 So Silverman 1980, 9. 56 Faulkner (1962, 36; 1973–78 vol. 1, 30) and Grieshammer (1970, 117) interpret the first question as “have ye said/habt ihr gesagt?” (i.e. as in Dd.n=Tn with a sDm.n=f). As the tribunal is not indicated as saying anything at any point, this is unlikely to be correct. Faulkner’s rendering of the second question as “Do you say that I may bring you his utterance?” (see also Faulkner 1979, 2) makes little sense and ignores the w after int in B12Cb. Grieshammer’s “Habt ihr gesagt, dass ein Ausspruch von ihm mich wegführt zu euch?” (1970, 117) is better, but the tense of Dd=Tn and the translation of in.tw=i (the verb ini is translated thus recurrently) seem incorrect. 57 This clause, due to B16C, appears wholly spurious. 58 Because of the particle rf, Faulkner (1962, 36; 1973–78, vol. 1, 30) translates as imperative “Transfer to me thy power”, “despite the preceding r” (1962, 38 n.o). Gries-
affirmative object complementation
69
may revert to me, that I may take for myself your offices and also that I may say thus: “My power is equal to yours”? (CT I 160g–161b)
The father replies to this by claiming that the usurpation plotted by the son will result in a harmful situation where the father’s enemies will be given a free rein to exult over him and his house will face destruction. The son seems amazed of this attack and launches upon a forceful denial:59 (47)
Ax nTr(y) ir=k m imnt m tA pw Dsr nt(y)=k im=f bA=k n=k Axw=k Hna=k mr.n=k bA=k im=i tp tA in Dd=k in.t(w)=i r=i r tA pw Dsr nt(y)=k im=f r sD rf pr=k r whn rf arrwt=k r wSr rf iwa=k r nHn xftw=k im=k isT wi aA m tA pn Hr DbA nst=k Hr sAq bdSy=k You have divine might for yourself in West, in this holy land where you are; you have your soul and your powers are with you. You wanted your soul in me upon earth. Do you (really) say that I am brought to this holy land where you are only to destroy your house, to break up your gate and despoil your inheritance, so that your enemies will exult over you? Why, I am here in this land for the sake of reasserting your seat and (CT I 162d–163d) for pulling together your weakness!
The son is not asking whether his father said the complement proposition since he himself has just heard it, but rather what is his justification for doing so. He appears genuinely offended by the accusations and shows that he does not accept them at all. There then follows a lengthy section where the son protests his righteousness as an heir, after which he again turns to the tribunal and asks could they not60 acclaim his rights, seeing that he is now part of the community of the blessed dead:
hammer has “damit dann die Verklärten, deinesgleichen, über dich sagen…” (1970, 119). 59 Grieshammer’s translation in particular is notably different here (1970, 120–01). 60 See Silverman 1980, 88–93 on the particle rr as marking the question as rhetorical and expecting an affirmative reply; but cf. Junge 1983, 548.
70 (48)
chapter two in rr Dd=Tn in.t(w)=i r st nt it=i pf... isk rx.n=i qd=Tn mA.n=i Xnw=Tn Should you not say that I be brought to the status of my father…for I know your nature and have seen your abode? (CT I 168d–169a)
Once again, his words show careful avoidance of presenting the complement situation in such a way as to imply that he himself is in a position to assert it. Then, without expecting an answer the son continues with a further denial of his father’s accusations: (49)
Dd n=Tn Ssp=i saH=f nHm=i st=f r nHn xftw=f im=f... Dd n=Tn whn(=i) rf msxnt=f m iw nsrsr... isk ink is irr DbA=i nst=f tpt tA nt (sic) anxw m iw nsrsr imi wS=s m tp tA You have been told61 that I would receive his status and take his seat only that his enemies may exult over him… and you have been told that I would overthrow his abode in the Island of Fire… But I am, rather, one who acts, and I will replace his seat upon the land of the living and on the Island of Fire lest it be destroyed upon earth. (CT I 169d–170f)
Here the son presents a report of his father’s accusations and signals that he is not the least committed to their correctness; quite the contrary. After some further argument there follows the final example from this most extraordinary text, where a bare non-geminating sDm=f and wnt appear closely together after Dd. First the son appears to say grudgingly to his father that at least his right to be brought to be aside his father should be recognised,62 but then reminds him of what he had said earlier: 61 Again the construction involves not object- but subject complementation after a passive Dd. Faulkner (1962, 40; 1973–78, vol. 1, 32) and Grieshammer (1970, 123) have “you have said/Ihr habt gesagt”, which entails an assumption that the tribunal said the negative accusations. It did not—the father did, and the son now seeks to denounce them. 62 This seems to imply that the son actually concedes defeat and pursues the second-best option, which is to be with the father in the hereafter, although not in a superior position.
affirmative object complementation (50)
71
mk Dd.t(w) in.t(w)=i n=k r wnn Hna=k m tA pw Dsr nt(y)=k im=f Dd.t(w) wnt mdw.n=k r int=i Ds=k r DbA nst=k Ssp=i saH=k r nS=i Tw m st=k tw m tA pw Dsr nt(y)=k im=f Look, may it be said63 that I am brought to you to be with you in this holy land where you are, although it is said that you yourself have spoken against bringing me to occupy your throne, as I would assume your status only that I may oust you from this your seat in this holy land where you are. (CT I 174j–175a)
Assuming that the interpretation of the son’s intentions is correct, the complement of the first Dd parallels closely the Spanish use of the subjunctive in the following example:64 (xxxviii) Admito que sea (SUB) inteligente, pero qué pesado! “I admit that he’s intelligent, but what a bore!” In Spanish, the verb admitir ‘admit’ is normally used to assert the complement and the indicative is employed. But this can be overruled if the speaker is ‘forced’ to admit something that he does not really accept, even when this may be something very real and ‘true’. Similar ‘grudging acceptance’ seems to be in question in the first Dd-complement in example (50) above. But in the second instance the son is merely recalling what he is told that his father has said,65 and recognises the fact that his father indeed expressed reluctance to hand over his position, referring to the various excuses cited. It is only here that wnt makes its appearance. Accordingly, the text does not present all its indirect speech without this element. Yet, in all but one instance the speaker is keen to disclaim responsibility of his words and avoid 63 Here, given the date of the text, analysing the complement as an object of the main verb Dd with an indefinite subject .t(w) ‘one’ seems legitimate. See 4.1 below for discussion. Faulkner (1962, 42; 1973–78, vol. 1, 33) has “it is said” and Grieshammer (1970, 127) “es ist gesagt worden”. 64 See Klein 1975, 360, 364 n.8; cf. Bergen 1978, 224. 65 By whom? The reference is less than lucid, but the interpretation offered here appears better than that suggested by Faulkner’s (1962, 42; 1973–78, vol. 1, 33) translation “See, it is said that I am brought to you to be with you… a word to you yourself is indeed spoken to bring me to occupy your seat…” which is dubious both semantically and grammatically. Grieshammer (1970, 127) has “Es ist gesagt worden, dass du selbst gesprochen hast…”
72
chapter two
presenting them as something to which he is committed, or there are factors mitigating his commitment and acceptance of the complement propositions, for which reason they appear as ntt/wnt-less irrealis. Another instance of a wnt-introduced complement and one with a bare sDm=f appearing close to each other after Dd in reported speech occurs in the following example, where the context is unfortunately both damaged and obscure, and the motivation for the variation unclear: (51)
Ra is evoked on behalf of the deceased:
i ra iw Dd.n x[ftw] i[p]f nw wsir N nHm=s[n] HDt aAt tp=k Atfw imw wp=k iw Dd.n=sn wnt=sn r HDt tpw r Xnn r[...] m-bAH iw Dd.n=sn wnt=s[n] r Xnn mAat O Ra, those enemies of Osiris N have said that they will take away the great White Crown on your head and the Atef-crown upon your brow. They have said that they are going to destroy heads and to confuse […] previously. They have said that they (CT VI 277q–278d) are going to disturb Order. Nevertheless, ntt/wnt-introduced and un-introduced complements after Dd do appear side by side, and there seems to be a definable set of pragmatic parameters for their respective uses, which relate directly to the real speaker’s knowledge, commitment and acceptance of the complement state of affairs. Un-introduced non-geminating sDm=f’s do not ‘substitute’ for realis ntt/wnt-complements or vice versa. Their distribution is complementary. The un-introduced sDm=f forms clearly function as ‘echter Modus’ in Earlier Egyptian indirect speech and are used for non-assertion under exactly the same conditions as in other languages.66 It is of some interest that the sole form of indirect speech after Dd that survived after the Old Kingdom and the stage of language represented by the CT is the one without ntt/wnt. Thus, in Classical Middle Egyptian, with its system of quoting speech directly, the Latin-style ‘highly developed indirect speech’ referred to by Gardiner is an indication of irrealis modality and speaker non-assertion of the complement. This same situation pertains also to the other
66
Contra Polotsky 1964, 272.
affirmative object complementation
73
potentially ntt/wnt-combining predicates of Earlier Egyptian, whose complements similarly display a split between asserted/assertable and irrealis complements and with the same grammatical division into introduced and un-introduced object complements respectively. 2.2 Verbs of Cognition and Perception 2.2.1 Seeing is believing and other truisms: realis, irrealis, knowledge and perception Besides locution, in Earlier Egyptian the division between ntt/wntintroduced and bare suffix-conjugation object complement clauses is primarily found after predicates that describe mental states and perception. Unlike with Dd, this feature remains constant throughout the stage of Egyptian language discussed in the present work. The verbs in question are rx ‘know’, mAA ‘see’, siA ‘perceive’ (visually) and sDm ‘hear’. There are certain differences between the said verbs and verbs of locution with respect to the indication of speaker commitment and acceptance. Nevertheless, the principles governing the grammar of their object complement clauses are for the most part the same. The differentiation between introduced and un-introduced complements is, again, reflective of their status as assertions and non-assertions respectively. Although there is a considerable quantitative bias in favour of rx,67 affirmative non-future verbs of cognition and perception with ntt/wnt– introduced object complements are common in Earlier Egyptian: (52)
Iyni-hor says concerning his tomb:
iw rmT nb rxy wnt iri.n(=i) s[w] Everybody knows that I made it.
(EAG § 1022)
67 Further similar examples of rx + ntt/wnt not quoted elsewhere are CT I 140g (wnt=f sr/wnt nts srt); CT IV 84j (ntt iT.n N Hw); HPB III, pl. 6b (wnt sp.n n=f […]); MFA 04.2059, vso. 6 (nt(t)=s[…]); pBerlin 10036, 15 (ntt st SA); Urk IV 429, 5 (nty (sic) wbA Hr=i); Urk IV 592, 1 (ntt sA=f pw); Urk IV 993, 9 (ntt anx=f); Tb 78, 37/Nu pl. 41, 42 (ntt it.n=f). CT VI 382e–f has rx=i ntt wn di=i sA n it, which may mean “I know that I (used to?) give son to a father”, but this sounds rather strange.
74
chapter two
(53)
It is said of certain divine beings:
iri.n=sn rn n N pn m bik nTr(y) sk sn rxy ntt wa im pw They have made the name of this N as a divine falcon, because (CT VI 312d–f) they know that he is the one therein. (54)
Hapdjefa reminds officials that cancellation of funeraryarrangements is unacceptable:
mTn rx.n=Tn ntt ir sr nb nDs nb dd tp n Smw=f r Hwt-nTr n nDm.n n=f xtxt im Look, you are aware that as for any official or any ordinary citizen who gives the top-ration of his harvest to the temple, (Siut I 310) reversal therein is not pleasant to him.68 (55)
The author reminds her dead addressee of the protection promised by another dead person:
i(w)=k rx.t(i) ntt Dd.n idw r sA=f ir wnnt wnt im nn di(=i) nqm=f n nqmt nbt iri m n(=i) mitt irt You know that Idu said of his son: “Whatever there might be in the hereafter, I will not allow him to be afflicted by any mishap”. Please do the same for me! (Haskell Museum 13945, 1–2)
(56)
A reply to an inquiry concerning the addressee’s sister:
mk tw rx[.ti] ntt sy mn.ti Hr wpt [gs]-iAb Look, you know that she is fixed on the listing of [Ges]-Iab. (pUC 32126, fragment ii, 3–4)
(57)
The king is the lord of Egypt, thus:
nn tA nb r ib=s n rdi(.n) n=s Hr=sn rstw n Dar.n st mHtyw rx.ti ntt nxy=s wn mi min qA-a nsw-bity mn-xpr-ra There is no land against her will: southerners do not aspire for such, northerners do not seek such, for they know that her protector exists like Min, one high of arm, the dual king (Buto stela of Thutmosis III, 6–7) Menkheperra.
68
Rather similarly Siut I 280–01.
affirmative object complementation (58)
75
A legend under the cartouche of Thutmosis III
qmA.n=sn m Hwt-aA m nfrw Ha=sn rx ntt iri=f nsyt wAHt xr nHH One whom they (the gods) created in the palace from the splendour of their own flesh, and they know that he exercises (Urk IV 593, 4–5) kingship that lasts beyond eternity. (59)
An instruction concerning the vizier’s conduct in land-disputes:
ir sprty nb n[t]y r Dd mnmn tAS=n xr mAA.t(w) ntt st Hr xtm n sr iry xr=f Sd=f Sdwt n tA DADAt smnmnt st As for any petitioner who shall say: “Our boundaries have been moved”—when it has been seen that69 they (i.e. land-register documents) carry the seal of the relevant official, then he (the vizier) can confiscate the Sdw-lands of that council which caused (Urk IV 1111, 9–13) them to be moved. (60)
The postscript of Tb Spell 148 tells of its discovery by prince Hardjedef:
in.n=f sw mi biAw n nsw xft mAA=f ntt StA pw aA He brought it (away) like a royal marvel when he saw that it (Tb 148, 21–22) was a great mystery. (61)
In the debate between father and son, the son claims the primeval god having been informed of his acceptance to the hereafter:
sDm.n=f m r irrw wnt Ts.n wi it=i pf imy imnt He has heard from the mouth of the rite-performers that my yonder father who is in the West has elevated me. (CT I 167a–b/B12Cb)
In these examples the subordinate clauses report what the main clause subjects know or perceive. However, as seen, they do not amount to assertions from the subject’s part inasmuch as the subject is not necessarily saying, or has not said, anything. The same may also happen with main verbs of locution, but is always the case with other 69 Not ‘whether’ as e.g. in Van den Boorn 1988, 185, 190. The prerequisite for the seizure of lands by the vizierate is that official records, which the local authorities have ignored, support the petitioner’s claim.
76
chapter two
verbs.70 This matter is less of an issue in the numerous instances of rx and the verbs of perception where the speaker indicates his own knowledge or sensory experience in the first person and is the complement ‘original speaker’:71 (62)
The king tells his addressee what he thinks of the latter’s skills as an organiser:
sk Hm Hm(=i) rx wnt Haw nb Hr nfrw=f Verily, my majesty knows that every ship is on an even keel.72 (Urk I 61, 9)
(63)
Queen Hatshepsut explains why she has favoured the temple of Amun:
iw=i rx.kw ntt Axt pw ipt-swt tp tA I know that Karnak is a horizon upon earth. (Urk IV 364, 1–2)
(64)
Ptahmose says his conduct in life was flawless:
iri.n=i m mAat mr nsw rx.kw ntt anx=f im=s I acted in righteousness, as the king would desire, because I knew that he lives of it. (Urk IV 1531, 15) (65)
The king tells why he has chosen Ichernefret to carry out a mission on his behalf:
iw is [hA]b Tw Hm=i r irt nn siA.n [Hm]=i wnt nn ir.t(y)=f(y) st Hr-xw=k Now, my majesty sends you to do this because my majesty has perceived that there is no-one who will (be able to) do it, except you. (Berlin 1204, 8–9)
70 Cf. chapter 1 above; this is particularly clear in instances such as example (52) where the subject is generic. 71 Further similar examples: Red Chapel 188a, 13 (ntt mrt.n=f xpr); Urk IV 835, 16 (ntt Htp=f Hr=s; see example (105) below); Urk IV 1074, 12 (ntt wpwt aSA). Urk IV 346, 3–4 is corrupt. See example (226) below for sDm + iwt, the negative equivalent of ntt/ wnt. 72 Following Wente’s (1990, 19) splendid rendering. For the text, see Brovarski 2001, text figure 1.
affirmative object complementation
77
Yet, it is not necessarily immediately clear whether the complement clauses here, and particularly when the real speaker and the main clause subject differ, are truly asserted (rather than simply reported) by the former participant either. With the verb Dd, this could be shown to be the case by studying examples where the complement was not accepted or believed by the real speaker, even though it was reported as having been said by the original speaker. As seen, in such instances ntt/wnt turned out to be systematically absent, signalling the nonasserted irrealis status of the complement. Here, however, the situation is somewhat different. When occurring in affirmative reporting main clauses, verbs of cognition and perception make a strong claim for the ‘reality’ of their object clauses, regardless of possible differences between the original and the real speaker.73 Consequently, unlike with Dd, instances where the real speaker could ‘dismiss’ the reported complement information as unacceptable or unreliable hardly occur with these verbs. Nevertheless, various modifications of the context that remove the main clause from being a report again show that the real speaker’s knowledge and commitment to the complement situation are key factors in deciding the latter’s grammatical form. As with Dd, the context in which the verbs of cognition and perception appear are often such that it is impossible for the main clause subject to have perceived or been aware of the situation described in the complement clause prior to the uttering of the sentence itself. Yet this in no way affects the employment of ntt/wnt, whose use pertains to the real speaker’s modal stance: (66)
Sinuhe begs understanding for his long hiding from the king:
nb siA siA rx(y)t siA=f m Hm n stp-sA aws wnt bAk im snD Dd st O lord of perception who has insight with plebs; may he perceive in the majesty of the palace l.h.p. that this servant was afraid to say it. (Sin B 214–15)
73
See chapter 1 n.30.
78 (67)
chapter two Horus quotes what Ra told him to say concerning certain deities:
iw=sn Hna=i kA=k Dr.kA=sn Hna=k r rxt stX wnt=sn Hna=k nx=f “They are with me”—so you shall say. Then they will end up with you well before Seth knows that they are with you and (CT II 359c–60a/S2P) complains. (68)
Sabni relates what he did prior to his journey to the hostile hillcountry:
iri.n(=i) [ig]r mDAwt r [r]di[t] rx.t(i) ntt w(i) pr.k(i) r int it(=i) pf I wrote some letters to let it be known that I had gone to bring back my late father. (Urk I 136, 9–11) (69)
The writer informs his superior of a failure in work-assignments:
Dd bAk-im di rx (i)r(y)-mDAt nt(y) m DADAt wnt iqdw n-sprt=f r rwDt r irt wAt HqA n dmi-iw Yours truly is writing; inform the courier who is in the council that, as for the potter, he has not yet arrived at Rewdet to prepare the journey of the chief of Demiy. (Balat-aAyn Asil 3686, 1–3)
In the epistolary expressions akin to those in examples (68)–(69), the complement of rx is most commonly introduced by ntt or by r ntt. The latter is originally a particular idiom of these contexts, but its use spreads also to other object complement contexts in early XVIII dynasty texts:74
74 r ntt will be further discussed in 6.3 and 10.2 below. Further examples of this type with ntt/wnt are Urk I 126, 10 (broken); Urk I 128, 8 (wnt=k hA.t(i); Urk I 137, 12 (broken); Heqanakhte III, rto. 4 (ntt rdi.n(=i); Meketra letter, 5–6 (ntt N iy); Sin B 181 (ntt dbn.n=k); pBrooklyn 35.1446, insertion c, 4 (ntt spr.n N); Urk IV 80, 8–9 (ntt Hm=i aws Ha). Examples with r ntt are Michaelidis Lahun letter, 1 (r ntt hAw=k nb aD wDA); pBerlin 10003, 1 (r ntt xpr prt-spdt); pBerlin 10016, 1 (r ntt wAgi r xpr); pBerlin 10022, 3 (sDm. n=i); pBerlin 10023A, 1 (spr.n N); pBerlin 10033, 1 (r ntt sDm.n=i); pBerlin 10036, 1 (r ntt sDm.n=i); pBerlin 10038B, 1 (r ntt rdi N, a passive sDm=f); Lesestücke 98, 6 (r ntt rdi.n N). See also 6.3 below. An example of r ntt after siA is Urk IV 1381, 4 r ntt TAty pH=f Tni “that the vizier reaches senility”, with an anticipated subject TAty; see 2.4 below.
affirmative object complementation (70)
79
The king tells what he told his followers when he was about to enhance the temple of Amun:
di=i rx=Tn r ntt ib n Hm=i r mnx [...] I will let you know that my majesty’s mind will be fixed (‘to (Urk IV 181, 10–11) build this temple’ or the like) The only participant who is constantly wholly aware of the situations described in all the complement clauses above is the real speaker. In the following example the speaker and the subject of the complement clause are the same individual, but what is actually said is that the speaker now (at the time of writing) knows the complement state of affairs to obtain. Hence the clause is introduced by wnt:75 (71)
In a letter from the king to one of his officials:
[iw mA.n Hm(=i) mDAt=k tn irt.n=k] r ist r rdit rx Hm(=i) wnt in n=k wD n nsw [My majesty has seen this letter which you made] for the palace to inform my majesty that a royal decree has been brought to (Urk I 61, 17–18) you. In the next instance the main clause subject is referentially generic and semi-hypothetical, but the actuality and assertability of the complement situation remain the same for the speaker, and ntt appears: (72)
The king smites his enemies ignorant of his might; however:
mar sp r Tnw rxw ntt sA=f pw mAa Fate is lenient to multitudes that know that he is his (Amun’s) (Urk IV 1293, 1–3) true son. Instances where the real speaker would be rendered not ‘entitled to assert’ the complement due to lack of knowledge about its validity are rare with verbs of cognition and perception. As for interrogatives, with these verbs there are no examples of the type “Do you know whether Jill is there?”, i.e. where the complement situation would 75
Cf. chapter 1 above. The situation is rather complex: the king is writing back to his addressee to inform him that the latter’s previous letter, in which a reception of royal decree is confirmed, has arrived at the palace. The king now knows that his addressee has received his orders and tells him that. The restoration of the beginning follows Urk I 60, 16–17, another letter from the same king to the same recipient.
80
chapter two
also be questioned and fell under the interrogative scope, in Earlier Egyptian. In yes/no-questions involving these verbs, the interrogative scope invariably invests the main clause only:76 (73)
An utterance of a workman to his fellow whilst manufacturing a vessel:
in iw=k Hr mAA ntt n aHa.n pA mnw Do you see that the cup cannot stand up? (74)
(Meir I, pl. 5)
The author asks if his dead correspondent is ignoring something essential in his failure to protect a member of the household:
(i)n-wnn n rx.n=k ntt in tA bAkt irr pr=k m rmT Could it really be that you do not know that it is this maidservant (Cairo Bowl 7–8) who runs your house among the people? In both these examples, the first of which is a fine instance of ntt n, a later isomorphic version of iwt, (the negative equivalent of ntt/wnt77) the matter inquired is whether the addressees see or know something. Yet, the speakers are not at all unaware whether or not this something holds and are not asking anything about it. They themselves clearly do see and know that the situations in the complement clauses obtain, and ntt/wnt mark this. Also in case of negations, in most instances the negative scope invests only the main clause. What is denied in the following examples is that the subjects knew or are allowed to know about the state of affairs described in the complements, whereas it is of course not the case that the real speakers are, or ever were, unaware of their ‘truth’: (75)
The king’s superiority in battles against foreign foes is attributed to higher forces:
ii.n=sn m HHw n rx=sn ntt imn-ra Hr mw=f Although they (the foreigners) came in millions, they did not (Urk IV 1291, 1–3) know that Amun-Ra is his ally. 76 Another example (with some Late Egyptian influence) is Tb 153A/Nu pl. 58, 29– 30: in iw=Tn rx.twny ntt {t}wi rx.kw rn n N “Do you know that I know the name of N?” This sentence is repeated 11 times (spelled r ntt {t}wi rx.kw) in Tb 153B, 2–11/Nu pl. 59, 3–11. For a detailed study of example (74), see Uljas 2004. For further comments on example (73), see 5.2 below. 77 See 5.2 below.
affirmative object complementation (76)
81
The king instructs the high steward concerning the treatment of the peasant:
wnn=k Hr rdit di.tw n=f aqw nn rdit rx=f ntt ntk rdi n=f st You are to keep on making sure that he is given rations, but without letting him know that it is you who has given them to (Peas B1, 114–15) him. Yet, there is one unique example where the real speaker states that at the time of speaking he does not know whether the complement situation holds or not, and, as expected, ntt/wnt is not employed (the version BH3C which provides the first person is badly damaged and presented here in conjunction with the better preserved T1L):78 (77)
The deceased says: T1L: N pn Htm=f n rx N pn hAy N pn i(w)=f Ha BH3C: ink Htm n rx=i [...] iw(=i) Ha.ki
I am the destroyer, but I do not know whether I must descend, (CT VI 260b–d) as I have appeared in glory. Here the real speaker is not in a position to assert the complement—he in fact says this explicitly. The subordinate clause is semantically an indirect question, as is the whether-clause in the English translation, and remains un-introduced. Rather than the geminating sDm=f, in this expression of ignorance about the complement state of affairs use is made of a non-geminating form, (which also here displays the distinctive writing hAy) just as after the verb Dd.79 Somewhat similar 78 In Urk IV 365, 10–11 one reads sAw Dd=Tn m n rx=i sp-sn iri.n=Tw nn Hr m. However, the clause headed by iri.n.Tw is not a complement of n rx=i, but an independent appositional second tense sentence: Take care lest you say: “I don’t know, I don’t know; why has this been done?” 79 Here a mention should be made of non-verbal indirect questions after rx. Strangely, in these instances the complement consists of a bare noun + interrogative adverbial. Examples are of extreme rarity, but the following instances after interrogative and conditional rx may be quoted: (Sin B 126–27) Sinuhe expresses his belief in preordination with a rhetorical question: in iw nTr xm Sat.n=f rx nt pw mi m Is god ignorant of what he has ordained or aware of what the facts are like? (Admonitions 5, 3) The sage quotes the words of a hypocrite: ir […]=i rx.n=i nTr tn kA iry=i n=f If I… knew where god is, I would serve him.
82
chapter two
are instances where some contextual factor cancels the speaker’s ability to assert. The following examples, which similarly represent indirect questions, are particularly instructive: (78)
Neferty describes Egypt in turmoil:
iw ra iwd=f sw (r) rmT wbn=f wn wnwt nn rx.tw xpr mtrt Ra separates himself from mankind; he rises in due time, but one will not know whether noon has occurred.80 (Neferty XI d–e)
(79)
An instruction on determining the effects of an ulcer on the health of a bull from the discharge:
siA=k snb=f Hr iw HsA You will be able to perceive whether he is healthy on the basis (pUC 32036, 1.x+14) of how semi-solids emerge. (80)
An instruction concerning the treatment of a certain type of skullfractures:
srwx=f pw Hmst iri n=f mkAty nty Tbt r rx=k spr=f r xt His treatment is rest (lit. ‘sitting’); provide two brick-supports for him so that you may establish whether he attains critical (pEdwin Smith 2, 7–8) condition.81 Although the complements here are not from verbs with mutable roots, it is nonetheless noteworthy that no ntt/wnt is employed, as is the clear relative past sense of xpr in example (78). In this said example the speaker’s key idea is that in the circumstances described, nobody will know whether or not the noon has arrived. In examples (79)–(80) the texts describe conditions in which it can be ascertained whether some state of affairs is the case and procedures to be carried out to (Note also the use of sDm.n=f in a counterfactual sense in the second example and see 7.1 below). The nt pw in the first example is a crux. Seeing that it occurs also after the verb gmi after which ntt/wnt are not licensed, (2.4 below) nt cannot be an abbreviated form of ntt, but neither can it be a lexical expression ‘what exists/fact’ serving as a predicate of pw since mi m could then hardly be accommodated in the sentence. 80 Somewhat similarly Faulkner in Simpson 1972, 238 and Parkinson 1997, 138; Allen (2000, 376) has “that noon has happened”. 81 Following Wb IV, 102, III; the same expression appears also in pEdwin Smith 2, 10; pEdwin Smith 2, 23; pEdwin Smith 3, 8; pEdwin Smith 3, 15; pEdwin Smith 4, 9; pEdwin Smith 8, 9 and pEdwin Smith 12, 18–19.
affirmative object complementation
83
verify whether some situation results, but in both instances the latter are mere possibilities among various others. A somewhat later text provides an example of this with the distinctive form msy:82 (81)
Title of a birth-prognosis:
ky mAA msy st nn msy=s Another (method) of ascertaining whether a woman will give (Berlin medical papyrus rto. 2, 2) birth or not. In none of the above instances does the ‘speaker’ indicate any commitment or certainty regarding the situation described in the subordinate clauses, which are consequently not introduced by ntt or wnt and whether/if is used in translation. It thus appears that in object complement clauses of cognition and perception verbs, the employment of ntt/wnt follows the same modal parameters as after Dd. These elements are used whenever the conditions for speaker assertion of the complement are fulfilled—primarily when the information therein is vouched for through knowledge and commitment.83 When these prerequisites are not fulfilled for whatever reason, the complement is un-introduced and modally irrealis. In the latter case the morphological evidence is rather weaker with verbs of cognition and perception than with Dd, but weak verbs use non-geminating sDm=f, including forms with the ending -y. Again these complement types do not stand in any mutual substitutionrelationship, and the role of the speaker dominates the grammatical coding. Incorporating the data from examples with Dd, the respective ‘positions’ of the real speaker and the subject of the governing clause are charted in the diagram below. It can be seen that the use or non-use of ntt/wnt reflects the knowledge, acceptance etc. of the real speaker in every instance:
82 As pointed out in WGMT § 204, the relation of the negation nn msy=s and the verb mAA is unclear. 83 As discussed, speaker acceptance plays only a minor role here. For the third parameter involved, relevance, see next section.
84
chapter two
asserted / assertable Main clause speaker +
Complement speaker subject + +
→ ntt/wnt: “You said that X”
-
+
-
→ ntt/wnt: “If only someone said that X”
-
+
-
→ ntt/wnt: “Do you see that X?”
-
-
+(?)
→ no ntt/wnt: “If you have said X”
-
-
-
→ no ntt/wnt: “I don’t know whether X”
+
-
+
→ no ntt/wnt: “He said he would fight with me”
2.2.2 The ‘Evaluative’ Use of Irrealis In the instances above the decision between assertion and non-assertion has been conditioned by the speaker’s possibilities and willingness to express commitment towards the complement proposition. However, languages use modality also as a discourse-organising device to indicate degrees of information relevance. As seen, negligible information-value is one of the basic reasons for non-assertion. Also in Earlier Egyptian complement clauses, irrealis and realis are mobilised to signal differences in the relative salience of the information conveyed as evaluated by the speaker. The principles of this use of modality are apparent in objectclauses after verbs of cognition and perception. Here both geminating and non-geminating sDm=f’s and their functional counterparts occur extensively in opposition to ntt/wnt, and indicate that the complement proposition consists of information that belongs more to the discourse background. To understand the motivations for assigning these values to complements it is necessary to take into consideration the wider context, both the immediate sentence frame and beyond. Purely numerically, of all morphologically revealing object complement clauses after verbs of cognition and perception, those with the geminating sDm=f are most common, and this use of the form has prompted many a syntactic and semantic hypothesis. Of the former, one may mention e.g. the analysis by Loprieno according to which the form after the said verbs signals a particular syntactic conversion
affirmative object complementation
85
of “substantivization of a jw-sentence” having taken place.84 Thus, e.g. in an example such as (82)
The king tells his henchman of the high regard in which the latter is held:
rx.n(=i) Hm mrr w(i) ra Hr rdit=f n(=i) Tw I truly know how Ra loves me because of his giving you to (Urk I 180, 7) me.85 the underlying form of the sentence would be †rx.n(=i) [iw mr wi ra]. However, as with the ‘thetic’ hypothesis of ntt, there does not seem to be any principle restricting the use of such ‘embedded iwsentences’ to the ‘object-slot’ of verbs of perception/cognition only, or, conversely, prevent them from following other predicates.86 The derivative mechanism postulated must similarly be viewed with some suspicion, particularly as it seems to be at work only in complement clauses.87 Furthermore, bare geminating sDm=f object complements after verbs of cognition and perception are often best translated in a manner never applicable to either iw- or ntt/wnt-clauses. Rather than informing that something is known or perceived—as is always the case in ntt/wnt-introduced object complements—often the most suitable rendering of a complement geminating sDm=f after the said verbs corresponds to a description of the character (manner, degree etc.) of the complement situation.88 For example:
84
Loprieno 1991a, 214 and particularly n.55; cf also Gilula 1971, 16. Similarly in Urk I 62, 12 (see Brovarski 2001, text figure 2). 86 See also Uljas 2000, 129. Loprieno analyses the use of the geminating sDm=f after verbs other than locution, cognition and perception to signal a different clausal conversion; see 3.1 below. 87 Indeed, it may be asked should clauses with a bare initial geminating sDm=f, that is, second tenses, be similarly analysed as somehow derived from iw-sentences in turn? 88 Contra Allen 1984 § 231. Further examples after mAA are Urk I 39, 1 (irr.t(i); Urk IV 267, 17 (mss sw N). CT I 169f shows iw=f and pRamesseum I, Bii, 10 the interesting form iy of the anomalous verb ii, ‘come’. There are no certain examples of the geminating sDm=f after siA or sDm in the corpus studied for the present work. A possible instance of the latter is CT VII 32c/T1C, but whether the prr=f is a complement of the preceding sDm is unclear. However, examples occur in the PT; e.g. PT 1775b sDm. n=sn dd N mAat [m st izft] “they have heard how king N places right in the place of injustice”. Further examples after rx not quoted below are Urk I 61, 14 (mrr(=i); Urk I 63, 11 (mrr(=i); Urk IV 1270, 1 ([A]bb=i, restored); Urk IV 1673, 9 (wnn=sn); Urk IV 1676, 12 (xaa N); Urk IV 1830, 10 (Haa=f); Urk IV 1833, 9 (wnn=f). 85
86 (83)
chapter two A caption of a scene depicting a man delivering a cow:
mAA mss Hmt Monitoring how a cow gives birth (84)
(Deir el-Gebrawi II, pl. 28)
A magician suggests a form of entertainment guaranteed to delight the king:
Hwy A wDA Hm=k r S n pr-aA aws apr.n=k bAw m nfrwt nbt nt Xnw aH=k ib n Hm=k r qbb n mAA Xnn=sn Xnt m xd m xnt I propose your majesty proceeded to the lake of the Great House l.h.p., having equipped a barque with all the pretty ladies of your palace; your majesty’s heart will be refreshed through seeing how they row hither and thither. (pWestcar 5, 1–5)
(85)
Ahmose tells of his valour in battles against foreign foes:
nn tnwt m sqr-anxw in.n Hm=f m nxwt=f ist wi m tp n mSa=n mA.n Hm=f qnn=i Countless were the living captives whom his majesty brought from his victories; I was at the helm of our army, and his majesty (Urk IV 9, 14–16) saw how brave I was. (86)
The king tells his correspondent:
iw Hm Hm(=i) rx mrr=k Dd xt nb mrrt Hm(=i) My majesty most certainly knows just how you love to say (Urk I 180, 1) everything my majesty desires. (87)
Irtysen describes his skills as a sculptor, including:
iw(=i) rx.kw... dgg irt n sn-nwt=s I know… how an eye looks at its fellow. (88)
(Louvre C14, 9–10)
Sinuhe relates how he gained his host’s respect through warlike exploits:
Ax.n(=i) m ib=f mr.n=f wi rx.n=f qnn=i89 He came to see me as useful and grew to love me, having learnt (Sin B 106–07) how brave I was. 89 The second n is written after the determinative; the H-version (pHarageh I) has it correctly before. Ashmolean Ostracon has qnn{n}.
affirmative object complementation (89)
87
Ptahhotep advices on successful strategy in debate:
ir gm=k DAisw m At=f m Hwrw n-is mitt=k m Ad ib=k r=f rx.ti Xss=f If you come across a disputant making his case, but one who is a dilettante and no match to you; do not show hostility against him even when you know how feeble he is.90 (Ptahh 74–76) (90)
Instruction concerning the use of certain ingredients in preparing a potion:
ps Hr qd r mnx rx=k pss=sn Hr axx mw iry Hr Sww=sn Boil completely and thoroughly. You may tell how cooked they are on the basis of how water evaporates and how dry they are.91 (pEdwin Smith 21, 17–18) (91)
A further instruction following the previous one:
ir isk st qb rdi.xr.tw ø r anDw r iat st Hr itrw iaw.xr.tw ø r mnx rx.tw iaa=sn Hr dp.tw dpt nA.n mw nty m pA anDw nn dhr nb r=s Now, when they have cooled off, they are to be put into an anDw-jar so that they can be washed in the river. Then (they) have to be washed thoroughly. One may tell how clean they are by tasting the taste of the water in the said anDw-jar: there should be nothing bitter therein. (pEdwin Smith 21, 19–22, 1) (92)
Queen Hatshepsut claims to have abided to Amun’s will at all time:
n mh=i Hr sp n Sat.n=f iw Hmt=i rx.ti ntrr=f I did not neglect a matter of his ordering; my majesty knows just how divine he is. (Urk IV 363, 5–6)
90
So pBM 10509; the version in pPrisse has xft Xss=k. Rather similarly in pEdwin Smith 22, 3, but the subject of pss is omitted. Reintges (1997, 381) analyses the pattern in this and the next example as ‘intensive stems’. 91
88 (93)
chapter two Ramose extols his virtues before Osiris:
iw iri.n=i mAat tp tA iw=i rx.kw Hss=k mAa-ib tm irt spw n Dat I behaved justly upon earth, for I know how you favour righteous-minded who do not commit acts of evil. (Urk IV 1776, 14–16)
However, the apparent suitability of this translation here cannot arise simply from the mere simultaneity of the governing verb and the complement situation or the generality of the latter.92 There are examples after rx where the geminating sDm=f is neither generic nor simultaneous with the main verb,93 and, conversely, ntt/wntclauses may also describe simultaneous or generic instantiations. The apparent suitability of translating ‘how’ here formed the cornerstone of Callender’s thesis of all ‘nominal forms’ as actual ‘manner nominalisations’ specialised to indicate “the way an action is performed, or the circumstances under which it occurs”.94 Callender’s theory has not received wide acceptance among Egyptologists, and for good reasons.95 The ‘manner’ sense does not suit non-geminating sDm=f complements and its applicability even with the geminating form is often a matter of degree.96 Thus the sense is clearly not an 92 Nevertheless, there is no doubt that this is indeed often the best way of rendering geminating sDm=f complements of the above sort. For example, translating example (84) as “that they are rowing” would seem to miss an instance of Ancient Egyptian humour. Certainly the main attraction for the male king is watching how the semi-naked girls row, rather than ‘that’ they do it! 93 See example (100) below. 94 Callender 1975, 47–48, 72–77. This idea is anticipated in Assmann 1974, 63 and Schenkel 1975, 56. See further Callender 1977, 306 n.3; Depuydt 1983, 36–37. 95 Satzinger (1987, 620) dubs Callender’s thesis ‘unfortunate’ but against his later claim that the sense arises from the ‘nominal form’ being embedded in an ‘Adverbialsatz-Konstruktion’ (1993, 205 n.27) one may refer to the arguments concerning Loprieno’s theory discussed above. Junge (1978b, 32–33) appeals similarly to the role of the geminating sDm=f in a ‘subject position’ of an adverbial sentence, but his statement that the concept ‘manner nominalisation’ alone does little to distinguish the patterns used after verbs if applied to all forms appearing there without taking into consideration time-reference etc. is sound. The same holds also for his remark that the term ‘manner nominalisation’ suggests a curious hybrid between syntax and semantics whose precise nature is difficult to grasp. For example, Callender speaks of what he calls ‘prospective’ as a “nominalization of the optative mood” (1986, 10). This begs the question is ‘optative mood’ a ‘manner’ and how can it be ‘nominalised’. Polotsky (1984, 119) notes the impossibility of ‘manner’ reading in second tenses where the vedette is not a manner adverbial. 96 See Uljas 2003, 396 n.52.
affirmative object complementation
89
inherent property of any of the bare suffix-conjugation forms found in complement environments. Callender’s hypothesis confuses what is actually a mere translation strategy of one form in one syntagmatic environment with its wider functional profile.97 Yet, the said translation here is nevertheless a manifestation of the more general character of the geminating sDm=f, and more particularly of its value as expressive of a standard type of irrealis modality.98 Several commentators have noted that the difference between the geminating sDm=f ‘how’- and the ntt/wnt-introduced ‘that’-translated complements is that communicatively it is not the situation per se that is at stake in the former case, but merely its internal character.99 This notion not only provides a fine characterisation of the pragmatic contour of these complements. It also reveals them as examples of the phenomenon of using irrealis modality to indicate that the information presented is considered as background and low in relevance. In this instance, this information is the actuality of the situation predicative nexus. From a discourse-perspective, nothing hinges on the reality of the complement situation in e.g. “his majesty saw how brave I was”, (example (85) above) which is uncontested information taken for granted and already assumed in the context. That is, the speaker’s interest centres rather on making a judgement and assessment on the degree and character of the bravery, and in this [I was brave] represents merely a necessary point of departure. Any subsequent discussion would centre on how brave the speaker truly was, rather than on the acceptability and reliability of his claim of actually having displayed such qualities. This contrasts sharply with e.g. “his majesty saw that I was brave” where the issue presented for attention and comment is precisely whether [I was brave] is true or not. The information communicated by bare non-geminating sDm=f complement propositions above thus approaches presupposed, which occasionally is very nearly at issue, e.g. due to the nature of the preceding co(n)text: 97 However, this does not imply that the same strategy cannot be used to translate the geminating sDm=f elsewhere: as noted by Polotsky (1984, 119) and Borghouts, (1985, 35) it is often quite suitable after prepositions with a ‘correlative’ sense—see 6.2 below. 98 The hypothesis presented here appears in an embryonic form in Uljas 2000 and 2003, 396–97. 99 See Borghouts 1985, 37, particularly n.34; similarly Reintges 1997, 107 and already EAG § 494–95.
90 (94)
chapter two From a dialogue between the deceased and a god:
wDa n(=i) sbA pry(=i) im=f di mA=i prr=k di mA=i aq[=k] D: A gate has been opened for me that I may exit through it. G: Let me see you exit; let me see you enter. (CT VII 228k–l) In this and the examples above, the geminating sDm=f thus signals that the actuality of the complement situation, or the predicative nexus between the verb and its subject, is not intended as the communicative focus of what is said. Here, then, we have the other basic cross-linguistic motivation for irrealis modality besides lack of commitment in complement clauses after verbs: the relative lack of optimal discourse relevance of the information presented.100 This sort of ‘evaluative’ use of irrealis/ non-assertion is a wide phenomenon both typologically and crosslinguistically.101 As seen, “speakers do not have to assert assertable information”102 if e.g. they disagree with it, but also if they assess it to be of limited interest and usefulness to the hearer. In many languages irrealis can be used when there is nothing ‘objectively’ barring assertion and the ‘unwillingness’ to do so does not relate to rejection or nonacceptance of the information but rather to speaker assessment of it as somehow lying “at the bottom of the scale of ‘relevance’”.103 In Earlier Egyptian the geminating sDm=f after verbs of perception and cognition is similarly a mode of refraining from asserting for this reason. The unquestioned ‘reality’ of the situation in such instances is less than optimally relevant background. For such information languages typically employ “forms associated with a lower degree of assertiveness, and even forms designated as irrealis”.104 Also the non-geminating sDm=f forms are mobilised for this general 100 The word ‘optimal’ signals that in communicative terms the information is located, as it were, somewhere on the continuum between maximal relevance and total irrelevance; see Lunn 1989a, 251. 101 See Wallace 1982, 209; Lavandera 1983, 232–33; Lunn 1989a, passim and 0.1.2 above. 102 Lunn 1989b, 695 (emphasis in the original); see chapter 1 and Lunn 1995, passim. 103 Lavandera 1983, 231, referring to Spanish subjunctive. 104 Hopper 1981, 216 (italics by SU).
affirmative object complementation
91
function, and as such play a very significant role in object complements after many notionally non-assertive verbs.105 However, after verbs of cognition and perception, mutable forms occur only in ‘future-inpast’ clauses after rx, and nearly all these examples show the same complement verb iri:106 (95)
Wepwawetaa says he never failed to satisfy royal expectations:
xd.n=i xnt.n=i m Xnw rx.n=i iri=i HAw Hr mdwwt Whenever I travelled north or south from the residence, I knew I would exceed what had been asked. (Lesestücke 72, 20–21) (96)
Amun expresses his satisfaction with Queen Hatshepsut’s performance as a ruler:
ist Tn m imt sS=T rx.n(=i) iri=T n(=i) mnw When you were still in your nest, I knew that you would make monuments for me. (Deir el Bahri pl. 131, right) (97)
Having poured out her woes, the author of a Letter to Dead turns to her addressee:
srs it=k iy r bHzti Tz Tw wn Tw r=f i(w)=k rX.t(i) ii(=i) n=k aAy Hr wDa-mdw Hna bHzti AAi sA a-nanxi Awaken your father Iy against Behezti; rouse yourself and hurry against him! You must have known that I would come to you here for the sake of litigating against Behezti and Aai’s son Anankhi. (Cairo Linen 9–10) Here the speakers report (and perhaps make assessments) of their own or someone else’s past commitment to some then future situation. The complements are future vis-à-vis the main clause state of affairs 105
See 3.2 below. Further similar examples are White Chapel 170 (rx.k(i) iri=k); Urk IV 1293, 7 (rx.n=f it=i); Urk IV 1425, 13; (rx.n=f iri=i); Urk IV 1468, 1 (rx.n=f iri=f) and Urk IV 1579, 13 (rx.n=f iri=f). See also example (106) below. In example (97) the reading of the penultimate word as ii(=i) (cf. Gunn 1930, 149) seems better that analysing it as a participle “the one who comes to you” (Gardiner & Sethe 1928, chapter 1, 1). If the former reading is adopted, the form contrasts with writings such as the iy noted in n.88 above and showing two i-reed-leaves. In this example also, a translation as epistemic necessity “you must have known” seems to suit the overall sense particularly well. 106
92
chapter two
and there is no temporal overlap between the two. This will be seen to be a characteristic temporal feature of non-geminating sDm=f complements, but it is not the sole motive for the choice of form, nor for the absence of ntt/wnt, and thus for non-assertion. This is rather the lack of discourse relevance of the complement propositions in the current speech situation. The latter does not equal the time of the main verb, but again refers to the same point of perspective from which all modality is assigned, namely the actual time of speaking, i.e. the ‘here and now’ of the real speaker. Due to the inherent factivity of the verb know, future-in-past complement situations of this verb must have occurred, and are known to have occurred, prior to the time of reporting.107 That is, e.g. “I knew that you would make monuments” entails “you made monuments”, which is a non-issue and already clear by default to all participants involved in the discourse context in which the report is made.108 Thus, in examples (95)–(97) above the subordinate clauses are future relative to the main verb, but in the current speech context they also contain no newsworthy information whatsoever. For this reason they remain non-asserted and non-geminating sDm=f forms appear. The same motive for the use of irrealis as in examples (95)–(97) above pertains also to the following instances of negated perceptionverbs with immutable, but, notably, ntt/wnt-less complements: (98)
Amenemhat I reflects posthumously on his assassination to his son:
mk stAw xpr iw=i m-xmt=k n-sDmt Snyt swD(A)=i n=k See, the mishap took place when I was without you and when the entourage had not yet heard that I would hand over to you (the kingship). (Amenemhat VIIIa–b) 107
See chapter 1 n.20 and n.30 above for ‘factivity’. This is clear from the impossibility of cancelling this entailment: †“I knew that I would do it but I didn’t”. The same does not, of course, hold with verbs such as e.g. ‘order’ or ‘say’: “I said that I would do it, but I didn’t” and “He ordered me to do it, but I didn’t” are perfectly possible, since ‘say’ and ‘order’ do not in any way entail that the thing said or ordered was realised somewhere between the time of saying/ordering and the time of reporting. Similarly, the slight strangeness of “?I knew that I would do it and I did it” results from the obviousness of [I did it]. Additional confirmation of the latter is communicatively redundant and it could be said that there is never any need to further verify the realisation of the situation in the current context of discourse. As will be seen in chapter 3 below, this constitutes one of the key uses of bare non-geminating sDm=f complements. 108
affirmative object complementation (99)
93
The deceased says that his rebirth is a mystery even to the divine:
mA.n wi nw xpr.k(w) n rx=f bw xpr.n=i im n mA=f xpr=i m Hr=f Nu has seen me after I came into existence, but he does not know where I came to be because he did not see with his own eyes that I had come into existence.109 (CT I 334/335a–c) Due to the negations, in these examples there is no similar ‘automatic’ entailment involved as in (95)–(97) above. But again the speakers need not phrase the complement situation descriptions as optimally ‘informative’ to their audience if the latter can be assumed to be already quite familiar with them. This can also depend on the context. In example (98), at the time of speaking it is of course well known to the addressee (Senwosret I) that the transition of government took place, seeing that he himself was the person upon whom it was bestowed.110 In example (99) the complement xpr=i is past in relation to mAA, with the speaker reporting that the matrix verb subject did not personally witness the ‘coming into existence’. Again, this shows most clearly that relative futurity time-reference is not the sole factor conditioning the choice of complement construal. Yet, that the event described took place is obvious from the context to everyone involved: the speaker, the subject of mAA and the audience. There is thus no need to assert and, consequently, no ntt/wnt introducing the complement. For the present analysis, examples (95)–(97) above with nongeminating sDm=f forms are of some importance. In these examples the ‘irrelevance’ of the complement information arises from a ‘logical’ necessity and is not based on the decision and assessment of the speaker, as is the case with the geminating sDm=f. The situations described are, of course, currently both accepted and realised— or better, completed.111 However, exactly because of its ‘automatic’ 109 So the variants S1C, M3C, M5C, M20C, M28C and M-Ann. B1Bo and B3C have n mA.n=f xpr=i/N tn whereas T3C, G1T, A1C and BH2C have understood xprw ‘forms’. M4C, B1C and B2C have xpr.n=i. See 7.1 below for further discussion of this example. 110 An alternative translation could be “whether I would hand over”, which leaves these matters open, but can be used in similar circumstances when there is no need to assert. 111 This issue and its significance to the modal profile of the non-geminating sDm=f forms will be discussed in 9.2 below.
94
chapter two
obviousness, as information their ‘reality’ is also maximally lacking in discourse relevance. Accordingly, the use of non-geminating forms in (95)–(97) does not represent a violation of their functional profile proposed earlier, quite the contrary. It shows that just as complements with bare forms of this sort are associated with the lower degree of acceptance and commitment of the situation described vis-à-vis the bare geminating sDm=f, the mutual relationship of these two types of Earlier Egyptian complement non-assertion is identical also in what pertains to the relevance of the information conveyed. The role of nongeminating sDm=f complements as descriptive of maximally ‘obvious’ information is relatively marginal after notionally assertive verbs. However, it will be seen to be a major feature of their use elsewhere, particularly in clauses after the verb rdi, ‘give, cause’.112 In sum, all the ntt/wnt-less complements after verbs of cognition and perception discussed in this section represent information and “describe ‘states of affairs’ that could easily be denied or affirmed, but is instead left unasserted”113 due to lack of need to assert. They do not ‘substitute’ for ntt/wnt-clauses, nor does e.g. the geminating sDm=f simply describe ‘indicative’ or ‘objective fact’.114 Yet, as seen, the speakers’ decisions as to how much and what sort of information they think the audience requires are mostly based on highly subjective criteria.115 The marking of complement propositions as realis or irrealis provides merely a general sign “about the presence or absence of ‘assertiveness’… whereby the hearer is instructed on how to interpret the content expressed in a message”.116 For this to succeed “it is necessary to know quite a lot about the discourse context”, as “the sentences themselves do not
provide enough context to explain these choices”.117 As an illustration of this complexity, one may consider the following excerpt from the CT, which also furnishes a rare example of wnt after a verb other than one of locution, cognition or perception:118 112
See 3.2 below. Lavandera 1983, 211. 114 Contra Gilula 1971, 16; Allen 1984 § 364; Doret 1986, 23, 40, 41, 49; Malaise & Winand 1999 § 908. 115 Cf. e.g. 0.1.2 above for the Spanish subjunctive. 116 Lavandera 1983, 232. 117 Lunn 1989a, 250. 118 See Griffiths 1960, 52–53; Faulkner 1968; Polotsky 1969, 481; Gilula 1971 and Allen 1986b, 25. The interpretation here is rather similar to that by Griffiths, Gilula and 113
affirmative object complementation
95
(100) Isis has claimed that her unborn child is ‘Osiris’ seed’ and divine; a dialogue ensues:
ihi i.n ra-tm sA ib=T Hmt i.rx Tn rf mi iSst nTr is pw nb iwa psDt... ink Ast Axt Spst r nTrw iw nTr m-Xnw Xt=i tn mtwt wsir pw Dd.in ra-tm iwr=T sdx=T Hnwt pw mss=T iwr=T r nTrw wnt mtwt wsir is pw imi iw rqw pw smA it=f sD=f sw(H)t m-Xnw nxn=s “Well”, said Ra-Atum; “Be prudent, woman! How do you know if he is a god, lord and heir of the Ennead?” “I am Isis, more spiritual and noble than (other) gods; a god is inside this my womb, and he is Osiris’ seed”. Then said Ra-Atum: “Since you are pregnant, young lady, you should conceal from the gods that you are pregnant and about to give birth, and that he is Osiris’ seed, lest that enemy who would even slay his father come and break the egg in its early (CT II 215b–217g) stage”. Here the variation between the introduced and un-introduced complements is again based on commitment and the evaluation of information salience. Ra-Atum asks how Isis can know that her unborn child is of divine offspring, and she answers this.119 He then suggests that she conceal her pregnancy and delivery, and chooses this incontestable state of affairs as his starting-point “since you are pregnant”. When this presupposed situation, and Isis’ imminent delivery occur again in the complement, they appear un-introduced/as geminating sDm=f for the same reason.120 It is noteworthy that the situation mss=T ‘you give birth’ is not ‘generic’ or ‘simultaneous with the main verb’, but lies still in the (apparently immediate) future. The geminating sDm=f seems to have been chosen instead of a perhaps more ‘expected’ nongeminating one due to the different degrees of information relevance Polotsky. Faulkner (ibid, 40 & 42 n.15) translates wnt as “seeing that(?)”, viewing it as being used “after the manner of n ntt”. 119 Gilula (1971, 17 n.14) assumes the first utterance to be Ra-Atum questioning other gods “on Isis’ side.” This seems unlikely; more probably, and as assumed by both Faulkner (1968, 42 n.11) and Griffiths, (1960, 53 n.1) the question is addressed to Isis, although the pronoun Tn is strange. 120 The presupposition does not arise from the verb ‘conceal’; matters ‘concealed’ can be asserted.
96
chapter two
expressed by these two alternatives. That is, modality appears to have guided the complement selection over considerations of tempus. But the divinity of the unborn child is not similarly taken to be obvious and accepted by everyone involved—in fact, it was initially subject to doubt by the first speaker (Ra-Atum). Notably, wnt does not appear in “how do you know if he is a god” when the speaker was not yet committed to the complement proposition, but only when he has accepted it in “you should conceal… that he is Osiris’ seed” (=god).121 Both these complements also involve nominal sentences, which shows clearly that the use of ntt/wnt does not depend on the following construction, but on what is or is not acceptable as, and worthy of assertion. Deducing the speaker’s motives for the complement form choices and following his intentions requires meticulous attention to the context from the audience. Yet, this is an inseparable part of interpreting modality; modal expressions are not self-contained ‘meaning-chunks’ to be studied in isolation. The ‘evaluative’ use of irrealis in Earlier Egyptian is the context-sensitive class of employment par excellence and ultimately serves the same discourse organising and information structuring function as in all other languages. There are also other such strategies available in complementation. In the following example the geminating sDm=f appears in variance with ‘object-raising’, where the complement clause actor is both syntactically and pragmatically ‘foregrounded’ as the primary object and first-order focus of interest, followed by an oblique object describing the action in which it is engaged:122
121
Cf. Griffiths 1960, 53, who notes “Re-Atum’s doubt appears to be removed”. See Givón 2001 vol. 2, 272; Collier 1991a, 48–49; Uljas 2003, 395. ‘Object raising’ is in Earlier Egyptian restricted to verbs of perception involving visual contact between the governing clause subject and the complement object, namely mAA, siA and xpi ‘encounter’. Other verbs of perception such as sDm do not allow this construal. Also rx and gmi ‘find’ belong to the ‘raising’ group, but with the latter the situation is rather more complex; see 2.4 below. The verbs rdi and iri may also be followed by object + adjunct (GEG § 84) when direct manipulative contact is suggested (see 3 n.52 below). In PT there are examples that look like ‘raising’ after wD ‘order’, see Allen 1986b, 15. 122
affirmative object complementation
97
(101) The deceased says of hostile serpents in the underworld: S1C,123 B1Bo, M23C, G1T: iri=sn n=i wAt nfrt mAA=sn prr=i m kAr S2C, T2C:
iri=sn n=i wAt mAA=sn wi pr=i m kAr They make (a good) way for me when they see how I come out of the shrine/me coming out of the shrine. (CT I 390c–91a) This does not indicate e.g. some mutual syntactic ‘equivalence’ [prr=i + adverbial] and [wi + adjunct clause]. Instead, both the use of irrealis and ‘object raising’ alter the shape of the information conveyed in the complement. They reflect the speaker’s perspective to the situation described and his decisions on what to highlight and what to leave to the background in the process of structuring the information communicated. 2.3 Interim Summary Evidence of assertion with ntt/wnt is sparse after verbs other than locution, cognition and perception. Apart from sdx in example (100) above, there is only a single instance of an ntt-complement after some other predicate in the corpus studied, namely swD, ‘inform’: (102) After a directive outlining offences for which individuals are entered to the criminal dockets, it is decreed:
[ir iw] sp=sn ky sp xr.t(w) smi.t(w) swD.t(w) ntt st Hr Sfd n xbnty sSr mdwt wAH st Hr=s Hr pA Sfd r DAt qnw=sn [Should] a case involving them [arise] again, then a report is to be made and it is to be informed that they are on the criminal docket, with a statement on the matters because of which they were entered on the said dockets, in accordance with their offence. (Urk IV 1109, 5–8) However, this is presumably merely a matter of survival; there is one 123 This is the variant followed; the versions differ slightly in their phrasing of the passage.
98
chapter two
possible instance of the verb ib ‘think’ followed by iwt, the negative equivalent of ntt/wnt.124 In the PT there is also an example of nttclause after the verb sr ‘advice’,125 which further confirms that the use of these elements is not restricted to complements of Dd, rx, mAA, sDm and siA.126 Indeed, it would be strange if this were so, seeing that the pragmatic parameters determining the use of object complements introduced by ntt/wnt and of those without these elements are of very general character and may now be summarised. In Earlier Egyptian, the basic division between ntt/wnt-introduced and ntt/wnt-less complements is organised as follows (S = Speaker, H = Hearer, p = proposition and the information it conveys):127
ntt/wnt
gem. sDm=f/ non-gem. sDm=f/128 no ntt/wnt
It is possible for S to assert p
S is willing to assert p
→ S knows p
→ S believes p → S accepts p → S thinks p is relevant to H
It is impossible for S to assert p
S refrains from asserting p
→ S does not know p → S knows p is not relevant to H
→ S does not believe p → S does not accept p → S does not think p is relevant to H
128
Starting from the bottom half of the above diagram, non-assertion is an expression of the speaker’s ignorance, lack of commitment and acceptance of the complement proposition as well as of its known or perceived lack of discourse relevance. The overall formal signal of such 124
See example (123) below. PT PII, 1055 +30 sr n=k n ra ntt P [iw=f] “Advice Ra that king P comes”; Nt 40– 41 has changed this into Dd n=k n ra ntt N iw=s, “Say to Ra…” etc. 126 Anticipating the discussion in chapters 4 and 6, it should also be noted that the same elements are used to mark assertions also after prepositions and in subject complements of verbs. 127 Cf. here Searle 1969, 66. 128 The labels ‘gem. sDm=f’ and ‘non-gem. sDm=f’ should be understood as also inclusive of the functional counterparts of these forms of the ult. inf. verb-class, such as the doubling form of doubling roots in the first instance. 125
affirmative object complementation
99
a modal status of these clauses is the absence of the elements ntt/wnt and the use of bare sDm=f forms. However, in clauses involving verbs with mutable roots a further distinction is made. Complements with the geminating sDm=f, or some of its functional counterparts from verb-classes other than ultimae infirmae, describe situations that are both ‘objectively’ (mostly in terms of current ‘reality’) and/or ‘subjectively’ (accepted with commitment) ‘near’ to the speaker. Yet, asserting their reality by presenting this as optimally relevant information is not the speaker’s illocutionary intention. As opposed to its functional counterparts, which also have other values, the specialised or marked formal representative of this proximal irrealis is the geminating sDm=f. As a typological class the proximal contrasts with the kind of non-assertion expressed by non-geminating sDm=f forms. These demonstrably cluster to complement clauses describing situations that are both temporally separated from the speaker—i.e. are future or past—and whose assertion is most clearly blocked by the ‘objective’ factor of lack of knowledge. But lack of acceptance, rejection and denial, i.e. when p is most strongly ‘subjectively’ un-assertable to (and not just un-asserted by) S and where there is the greatest motivation for S to distance himself from p, seem also to be domains of non-geminating sDm=f forms. The same holds also for complements known to convey information most lacking in discourse relevance, whose assertion is not merely deemed unnecessary, but actually blocked. The marked forms of this distal irrealis are the sDm=f forms of mutable verbs displaying the endings -w and -y. Both these hierarchically organised categories of irrealis contrast with complements introduced by ntt/wnt, whose distribution is complementary to the former two. Semantic-pragmatically, in such clauses p is presented as information of whose reliability the speaker is prepared to take full responsibility, which he accepts and which is offered to the consideration of the audience as something new, relevant and worthy of their full attention. Grammatically, the status of such realis and asserted p is clearly signalled by the introducing element. Here it may be noted that just as the presence or absence of ntt/wnt is neither random nor based on syntactic criteria, the role of these elements as modal operators and markers of assertion provides a fine example of the principled manner in which grammaticalisation of such function-words occurs. The etymological origin of ntt/wnt in
100
chapter two
expressions referring to existence and reality is not coincidental. It has been postulated that ntt and wnt may have originally represented abstract lexical expressions meaning something like ‘what is/exists’, i.e. ‘fact’, serving as the actual object or subject of the governing verb and followed by a paratactic or appositional clause.129 Later they were probably reanalysed as functional elements forming a clausal unit with what follows in a manner known from other languages.130 If true, the later functional role of the erstwhile lexical ntt/wnt not only as ‘syntacticised’ clause subordinators but also as ‘pragmaticised’ indicators of realis modality seems perfectly natural.131 Nevertheless, the expression of realis is more restricted in Earlier Egyptian object complementation after verbs than irrealis. The speaker’s motives for non-assertion are extremely mixed and only one of them suffices to render the proposition unasserted. In contrast, assertion only takes place when all the relevant conditions and prerequisites are fulfilled and even then it is ultimately the choice of the speaker whether or not he wishes to profile the complement proposition thus. More common overall is distal or proximal non-assertion with the geminating and non-geminating sDm=f’s (or, more generally, without ntt/wnt) which is ubiquitous after nearly all governing predicates. Hence assertion is not the ‘default case’ in the modal system of Earlier Egyptian object complement clauses after verbs, but non-assertion.132 Atypically, but in accord with the previous point, asserted complements are also the more marked option, and non-asserted the less marked. Interestingly, this relationship between ‘markedness’ and assertion/non-assertion seems to have an exact parallel in main clauses. It is noteworthy that
129
See Goedicke 1955; LGEC §§ 701, 705 and Gunn 1924, 176–77. See Heine, Claudi & Hünnemeyer 1991, 205–06; Hopper & Traugott 1993, 14– 16, 185–89; Heine & Kuteva 2002, 211–12, 295; Cristofaro 2003, 96–97 for discussion and examples of such a process. However, it must be stressed that throughout the historical period ntt/wnt are clearly grammaticalised elements and are not be treated as lexical ‘objects’ or ‘subjects’. This honour belongs to the clauses that the elements introduce, following the assumed ‘re-bracketing’ of e.g. the hypothetical prehistoric *rx=f [ntt] [sDm.n=f] as rx=f [ntt sDm.n=f]. 131 Yet as is typical for such function words as ntt/wnt, some remnants of the original lexical use persist: bare ntt functions as a lexeme for “what exists/is”; cf. e.g. the expression ntt-iwtt “what exists and what does not” (GEG § 203.4). 132 As will be seen, this state of affairs finds an exact parallel also in the sphere of subject- and predicate complements as well as in complement clauses of prepositions, and is thus a general principle of all complementation in this language 130
affirmative object complementation
101
sDm=f main clauses with a seemingly ‘indicative’ profile are preceded by some auxiliary element such as iw, whereas clauses of most nonassertive sort (wishes, exhortations) appear un-introduced.133 The hypothesis above is—and could only have been—a result of an examination to the motivations for the use and distribution of the different Earlier Egyptian object complement patterns of verbs as a whole, and, in case of ntt/wnt-clauses, without regard to their internal build-up. However, as noted in the introduction, Egyptological researchers have hitherto preoccupied themselves rather with the question of what sort of forms and constructions are to be found after these elements. This issue must now be addressed, because closer examination of the paradigm after ntt/wnt actually supports the modal analysis of these elements proposed here. The phenomena to be observed allow one to define the modal character of the bare forms and constructions following ntt/wnt and provides an initiative for studying them from this vantage point also in other grammatical environments of Earlier Egyptian. 2.4 The Paradigm after ntt/wnt: Form and Function Revisited In addition to their role as markers of realis modality, the elements ntt and wnt also serve to identify the clauses they introduce as complements. Consequently, besides modal operators they may also be characterised as complementisers.134 Rather in agreement with the earlier Egyptological analyses of ntt/wnt, in linguistics the defining feature of elements assigned to this grammatical category was traditionally taken to be the syntactic function of a subordinator of sentential objects (and subjects).135 More recently, however, complementisers have been increasingly interpreted also as one of the means in which languages encode modality.136 For example, the variation in Japanese of the 133
See conclusion for further remarks on this issue. Barta (1986) argues that ntt/wnt and iwt introduce adjunct clauses in the grammatical idiom of New Kingdom Unterweltsbücher. No comparable uses are observable in earlier material. 135 Generative linguistics forms a notable exception here. 136 In Egyptology this point is most clearly expressed by Ritter 1995, 97–98, who speaks of complementisers as bringing “die innere Einstellung des Sprechers zu seiner Aussage (das sogenannte commitment of the speaker towards the truth value of the 134
102
chapter two
complementisers to and koto has been analysed as expressive of whether or not the complement proposition is presented as ‘psychologically distant’, i.e. untrustworthy, unacceptable, representing someone else’s beliefs, or seen from a ‘wholly objective’, ‘detached’ perspective.137 In Earlier Egyptian, the dual role of ntt/wnt as both syntactic and modal markers is most apparent from the distribution of the clauses that they usher in, but the same holds also with the paradigm that follows them. There are clearly but few restrictions on what forms and constructions can actually occur in this syntactic position. On basis of the examples cited thus far, at least the following patterns are combinable with ntt/wnt:138
The nominal sentence The adverbial sentence Subject + stative Subject + Hr + infinitive Subject + r + infinitive Cleft sentence Existential sentence Ante-position with/without ir + sentence sDm.n=f Active sDm=f with or without subject ante-position (ntt sDm=f/ntt=f sDm=f ) Past passive sDm=f n sDm.n=f
ntt can also introduce second tenses, although in complements after verbs no examples headed by geminating sDm=f are attested. Instances of sDm.n=f apparently in this function are late and far and wide apart:139 proposition) zum Ausdruck”. See also e.g. Palmer 1986, 148–49; Ransom 1986, 87– 91; Wierzbicka 1988, 164–65; Langacker 1991, 446–47; Frajzyngier 1995 and ibid pp. 474–75 for a bibliography on past research on this topic. Cf. also Givón 2001 vol. 2, 72–74. 137 Suzuki 2000; cf. also Palmer 1986, 148. 138 Cf. Johnson 1984, 81–83; Collier 1991a, 29–30 (on ‘preposition-ntt converters’); Uljas 2000, 130. In addition, the negation n sDm=f is attested in ntt-introduced complement clauses of prepositions—see 6.3 below. 139 Cf. Polotsky 1944, § 28 end. Allen (1979, 8) and Silverman (1985, 272; 1986b, 317) argue that the element is is required for second tenses to be able to appear as object
affirmative object complementation
103
(103) Thutmosis III is told that Amun favours him: rdi.n=f n=k tA nb xrp n=f sw rxw ntt pr.n=k xnt=f He gave to you all the land. Lead it for him, for he knows that you have come before him. (Urk IV 1241, 5–6) However, abundant examples of unambiguous second tenses with geminating sDm=f occur in ntt-clauses that serve as complements of prepositions.140 Contrary to claims made by various researchers when attempting to find a syntactic explanation for the paradigm above, it is obvious that the use of ntt/wnt does not depend on whether the construction following these elements is verbal or non-verbal,141 or inherently ‘initial’ or ‘non-initial’. 142 For Junge, ntt/wnt is required if the construals subordinated are to retain their status as sentences rather than clauses.143 However, this claim entails the assumption that e.g. a bare non-geminating sDm=f cannot alone constitute a ‘sentence’, a view to which Junge himself does not readily subscribe.144 Further, the occurrence of ntt/wnt with verbal clauses is not governed by whether or not the verb stands first in the following construal or, more narrowly still, in case of sDm=f, whether or not there is an ante-posed subject.145 Much ingenuity has also been applied, particularly among the ST, on attempts to segment the paradigm internally on basis of the assumption of ntt and wnt as ‘subordinate equivalents’ of the initial elements iw and mk, and their paradigms as somehow mutually correlative.146 complements, even when ntt is present (see 7.2). Yet, in view of the evidence from preposition complements, this claim seems to be incorrect (see 6.3 below). A further example of a sDm.n=f second tense in an object clause of a verb is Urk IV 1531 11/14, where the E-variant has rx.kw ntt wD.n.tw n=f st “knowing that it has been ordained to him” with a sDm.n.tw=f passive. The damaged A-version agrees, whereas the late XVIII dynasty variant D has ntt wD.tw. Wente (1969, 3 n.18) states categorically that the sDm.n=f of verbs of motion (henceforth VOM) is not used in Old Egyptian after ntt/wnt. 140 See 6.3 below for examples. 141 Frandsen 1975, 13; cf. also Allen 1986b, 33. 142 Silverman 1985, 272; 1986a, 38. 143 Junge 1979, 83–84. 144 See Junge 1989, 84; cf. also 3 n.21 below. 145 EAG § 1018. 146 Thus, for Polotsky (1969, 480–81) wnt was originally (in Old Egyptian) a counterpart of iw and followed by the same paradigm as the latter. ntt, compatible with all construals, corresponded to mk, with the later appearance of e.g. wnt-introduced nominal predicates being explained as reflecting the merging of functions of ntt/wnt. Gilula (1971, 16) added ntt to be confined mainly to “adverbial sentences… which are
104
chapter two
Sometimes these speculations have resulted in certain quite correct characterisations of the semantic-pragmatic properties of e.g. wntclauses, which have been dubbed ‘indicative’ due to their assumed parallelism with iw-sentences.147 In general, however, hypotheses of this sort have been plagued by numerous anomalies arising from the impossibility of dividing the patterns following ntt/wnt as the respective paradigms of iw and mk.148 Also where no such detailed divisions have been attempted, the idea of ntt/wnt as ‘syntactic counterparts’149 of iw and/or mk has nevertheless been often raised. This is somewhat surprising in view of the fact that syntactically these pairs have nothing to do with each other: iw and mk introduce independent initial main clauses, whereas ntt/wnt usher in dependent non-initial subordinate clauses. That is, the most basic syntactic characteristics of these two sets of elements are in fact polar opposites. The last point seems to have been recognised in the more recent accounts of ntt/wnt where no correlations with iw and mk are proposed.150 However, this has not disturbed the wider consensus on the syntactic character of the paradigm following the former two elements. As seen, the ST and post-Polotskyan approaches alike have embraced the assumption that it consists of patterns ‘converted’ or nominalised by ntt/wnt.151 One of the more persuasive arguments seemingly favouring this interpretation is the conspicuous absence of the bare geminating of the pattern of non-verbal sentences with iw”. wnt was used for this function in Old Egyptian, where it could also occur in “verbal constructions with iw”, but not in sentences “without iw” (cf. also Gilula 1978, 47). For Frandsen, (1975, 12) ntt subordinates non-verbal sentences with iw, whereas the particle is (see 7.2 below) serves as a subordinator of such sentences without iw. 147 E.g. Malaise & Winand 1999 § 912. 148 For instance, Polotsky’s hypothesis raises the old problem of how can the allegedly ‘adverbial’ ‘pseudo-verbal’ or ‘circumstantial’ patterns substitute for e.g. nominal sentences after elements such as mk/ntt (see Collier 1990a; 1991a). In Gilula’s case, Middle Egyptian examples of ntt preceding “verbal constructions with iw” (even excluding ‘pseudo-verbals’) are common, as are non-verbal sentences without iw (contra Frandsen). Second tenses also belong to this latter group in orthodox ST, but seem to follow ntt regardless. Johnson (1984, 81) argues that with ‘circumstantial’ patterns, iw is deleted after ntt, but still performs its function of assigning the forms/construals a ‘predicative’ value. It may be asked, pace Collier (1990b, 87) why should iw be deleted here if its function is so decisive? 149 Allen 1986a, 11; cf. also Satzinger 1986, 299, 307; 1989, 216. 150 Indeed, Malaise & Winand (1999 § 912) speak merely of a ‘resemblance’ between iw- and wnt-clauses. 151 See 0.1.1.
affirmative object complementation
105
sDm=f, the quintessential ‘nominal form’, from the paradigm. Yet, there is no need to assume that this restriction is syntactically founded. There are also other absentees from the paradigm, and these share features in common mutually and with the geminating sDm=f. It is notable that similarly missing from the set of forms and constructions after ntt/wnt is the thoroughly modal imperative. This constraint is unaccountable if the function of these elements is simply to ‘convert’ main clause construals into complements, particularly as there hardly is a verbal pattern less inherently ‘subordinate’ than the imperative.152 Further, there are no examples of any sort of interrogative constructions after ntt/wnt either. This again would represent a strange gap in the ‘converting’ facilities of the said element—particularly as formally many Earlier Egyptian interrogatives are in fact nothing but nominal sentences, which of course otherwise do occur after ntt/wnt.153 Yet, from the perspective of modality these omissions make perfect sense. Imperatives and interrogatives do not assert, but command and inquire, and this sets them into the same irrealis category with the geminating sDm=f, which as a group is incompatible with the realis-markers ntt/ wnt. The absence of the geminating sDm=f after ntt/wnt thus finds a straightforward explanation from the respective modal profiles of this form and of the said elements. Both indicate the modal status of the clause in which they appear: the geminating sDm=f that it is unasserted and irrealis, ntt/wnt that it is asserted and realis. Although speakers may withhold from asserting almost at will, the reverse is not possible.154 In Earlier Egyptian, the complementisers ntt/wnt are a device of modal marking, just as are particles and, in case of the geminating sDm=f, inflection. Marking a bare [hAA=f] non-assertion 152 However, it is not the case that imperatives can never be subordinated; e.g. in Late Egyptian they occur in adjunct clauses introduced by the subordinator iw. 1erný & Groll (1993, 352) note that in such instances iw “cannot be classified as a subordinizer”, which seems to contradict their own assertion “every initial main clause (both nominal and verbal) may be preceded by the dependent iw, which converts the initial main clause to a non-initial subordinate clause” (ibid. 422). 153 One thinks e.g. of sentences such as pty rxt=f “What is its amount?” (pRhind no. 46, 1) or sy ty pw “Who is he?” (CT IV 191c). 154 This asymmetry follows from the most basic principles of modality. For example, Lyons notes that “it may very well be the case that in the vast majority of the world’s languages it is impossible for a speaker to assert the objective existence of either epistemic or deontic possibilities” (1982, 112). The reason for this is obvious: a mere ‘possibility’ of something lacks precisely the properties that render that something assertable as a proposition.
106
chapter two
as an assertion with ntt/wnt is ungrammatical because the resulting complex would be modally contradictory and un-interpretable. The same holds, of course, also with imperatives and interrogatives, and not only in Earlier Egyptian, but in all languages. The bare geminating sDm=f does not need, nor allow ntt/wnt. On its own this form is capable, when required, of forming complement clauses, but the character of these is exactly the opposite of those with the latter elements. The relationship between ntt/wnt and the other types of active sDm=f is also modally based, but here again arise the same problems of morphology discussed in the introduction (section 0.2). According to some commentators, at least one of the forms encountered after ntt/wnt is what they call the ‘subjunctive’ or ‘prospective’ sDm=f.155 Yet, there is not a single example of these elements introducing a sDm=f displaying the endings -w and -y, which most researchers (including the present author) would probably accept as some sort of ‘prospective’ forms.156 Unless these morphological characteristics were for some unknown reason systematically omitted after ntt/wnt, this speaks strongly against the use of any ‘prospective’ form(s) after these elements. In addition, their inclusion in the paradigm would in effect demolish the argument that the latter consists of forms and patterns that need to be ‘nominalised’ by ntt/wnt, seeing that the ‘prospective(s)’ can and do occur as complements also without the latter. Arguing that the issue is not that of ‘nominality’ plain and simple but rather that the ‘converters’ introduce forms that are merely not specialised for nominal use does not remove this problem either.157 This would, of course, account for the ‘adverbial’ final ‘so-that’-clause use of the assumed no-ending ‘prospective’ or ‘future’ sDm=f, but would still not answer why should the ‘converters’ ntt/wnt be used with this form that nevertheless could form complement clauses alone and without their help. Similarly, the complete absence of forms with the endings -w and -y after ntt/wnt does not support the hypothesis of the presence of some ‘unspecialised future sDm=f’ any more than it does that of a ‘nominal prospective’ after these elements. Yet, as with the geminating sDm=f, their absence is not at all peculiar from the vantage point of 155 Johnson 1984, 81 and Collier 1991a, 29, on basis of examples after r ntt (see 6.3 below) and Collier, ibid, 29 n.35 specifically for ntt/wnt. 156 But cf. chapter 6 n.72 and see section 9.2 for discussion. 157 See Collier 1991a, 29–31; 1999, 54.
affirmative object complementation
107
the present analysis. Forms with these morphological characteristics are similarly specialised for non-assertion: they are marked for distal irrealis function through their inflection and are thus modally incompatible with the assertion-markers ntt/wnt. Other researchers have argued for the appearance of the alleged ‘circumstantial’ sDm=f, a truly ‘non-nominal’ form, after ntt/wnt.158 In support of this claim one could refer to example (58) above showing ntt iri=f, example (28), where ntt is followed by an anticipatory subject, (ntt wDa Ad=f wi) and the following instance in which wnt hosts a subject-anticipating suffix-pronoun in a fashion similar to iw:159 (104) A damaged text narrates the courtiers’ reaction to Wašptah’s sudden seizure before the king:
[...] Dd=sn xr Hm=f wnt=f dbAx=f […] with them saying to his majesty that he was unconscious. (Urk I 42, 11)
However, arguments against the entire category of ‘circumstantial’ sDm=f have already been put forth earlier,160 and these examples give no reason to revise them. As for the writing iri in example (58), this is not distinctive of any ‘circumstantial’ sDm=f but has exactly the same appearance as iri e.g. in examples (95)–(96) above. The former instance can just as well be interpreted as involving a ‘circumstantial’ or a ‘prospective’ and is useless as evidence if these terms are seen as referring to forms rather than functions. The problem is further exacerbated when ntt/wnt-introduced and ntt/wnt-less clauses with immutable verbs are contrasted: (105) Thutmosis III explains why he has built a temple for Amun: rx.kw ntt Htp=f Hr=s (Urk IV 835, 16) I know that he is content with it.
158 E.g. Doret 1986, 34 n.264; cf. also the works cited in n.146, n.148 and n.156 above. 159 Cf. Polotsky 1969, 481; one may also again cite Nt 40–41, ntt N iw=s and Urk IV 1381, 4 with r ntt TAty pH=f tni. 160 See 0.2 above.
108
chapter two
(106) Amun gives thanks to Thutmosis III for his building enthusiasm: rx.n=f Htp(=i) xnt=f He knew that I would be content before him. (Urk IV 883, 14)
Here there is no morphological (or phonological) support whatsoever for the claim that Htp in (105) is a ‘circumstantial’ rather than a ‘prospective’, ‘subjunctive’ or in general a different sDm=f form than Htp in example (106). The only visible difference is that in examples (58), (104) and (105), and unlike in (95)–(96) and (106), iri and Htp are preceded by ntt/wnt. This latter very ostensible signal is of decisive importance: it indicates that the clauses and the forms following the introducing elements have a particular modal value, and this value is assertion. In complementation the opposition between, say, ntt iri=f versus bare iri=f is that between the non-geminating sDm=f form of the verb iri in an assertive versus non-assertive (irrealis) function. Clearly this is not a matter of morphology since the verb itself is written iri=f throughout. The same holds, of course, also with immutable roots, whose function is similarly indicated by the presence or absence of the complementiser.161 As regards the anticipated subjects in examples (28) and (104), this phenomenon does not indicate the presence of any subordinated ‘circumstantial’ sDm=f form, but simply shows that with the active sDm=f in an assertive function, similar subject-anticipation as after iw can occur also after the elements ntt/wnt.162 Subject-anticipation is hardly a signal of any ‘circumstantial’ sDm=f after iw either. As is well known, iw-introduced sDm=f passivised with the element .tw do not readily show this feature,163 and the same holds (in texts predating early New Kingdom) for verbs of state and quality as well.164 Rather than form, subject-anticipation thus clearly depends on the semanticpragmatic profile of the situation described. If anything, the similarity 161
With the exception that, as seen, immutable roots do not distinguish between distal and proximal irrealis. 162 Contra the earlier arguments by the present author (Uljas 2000, 130). 163 This is regardless of which element actually represents the subject in such construals. In sDm.tw=f the subject may in principle be analysed to be either =f or an indefinite .tw (see n.15 above and 4.1 below for this question). Yet, there is no †iw=f sDm.tw=f at all and iw.tw sDm.tw does not occur in pre-XVIII dynasty Egyptian. 164 See Vernus 1984.
affirmative object complementation
109
in the positions of the subjects suggests that the pragmatic function of this ante-position might be the same after ntt/wnt and iw.165 But this does not indicate anything about the form of the sDm=f involved and is not an acceptable definition for the latter. In object complementation after verbs there is also one notable exception to the axiom that introduction by ntt/wnt is mandatory with the allegedly ‘non-nominal’ patterns. This exception is the verb gmi, ‘find/discover’ after which ntt and wnt never appear in Earlier Egyptian, and construals such as the bare nominal sentence (examples (107)–(109) below) and past passive sDm=f (examples (110)–(111)) function as objects without any introducing element:166 (107) The sailor encounters the magnificent serpent for the first time: kf.n=i Hr=i gm.n=i HfAw pw iw=f m iit As I unveiled my face, I found that it was a snake coming.167 (Sh.S. 60–62)
(108) Amenemhat I tells from beyond the grave how he met his assassins: nhs.n=i n aHA iw=i n Haw=i gm.n=i Hw-ny-r-Hr pw n mwnf When I woke up to the battle, I was alone, and I discovered that it was an attack of the bodyguard. (Amenemhat VIIa–b)
165 See the discussion in 10.3 below. For remarks on the pragmatic role of subjectanticipation after iw/ntt/wnt, see section 8.2. 166 Cf. Allen 1986a, 20 n.50. For a further instance of a nominal sentence after gmi, see example (113) below with the possessive pattern n(y). For the analysis of this sentence pattern as a nominal rather than an adjectival sentence, see Wycichl 1954, 367–72; Loprieno 1995, 118 and Jenni 2004, 123. For gmi + sDm.n=f, see 7.1. For an example of a passive sDm=f as a subject complement of passive gmi, see (211) below. 167 Cf. GEG § 186.3. Junge (1978a, 42) and possibly Satzinger (2001, 241 n.14) analyse the object in this example to be HfAw only, which functions simultaneously as the predicate of a following pw-sentence [HfAw pw]. However, this would make the construction structurally similar to ‘object raising’ construals and result in pw (or HfAw pw or even pw) ending up in the same paradigmatic slot with the ‘circumstantials’ following raised objects (cf. Junge 1991, 409). The clause-linkage is clearly neither a paratactic ‘direct quote’ (I found: “it is a snake coming”) as argued e.g. by Jansen-Winkeln (1996b, 47) and Grandet & Mathieu (1997, 446 n.1).
110
chapter two
(109) A remark after the presentation of the solution to a mathematical problem:
g[m]=k nt(t) pw You will discover that this is what (it) is.168 (pMoscow math. no. 9, 29)
(110) An account of juridical proceedings notes of a certain document: gm.n.Tw in snn m xA n wHmw n wart mHtt m xA n TAty It was found that a copy had been brought from the office of the herald of the northern district and from the bureau of the (Stèle Juridique 15) vizier. (111) The preparations for a display of royal prowess in archery are narrated:
aq.n=f rf r S=f mHty gm.n=f smn n=f stw 4 m Hmt styt Having entered into his northern garden, he (the king) found that four shooting-targets of Asiatic copper had been prepared (Urk IV 1280, 12–13) for him. The regular absence of ntt/wnt after gmi shows that it is impermissible to assume a covert ø-complementiser in the structure.169 Certain semantic factors could also be argued to favour a differentiation between ‘object raising’ [object + adjunct] and bare adverbial/pseudo-verbal sentence complements of gmi. As noted, in Earlier Egyptian ‘objectraising’ presupposes direct visual contact between the main clause subject and the complement actor, which is not always achieved after gmi:170 168
Analysing nt(t) here as a complementiser is, of course, syntactically impossible. The object clause is rather a bare bipartite nominal sentence [nt(t) pw], where the predicate of pw is lexical nt(t), ‘what exists/is’. 169 Pace Sweeney’s, (1986, 343–44) analysis of Late Egyptian constructions of gmi+ iw-clause (see below). Examples with an overt variant should exist for this to be a viable analysis. Alternatively, one could try to demote the passive sDm=f to an adjunct status by postulating some zero ‘real’ object for gmi (following the attempts of Junge (2001, 218) and Satzinger (2001, 241–42) again to explain away the ‘adverbial’ iw-complements of gmi in Late Egyptian). However, this is semantically unacceptable: it would be difficult to assign any content to such a ø that made sense in the overall situation. Clearly it is not the case in e.g. example (111) that the king found ‘ø after targets had been set up’ or ‘found [the targets] after targets had been set up’, with ‘raising’ (followed by deleting!) of ‘targets’, but simply that he found [targets had been set up]. 170 See n.122 above. Indeed, ‘discovering’ something after being told of it, or in a pitch-dark room involves no vision.
affirmative object complementation
111
(112) Harkhuf tells of an abortive rendezvous with a foreign chief: gm.n(=i) HqA iAm Sm r=f r tA TmH I found that the ruler of Yam had gone to the Tjemehi-land. (Urk I 125, 15–16)
To quote Allen, here “the sense is clearly ‘I found (that) the Ruler was gone’ and not *‘I found the Ruler, he being gone’”.171 Given the ban on ‘alienation’ of the main clause subject and the object in Earlier Egyptian ‘raising’, examples such as (112) could be analysed as bare adverbial sentences instead.172 In Late Egyptian gmi can be followed by bare iw-clause complements (which are ‘circumstantial’ by form but not function) as well as complements consisting of a noun/pronoun + an iw-adjunct. This could reflect a process of grammaticalisation resulting in formal differentiation of complements with and without ‘raising’:173 MEg LEg
gm.n=f [sw][Hr sDm] gm=f [sw][iw=f Hr sDm]
gm.n=f [sw Hr sDm] gm=f [iw=f Hr sDm]
171 Allen 1986b, 15; similarly Allen 1991, 11; cf. also Malaise & Winand 1999 § 918. A further similar example is pUC 32201, rto. 5 gm.n=i nb aws xnt “I found that the lord l.h.p. had sailed south”. One may also consider here examples from medical texts where body-parts certainly not ‘lost’, are ‘found’ with injuries etc., i.e. where the discovery clearly consists of an object-in-situation. 172 With e.g. the verb mAA, ‘see’, there are no examples of the type †mA.n=i N Sm comparable to gm.n=i N Sm of (112) above, clearly due to the requirement that a ‘raised’ object be visually accessible to the main clause subject, which is always the case with mAA but not gmi. No such construal as †“I saw X being gone” exist in English either, but it is grammatical e.g. in Finnish: Näin miehen poistuneen; literally “I saw a/the man being gone”, where the object miehen ‘man’ is in the same genitive case as in Näin miehen, “I saw a/the man”. 173 However, gmi has also a number of other object clause patterns in Late Egyptian (see n.180 below) and is not the only verb compatible with bare iw-complements. Object clauses of the latter sort are also attested at least after rx, (LEM 47, 12; possibly KRI III 145, 16–146, 1) smAa ‘beg’ (oDeM 1406II, x+3) and probably mri (KRI I 322, 9); see Kruchten 1997, 59. For a bare iw-complement of mri from a text falling within the scope of the present work, see example (425) below. According to Sweeney (1986, 343), also ptri ‘see’ may in Late Egyptian be followed by iw-complements, but her only example oDeM 133, vso. 5 is broken and unreliable and, as she notes, (ibid, 362 n.28) instances such as KRI VI 828, 15–16 (pBM 10403 1, 4–5) or e.g. KRI VI 843, 10–11 (pBM 10403, 3, 19–20) are more probably adjuncts. KRI VI 816, 5–6 (pMayer A vso. 6, 16– 17) is wrongly cited by Groll (1969, 190) as an object clause of ptri and reads bwpw=i ptr=w iw wn=w tAy xtm “I did not see them having opened this seal”.
112
chapter two
Yet, although in Earlier Egyptian the complements of gmi such as those above may differ semantically, there are no formal indicators of the presence of adverbial/pseudo-verbal sentences here. In fact, syntactic evidence seems to speak against such an analysis. Often what follows the first (pro)noun of the complement is a fully clausal sDm=f/sDm. n=f or the status of the former as an adjunct is betrayed by word order.174 Thus, rather than to conjure up ‘invisible’ adverbial and pseudo-verbal sentences after gmi in Earlier Egyptian, it is better to note simply that at this stage of the language the said constructions seem to be included in the object paradigm of this verb merely as ‘meanings’ and as ‘forms’ only in Late Egyptian, if at all. Once again, form does not correspond to semantic function. In any case, the argument that some Earlier Egyptian patterns are unsuitable for complement use and require ‘conversion’ by ntt/wnt is undermined by the use of the bare ‘non-nominal’ forms after gmi.175 Ultimately, however, this phenomenon is not syntactic but modal by nature. After gmi all the modal markers and marked forms—i.e. ntt/ wnt, the geminating sDm=f and forms with the endings -w and -y—are conspicuous of their absence. 176 Instead, the bare ‘non-nominal’ forms occur alone, and in fact quite irrespective of the semanticpragmatic character of the complement. In examples (107)–(111) above, the speakers are in every instance clearly aware of and committed to the complement situations, which in notional terms have all the characteristics of realis. Yet, there is no grammatical sign of this in the
174 For example, consider a hypothetical sentence such as *gm.n wi Hm=f Hms.kw m pr=i “His majesty found me living in my house”. For wi Hms.kw m pr=i to be a pseudoverbal sentence complement, its subject wi, a dependent pronoun, should have had to abandon its clausal domain in order to precede the noun subject Hm=f of the governing clause (cf. Collier 1991a, 48 n.101). 175 As no ‘conversion’ is required with nominal sentences or the past passive sDm=f, there is no reason why it should be with the other ‘non-nominals’. Of course, there are no examples of e.g. bare cleft sentences after gmi, but this is hardly more than a matter of survival of an unusual expression in the textual corpus. Agatha Christie’s scribbles aside, what is the likelihood of a sentence such as “He discovered that it was he who heard” to appear even in (good) written English? Moreover, there are neither any examples of e.g. the negation n sDm=f after ntt/wnt following verbs. Does this indicate that such a combination is ‘ungrammatical’ or rather that all the possible permutations are not present in the evidence? 176 The form i.irr=sn in CT V 119/99c occurs in only a single variant (Sq7Sq) and is strange anyway.
affirmative object complementation
113
form of ntt/wnt. By contrast, in the following example a bare nominal sentence appears in what is clearly a semantically hypothetical and thus notionally irrealis complement: (113) A Machiavellian advice on how to deal with a dangerously popular local potentate:
ir grt gm=k n(y) sw nwtw [... sp]=f swA Hr=k sxr sw m-bAH Snyt dr sw Now, if you discover that he has townspeople (i.e. urban support) and his deeds surpass yours; accuse him before the entourage and purge him. (Merikara E 25–26) Examples of bare non-geminating sDm=f of mutable verbs without the endings -w/-y are also forthcoming. Sometimes, as elsewhere, assertion of these clauses may be clearly contextually cancelled, as in the following negative examples:177 (114) It is said that Senwosret I prepared a chapel in Karnak: sT n sp gm Hm=f iri.t(w) mitt m r-pr pn Dr-bAH For never did his majesty discover that the like would have been done in this temple since time immemorial. (White Chapel 59)
(115) The deceased describes the fate of fallen stars: n gm.n=sn Ts=sn st They find no way to raise themselves.
(Tb 99, 10)
Yet, and unlike after other object complement-taking verbs, after gmi the bare form may also equally well carry all the semantic and pragmatic hallmarks of assertion:
177 See also Urk I 182, 15 (ny gm.n rmT nb Sm=sn); CT VII 202i (n gm=Tn iri=Tn); pEdwin Smith 1, 25; pEdwin Smith 2, 4-5 and 14; pEdwin Smith 3, 4–5 (all n gm.n=f dgA=f). dgA is possibly an idiosyncratic spelling of the non-geminating form of the verb dgi, ‘look’, as opposed to dgg; cf. chapter 4 n.42.
114
chapter two
(116) A maidservant has told her mistress about a strange sound in a room; she goes to investigate:
wn.in=s Hr dbn tA at n gm.n=s bw irrw st im aHa.n rdi.n=s mAa=s r pA XAr gm.n=s iri.tw ø m-Xnw=f She kept going around the room but could not find the place where it was made. But then she pressed her temple against the said sack, and discovered that (it) was made inside of it. (pWestcar 12, 3–4)
(117) Thutmosis III tells that he arranged the provisioning of offerings to the Mansion of Amun…
m-xt gmt Hm=i iri.tw xt im …after my majesty found out that rituals were performed (Urk IV 751, 2) there. (118) Instruction concerning the use of pulverised eggshell in treating a wound:
sSwt wbnw pw rdi.xr=k n=f HAyt Hr=f nt Xn swnw kf.xr=k sw 3-nw xrw gmm=k Ts=f pqt irtyw mi swHt nt niw To desiccate a wound: you apply on it a HAyt-bandage of a surgeon’s knife and cover it for three days. You find that it (the powder) reconstitutes the flesh, its appearance being exactly as (pEdwin Smith 4, 21–5, 1) (that of) an ostrich-egg. Notably, here the bare non-geminating sDm=f without the endings -w/-y not only carries none of the properties of the distal irrealis function outlined above, but also has a concomitant or generic sense instead of future or past. This shows that the non-assertive and temporal functions of the -w/-y-less form go hand-in-hand.178 It is unknown why gmi should differ so radically from all other Earlier Egyptian verbs in the grammar of its object complement clauses. In other languages the speaker’s role vis-à-vis the propositions following ‘find out/discover’ and its grammatical coding parallel that of verbs of cognition and perception.179 Yet, the complement selection of 178
The relevance of this notion will be seen in 9.1 below. See Hooper 1975, 115–21. Thus one has “I cannot find out †that/whether it is here”
179
affirmative object complementation
115
gmi shows unique characteristics also in Late Egyptian, and the same applies to Coptic qine. The problem of the complement taking of gmi → qine thus requires a diachronic study encompassing all the stages of Ancient Egyptian.180 However, it is clear what these phenomena represent—namely a unique lapse or ‘deactivation’ of the modal system seen elsewhere in Earlier Egyptian complementation. Paradoxically, this both confirms the analysis of ntt/wnt and the geminating and -w/y-ending sDm=f forms as specially marked for modality and reveals the modal-functional character of the bare forms and constructions that constitute the paradigm of ntt/wnt. In the one environment where the normal organisation of complement modality is not active, i.e. after gmi, the modal markers and inflectionally marked forms are not used. Instead, the bare ‘non-nominal’ patterns occur alone and express what are notionally both realis and irrealis meanings. Thus the latter clearly have no fixed and inherent modal value. That is, in contrast to both the irrealis-marked geminating sDm=f and the forms with the endings -w/-y, the said forms and constructions are modally unmarked.181
“He cannot find out that/whether it is there” “I did not find out that/whether it was there” “Did you find out that/whether it is raining?” (‘Find out/discover’ have punctual Aktionsart incompatible with indicative present tense: thus †“I find out/†am finding out” (the second sentence should be understood as non-epistemic). As before, the first sentence is fine with whether but contradictory with that because by framing the complement as assertion the speaker claims to know something and simultaneously denies this. The second sentence is fine with that if the speaker knows that the complement holds, as is the third if he knows this now. Whether signals the opposite, or omission to confirm this. In the last sentence the scope of interrogative is ambiguous with that; the speaker may ask if the addressee found out something he himself already knows or both whether something was ‘found out’ and ‘is it raining’. Whether unambiguously signals the latter as intended. 180 Besides bare iw- and r-Dd-clauses, (for the latter, see 7.3 below) in Late Egyptian only gmi can take complements of the form [noun/pronoun + ‘circumstantial’ without iw] as well as [object + r-Dd + sentence]; see Sweeney 1986, 353–55. The former type of construal may represent a vestige from Earlier Egyptian. In Coptic, qine is the sole verb that, when negated and appearing in the status constructus form qN can take bare second tense and interrogative object complements of the type anG nim “who am I?” and einaR ou, “What shall I do?” See Spiegelberg 1923 and Layton 2000, 363, 426 for examples and discussion. 181 Also in Uljas 2000, 131 the patterns following ntt/wnt are argued to be modally neutral and to contrast with modal forms that can occur alone as complements. Collier (1994, 79) interprets the ‘circumstantial’ sDm=f/sDm.n=f as modally unmarked and the ‘future sDm=f’ as marked.
116
chapter two
One may now formulate the following general principles of construing clausal complements in Earlier Egyptian: 1. Earlier Egyptian complement clauses contain modally marked and unmarked verbal and other patterns 2. In modally realis asserted complements the complementisers cum realis operators ntt and wnt introduce unmarked forms/ constructions and assign the latter the modal value +asserted.182 The unmarked group encompasses all verbal and non-verbal constructions except active sDm=f forms with gemination or the endings -w/-y.183 3. Active sDm=f forms displaying gemination or the endings -w/-y are marked by their inflection for proximal and distal irrealis respectively. They are used for non-assertion and are hence incompatible with the realis markers ntt/wnt 4. However, unmarked active suffix-conjugation forms,184 i.e. ones that do not show the formal characteristics of marked irrealis and the most important of which are the unmarked forms of sDm=f, may also be used alone in non-asserted complements. There are two reasons for this. Firstly, the vast majority of verb-classes simply do not have the irrealis-marked (geminating and -w/-yending) sDm=f forms and instead use the unmarked options as their functional counterparts. Secondly, the endings -w/-y are fast disappearing (probably due to a phonetic collapse185) with the result that also ult. inf. and caus. roots that in principle are 182 ntt/wnt thus function as ‘illocutionary force indicating devices’ (Searle 1969, 30; Junge 1989, 102). 183 However, here a proviso is in order on the forms iwt=f, int=f and iit=f of the anomalous roots iwi, ini and ii, analysed as specialised uses of the sDmt=f form for distal irrealis (see 0.2 above). It will be postulated in 10.3 below that the sDmt=f may be analysable as an irrealis-marked form. 184 The expression ‘unmarked active suffix-conjugation forms’ is used here in anticipation of the discussion in 7.1 below, where it will be shown that besides the unmarked forms of the sDm=f, also the sDm.n=f belongs to the class of modally unmarked forms that have the capacity of forming irrealis complements on their own, i.e. without an introducing ntt/wnt. 185 Cf. Eyre 1994, 118.
affirmative object complementation
117
capable of forming marked distal irrealis often have recourse to the bare unmarked forms such as hA=f, pr=f, sxA=f etc. 5. Except for the unmarked suffix-conjugation forms, whose irrealis function is indicated by the absence of ntt/wnt, no form or construal without a specific modal value/marking may occur in an irrealis function, i.e. alone and without ntt/wnt.186 Instead, adjectival and ‘pseudo-verbal’ sentences are replaced by the sDm=f of the relevant verb (i.e. nfr sw → nfr=f and N Hr sDm → sDm N). With adverbial and nominal sentences the sDm=f is formed of the auxiliary wnn (N m pr → wn(n) N m pr and sS pw → wn(n)=f m sS187). The passive sDm=f is replaced by sDm.tw N. The sDm=f thus provides a functional counterpart for all these patterns in irrealis. The remaining construals, namely second tenses and cleft constructions are not only modally unmarked, but also ‘emphatic’ ‘marked focus’ statements, which, as a principle, are not non-asserted.188 This explains why they, and sentences with ante-posed elements, do not have an irrealis counterpart in complementation like the other modally ‘neutral’ forms/constructions. The replacement of the unmarked patterns appears to be a grammatical reflection of the differences in canonicity of realis and irrealis situations. A recurrent property of Earlier Egyptian is a tendency to reduce the grammatical coding of fine-grained semantic-pragmatic features and the number of patterns used for this the further one moves away from describing ‘canonical’ or ‘cognitively simple’ situations. This has been noted with affirmative versus negative and active versus passive sentences, where the first alternative represents conceptually a more ‘basic’ scenario.189 The same is clearly the case also with indicative 186 The verb gmi forms only a partial exception here. Although the unmarked forms may occur alone as its complements, they do not function exclusively as irrealis. 187 The latter option is also open for the ‘pseudo-verbals’ (e.g. wn(n)=f Hr sDm). 188 Hooper 1975, 99; cf. also Lambrecht 1994, 206–07, 213 and passim; Givón 2001 vol. 2, 222–23. 189 Collier 1994, 76–77. Thus, in present affirmative a difference is made between iw(=f) sDm=f and iw=f Hr sDm, but these share the negation n sDm.n=f. Past differentiates between iw sDm.n=f and iw=f prw, but both are negated by n sDm/pr=f. In the future affirmative one has either sDm=f or iw=f r sDm, but the negation is characteris-
118
chapter two
(realis) versus non-indicative (irrealis) complements. In the former the original constructions and their particular semantic-pragmatic characteristics are preserved, whereas in irrealis just a handful of counterpart forms are used. The reason for this is clear: states of affairs not seen as subjectively or objectively wholly ‘real’ or worthy of full attention are less canonical than ones that are so viewed. The grammatical organisation of Earlier Egyptian object complement clauses may thus be expressed as a tripartite system of form and function: modality marked realis ntt/wnt + hA=f/mAA=f/stp=f/any unmarked pattern irrealis Proximal: hAA=f Distal: hAw=f, hAy=f, mA=f form
modality unmarked
Proximal: mAA=f Distal: hA=f Generic: stp=f (stp.n=f)190 function
190
The ‘marked realis’ is an analytic construal consisting of the marker ntt/wnt + unmarked form/construction. Irrealis is always morphosyntactically synthetic. This analysis is true to morphological evidence: it assumes that writing such as e.g. [stp=f] does not hide anything that is not accessible to readers ancient or modern. It is also non-reductionist in that it begins with the larger units (the entire clause) and defines their constituent smaller units vis-à-vis their relation to the larger units.191 With the exception of the marked irrealis sDm=f forms, in the present analysis the basic units of syntactic organisation are essentially the constructions into which different forms enter, rather than the forms themselves. It is the syntactic environment that determines the function of the latter, not the other way round. Regarding these functions, apart from the fact that in languages with
tically just nn sDm=f. As for active and passive, both iw(=f) sDm=f and iw=f Hr sDm are passivised by iw sDm.tw=f and in the future bare sDm=f and iw=f r sDm by sDm.tw=f. Only in the past does one find iw=f sDmw and iw sDm N for the active iw sDm.n=f/N. 190 The sDm.n=f is included here for the sake of completeness; its role in the modal organisation of complement clauses in Earlier Egyptian will be discussed in 7.1 below. 191 See Croft 2001, 47–48.
affirmative object complementation
119
a grammaticalised realis-irrealis distinction, a third category of modally unmarked forms and constructions is anything but uncommon,192 the above division based on evidence from object complements finds its exact parallel also in subject- and preposition complement clauses.193 In addition, in case of ntt/wnt-complements it is of course the entire complement clause [ntt/wnt + p] that constitutes the assertion, but from this it does not follow that the pattern coding p must itself be ‘assertive’, although it must not be non-assertive (irrealis) either. Second tenses and cleft sentences after ntt/wnt are no exception, even though the former may contain the geminating sDm=f and in the latter the predicate (participle) is presupposed and most certainly nonasserted.194 The minimal clause e.g. in a second tense is not just the verb, but includes also the vedette, and what matters is the modality and grammar of this entire unit, not just a part thereof.195 Also, if the forms and construals constituting p in ntt/wnt p were realis-marked, there would be two markers of the same modal function in the clause. Such combinations are excluded inasmuch as they would lead to repetition and ‘over-encoding’ of that modality in one clause.196 Just as encoding modality must not result in pragmatic conflict, (cf. here † ntt mrr=f) the same holds for tautology. Notably, if the function of iw is the same as that of ntt/wnt, i.e. to mark the clause as asserted, the exclusion of †ntt iw follows directly from the ban on modality ‘double-marking’.197 ntt/wnt is also regularly followed by non-verbal patterns such as nominal and adverbial sentences. With such construals it is quite legitimate to speak of assertion or non-assertion in notional terms, but the grammatical expression of modality is carried out by either lexical-functional expressions (particles, operators etc.) or verbal inflection. Non-verbal construals in general stand outside the system of tam-marking in Earlier Egyptian.198 When modal and temporal 192
Palmer 2000, 161–63. See chapters 4 and 6 below. In 6.3 it is suggested that this organisation finds a parallel also in the organisation of adjunct clauses. 194 Levinson 1983, 182–83; Loprieno 1995, 115; Givón 2001 vol. 2, 234. 195 Cf. Collier 1994, 79 n.61. 196 Frajzyngier 1995, 477 and passim; for this reason e.g. English allows no accumulation of modal verbs in a clause: †“I think he may must be there”. In addition, as here, the modality expressed would also mostly be repeated in different degrees, against the Gricean maxim of Quantity (see Levinson 1983, 106–07). 197 See 10.3. 198 But see chapter 8 below. 193
120
chapter two
features are explicitly communicated, these patterns combine with operators or are replaced by verbal suffix-conjugation forms. Finally, and anticipating somewhat the discussion in 6.3 below, it is notable that the unmarked character of the sDm=f forms used after ntt/wnt manifests itself as non-assertive use also elsewhere, for example in adjunct clauses. The use of the bare sDm=f forms without gemination or -w/-y in final ‘so-that’-clauses is of course their most notable irrealis use outside complementation. Most sDm=f relative present circumstantial clauses can be characterised as assertions in notional terms, but there are other, apparently non-assertive/irrealis meanings in the same category, such as e.g. the ‘virtual clauses of condition’:199 (119) Ptahhotep stresses that a successful speaker must also know how to listen:
sDm rk mr=k smn.t(w)=k m r n sDmyw Listen—if you want yourself to be well-established in the mouth(s) (Ptahh 613–14) of listeners. Most interestingly, however, the irrealis meaning of these clauses receives formal recognition in negation.200 However, many questions remain concerning the modal status of the constructions following ntt/wnt. For example, in second tenses aside from the ‘grammatical’ predication between the verb and its argument(s) there is also the ‘logical’ predication between the verb and the focal adjunct at stake, but the relationship of this linkage to assertion and modality is as yet unknown.201 The status of the bare unmarked forms and construals in the wider system of Earlier Egyptian modality is also inherently linked to the semantic-pragmatic role of auxiliaries and other initial elements with similar functions. Complementation provides suggestive insights into the modal profile of these elements, but clarifying their exact semantic-pragmatic character requires an extensive study of Earlier Egyptian main clauses.202 199
GEG § 216. See 5.3 below. 201 But see 10.3. 202 This question will be further touched upon in 9.1 below and in the conclusion (section 10.3). 200
complementation after notionally non-assertive verbs 121
CHAPTER THREE
AFFIRMATIVE OBJECT COMPLEMENTATION AFTER NOTIONALLY NON-ASSERTIVE VERBS In addition to verbs attested with complement clauses introduced by the assertion-markers ntt/wnt, there are a number of transitive predicates in Earlier Egyptian that are never followed by these elements, but only govern object complements with bare sDm=f. For the latter, verbs with mutable roots, which here are the sole type warranting examination, show forms both with and without gemination, doubling and the endings -w and -y. However, the number of matrix verbs constituting this group and particularly the extent of the said form-variation is difficult to determine, as evidence of object complementation after verbs other than locution, cognition and perception is very uneven. The number of attestations of many verbs with finite object clauses, let alone involving mutable roots, is minimal. For instance, the following example is the only occurrence of snD, ‘fear’ with such a complement: (120) Sinuhe tries to avoid being detected by frontier-guards: Ssp.n=i ksw m bAt m snD mAA wrSyw tp Hwt imt-hrw=s I took a crouch in the bushes in fear that the guards on duty (Sin B 17–18) upon the ramparts may see. Fear is the opposite of hope: it is an emotion felt towards situations not known to occur or to be occurring for certain, but suspected with apprehension, negative orientation and ‘epistemic anxiety’.1 Thus irrealis complement constructions and forms regularly follow verbs of fearing across languages, and Earlier Egyptian seems to behave similarly.2 Yet, whether forms expressing different types of irrealis alternate after snD, and what are the conditions in which this occurs,
1 Cf. Givón 1994, 280; Palmer 2000, 134–35; see also 6.2 below for the preposition n-snD. 2 See Palmer 2000, 133–34 for examples from Spanish, Classical Greek and Latin; cf. also Lichtenberk 1995.
122
chapter three
cannot be determined on basis of a single example.3 Other poorly attested verbs are sxA, ‘remember’ and nHi ‘pray’ of which the following are the sole attestations with mutable roots:4 (121) A question in an incantation: in iw trw sxA.n=k it.tw Hr Hna stX r ist aAt nt iwnw Do you remember if/that Horus was taken with Seth to the (pEbers 2, 3–4) great palace of Heliopolis? (122) The peasant, believing he is doomed, utters in despair: xsfw n ib m mw DAt-r n Xrd n sbnt m irtt ntf mwt nHy mA=f n iy=f ii wdf mwt=f r=f The approach of a thirsty man for water, the reaching of the mouth of an infant for milk—death is their lot. But as for one who prays that he may see it coming, tardily comes his death. (Peas B2, 119–22)
Perhaps the most unfortunate is the situation with the verb ib, ‘think’. This predicate is modally of great interest, as it describes a positive propositional attitude, but also one that is considerably more reserved and hesitant than ‘know’. There are tantalising signs that gradations of this attitude between almost full commitment and strong doubt that are modally indicated in the complement in many languages, also find expression there in Earlier Egyptian. Yet, here more than anywhere the evidence leaves much to be desired. There is one possible example of ib followed by a clause introduced by iwt, the negative equivalent of ntt/wnt:5 3 E.g. in Russian the verb fear may be followed by indicative or subjunctive depending on the degree of certainty with which the situation feared is assumed to be the case or expected. 4 In (121) the scope of the interrogative is unclear; the complement may or may not be questioned. Another, morphologically unrevealing example of sxA is Siut I 267: iw=i sxA=i spr=i r nTr hrw pf n mny gm=f wi “I used to remember (=kept in mind) that if I approached the god on the very day of mooring, he would find me”. For further comments on example (122), where the form iy=f appears as a complement of mAA n, actually ‘look at’ (n), see 9.1 below. 5 See chapter 5 and section 6.3 for negative complements. Edel (EAG § 1018 n.1) argues that Sethe’s edition of the text is incorrect. The first publication of Sabni’s ‘autobiography’ by De Morgan et al, (1894, 147–48) based on squeezes of the original, shows neither ib nor iwt, but is so riddled with inaccuracies as to be unusable. Breasted (1905–06, vol. 1, 164 n.f) based his translation of the inscription on the said
complementation after notionally non-assertive verbs 123 (123) The king greets Sabni for his successful mission abroad: iw(=i) r irt n=k xt nb iqr m-isw sm pn aA [ir.n=k] n int [it]=k ib(=i) iwt zp xpr mitt Dr bAH I will do all great things for you in reward of this mighty undertaking which you have carried out by bringing (back) your father, for I think nothing similar has ever happened before. (Urk I, 138 13–17)
Beyond this rather problematic instance, the only other example of ib with an object complement clause is the following, showing a geminating sDm=f: (124) After his final scornful words, the peasant rushes out but is forcibly brought back:
wn.in sxty pn snD ib=f irr.t(w) r xsf n=f Hr mdt tn Ddt.n=f Then this peasant became afraid, thinking that (it) was done to punish him for these words, which he had said. (Peas B2, 117–18)
Assuming that (123) is a genuine example, it would seem that in Earlier Egyptian as in modern languages ‘think’ could be followed by propositions that qualify as assertions also grammatically. Yet, the evidence hardly allows firm conclusions to be drawn as to under what circumstances this was or was not so. Nevertheless, the two examples above differ in that in (123) the speaker reports his own belief whereas in (124) this is assigned to a third person. In many languages speakers may use modal means to indicate their own attitudes and opinions also when reporting someone else’s thoughts and beliefs, seeing that what the subject of the main verb may think and believe need not be believed by the reporter.6 Interestingly, although the ‘thinker’ in squeezes collated with Erman’s collation of the original, but he has a lacuna here (ibid, 168). Sethe says having checked both the original and the squeezes once more. The reading ib + iwt does not yet appear in the first edition of Urkunden I, but has been inserted in the second. This then formed the basis e.g. for Roccati’s translation of the text (1982, 219) and in Hannig’s recent dictionary (2003, 64) this passage is cited as an example of the verb ib without further ado. 6 For example, in Kinyarwanda, a Bantu language, (Noonan 1985, 115) the variant complementisers ko and ngo are used to indicate speaker commitment: Yatekereìe ko
124
chapter three
example (124) is committed to the complement situation, the speaker (narrator) need not be. It is clear from the tale from which the example derives that the subject’s belief was incorrect. The bare geminating form here could be the real speaker signalling this information and again indicating his attitude and degree of (non-) commitment to the complement situation.7 Although well in keeping with the conclusions above, this suggestion remains a conjecture due to lack of examples. Neither are there sufficient synonyms of ib to aid in analysis. xmt ‘take thought of, anticipate’ resembles it semantically, but of this verb again only a handful of examples exist and only one with a mutable complement:8 (125) From a description of the divine birth of queen Hatshepsut: TAri.n=f awy=f(y) Hr swHt=f xmt.n=f iT=s idbwy He (Amun) clasped his hands upon his offspring, having anticipated9 that she would seize the Two Banks. (Red Chapel 166, 22–23)
But although evidence of object complementation with many verbs beyond those of locution, cognition and perception is at short supply in the surviving Earlier Egyptian textual material, the following predicates are nevertheless attested in sufficient numbers with such construals either alone or as a class of close synonyms:10 wD ‘order’ mri ‘want, wish, love, like’ rdi ‘cause, let, allow’
amazi yari mare-mare “He thought that the water was deep (no comment)” versus Yatekereìe ngo amazi yari mare-mare “He (misguidedly) thought that the water was deep”. See also (lxi) in 9.2 below for an English instance where the speaker attitude towards thoughts of a third party is indicated by changing the complement clause tense. 7 I.e. rather as in the second sentence in the previous note, the sense might be “thinking (misguidedly) that it was done to punish him for these words which he had said”. 8 A further example with an immutable sDm=f complement is Sin B 7 xmt.n=i xpr HAayt “I anticipated trouble to occur”. Sin B 111–12 xmt.n=f Hw{t}=f wi “He anticipated defeating me” appears slightly corrupt. 9 The Deir el-Bahri parallel-text reads xm.n=f “he was ignorant”, which is certainly wrong. 10 For the verbs dr, perhaps ‘remove from doing something’ and nxm ‘obstruct, prevent’, see 3.4 below.
complementation after notionally non-assertive verbs 125
xwi ‘prevent’ xsf ‘prevent’ sAw ‘take care that not’ dbH ‘request, ask’
Taken individually, the verbs xwi, xsf and sAw are relatively rare, but they have a closely similar meaning and their complement-taking properties can be studied as a group. In contrast, wD and mri are very common indeed and rdi is by far the most common of all the verbs above. Notionally, the verbs of preventing and causing form a group of manipulative verbs that refer to dynamic and prototypically physical acts of overt manipulation, namely obstructing or causing. The directive wD and the requesting dbH describe attempted manipulation of some other individual through speech acts,11 whereas mri has volitive, desiderative and ‘emotive’ sense(s). With these verbs, either individually or as a group of (near-) synonyms, there is reasonable certainty that all the possible types of complement clauses are present in the data and that this is not greatly distorted by accidents of preservation. In the entire corpus of Earlier Egyptian texts, there is not a single instance of any of them with complements introduced by the elements ntt/wnt, but only clauses with bare sDm=f, which often show gemination or the endings -w and -y. This restriction is of central significance to the present analysis and provides strong empirical support for the hypothesis of ntt/wnt as realis modal operators and their absence as a signal of irrealis modality. For also in all other languages, the above verbs do not combine with asserted complements, possible differences in identity between the subject and the real speaker notwithstanding. The absence of ntt/wnt here in Earlier Egyptian is thus not an anomaly. It follows directly from the function of these elements to mark complements as assertions, which is never the case after verbs such as e.g. ‘prevent’, ‘want’ or ‘order’, regardless of the particular language in question. This notion thoroughly undermines the thesis of ntt/wnt as syntactic ‘nominalisers’ or ‘converters’ and of the bare complement sDm=f forms as some sort of ‘variants’ or ‘substitutes’ thereof. Once again, whereas the ‘nominal hypothesis’ assumes Earlier
11
Givón 1975a, 66 and passim; 1994, 272; 2001 vol. 1, 152.
126
chapter three
Egyptian to differ from all other linguistic systems in its grammatical organisation of complement clauses, the present analysis views the latter as fundamentally similar to other languages and as based on the same underlying principles. The occurrence of the different forms and constructions is again driven not by autonomous structural rules arising from their inherent syntactic profiles, but by the nature of the information communicated by human language users and their subjective stance towards it. As in other languages, in Earlier Egyptian the irrealis modality of the object complement clauses of these non-assertive verbs is determined by the notional properties of the latter. That is, the primary vantage point from which the modal status of their complements as nonasserted is assigned is that of the original speaker and speech context. However, as seen, the actual speech context of the real speaker need not be identical with that of the original speaker.12 In most languages the former vantage point is not taken into account in the grammatical organisation of complement clauses of non-assertive verbs, but there are others where it is, and Earlier Egyptian seems to belong to this latter group. The alternating between the geminating and non-geminating sDm=f’s after many of the verbs above or the exclusion of the former pattern represents a strategy by which the real speaker’s perspective is variously included in the complement situation-description. In addition, this variation will again turn out to be a concrete manifestation of the basic cognitive abstract of proximal and distal irrealis. The discussion below will first examine the group of verbs referring to attempted manipulation. This is followed by a detailed look of the verbs of prevention and causation and, finally, of the volitive mri. The modality of the complement clauses of all these verbs will be seen to reflect the same principles of expressing the speaker’s attitude towards and assessment of the information they convey. 3.1 wD and dbH: Between ‘Ontological’ and ‘Attitudinal’ Modality The verbs wD ‘order’ and dbH ‘ask, request’ can be characterised as archetypal predicates of social coercion. Notionally, they both describe deontic speech acts aimed at affecting the addressee and ‘imposing 12
See chapter 1 above.
complementation after notionally non-assertive verbs 127 the speaker’s will’ upon the latter. As such, requests and orders equal attempted manipulation in that from the semantics of the situation itself it is not clear whether or not the action or state, which the speaker is or was trying to get someone to perform or assume, will or did take place, i.e. whether the manipulation was, is, or will be successful. In Earlier Egyptian object complement clauses of verbs wD and dbH show much variation between sDm=f forms with and without gemination (as well as wnn/wn). After dbH, the ratio of surviving examples with gemination/wnn versus no gemination/wn is approximately 1:2 in favour of the latter. After wD, the evidence appears to be slightly distorted by accidents of preservation. In Old Egyptian data there is but one example of a non-geminating form (example (144) below) whereas gemination dominates. Conversely, in the early XVIII dynasty material there are no geminating examples after wD. Yet, with this verb there are many Old Egyptian examples from ult. inf. roots ending at -t that can be read either as infinitives or as passives with the ending .t(i). The number of post-classical examples of wD + mutable sDm=f-complement is also relatively low. In any case, the numbers are again in favour of lack of gemination also with wD, as noted by some authors and contrary to an earlier claim by the present writer.13 These numerical differences in frequency of the two types of sDm=f already suggest that the geminating form represents a somehow more ‘special’ option after all these verbs. As elsewhere, it is perilous to assume that the form-variation were merely a matter of statistics and to treat the evidence as if the examples were similar, without considering their overall semantic context. Such an approach has little hope for explaining the occurrence of the geminating and nongeminating sDm=f forms after these two verbs, which at first sight appears quite perplexing:14 13
Silverman 1985, 282; Loprieno 1995, 201; cf. Uljas 2000, 128; 2003, 392. Further examples of wD with mutable sDm=f complements not quoted below are: CT I 131b/L2Li (wD.n N sipw N, see example (3) above); CT III 207d (wD.n N… ssnw N); CT VI 348p (wD.n N rdi N); CT VI 371o (swt wD wnn N); CT VI 393h (wD.n N wnn=k); Wadi el-Hudi 21, 8 (xft wD N pr=f, if not prt=f); pUC 32157 h/v left, page 3, 2 (wddw iri=f); Tb 144/Nu pl. 75, 33 (wD.n N wn N). dbH is relatively seldom met. In addition to the examples here and further below, there are Old Kingdom instances where it is unclear whether one should understand the object as an infinitive or a t(i)-passivised non-geminating sDm=f, including Urk I 99, 10–11 (dbH.n(=i)… in.t(i) or int) and Urk I 146, 6–8, (dbH.n(=i)… Sd.t(i) or Sdt) which Edel (EAG § 711) interprets as infinitives and Doret (1986, 47) as ‘subjunctives’. CT V 241a has dbH=f Sm=f iw=f (S10C 14
128
chapter three
(126) An order by the king in a decree for one of his officials: iw wD.n Hm(=i) dd=k iri.t(i) wiA n igAi n mH [...] My majesty has ordered you to have a boat of X cubits be made (Urk I 298, 8) for Igai. (127) Thutmosis III describes his pious donations to Amun: [iw grt] wD.n Hm=i di.tw iry[.tw n=f twt]15 My majesty further ordered that there be caused to be made a (Urk IV 1257, 3) statue for him. (128) An unidentified deity is addressed:16 n N tn is dbHt mA=s Tn m qd=T pw wn=T im=f in Hr dbH mA=f Tw17 m qd=k pw wn=k im It is not this N who asks if she may see you in this your form in which you will be; it is Horus who asks if he may see you in (CT VI 353l–m) this your form in which you will be. (129) Isis is on her way to meet her son in the entourage of gods: hA Ast r wHa in Hr dbH.n Ast wnn=f m wHa m sSmw nHH Isis goes down to wHa18 who has brought Horus, for Isis had asked that he might be with wHa as the leader of eternity. (CT II 222b–c)
A syntactic explanation for this variation has been proposed by Loprieno, who, pace Borghouts, interprets the appearance of the geminating sDm=f as an object of a verb of wish and command to signal direct embedding of entire second tense macro-sentences, and only Sm=f)“He asks whether he may come and go” (Sm=f iw=f is a compound expression; see Wb IV 462). 15 The restoration is based on the rather similar Urk IV 1255, 13, where the spelling rdi.tw occurs. 16 The PT original (1128a–29b) is rather more extended, but writes mA=f in all of the four instances of dbH + mAA. A further PT instance is PT 149a–b: dbH=k pr=k r pt prr=k… dbH=k hA=k (var. hAw=k, PT PII, 709, +57) ir nwt hAA=k “Whenever you ask that you may go forth to heaven, you go… whenever you ask that you may descend to Nut, you descend”; cf. also PT 1275b, (dbH=f hA=f) PT 1276b (dbH=f pr=f) and Allen 1984 § 236. 17 Changed from the feminine Tn. 18 Judging from the determinative, some divine being (maybe a scorpion deity) is meant.
complementation after notionally non-assertive verbs 129 that of a non-geminating form of normal clausal subordination.19 Thus, a complement such as that in example (126) above would be structured as follows:20 iw wD.n Hm(=i)
[[dd=k iri.t(i) wiA n igAi]NP[n mH X]AdvP]
Various scholars have disputed this analysis, and in spite of the somewhat suspect character of many of these counterarguments,21 the hypothesis is indeed not without problems. For Loprieno, the alleged second tense objects are a prerogative of epistemic and deontic predicates, whereas after other types of verbs the geminating sDm=f signals embedding of ‘nominalised iw-sentences’.22 Yet, as seen, the reason for making this division is unclear—why would such embedding be restricted to this particular set of verbs? Further, and anticipating the discussion of the verb mri below, there are numerous examples of this verb with a geminating sDm=f complement in the so-called ‘Appeal to the Living’, where the postulated second tense seems to lack the essential adjunct.23 In example (126) there are of course various suitable candidates for the latter, but these would all have to be located within a complement clause of dd=k:24 iw wD.n Hm(=i)
[[dd=k [iri.t(i) wiA n igAi]NP[n mH X]AdvP]]
The ST has made short work of the notion of constituency, and its advocates would probably deem this irrelevant. Yet, for the ‘emphatic’ scope of the assumed head second tense verb to cut across and into its own object complement clause seems rather improbable. Far-ranging 19
Loprieno 1988a, 68; 1991a, 214; 1995, 201; Borghouts 1985, 37 n.33. Adapted from Loprieno 1991a, 214 n.54; alternatively, the AdvP might be taken to be n igAi. 21 Silverman (1986b, 318) argues—incorrectly, as will be seen in 6.3 below—that second tenses can occur as objects only if accompanied with the particle is (so too Satzinger, (1989, 217 n.81) who also questions the semantic aptness of ‘emphasis’ in many instances). Junge (1979, 83–84) excluded this analysis by appealing to his view that in Earlier Egyptian ‘direct’ subordination of sentences does not take place inasmuch as their predicates lose their predicative force in the process without a ‘converter’ such as ntt/ wnt. However, as seen, (2.4 above) this would deny bare non-geminating sDm=f forms a ‘sentential’ status. 22 See 2.2.2 above. 23 See 3.3 below; cf. also Uljas 2000, 129. 24 Cf. Uljas 2004 97–98. 20
130
chapter three
relational arguments of the sort should respect or at least recognise clausal domains. It seems thus that the form-variation above is more likely to have a semantic-pragmatic rather than syntactic motivation. Seeing that both the non-geminating and geminating forms have a relative future reference in the examples cited above, it is clearly incorrect to argue that the latter form marks the complement situation as “simultaneous with the main verb”.25 Doret’s proposal of the geminating sDm=f as signalling the situation being viewed as “objective fact” and the lack of this morphological feature as “possible action” that can entail an “element of unreality” may be mentioned again as a rare attempt to explain the use of the different forms on basis of modality.26 It is not altogether clear from this formulation, who exactly is supposed to treat the complement situation as more or less ‘unreal’. It seems that the main clause subject is meant, but this raises the question why would, for example, the king ordering in example (127) above have viewed the outcome less ‘objectively’ than his royal colleague in (126)? After a verb such as ‘order’, at least in affirmative environments the matter ordered can hardly be profiled as involving “an element of possibility or doubt”.27 Doret’s suggestion points to the right direction, but here the choice between the forms is in fact determined by rather more intricate factors that once again reflect the way in which the real speaker views the subordinate state of affairs and communicates modal information both attitudinal and evaluative in character. In Earlier Egyptian object complements after verbs of attempted manipulation the variation of the clause form concerns primarily the indication of whether or not the manipulation was successful, or whether this was not necessarily so—a situation that may, of course, only arise in past reported environments. As seen, some languages can indicate that reported attempted manipulation was successful by selecting a particular complement form; the following examples from Bemba were quoted as (xxvi) above and are repeated here for convenience:
25 26
So Doret 1986, 23, 49. Doret 1986, 23, 39–41, 49–50 and 0.1.3 above; see also Malaise & Winand 1999
§ 908. 27
That is, ‘orders are orders’ and are not presented with any real ‘choice’ for the addressee.
complementation after notionally non-assertive verbs 131 (xxxix) Jack a-à-koonkomeshya Jill a-à-boombele (FIN) “Jack ordered Jill to work” (and she did) Jack a-à-koonkomeshya Jill a-boombe (SUB) “Jack ordered Jill to work” (but maybe she didn’t)
Exactly the same distinction is grammaticalised in Lango, a NiloSaharan language. Here complements of past reported verbs of attempted manipulation appear in the subjunctive if no comment is made concerning the realisation of the situations they describe, whereas what the original analyst calls ‘indicative’ occurs if it is specifically signalled that the target of the manipulation-attempt was reached:28 (xl)
Dákó òdiò lócà nί ‘tét (SUB) kwèrί “The woman pressed the man to forge the hoe” Dákó òdiò lócà òtètò (‘IND’) kwèrί “The woman forced the man to forge the hoe”
In both instances, the main verb remains the same and the only difference is the additional information provided by the real speaker concerning the outcome (note the change of meaning in the second sentence in (xl)). The manipulation, in other words, is no longer attempted, but equals causation in such instances. It is noteworthy that in all these instances the difference is modally marked: by subjunctive versus ‘neutral’ pattern in Bemba and by subjunctive versus ‘indicative’ in Lango, the former alternative always representing the not necessarily successful option. However, the use of the subjunctive does not have to mean that the situation was not successful and that the complement situation remained unrealised. This can be seen from the second Bemba sentence above, which carries the entailment “but maybe she didn’t (work)”. The manipulation may or may not have been successful, but this is simply left open, most likely because this is not relevant in the
28 Noonan 1985, 126–27. However, it is extremely unlikely that the form used in the second sentence is truly ‘indicative’ (=assertive). This would represent the sole attested instance of such a form after a notionally non-assertive verb anywhere. One suspects a particularly unfortunate terminological confusion here.
132
chapter three
current co(n)text or perhaps because it is clear from the latter whether it was or not. In other words, indication of the success/non-success is a modal speaker choice based on evaluation of communicative value of providing this information. Earlier Egyptian resembles Bemba and Lango in that here geminating sDm=f complements of verbs of attempted manipulation turn out to refer to successful manipulation, whenever this is contextually possible. In present and future, it is instantly obvious from the context that the complements of these verbs do not refer to actualised states of affairs, but in past reported instances they may do. The geminating sDm=f (and its functional counterparts) confirm this possibility when it is an option and indicate that the manipulation was successful. By contrast, nongeminating forms make no comment regarding success/realisation. The situations described may or may not have been actualised, but this is not in any way indicated. Instead, this possibility is left unconfirmed and without comment. This difference between the use of the two types of sDm=f can sometimes be verified from the context. Successful manipulation is certainly at issue in the following examples with dbH and wD, all of which show the geminating sDm=f/wnn: (130/129) Isis is on her way to meet her son in the entourage of gods: hA Ast r wHa in Hr dbH.n Ast wnn=f m wha m sSmw nHH Isis goes down to wHa who has brought Horus, for Isis had asked that he might be with wHa as the leader of eternity. (CT II 222b–c)
(131) Amenemhat tells of his expedition to the quarries in a rockinscription:
iw grt wD.n Hm=f prr(=i) r xAs[t tn] His majesty ordered me to go to this hill-country. (Hammamat 113, 10)
complementation after notionally non-assertive verbs 133 (132) It is said of an act of adornment: apr.n Hr it=f wsir m nbyt skr Ds=f wD.n ra irr=f sw Horus himself adorned his father Osiris with a collar of Sokar (CT VI 210h–i) because Ra ordered him to do so.29 There are not many examples of object clauses after wD where the possibility of success in the attempted manipulation could definitely be ruled out. The following example, where this can be done and where the form is ominously a non-geminating one, the complement is grammatically a subject- rather than object complement:30 (133) The deceased says to the keepers of the mounds of the Field of Reeds:
swab iAt=tn wddt iri=tn pw in wsir n Dt “Purify your mounds!” That is what you have been ordered to (Tb 149, 19–20) do by Osiris forever. However, that the distinction suggested actually is the one grammaticalised herein may be further tested e.g. by replacing the word ‘order’ by ‘force’ or ‘cause’ as the translation of wD. This fits well the above examples with the geminating sDm=f, but not many non-geminating instances, e.g.: (134) Intef relates his outstanding performance in royal service: ir grt xt nb wD.n Hm=f iri(=i) n=f st iw iri.n(=i) st mi wDt.n Hm=f irt Now, as for anything that his majesty ordered/†caused that I do for him, I did it according to what his majesty had ordered to (MMA 57.95, 8–9) be done. In the first instance, with the complement of wD referring to attempted manipulation, the second clause iw iri.n(=i)… is interpreted as adding that the manipulation was successful and the sentence is fine. But if the first clause already carries an indication of the success of the 29 Faulkner (1973–78, vol. 2, 191) has “R¿a has commanded that he make it”; however, the masculine sw can hardly function as an anaphoric pronoun to nbyt, a feminine noun. 30 See chapter 4 below for discussion.
134
chapter three
manipulation, there is no point in further reporting that this was so. The second reading is infelicitous because saying that one was caused to do something and adding that one did it is tautological and uninformative.31 This principle is exactly the same as that seen in 2.2.2 above in connection with the use of non-geminating forms after the verb rx in ‘future in past’ complements. There, the forms were used because the complements were future relative to the main verb, but also because the situations they described were known by all to have occurred prior to the time of reporting. However, in the latter case the redundancy of the information resulted from a default ‘logical necessity’. In the examples above, the marking of realisation depends more on the decisions of the speaker and his assessment of the need for providing information about the realisation of the event; it does not follow automatically from the status of the complement situation as actualised. Thus, for example with wD, the indication of success of the reported manipulation may be omitted, just as in Bemba or Lango, if the latter is clear from the context or more generally not deemed relevant enough to warrant a specific comment: (135) Amenhotep II says the subjugation of Mitanni was destined upon him:
in it=i wD iri=i st [i]mn qmA nfrw It was my father who ordered that I do it: Amun, the creator (Urk IV 1326, 13) of splendour. (136) Ra-Horakhty declares to the king: Di(.n)=i n=k nHH m nsw tAwy Awt-ib xnt anxw mi wD.n(=i) iri st Hm=k anx Dt I have given you eternity as the king of the two lands and joy before the living, according as I ordered your majesty, living (Urk IV 1349, 17–18) forever, to do.
31 The sentence would be more acceptable with ‘cause’ if the adjunct mi wDt.n Hm=f irt were to be assigned the status of ‘logical predicate’ and the whole passage (badly) translated as “As for anything his majesty caused me to do for him, it was exactly according to what his majesty had ordered to be done that I did it”. However, this is disallowed by the Egyptian original iw iri.n(=i) which is not a second tense, but an iwsentence, whose primary message is clearly “I did it”.
complementation after notionally non-assertive verbs 135 (137) Irtysen says his son has the power to make manifest the secrets of artisanship:
wD.n nTr iri=f pr n=f Xr=s God has ordered that he act for him as a revealer of it. (Louvre C14, 13–14)
(138) It is said of Amun’s ordinance of the kingship to the king:32 wD.n=f it=f tA nb dmD nn nhw=f He ordered that he seize each and every land without excep(Urk IV 1278, 4) tion. But the system works also in reverse. By ‘flagging up’ the realisation it is possible to communicate more than what seems to be actually said; for example: (139) Nebipusenwosret assures prospective visitors to his stela: iw wD.n nTr aA wnn=Tn tp tA Xr Hswt=f The great god has ordered you to be upon earth under his (BM 101, horizontal 4) favour. If the complement clause here again refers to a situation actualised as a result of the ‘great god’s’ successful manipulative speech act, the speaker will then not be promising that his audience will come to be favoured by the god upon earth, but continuity and perpetuation of this existing state. As will be seen below, indicating this particular difference seems to lie behind much of the enigmatic form-variation after the verb mri in the so-called ‘Appeal to the Living’, whence originate the vast majority of examples of this predicate with mutable clausal complements. The organisation described above accounts for most of the morphological variation in Earlier Egyptian object complement clauses of notionally non-assertive verbs. Yet, there are some instances of the geminating sDm=f after the verb wD, where the occurrence of this form does not seem to signal success of manipulation. For example, in (140) below the order is not reported and no success in manipulation is possible, but the geminating sDm=f appears regardless: 32
Similarly Urk IV 1286, 19.
136
chapter three
(140) The deceased addresses Thoth: i.nD Hr=k DHwty imy Htp nTrw Hna DADAt nb(t) ntt Hna=k wD=k prr=sn33 m xsfw wsir N Hail to you Thoth, in whom resides the peace of the gods, and the entire Ennead with you! Command that they come forth at the approach of Osiris N. (CT I 27c–28b) However, although here the speaker does not pass an order himself but asks for this to be done by an intermediary, his intent is clearly manipulative. Put another way, he is truly attempting to exert his influence over others and bring about the situation described in the complement. Closer inspection reveals the same feature also in the following instances with wD, the first of which was cited as example (126) above: (141/126) An order by the king in a decree for one of his officials: iw wD.n Hm(=i) dd=k iri.t(i) wiA n igAi n mH [...] My majesty has ordered you to have a boat of X cubits be made for Igai. (Urk I 298, 8) (142) A royal decree to an official concerning the latter’s son: iw wD.n Hm(=i) srr=f irr=f qd m spAwt (i)ptn xft wD=k irr=f m wHm=k My majesty has ordered that he be an official and gain reputation in these nomes in accordance with your order, and that he act as your herald. (Urk I 301, 3–5) (143) A further order to the same official: [iw wD.n]34 Hm(=i) dd=k iAwt nHm=k [...] My majesty has [ordered] you to award an office and take away […] (Urk I 298, 16) These sentences could of course be taken to describe past reported and subsequently fulfilled orders, but then one would have to treat the 33 So the variants B3Bo, B2Bo, B4Bo, B1P, B6C, B4C and S10C. The versions M.C.105 and T9C have pr=sn. 34 The restoration is certain; see Goedicke 1967, pl. 24.
complementation after notionally non-assertive verbs 137 texts from which they derive as intended to be purely ‘commemorative’ rather than to represent faithful reproductions of original royal orders. That is, it would have to be assumed that they were scripted anew after the matters ordered had been carried out rather than repeated verbatim what had been ordered originally at the time when the orders could not possibly have been fulfilled. This is certainly not the case: there is ample evidence that the stelae from which the above utterances derive represent exact copies of original (portable) royal decrees.35 Their purpose is not to describe past injunctions but to actually pass orders, but although in grammatical form the sentences appear to be simple reports, they serve a rather different communicative function. The speakers’ secondary illocutionary intention is the one suggested by the outward form of the sentences as simple statements, but their primary intent is manipulation and imposing their will upon the addressees or someone else. These sentences thus perform indirect speech acts, which the audience is expected to recognise and infer from the context.36 This requires no major effort, provided that the audience knows the complement situations to be as yet unrealised and the circumstances and the overall co(n)text generally are such that the utterance can be recognised to be intended as an order.37 Both these conditions are fulfilled with the examples in question, whose intent is, accordingly, to manipulate. However, this is not the case in any of the instances of wD + a non-geminating sDm=f form quoted thus far, nor in the following examples, where orders are merely related or ‘talked about’:
35 Cf. Hayes 1946, 6; Goedicke 1967, 235 and Strudwick 2005, 37. Many of the decrees name the official delivering the document as well as include instructions to set up copies thereof at the gate of the temple; see examples (204) and (241) below. 36 Searle 1975; Bach & Harnish 1979, 70–76; Levinson 1983, 263–76; Allan 1986, 204–38; Saeed 1997, 214–20. One may compare this with e.g. many ‘Appeals to the Living’ (see 3.3 below) where the standard Dd=Tn “may you say” just before the offering formula, is replaced by the construction iw=Tn r Dd, which in other contexts would hardly be understood as a request but rather a statement “you are going to say”. 37 What this latter condition may involve of course varies enormously. Palmer (1986, 30) quotes a most interesting and directly comparable example of the sentence “Private Jones will report at 18.00 hrs” used to pass an order in a military context. That it is understood as a directive is a sum of many things, including the rank of the people issuing and receiving the ‘order’, the overall context (army) where orders are routinely given and received, and so forth. For details, see Searle 1975.
138
chapter three
(144) In a broken context: wD Hm=f pr zS-qdt His majesty ordered the painters to go. (Hassan 1975 vol. 1, pl. 51A)
(145) The king is characterised: iri=f tAS=f r mrr=f nn xsf a=f mi wDt.n it=f imn nb nswt tAwy iri n=f sA n Xt=f mr=f imn-Htp HqA-iwnw-nTry He sets his boundary as he pleases without opposition, according to what his father Amun, lord of the thrones of the Two Lands, ordered his beloved bodily son Amenhotep Heka-Yunu-Netjeru to do for him. (Urk IV 1298B, 9–14) (146) The creator says about the mankind: iw iri.n=i s nb mi sn-nw=f n wD=i iri=sn isft in ibw=sn HD Ddt.n=i I made every man equal to his brother. I did not order them to do evil; it was their minds that distorted what I had said. (CT VII 463f–464b)
It seems thus that in Earlier Egyptian a grammatical difference is also made between use and mention of attempted manipulation.38 Only in the first instance is there any real attempt by the real speaker to exert influence over events and an addressee. By contrast, the sentences in (144)–(146) are patently mere recollections or ‘mentions’ of orders—passed by third parties or, as in the last example, even ones that were not passed at all—from a retrospective vantage point. They derive from texts whose purpose is not to address anyone in particular or impose the will of the speaker on others, but to record and to ‘look upon’ orders, actual or non-actual.39 This distinction is actually an additional ‘meaning’ of complements of verbs of attempted manipulation only insofar as the geminating sDm=f is concerned. In general, in sentences where non-geminating sDm=f forms occur after such a verb, no element of speaker will is
38 This differentiation may be assumed as having been made with all verbs of attempted manipulation, although of dbH there are no examples of the former kind. 39 Cf. here the discussion in Givón & Young 2002, particularly 49–51.
complementation after notionally non-assertive verbs 139 expressed, there is no intent to influence anyone and nothing beyond a reference to an act of manipulation. Further, such sentences provide no information concerning the possible success of the manipulation ‘mentioned’. The pragmatic function of such utterances is nothing like that of examples with the geminating form such as (140)–(143), which represent actual manipulation attempts, or of (130)–(132), where it is indicated that the manipulation referred to was successful. Although the geminating sDm=f may have these two alternative interpretations, there is no danger of confusion. The option of this form signalling that the speaker’s intention is to carry out manipulation is open only when the possibility of it indicating success in manipulation is excluded by the co(n)text. As seen, the latter is equivalent to the complement referring to a situation that cannot as yet have been realised, whereas in true past reported contexts success is a possibility, but manipulative intent is not. Using the predicate wD as an example, the semanticpragmatic system behind the complement form-selection after verbs of attempted manipulation in Earlier Egyptian may be expressed schematically as follows:
wD + non-gem. sDm=f
Nothing is indicated of the success of manipulation or speaker attitude
wD + gem. sDm=f
Something is indicated of the success of manipulation or speaker attitude
→ Can the situation be currently realised?
Yes
The sentence refers to successful manipulation
No
The speaker’s intent is manipulation
These functional profiles of the forms involved also explain the complement clause selection after verbs of actual rather than attempted manipulation, namely those of preventing and their opposite counterparts of causing, to be discussed next.
140
chapter three 3.2 Verbs of Preventing and rdi
The verbs of preventing xwi, xsf and sAw form a group of predicates with closely similar semantic-pragmatic profiles and, as a corollary, their complement-taking properties are also identical. However, in the surviving Earlier Egyptian data, the occurrences of these somewhat infrequently encountered verbs are restricted not only in quantitative, but also in qualitative terms. In the corpus of texts studied for the present work, there seem to be no instances of a verb definitely with this meaning in a simple non-prospective use. Instead, with object clauses, verbs such as xwi and xsf in the following examples, occur only as promises, imperatives and exhortations: (147) The mortuary priest promises to see to the safe journey of the deceased to the hereafter:
xw(=i) mHw=k I will prevent you from drowning.
(CT I 297a/B10c40)
(148) Coffin Texts Spell 859 incorporates excerpts from the Pyramid Texts offering ritual:
wsir N pn mi n=k a n nbt-Hwt xw di=s sw r=sn Osiris N; take the arm of Nepthys. Prevent her from using it (CT VII 62r–s) against them.41 (149) The deceased says to a group of divine beings: xsf=Tn pr sdb nb42 Dw m r n nTr [nb nTrt] nbt You should prevent any (?) evil impediment from issuing from (CT VI 92p) the mouth of any god or any goddess. The verb sAw is rather peculiar in that it occurs only in the imperative:
40
The other variants (T2C, T9C, Sq3C) have the form mH=k. The Children of Horus are meant. This sentence is repeated twice in CT VII 61m–p, but spelled carelessly rdi=s{n}. It is not absolutely excluded that xw might be a passive sDm=f, in which case di=s would represent its subject complement: “It has been prevented that she use it against them”. 42 The r in the original is surely an error for nb. 41
complementation after notionally non-assertive verbs 141 (150) Ptahhetep stresses the importance of incorruptibility of the elite before their fellows:
sAw Sm nma=k im m sf rn=k xr srw Take care that you do not pass on partiality therein; do not maim your name before the officials. (Ptahh 438–39) (151) The peasant reminds the high steward of his duties as an official: rdi.n.t(w)=k r dnit n mAir sAw mH=f You were appointed as a dam for a pauper; take care that he does not drown. (Peas B1, 268–69) There are also signals of eroding of the lexical status of sAw in Earlier Egyptian: occasionally it seems to be used to introduce negative final clauses:43 (152) The author reports to his superior: iri.n bAk im wdt nbt nb aws... sAw Dd nb aws iw=f gr Hr Ddt. n=i n=f Yours truly has carried out everything that the lord l.h.p. ordered… lest the lord l.h.p. say: “He is unresponsive concerning what I said to him”. (pUC 32198, 13–17) In all these examples non-geminating sDm=f forms are used, with example (147) showing a distinctive distal w-form mHw of the verb mHi ‘drown’.44 In stark contrast with the verbs of preventing, the verb rdi, with the basic lexical meaning ‘give’, is by far the most widely attested of all Earlier Egyptian main predicates with object complement clauses. Copious examples abound in texts of all sorts throughout the period of
43 As noted by Allen (2000, 253; 2002, 23). Further certain examples of similar sort are Ptahh 223 and Hardjedef I3. Allen extends this analysis also to the passage cited as example (151) above. 44 In (150) the form Sm appears after sAw. The old view of Sm as a biliteral root (Sethe 1899, § 366; GEG § 278) should perhaps be revised; given its t-final infinitive and such (admittedly late) geminating examples as Urk IV 480, 3 (Smm=t Smm N “you go, N goes”) it looks rather like a weak verb. However, the Late Egyptian form Si (e.g. LES 70, 12 and often) and the Coptic še show that Sm must have lost this status by the end of the New Kingdom.
142
chapter three
language history discussed here. The reason for its prolific occurrence is the use of rdi, already in the earliest material available and as below, as a causative: (153) The teacher says social roles are under the dispensation of god: iw=f di=f sn kt[kt] wr[w]45 He makes the least surpass the great. (Man and Son § 4, 3) (154) Heqanakhte orders his correspondent to organise a trading trip: iri n=k rdit hAy Hty sA nxt Hna sA-nb-nwt r pr-hAA Take care that Hety’s sone Nakht and Sanebnut are made to (Heqanakhte I, rto. 3) go down to Per-haa.46 (155) Sinuhe appeals to the god to allow him to return to Egypt after a long exile:
smwn=k r rdit mA=i bw wrSw ib=i im Surely you will let me see the place where my heart dwells. (Sin B 157–58)
(156) The deceased is under the protection of Atum: dw.n=f wi r bant=f n rdi.n=f Hry=i r=f He has placed me on his neck, and he does not allow me to be (CT II 40b) far from him. (157) The deceased says to a demon: ir Hm tm=k rdit pry=i r xftw=i pf n aHa=k m-bAH psDt aAt If you do not let me to go forth against that my enemy, you will not stand before the great Ennead. (Tb 65, 11–12/Nebseni pl. 66, 10–11)
In the complement clause of rdi bare sDm=f form(s), which in case of mutable roots show neither gemination/doubling nor the ending -w, 45
So the variant BM 10258II; all the other versions write snn or snny. Noteworthy here is also the use of the verb iri, ‘do, make’ as a causative; a pattern much less common than rdi, (further examples: Siut IV 10 iri.n=k dy snD xt tA “you have had fear put throughout land”; Heqanakhte I, vso. 13 iri n=k grt rdit N Hr tA “have N thrown out”). Causative iri appears to take infinitival objects, although there are certain examples that seem to suggest otherwise (Urk IV 766, 11 and perhaps Peas B1, 112). 46
complementation after notionally non-assertive verbs 143 are employed.47 However, the ending -y is relatively common—also in persons other than the 1st singular, as in (154) above. The anomalous roots iwi ‘go’ and ini ‘bring’ show the forms iwt and int; rdi and wnn appear respectively as di and wn:48 (158) The author of a letter informs his superior of measures taken to evaluate some fish:
iw rdi.n bAk im iwt itA Hr=s Yours truly has sent (lit. “has caused to come”) Ita concerning (pUC 32205, 19) it. (159) The king says that he is sending his envoy Ichernefret to Abydos… ...r irt mnw n it(=i) wsir xnty-imntw r smnx bs=f StA m Dam di.n=f int Hm=i m xnt tA-st m nxt m mAa-xrw …to make monuments for my father Osiris Khentyamenty, and to make potent his secret image with the fine gold, which he caused my majesty to bring back from Nubia in might and (Berlin 1204, 3–4) in vindication. (160) The author of a letter tries to persuade his addressee to act on behalf of a property:
swDA-ib pw n nb aws Hr rdit di.tw ib xnt pr wAH This is a correspondence to the lord l.h.p. about having attention paid to (lit. “causing heart to be put before”) the estate of (pUC 32199, 5–7) Wah. (161) Thutmosis III reveals his reasons for depicting specimens of plants from conquered lands:
iri.n Hm=i nw n-mr(w)t rdit wn=sn m-bAH it=i imn My majesty has done this in order to cause them to be in the (Urk IV 776, 13–14) presence of my father Amun. rdi has a suppletive imperative form imi, negated by m rdi:
47 See n.89 in the Introduction. For apparent exceptions, see n.98 below. Ever since Erman 1884, 31 the assumption has been that rdi-causatives are followed by a finite sDm=f, not an infinitive. There do not seem to be any reason to dispute this timehonoured analysis. 48 Again, there are some counterexamples; see n.89 below.
144
chapter three
(162) The magician Djedi has just agreed to follow prince Hardjedef to the court:
aHa.n Dd.n Ddi imi di.tw n=i wa n qAqAw int{w}=f n=i Hrdw=i Hr sSw=i Then Djedi said: “Cause that there be given to me a boat so that it may bring me my children and my writings”. (pWestcar 8, 3–4)
(163) The peasant uses metaphors in trying to get the high steward to arrange the return of his stolen goods:
m rdi it msH Do not let the crocodile seize.
(Peas B1, 255)
rdi is a syntactic (as opposed to morphological) causative49 and may also be used to express what is known as ‘permissive’ sense of enablement. In the latter case, exemplified in (155)–(157) and (163) above, rather than strictly ‘cause’, rdi refers to ‘letting’ or ‘allowing’ something or someone to carry out the action referred to in the subordinate clause. These semantic shades of rdi + event complement are best described as a continuum of decreasing causer agentivity. Schenkel has convincingly shown rdi-causatives to describe indirect causation.50 Indirect causation implies no temporal simultaneity with the causing and the state of affairs caused or spatial co-occurrence or physical contact between the causer and the causee, (the subject of the caused event) although the former is still the ‘understood’ manipulator. Perhaps the clearest indicator of the correctness of this analysis is the absence in Earlier Egyptian rdi-causatives of the so-called ‘subject-to-object raising’, where the causee appears as the direct object of the causative verb, as in
49 On lexical, morphological and syntactic causatives and further divisions thereof, the most recent discussions are Dixon 2000, 33–37; Shibatani 2001, 1–11; Shibatani & Pardeshi 2001, passim and Kulikov 2001, 886–88. A completely different taxonomy is proposed in Song 1996. For cross-linguistic examples of the grammaticalisation of the verb ‘give’ as a causative/permissive, see Heine & Kuteva 2002, 152; for the cognitive ‘motivation’ of this development, see Newman 1996, 171–94. 50 Schenkel 1999; cf. also Winand 2006, 74–75. On direct versus indirect causatives, besides the works listed in the previous note, see also Shibatani 1976, 15–16, 31– 33. More detailed semantic classifications of causation are proposed e.g. in Talmy 1976 and Kulikov 2001, 891–94.
complementation after notionally non-assertive verbs 145 English “I caused him to go”.51 As seen, ‘raising’ in Earlier Egyptian entails direct (visual) contact between the main- and subordinate clause subjects, but this requirement is not necessarily fulfilled after rdi.52 Further, in indirect causation the causee is prototypically a conscious animate being, capable of executing the event caused apart from direct contact with the causer.53 This restriction seems to hold also in Earlier Egyptian rdi-causation stricto sensu.54 Yet, the causer may be semantically incapable of agentivity, as in the next example where the issue is rather of semi-independent and -voluntary action by the ‘causee’: (164) An excuse of an Asiatic servant found guilty of consuming stolen honey:
mk in bnrt rdi iry=i st Look, it was the sweetness that made me do it. (pUC 32124, fragment ii, 10)
Finally, the permissive sense may entail the absence of all agentivity from the part of the ‘causee’ if the issue is facilitating some state of affairs by inaction, i.e. by passive lack of interference with the course of events (see below). Nevertheless, as can be seen from the examples above, the form of the complement remains without gemination/doubling irrespective of the tense, illocutionary force and the causative or permissive sense of rdi.
51 The basic discussion of the syntactic treatment of the causee is Comrie 1976a; see also Givón 1975a, passim; Foley & Van Valin 1984, 102–06; Langacker 1987, 409–13; Dixon 2000, 41–61. 52 See 2 n.122 and 2.4. However, ‘raising’ does occur after rdi when direct contact is suggested, but then the meaning does not denote causation but ‘making something/ someone as something/someone’, e.g. in the sense of ‘appointing’ with a following m or r: aHa.n rdi.n wi Hm=f m HAty-a n mnat-xwfw “Then his majesty appointed me as a towngovernor of Menat-Khufu” (BH I pl. 44, 7). This construal occurs also after iri (e.g. Urk IV 119, 3 iri.n wi ary=i m rxxy “my pen made me famous”); cf. GEG § 84. 53 See Haiman 1985, 108–111; Shibatani 2001, 14. According to Schenkel, (1999, 319) for state verbs, whose subjects are necessarily non-volitional non-controlling patients, only morphological (s-prefix) causatives are attested in Earlier Egyptian. However, there are examples of rdi followed by state verbs both with a causative (e.g. CT VII 2d di=Tn Htp N “cause N to be content”; Heqanakhte II, rto. 37 dd=k Htp=f “cause that he is content”) and permissive sense (see e.g. (55) above, with nqm, ‘suffer’). 54 Yet, the same does not hold for the enabling sense: n rdi(=i) hA inr nb “I did not let any stone go down” (Edel 1981, fig. 27, 4).
146
chapter three
As noted, the combination rdi + complement underwent a gradual grammaticalisation and amalgamation into the morphological tcausatives of Coptic.55 In Earlier Egyptian, however, this development seems to be at a very early stage: both rdi and the verb describing the caused situation (the complement) are still inflected, showing that the complex remains syntactically bi-clausal.56 Thus far, the only signs anticipating the eventual merging of rdi and its clausal object are the old notion “that in combinations of the verb ‘to give’ with a sDm.f… it is only the former that can be negat(iv)ed”57 as well as relaxation of the rules on pronominal resumption in relative constructions, as in example (159) above.58 Consequently, the complement-taking characteristics of rdi can scarcely be viewed as mere fossilised ‘automated’ construing of a grammaticalised unit, but rather may still be expected to reflect conscious choices of semantic-pragmatic coding by language users. The reasons for the absence of the geminating sDm=f after verbs of preventing and rdi alike result from the particular conceptual properties associated with the notions that these verbs describe. To begin with ‘preventing’ and its exact opposite ‘causing’, i.e. leaving aside the permissive sense of rdi for the moment, the incompatibility of the geminating sDm=f with these notions arises from the status of this form as a signal of manipulative intent or of successful manipulation with verbs with which this is possible. Indication of manipulative intent is here straightforwardly excluded inasmuch as verbs referring to causation and prevention cannot be used to make (direct or indirect) performative speech acts. That is, even sentences such as “I hereby cause that you 55
See introduction n.89. Often the combined clauses of syntactic causatives share inflection as a unit and there may be a ban on breaking their syntactic adjacency. Such restrictions reflect synchronic and/or diachronic merging of these and morphological causatives; for examples, see Song 1996, 29–36. 57 Polotsky 1960, 404 n.1; cf. also 1erný & Groll 1993, 457 for Late Egyptian. 58 According to the ‘normal’ rule, objects of transitive clauses serving as object complements of relative forms should be overt, since syntactically the resumptive is not local to the agreement-carrying expression and semantically does not refer to the direct object of the relative clause; cf. e.g. Sin B 144–45 kAt.n=f irt st r=i, literally “what he planned to do it to me” (see Collier 1991b, 39 n.72). However, in (159) the object of ini, which serves as the complement of rdi, is not expressed—i.e. one has rdi.n=f int=i ø rather than the ‘expected’ †rdi.n=f int=i st. The same phenomenon occurs also e.g. with the negative function-verb tm (see 5.3 below); e.g. pRamesseum I, Bii, 15 tmt.n=k sDm “what you have not heard”. 56
complementation after notionally non-assertive verbs 147 go” (insofar as this is a feasible utterance at all) or “I hereby prevent you from going” cannot be understood as said with an intent to bring about or prevent the ‘going’ through verbal announcement, but must be interpreted as descriptions of actions.59 The ban on marking the success of manipulation with these verbs is a slightly more complex issue. As seen above, this function of the geminating sDm=f equals signalling the complement situation as realised subsequent to the manipulation and as its result. In future contexts the non-occurrence of such marking is obvious: if it is said that something will be caused or prevented, the success in this remains to be seen. In present and particularly past contexts, the situation is rather different. Notionally, both ‘cause’ and ‘prevent’ are a particular type of implicative verbs that in affirmative past and present contexts logically entail the truth and falsity of their complements respectively.60 In such environments, ‘prevent’ functions as a negatively implicative verb that binds the speaker to accepting that the complement situation does or did not occur:61 (xli)
Jack prevented Jill from leaving › Jill did not leave Jack prevents Jill from leaving › Jill does not/cannot leave
Put another way, when something is said to be or having been prevented, that something is most definitely irrealis—the speaker cannot hold it as a fact that it occurs or occurred. This is clear e.g. from the impossibility of arguing the contrary:62 (xlii)
†
Jack prevented Jill from leaving, but she left
By contrast, one of the defining features of causation is that they entail the truth (occurrence) of the caused state of affairs:
59 Cf. here Levinson 1983, 231–33. This holds in spite of the presence of the adverb ‘hereby’, an archetypal ‘signal’ of performative illocutionary force. 60 For what follows, see Karttunen 1971, particularly 357; Givón 2001, vol. 1, 153. Cf. also Givón 1975a, 77–81. 61 However, the implication does not survive in future environments: in e.g. “Jack will prevent Jill from leaving” no truth-value can be assigned to the complement situation-description. 62 In fact, this may be a matter of degree. For example, if the preventing is understood to have taken place habitually, one could say “On various occasions, Jack prevented Jill from leaving, but she did leave every now and then”.
148
chapter three Two events can be said to constitute a causative situation if the following two conditions hold: (a) the relation between the two events is such that the speaker believes that the occurrence of one event, the “caused event”, has been realized at t2, which is after t1, the time of the “causing event”; and if (b) the relation between the causing and the caused event is wholly dependent on the occurrence of the causing event; the dependency of the two events here must be to the extent that it allows the speaker to entertain a counterfactual inference that the caused event would not have taken place at a particular time if the causing event had not taken place…63
This definition does not concern itself with the character of causativity as a spectrum of manipulative contact, causer agentive control etc., but it does provide a description of a prototypical causative scenario. Most importantly, it emphasises the fact that in true causation an affirmative implication obtains: whenever causing is affirmed as happening or having happened, the same holds for the caused state of affairs:64 (xliii)
Jack caused Jill to leave › Jill left Jack causes Jill to leave › Jill leaves/is leaving †
Jack caused Jill to leave but she did not leave
However, although it is assumed that the audience may legitimately draw “a counterfactual inference that the caused event would not have taken place at a particular time if the causing event had not taken place”, negated causation does not actually entail that also the caused situation failed/fails to occur.65 “Jack did not cause Jill to leave” does not imply “Jill did not leave” (i.e. is not synonymous to ‘prevent’). She might or might not have left, but this is not obvious and, crucially, if she did leave, this in any case had nothing to do with ‘Jack’s’ agency and causation. Similarly, “Jack did not prevent Jill from leaving” does 63
Shibatani 1976, 1–2. This criterion for causativity is disputed by Song (1996), who provides a Korean example showing just the sort of following negation marked here as ungrammatical. However, the acceptability of the negation simply shows that Song’s sentences do not describe causation, regardless whether or not a ‘causative verb’ is involved. Exactly the same holds with e.g. Earlier Egyptian: for instance, example (163) above of course contains the ‘causative’ rdi, but semantically the sentence expresses a plea for negative facilitation, i.e. in fact prevention. 65 Karttunen 1971, 357. 64
complementation after notionally non-assertive verbs 149 not imply “Jill left”. Again, she might well have left, but all that is said is that ‘Jack’ took no measures to prevent this.66 Diverse as these semantic-pragmatic notions seem, their effect in complement selection after non-future verbs of causing and preventing in Earlier Egyptian is uniform and results in the sole acceptability of distal irrealis. As seen, in Earlier Egyptian the success of attempted manipulation may be confirmed by a geminating sDm=f complement, but the use of this grammatical marking-device is conditioned by considerations of redundancy. If the success is obvious without it, e.g. if this can be readily deduced from the context, specific marking thereof is superfluous. Causation affirmed as occurring or having occurred is not attempted manipulation; it is always successful, i.e. ‘implicative’, and the occurrence of also the caused event is self-evident. Accordingly, geminating sDm=f, which is a specific signal of this, need not be and is not used. The situation is, in other words, again identical to that seen in connection with the ‘future in past’ complements of the verb rx (2.2.2 above) where non-geminating sDm=f occur for the same reason. Further, in anticipation of the discussion of this topic later on, it might be noted that matters affirmed as having been caused are not merely actualised, but more properly completed.67 With verbs of preventing, one is handicapped by the lack of non-future examples. Yet, by the same token it may not be overly venturesome to suggest that the geminating sDm=f be similarly incompatible with the Earlier Egyptian equivalents of utterances such as “X prevents/prevented event E”, where the non-realisation of E is immediately obvious. In negated causation there is an inference—at least according to the traditional analysis—that the complement situation did not occur, and even if it did, this was not due to the causer. In the first instance the ‘success-confirming’ geminating sDm=f is of course excluded. In the 66
Comparable effects are to be seen also e.g. in interrogative contexts. In “Did Jack prevent Jill from going?” the interrogative may have scope over both the main and the complement verbs, so that the speaker is asking both whether Jill left, and if not, was this due to Jack’s interference. The situation is the same in “Did Jack cause Jill to go?” But the interrogative scope may also invest just the main verb, with the speaker knowing that Jill did (not) leave and merely inquiring whether this had something to do with Jack’s causing or preventing it. Not all implicative verbs behave in a similar manner. For example, in “Did Jack manage to see Jill?” and “Did Jack avoid seeing Jill?” the interrogative must have wide scope; cf. Karttunen 1971, 345. 67 See 9.2 for a detailed discussion of this issue and its relationship with irrealis in Earlier Egyptian.
150
chapter three
event of the situation having taken place, this will have occurred in spite of the agency of the causer; i.e. the causer did not actually affect the course of events by carrying out causative manipulation. Now, in descriptions of attempted manipulation such as with the verb wD the use of the geminating sDm=f may, in appropriate circumstances, be triggered by a desire to signal that the manipulation was successful and that the complement situation was realised as a result. It is easy to see why this does not occur in ‘causatives without causation’. Even if the complement event of “I did not cause him to go” might have taken place, this did not follow as a result of ‘my’ actions, and hence it is not marked with a geminating sDm=f.68 The same holds also for the hypothetical negated prevention. In “I did not prevent him from going” the potentially realised departure will have owed nothing to the actions of the main clause subject. The permissive enabling sense of rdi can also be unified with this account. In translating combinations of rdi + complement with this meaning into English, one normally uses ‘let’ and ‘allow’ rather indiscriminately. However, these two verbs are not synonymous. ‘Let’ is like ‘cause’ in that it implies the occurrence of its complement state of affairs, whereas ‘allow’ does not:69 (xliv)
Jack let Jill leave › Jill left Jack allowed Jill to leave (Jill may or may not have left)
In fact, ‘let’ is even more strongly implicative than ‘cause’ seeing that, unlike the latter verb, it implies the non-occurrence of its complement also when negated; interestingly, negated ‘allow’ has exactly the same implication: (xlv)
Jack did not let Jill leave › Jill did not leave Jack did not allow Jill to leave › Jill did not leave
68 Of course, the complement of a verb of attempted manipulation may be subject to similar ‘independent’ realisation. An example of this scenario would be “I order that you stay here”, where the ‘staying’ may be a currently realised state of affairs, but this is not due to the speaker’s manipulative speech act. 69 See Givón 2001, vol. 2, 43–44.
complementation after notionally non-assertive verbs 151 ‘Let’ also suggests active enabling from its subject’s part; ‘allow’ refers rather to ‘not-doing’ or refraining from action. For example, “I let the water flow out of the tank” is the more likely report when the speaker has pulled out a plug at the bottom of the tank, whereas “I allowed the water to flow out” means that the speaker did nothing to stop the flow (perhaps arriving when this had already started). Earlier Egyptian does not differentiate between ‘let’ and ‘allow’ lexically; instead, rdi is the sole verb used to convey corresponding meaning. This may suggest that the conceptual difference(s) between these two notions simply were not deemed significant,70 but even if the difference was felt, with ‘let’ the sole passable form could be expected to have been distal irrealis for the same reasons as after ‘cause’. With ‘allow’, the factors determining the use of the same form in all instances will or can be expected to have pertained to the character of ‘allowing’ as inactive non-manipulation. “I allowed him to leave” means that the speaker did not set up obstacles against ‘his’ leaving, and whether or not the latter departed again had nothing to do with the agency of the speaker, who merely provided an opportunity for this by refraining from exerting manipulative force. In a sense then, ‘allowing’ resembles negative causation with respect of the implications involved, and the complement-taking properties of rdi, which expresses both, do not change. Conversely, negated ‘allowing’ equals preventing—witness their identical implicative properties—and both combine only with distal irrealis complements. In sum, the grammatical organisation of Earlier Egyptian object complementation after verbs of prevention and causation/enablement is perfectly logical. Once again, the complement construal reflects the semantic character of the main verb, the intrinsic meaning and function of the form(s) involved and the mutual compatibility of these in different illocutionary contexts. In this particular class of verbs all factors conspire to render distal irrealis the sole acceptable type of nonassertion. The principles involved in the choice are exactly the same as
70 However, there may have been ways to express ‘allowing’ more particularly by employing some alternative expression. Particularly interesting in this respect is the n wD.tw irt mnt iry n tA awt Spst in pWestcar 8, 17, which seems to mean, “it is not allowed that such a thing be done to the august flock” rather than literally “it is not ordered…” See also example (181) below.
152
chapter three
with verbs of attempted speech act manipulation; only the character of the main verbs differs. Notionally, preventing and causation/ enablement refer to either non-verbal accomplished manipulation or refraining thereof, which entail various sorts of implications concerning the successful carrying out of the complement situation and of the main clause subject’s status as a manipulating entity. Using English translations for the sake of clarity, grouping the logically/semantically synonymous situations together, and including also some hypothetical construals, the ‘chain of causality’ determining the object complement clause form of Earlier Egyptian verbs of prevention and causation/ enablement may be illustrated thus: Situation Type
Complement Event
Prevention
“I prevented him from going” “I did not allow him to go” “I did not let him go”
→ NEG implied
Possible Prevention/ Causation/ Enablement
“Prevent him from going!” “Let/allow him (to) go!” “Cause him to go!”
→ unrealised
Enablement & Noncausation
“I allowed him to go” “I did not prevent him from going” “I did not cause him to go”
→ if realised, not as a result of manipulation
Causation & “I caused him to go” ‘Let’ “I let him go”
→ POS implied
→ geminating sDm=f cannot be used
→ geminating sDm=f need not be used
As can be seen, in the majority of instances the sentential meaning results in an outright ban on the use of a geminating sDm=f (or its functional counterparts in mutable classes) in the complement, but also in affirmatively implicative contexts there is never any need to use this overt signalling of success in manipulation. Finally, some comments should be made of the morphological particulars of the complement clauses of the verbs of preventing visà-vis rdi. It is unclear why the forms in -w should be disallowed after rdi, seeing that elsewhere in complementation these are functionally similar to bare non-geminating sDm=f and forms with the ending -y. One feature perhaps worth pointing out is that in the Pyramid Texts the complement morphology of rdi and the verbs of preventing
complementation after notionally non-assertive verbs 153 present a complete mirror image of each other. In these inscriptions rdi has the same complement morphology as later, whereas unlike in texts outside this particular corpus, the verbs of preventing regularly select forms with the ending -w.71 This divide might be more than a mere coincidence, but it is difficult to see what might have been the semantic difference between the two. In any case, in the data studied for the present work, non-geminating forms without -w have invaded the functional domain of the -w-forms also after verbs of preventing. The absence of the latter after rdi in the said material is not surprising. Forms with -w had never been part of the paradigm of rdi before, and neither were they introduced there during the long decline of these forms as a living member of the Earlier Egyptian modal system. 3.3 Pragmatics of Desire: The Verb mri The verb mri is a highly polysemic lexeme used to describe a variety of ‘desiderative’ and ‘volitive’ notions, including at least ‘wanting’, ‘wishing’, ‘hoping’ and ‘loving’. 72 Typologically ‘wanting’ is the prototypical deontic volitive attitude whereas ‘wishing’ and especially ‘hoping’ verge upon epistemic, and ‘love’ is more properly emotion than attitude.73 There are also differences in the modal ‘force dynamics’ of all these stances. For example, matters ‘hoped’ are expected with greater certainty than those ‘wished’. Thus in English, only ‘wish’ combines with the modally more unreal past complements:74 71 Thus, PT 1534a reads in Twt is xw nnw=sn “it is you who prevents them from becoming weary”, PT 1242a has N pw xw nnw nTrw “King N is the one who prevents the gods from becoming weary” and PT 828a, PT 835a and PT 838c have xw=s gAw=k “she prevents you from lacking”. PT 1439d/P and PT 1440b/P have the interesting negative xsf=k w hAw N pn m wiA=k pw “do not prevent N from going down into this your barque”. Nt 293 has xw stmw=s “prevent it from perishing”. Nt 323 is an earlier version of example (148) above and has xw rdi=s sw ir=sn “prevent her from using it against her”. 72 Otto 1969; Uljas 2003, 392; cf. also Jansen-Winkeln 2002. Further examples of mri + a mutable sDm=f complement not quoted below are CGC 20712, 10–11 mr.n=f wn=i m mAdw Hr irt xt[...] “He wanted me to be in Medamud making offerings[…]”; Urk IV 341, 8 “Queen Hatshepsut… mrt.n=f wn=s Hr nswt=f whom he (Amun) wanted to be on his throne” and Urk IV 1729, 6 “[I prepared for him a temple] in a splendid place mr=f wn=f im where he wants to be”. 73 For what follows, see Noonan 1985, 121–23; Palmer 2000, 131–35; Cristofaro 2003, 103–04. 74 See 9.2 below for discussion of the linkage between past and irrealis.
154
chapter three I wish Jack would/†will leave I hope Jack will/†would leave
(xlvi)
It is also argued that unlike ‘wanting’, both wishes and hopes may be expressed towards past situations of which the speaker is merely unaware: (xlvii)
I hope/wish that Jill has come ?I want that Jill has come
Yet, as with the verbs of attempted manipulation, this actually depends on the context. Past, present and future ‘wanting’ may well be or have been directed towards actual, currently realised situations:75 (xlviii) A: Why do you still work on that same report? B: ‘The boss wants/will want me to work on it. (xlix)
A: How did you end up in this miserable job? B: I wanted to do this already when I was a child.
Here, the issue is the reason and ‘point’ of currently actual states of affairs. With main verbs of this sort, restrictions concerning the possibility of the complement situation being realised are lexical in character and pertain to the character of the complement verb.76 For example, the sentence “The boss wants me to work on it” may well be said when actually working, but the same does not hold with “The boss wants me to go”, where the sense of non-realisation derives from the meaning of ‘go’ rather than from the governing volition verb. With desiderative and volition verbs, languages do not often differentiate grammatically whether or not the situation object to such attitude/emotion is realised or non-realised. Usually, indicating that the complement in e.g. “The boss wants me to work on it” describes an actual situation can only be carried out by adding more context to the proposition, such as a further sentence “and as you can see, I am already putting my best effort to it”. However, there are exceptions to this. For example, in Mojave, a Native American language, 75 76
See Cristofaro 2003, 112. Cristofaro 2003, 112.
complementation after notionally non-assertive verbs 155 the modal marking of the complement of a unitary ‘desiderativevolitive’ verb varies according to whether or not the former is a currently actual situation:77 - (IRR) “I don’t want to get fat”
(l)
(SS) “I don’t like being fat”
Here again, the marking of this distinction is modal, with irrealis signalling the non-realised status of the state of affairs described. Earlier Egyptian belongs to the small group of languages where the ontological status of the complement situation subject to volitive/desiderative attitude(s) can be indicated modally—again by variation of geminating and non-geminating sDm=f (and their functional counterparts). After the verb mri, the ratio of these forms is approximately 1:3 in favour of lack of gemination, i.e. again the marked character of the geminating sDm=f is already suggested by statistics. Indeed, it is this form that was used whenever it was deemed appropriate to indicate that the complement situation was currently realised, i.e. part of the real speaker’s ‘reality’. Just as after verbs such as wD and dbH the geminating sDm=f and its counterparts could be used to signal success in the attempted manipulation, after mri their occurrence tells that at the situation referred to in the complement clause was meant to be understood as actual, just as e.g. in Mojave. As with the verbs of attempted manipulation, the reality of this distinction can occasionally be deduced from the context. For instance, in the following example the complement state of affairs is most likely something presently actual to the speaker: (165) Ptahshepses explains (in third person) how he came to marry a princess:
rdi n=f Hm=f zAt nzw wrt xa-mAat m Hmt=f mr.n Hm=f wnn=s Hna=f r rmT nb His majesty gave him the great royal daughter Khamaat as wife, because his majesty desired that she be with him rather (Urk I 52, 2–3) than some/any other man. By contrast, in the following example with a non-geminating sDm=f 77
Palmer 2000, 157; notice the change of translation value of the main verb.
156
chapter three
it is obvious that the complement clause does not describe a currently actual situation:78 (166) The author tells her dead addressee that she would rather welcome death for his son than subordination to an adversary:
mr(=i) in=k n=k wn aAy r-gs=k r mAA zA=k xr zA izzy I would rather that you brought away to you the one who was here by your side than saw your son subordinated to Izezy’s (Cairo Linen 8–9) son. The same holds also to the following instances, which show nongeminating and w-ending forms of the roots xai ‘appear’ and ii ‘come’ as the object of mri: (167) A female harpist sings before the deceased: [i...]w nb tpw tA swA.ty=f(y) Hr [i]s pn mrr=k xaw nTrw=k gr pr=k anx msw=k bAqw Ha=k Sw=k m snD r Dd=k Htp-di-nsw... O […] upon earth, who may pass by this tomb; as you desire your gods to appear, your house will be well-founded, your children will live, your members will be hale and you will be free from fear, if you say: “An offering which the king (Rodin Museum 275, 1–5) gives…” (168) Appeal to the Living of Ii-ib near the Nubian fortress of Kumma: i anxw tp(w) tA sS Xry-Hbt nb wn(n).t(y)=sn m mnw pn m mr=Tn iw=Tn m Htp mi Dd=Tn Htp-di-nsw... O the living upon earth, every scribe and lector-priest who will be in this fortress; as you will desire to come (i.e. return home) in peace, say accordingly: “An offering which the king gives…” (Semna-Kumma R.I.K. 51, 3–6)
As after verbs of attempted manipulation, the geminating sDm=f (or 78 In fact, this example does not express what the speaker actually wants, but what she would rather if an even worse situation arose. I.e. the complement event is in reality not willed at all but constitutes something that the speaker would not normally hope to occur. For the interpretation of this example, see Willems 1991; Wente’s (1990, 211) rendering “I had rather that you should fetch [me] away to yourself so that I might be there beside you…” assumes an omission of a dependent pronoun.
complementation after notionally non-assertive verbs 157 its functional equivalents) could again be used for a rather interesting pragmatic effect; for example: (169) Henqu maintains his worth: ink Hm wabw.n nTr r=f rdiw snD r [gs]w=f i.mr wn i[mAx=f] xr=sn m bw [ntf im] I was truly one whose mouth the god purified, who inspired fear in his peers and who desired to have reverence before them (Urk I 79, 26a–28a) wherever he might be. (170) Idu claims to have lived according to superb standards: n zp Dd(=i) xt nb Dw iw XAb r rmT nbw n mrr(=i) hrt bAq. t(i=i) wnn imAx(=i) xr nTr xr rmT Dt I never said anything evil, unjust or crooked against anyone, because I desired happiness, vindication and having reverence (Urk I 204, 9–10) before god and men forever. These well-nigh identical reports of the speakers’ past desires differ in one decisive respect. In example (169) all that is said is that in past the speaker desired something, but it is not indicated whether or not this desire bore fruit. In (170) the speaker reports the same attitude, but by employing the form wnn also further indicates that now, at the time of speaking, he has ‘reverence before god and men’, and that the complement situation is part of his current reality. Indicating this difference seems to be also the factor conditioning the variation of forms of the object verbs of mri in the so-called ‘Appeal to the Living’, which provides the source of most examples of this verb with mutable complements. A recurrent feature of this standard mortuary text carved upon innumerable stelae and intended to persuade its readers to make an actual or invocation offering on behalf of the deceased, is a premise for the subsequent plea to recite an offering formula. In the most widely attested type of the Appeal, this premise is cast into a form of an oath. The actual request is typically made by the readers’ desire to be favoured by the gods and king, to be upon earth etc:79
79 See Lichtheim 1992, chapter 4 for a convenient summary of the structure and history of the Appeal. The objects of mri vary greatly, but mostly involve immutable verbs. The most important mutable complement verbs used are Hsi ‘favour’ and wnn.
158
chapter three
(171) Appeal to the Living of Seneb: i anxw tp(w) tA swA.ty=sn Hr is pn n Xrt-nTr m xd m xsfyt mrr=Tn Hs Tn nTrw=Tn nwtw swDA=Tn iAwt=Tn n Xrdw=Tn wAH=Tn tp tA Dd=Tn xA m t Hnqt... O the living upon earth who may pass by this tomb, going downstream or upstream; as you desire your local gods to favour you, you will pass your offices onto your children and endure upon earth if you say: “A thousand bread and beer…” (Wien ÄS 156, 28–30)
(172) An Appeal to the Living of another man named Seneb: i wab nb sS nb Xry-Hbt nb rmT nbt swA.t(y)=sn Hr wD pn m xd m xntyt swDA=Tn iAwt=Tn n Xrdw=Tn mr=Tn Hss Tn nTr=Tn nwty Dd=Tn Htp-di-nsw... O every wab-priest, every lector-priest and all people who may pass by this stela going downstream or upstream. You will pass your offices to your children. As you will want that your local god favour you, may you say: “An offering which the king (Turin 1903, 4–7) gives…”
Further examples of the Appeal to the Living showing mri + non-geminating sDm=f and not quoted below are Urk I 252, 3 (mrrw Hz sn N); Berlin 7311B, 2 (mrr=Tn Hs Tn N); BM 239, bottom 6–7 (mrr=Tn Hs Tn N); BM 504, bottom 2 (mrr=Tn Hs Tn N); CGC 20164, 2 (mr=Tn Hs Tn N); CGC 20335, 2–3 (mrt=Tn (sic) Hs Tn N); CGC 20540, 2–3 (m mri=Tn Hs Tn N); Florence 2561, 5 (mrr=Tn Hs Tn N); Geneve 19583, 3–4 (mr=Tn Hst (sic) Tn N); Naga-ed-Der pl. XV.2, vertical 1–2 (m mrr=Tn Hs Tn N); Ny Carlsberg Glyptotek ÆIN 964, centre, 4 (m mr=tn Hs Tn N); Oxford Queen’s College no. 1113, bottom 4–5 (m mr=tn Hs n (sic) Tn N); Turin 1546, 5 (mrr=Tn Hs Tn N); Turin 1547, 4–5 (m mr=Tn Hs Tn N); Tübingen University 458, 8–9 (m mrr=Tn Hs Tn N); Wien ÄS 166, 16–17 (mr=Tn Hs Tn N); Wien ÄS 186, back, 3–4 (mr=Tn Hs Tn N). The execrable stela Helck 1975, no. 49, 10 has mr=rn (sic) Hs rn (sic) {n} N. For the form wn=Tn/tn, see n.80 below. Further examples of the Appeal with mri + geminating form/wnn not quoted below are Urk I 268, 13 (mi mrr=Tn wnn=Tn); Bologna 2, 6 ([…]mr[...]=Tn Hss Tn N); BM 579, 4–5 (m mrr=Tn Hss Tn N); Florence 7599, x+2–3 ([m? mr/mrr]=Tn Hss Tn N); Turin 1447, middle panel, 4 (m mrr=Tn wnn imAx=Tn xr wsir); Dendereh pl. 2A, bottom left, 3 (mi mrr=Tn wnn=Tn m Sms n nTr=Tn nwty). There are also unclear examples with Hsi written without phonetic complements (e.g. Berlin 7732B, 2; BM 225, 2; BM 471, 2; CGC 20043, 2; CGC 20100, 4; CGC 20396. c3; Florence 2500, vso. 7; Leiden F95/83, 2; Felsinschriften no. 429, 7; Sinai 118, vertical 3; Tübingen University 479, vertical 9; Wien ÄS 168, 6; Zagreb 7, 3; Zagreb 8, 5 and numerous others).
complementation after notionally non-assertive verbs 159 (173) Appeal to the Living of lady Nefertut: i anxw tpw tA Hmw-nTr nw nTr aA m mrr=Tn wnn imAx=Tn xr nTr aA prt-xrw st=Tn n(=i) mw O the living upon earth and priests of the temple of the great god; as you desire that you have reverence before the great god (BM 152, 3–4) and invocation-offerings, pour water for me. It is argued here that the difference between these variants is that whereas the speaker e.g. in (171) simply appeals to his addressees’ desire to be favoured by their local god, the authors of (172) and (173) have included in their versions the additional and rather courteous suggestion that gods actually do favour them. The use of the different forms in the Appeal is clearly conditioned by personal taste and presumably also general and even local conventions.80 Various observations support this interpretation. There exists also a diachronically earlier version of the Appeal that makes no direct request for offerings but instead, in its usual form, declares those desirous of royal or divine favours to be ones who shall recite the invocation. The overall construction is that of a nominal sentence consisting of a participial subject plus a further participle or sDm.ty=fy serving as a predicate, with the copula pw usually omitted. In this variant, mutable complements of mri regularly show gemination/wnn:81
80 For example, it is notable that there are no variants of the Appeal showing the form wn as the object of mri. The apparent counterexample on the unpublished MMA 65.120.2 (line 3) is wrongly cited by Vernus (1990, 170) as “mrr.tn wn.tn tp tA”. The correct reading is mrr=Tn wn tp tA, with the infinitive (James P. Allen, PC). Occasionally sources from the same geographic area make curiously similar choices as regards the form after mri, perhaps indicating local traditions in the grammatical construal of the formula. For example, Appeals from memorials associated with the shrine of Heqaib on the island of Elephantine that refer to desiring ‘favours’ from the deified noble (20, 2; 52, e3 and 88, 2–3) are consistently cast in the form m mr(r)=Tn Hss Tn N, with gemination; Heqaib n. 49, d3 writes m mr=Tn mAA=Tn. 81 See Vernus 1976, 132, note (c). Further examples (all with the form Hss of the verb Hsi) are CGC 20046, 1; CGC 20523, 1–2; Berlin 1183, 2–3 and Sinai 519, 4–5. In two instances (Sinai 502, 1 and Sinai 510, 1) Hsi is written without phonetic complements; there is also one rather complex example with wnn (Tomb of Merefnebef, pl. 16/32, 2), and the damaged BM 1678, 5 ([…] mrrw wnn imAx=sn xr wsir) seems to belong here as well. The inevitable counterexample occurs in Wadi Hilâl N6, 8: mrrw Hs sn nxbt nbt nxb Ddw rn(=i) nfr “Those who desire Nekhbet, mistress of El-Kab to favour them are those who say my good name”.
160
chapter three
(174) Appeal to the Living of Amenemhat: mrr Hss sw wsir xnty-imnty xrp tAwy qaH n=i a=f dd xt m HtpnTrw Hr wDHw n wsir n twt=i pn nty r r-pr pn He who desires that Osiris Khentyamenty, the ruler of the Two Lands, favour him is the one who extends his arm for me and places offerings on the offering-table for Osiris and for this my (BM 462, right, 6–7) statue, which is in this temple. (175) Appeal to the Living of Qay: mrr Hss sw xnty-imntyw Dd.t(y)=f(y) xA m xt nb n imAxy qAy mAa-xrw He who desires that Khentyamenty favour him is the one who will say: “A thousand of everything for the revered Qay, (CGC 20567, a1–2) justified”. As in the geminating examples of the more common type, also in this version of the Appeal the function of the premise is to ‘flatter’ the audience to carrying out the plea. The initial mrr Hss sw varies commonly with mrrw nsw (pw) Dd/Dd.ty=sn “beloved of the king (etc.) are those who (shall) say” where the tense of the participle mrr shows the idea to be that persons so addressed are beloved of the king, and to appeal to their desire to remain so. In mrr Hss sw, by analogy, the actual sense intended would seem to be “he who desires to remain favoured by X is the one who will say”, with Hss referring to an actual state of affairs rather than a mere prospect. As further support of this analysis it may be noted that, as will be seen later on, the variation between gemination and non-gemination of the main verb mri in the Appeal is also determined by exactly the same parameter of realised/non-realised.82 In contrast, there is no correspondence between the complement type and the persons addressed,83 nor the presence or absence of the prepositions m/mi before mri. The formal properties of the main and the subordinate verb do not correlate and also the ‘embedded second tenses’-hypothesis founders due to lack of adjuncts following the geminating form in many
82
See 6.2 below. Lichtheim (1992, 163–64) argues for some degree of person/composition correspondence. 83
complementation after notionally non-assertive verbs 161 instances (cf. e.g. (172) and (175) above). In past it was suggested by the present author that just as in English the strength of the attitude expressed affects and conditions the grammatical form of complements of volitive and desiderative verbs, in Earlier Egyptian, with a single polysemic verb mri, some of the variation between the geminating and non-geminating sDm=f forms may have conveyed delicate gradations in the volitive/desiderative ‘force dynamics’.84 In particular, the geminating form was argued to perhaps having signalled stronger volition/desire being expressed vis-à-vis non-geminating sDm=f forms. However, this analysis must now be considered erroneous. Firstly, if the variation between the forms after mri truly reflected the suggested contrast, this would mean that the majority of the Appeals to the Living—namely those with a non-geminating form—spoke of their audience as desiring favours etc. only in weak and tentative terms. This is surely not the case. Secondly, the variation would then actually represent a vehicle for indicating gradations in the attitude expressed by the sentence subject, not the real speaker. Yet, in all instances of Earlier Egyptian complementation discussed thus far, the subjective stances conveyed are attributed to the real speaker. Finally, there are also syntactic objections to this analysis. In the Appeal to the Living, the premise for the plea for the readers to recite the offering formula is frequently expanded by additional clauses. The standard practise is to treat these as further complements of mri, as follows: (176) Appeal to the Living of Nehy: i anxw tp(w) tA sS nb wab nb Xry-Hbt nb Hm-kA nb rmT nbt swA.t(y)=sn m xd m xsfyt Hr Spss pn mrr=Tn Hs Tn nTrw wAH=Tn m anx mn=Tn Hr nswt=Tn Dd=Tn xA m t Hnqt... O the living upon earth, any scribe, wab-priest, lector priest, kA-priest and all people who will pass by this memorial going north or south; as you desire that (your) gods favour you, that you will endure in life and that you may be firm on your seats, say: (Florence 2506, 1–7) “A thousand bread and beer…”
84
Uljas 2000, 128; 2003, 391–92.
162
chapter three
There is no guarantee, however, that this interpretation of the additional clauses’ syntactic status is correct in every, or even most instances.85 When there is slight variation in the phrasing and sequencing of the clauses, sometimes those following the first one after mri cannot be understood as object complements, but rather have the appearance of wishes or assurances if the recitation is carried out:86 (177) Appeal to the Living of Intef: i anxw tp(w) tA mrrw anx msw=sn Hss tn nTr=tn niwty Dd=Tn t Hnqt xA... O the living upon earth, who desire their (=sn) children to live; your (=tn) local god favours you (tn) if you say: “A thousand (Leiden AP 72&73, middle 1–3) bread and beer…” (178) Appeal to the Living of Meh-ib: i anxw tpw tA wab nb sS nb rmt nbt Xry-Hbt nb aq.t(y)=sn r is pn mr=tn anx msd=tn xpyt swDA=tn iAt=tn n Xrdw=tn mi Dd=tn Htp-di-nsw... O the living upon earth—any wab-priest, scribe, all people and any lector-priest who may enter this tomb; as you will love life and hate death, you will pass your offices to your children if you say accordingly: “An offering which the king gives…” (CGC 20530, 1–3)
But when the additional clauses certainly represent further complements of mri and employ verbs from mutable classes, if the purpose of the
85
Cf. Lichtheim 1992, 168–71 who favours the universal interpretation as complements with Sethe (1927, 88 n.23) who suspects amalgamation with another version of the Appeal consisting of wishes. 86 Contra Lichtheim 1992, 173 who argues that this version appears only after the XVIII dynasty. There is a multitude of Appeals before this date where there is no premise (m) mr(r)=Tn, but which are of the type Hs(s) Tn N (mi) Dd=Tn and which are undoubtedly to be translated “X will favour/favours you if you say…” In the XVIII dynasty this is almost the sole form encountered. Some Appeals repeat the phrase m mr(r)=Tn, (e.g. BM 805, 4–6; CGC 20119, c3–4; CGC 20536, 4–6; CGC 20538 I, d3–5; CGC 20683, 2–4; BM 1213, vertical 6 has the sequence m mrr=Tn... mrr=Tn and Pushkin Museum I.1.b.32/UCL 14326, x+6 mr=Tn... m mr=Tn) but apparently only with noun/infinitive objects. The presence or absence of the preposition mi before Dd=Tn provides no guide to the organisation of the preceding clauses.
complementation after notionally non-assertive verbs 163 form variation was to indicate degrees in the strength of the attitude expressed, the complements should, a priori, be of the same type lest one and the same governing predicate be interpreted as a ‘source’ of two different modal ‘forces’ simultaneously. Yet, this prediction fails:87 (179) Nekhebu appeals to the visitors to his tomb: ir wn mry=Tn Hz Tn nzw wnn imAx=Tn xr nTr aA aq=Tn -w r iz pn zbt.tw[ny] If you desire your king to favour you and that you have reverence before the great god, do not enter this tomb in an impure state. (Urk I 218, 8–10) (180) Ankhmahor addresses those in charge of interring his mortal remains:
(i)n-iw mry=Tn Hzy Tn nzw prt-xrw (n)=Tn m Hrt-nTr wnn imAx=Tn nfr xr nTr aA dd=Tn n(=i) aA pn n qrs(=i) pn Hr mwt=f m bw mnx n rxt=Tn Do you desire the king to favour you, with your voice-offerings being in the necropolis, and that your reverence be good before the great god? Then put for me this lid of this my coffin upon its ‘mother’ in as careful a fashion as you can. (Urk I 205, 2–6) However, here again the variation is unproblematic if, as proposed, it equals indication versus no indication of actuality of potentially realised states of affairs. To conclude, the system by which the complement form is selected after the verb mri is a simpler version of that seen after verbs of attempted manipulation. Seeing that mri describes volitive and desiderative attitude rather than verbal (or real) manipulative acts, the ‘manipulation
87 In example (179) both Hz and wnn must belong to the protasis and the same holds also for the interrogative (180) (similarly Urk I 205, 12–14, with Hz + wnn). The sentence type exemplified in (180) should not be confused with examples such as in-iw mry n=Tn Hz Tn N “Is it desirable to you that N favour you?” (Urk I 217, 16 and Urk I 218, 16) which appear to be adjectival sentences with dative n and the clause headed by Hzi functioning as the subject; see section 4.2 and example (201) below.
164
chapter three
intended’ -function of the geminating sDm=f is not open here.88 For the same reason the realisation of the situation, which the form does signal here, does not have to represent a result of the attitude. Once again, the two types of mutable sDm=f have the same meaning(s) that they have elsewhere after notionally non-assertive verbs, but whether or not their entire functional scope is exploited depends on the more precise semantic-pragmatic character of the governing predicate. 3.4 Some Problems For the sake of completeness and to suggest that there might still be features to be discovered in the object complement selection of Earlier Egyptian notionally non-assertive verbs, a few problematic examples should be briefly noted. In the following instances of the verbs wD and dbH, the occurrence of the form wnn in the complement is not predicted by the present analysis: (181) An admonition in a royal decree protecting the mortuary foundation of a vizier:
ir rmT nb nw tA pn mi-qd=f Xnn.t(y)=sn xb.t(y)=sn xt xnt wabw=k... n gr wD.n Hm(=i) wnn=sn tp Axw [m] Xrt-nTr wp-r wnn=sn snHy nTTy m Xrw-mdw nw nzw wsir nw nTr=sn nwtyw As for any people of this entire land who will interfere with or jeopardise matters relating to your offering-arrangements… my majesty does not allow89 them to be among the blessed dead in the necropolis, but rather they shall be fettered and bound as ones under accusation of the king, Osiris, and their local gods. (Urk I 305, 8–306, 1)
88 It may be noted that in Earlier Egyptian present tense ‘want’ is not used as a directive in a fashion common to English (e.g. “I want you to go now”). There is also a curious dearth of examples of speakers expressing their own current volition by means of the verb mri + object complement. A rare example of this sort is Admonitions 4, 2 mr=i mt=i “I wish I died/were dead”. 89 See n.70 above for this translation.
complementation after notionally non-assertive verbs 165 (182) Isis says to newborn Horus: bik sA=i Hms rk m tA pn n it=k wsir m rn=k pw n bik Hr snbw Hwt imn-rn dbH=i wnn=k m Smsw ra Oh falcon, my son. Dwell in this land of your father Osiris, by this your name of ‘Falcon upon the ramparts of the mansion of the hidden name’, and I shall ask that you may be in the (CT II 221c–222a) following of Ra. In (181) no explanation for the use of this form here seems readily available.90 In (182) the main clause clearly has a future reference, i.e. the complement is unrealised, but no manipulative intent seems to be involved.91 In the latter instance, all the variants are also unanimous in their use of the form wnn.92 With verbs of preventing, the analysis proposed here forbids the occurrence of geminating sDm=f complements in all circumstances. However, there exist two counterexamples to this rule, the first of which is the only certain instance of the verb dr with a finite complement,93 the second also the apparently sole occurrence of nHm, ‘obstruct, prevent’ in a similar construct (with the verb xsf as a variant): (183) The deceased announces his resurrection intact: in anDw dd n=i a=f dr=f wnn=i m-m anDtw wsir It is Dawn who gives me his arm and prevents me from being (CT III 206c–d) amidst the executioners of Osiris.
90 One could possibly appeal to Allen’s opinion that in the PT the ‘prospective’ sDm=f is used after wD instead of the ‘subjunctive’ (1984 § 228) and argue that wnn is the former. However, although PT 1480c/P shows Hmsw N (M/N Hms) and PT 1596b nnw=T, in Nt 490 and Nt 781 one reads Hms=k, PT 467a has xa N, PT 967c i.iry=f and PT 1295a hAy=k. If the ‘prospective’ analysis is to be insisted upon, resorting to defective writings etc. is unavoidable and even this does not account well for spellings such as i.iry. 91 Faulkner translates this passage as “I ask that you shall always be in the suite of R¿a” (1968, 41; 1973–78 vol. 1, 126), noting that the form wnn is used “because the state of affairs envisaged is desired to last for ever” (1968, 43 n.27). 92 Thus S1P has dbH=i wnn=k and S1Chass. DbH(=i) wnn=k; S1Ca and S1Cb have dbH(=i) wnn kt, with the suffix =k apparently replaced by the word ‘another’. 93 pEbers 73, 20–21 has “various ingredients for treating a swelling… wt Hr=s r dr hA mw iry bandage it to prevent the water within from issuing out.” Yet given the overall sense and the fact that the variant pHearst 9, 8 has r rdit hA mw iry “to let the water within to issue out”, the correctness of dr is most doubtful.
166
chapter three
(184) The deceased threatens those who may impede his access to nourishment in the hereafter: B3Bo ir min nHm.t(w) hAA N tn r swr nn n mw T1C ir min nHm.t(w) hA=i r swr nn n mw B5C [ir] x[s]f[.tw] hAy=i r s[wr] m mw
sA.kA rf bikt wrt r tA Hrt-ib iwnt If today I am obstructed from descending to drink this water, then the great female falcon in the midst of Heliopolis will (CT V 21c–d) proceed to earth. Yet, for dr the translation ‘prevent’ in example (183) is merely a deduction.94 Dictionaries do not recognise this verb as taking verbal, but only noun objects, and then with a sense varying between ‘subduing’ and ‘removing’.95 In (184), besides B3Bo, there are two further variants showing the geminating form hAA, but of these Sq11C begins with an unfathomable ir minti N pn,96 and S10C actually has ir xsf.t(w) hi=s hAA=i r sA, apparently meaning something like “If her husband is rejected, I will descend to sA-water”. Further, as can be seen, even the less obscure versions do not agree on the complement form, and it is the ones preserving the supposedly original 1st person singular that have the non-geminating hA=i and hAy=i. Finally, there is one example of a geminating sDm=f after mri, where the nature of the event described by the complement clause verb excludes the possibility of it representing a currently actual state of affairs (cf. the discussion of the sentence “the boss wants me to go” above): (185) Appeal to the Living of Igai-hetep near the Nubian fortress of Kumma:
i anxw tp(w) tA sS Xry-Hbt wab nb wnn.ty=sn m mnw pn m mr=Tn xdd=Tn mi Dd=Tn Htp-di-nsw... O the living upon earth, any scribe, lector-priest or wabpriest who will be in this fort; as you will desire to sail north, say accordingly: “An offering, which the king gives…” (Felsinschriften no. 400, 1–2) 94 95 96
The translation follows Faulkner 1973–78 vol. 1, 175. See Wb V, 473–74; Faulkner 1996, 314–15; Hannig 2003, 1479. In CT V 19g.
complementation after notionally non-assertive verbs 167 There are two issues to be noted in connection with these problem cases. Firstly, three of the five problematic examples above involve the verb wnn. Secondly, by far the greatest number of ‘exceptional’ instances in fact occurs after the verb rdi after which—again—the writing wnn is relatively common, although mostly in early XVIII dynasty sources.97 Also the form mAA for ‘see’ is frequently met after this verb.98 The usual tendency among Egyptologists is to assume textual corruption in such instances.99 This is not to suggest that the examples cited in this section should all be dismissed as erroneous; merely that this possibility might account for at least some of these instances, particularly when the ancient authors themselves seem to disagree as to what is the correct form. In any case, the analysis of the character of the geminating and non-geminating sDm=f forms and their functional counterparts presented above is capable of accommodating all the other examples of these forms serving as object complements of Earlier Egyptian notionally non-assertive verbs. This is in spite of the possibility that there might still be discoveries to be made in the study of these constructions, as seems to be suggested by the problematic instances noted. 3.5 Interim Summary In Earlier Egyptian, the use of bare geminating and non-geminating sDm=f forms and their functional counterparts as object complements of notionally non-assertive verbs obeys quite ‘meaningful’ parameters. In these environments, the non-geminating form(s) have the same function regardless of the governing verb. They do not indicate the real speaker’s attitude towards the situation described, nor signal anything
97
A rare Middle Kingdom instance is Steindorff 1896, 4, 3 (rdi.n=s wnn=k, written wn=k in 10, 3). In the Book of the Dead and funerary stelae of the early XVIII dynasty, examples are too numerous to be listed here; from the texts collected together in Urk IV one may mention e.g. 1843, 18; 1844, 15 and 1939, 5. 98 Instances of rdi + mAA are common in the Book of the Dead of Nu; e.g. Tb 78, 27/Nu pl. 40, 31–32 (mAA=i, other versions mA/mAn=i); Tb 112, 4/Nu pl. 53, 4 (mAA=i, other versions mA=i); Tb 144/Nu pl. 76, 49 (mAA N); Tb 156/Nu pl. 79, 5 (mAA N). The same form occurs also in e.g. pEbers 43, 17 (repeated in 93, 12, but written mA); in 51, 19 one finds the writing hAA of the verb hAi, ‘descend’. 99 See locus classicus GEG § 458; a partial exception in this respect is Malaise & Winand 1999 § 611.
168
chapter three
particular about its ontological status as seen from the perspective of the same participant. The geminating sDm=f, by contrast, is bi-functional and marked in both these respects. Yet, it depends on the semantic nature of the governing verb whether or not the form can be used in both of its functions and whether it can be used in either of them, i.e. whether it can be used at all. With the non-geminating forms, no particular mental processing is required from the hearer’s part to decode the meaning of the form. Analytical ‘spelling out’ of their semantics e.g. after negative implicative verbs of preventing creates an illusion of undue complexity, whereas in reality the interpretation is automatic here and elsewhere where the forms occur.100 With the geminating sDm=f, two interpretations are potentially open, and the exact modal function and ‘meaning’ of the form is determined by the syntagmatic environment in which it appears; that is, it is construction-specific. However, this is actually the case only after verbs of attempted manipulation, and even then recovering the correct semantic interpretation of the form involves but one context-based deduction—namely whether or not the situation can be realised. This is considerably less than what is usual in Earlier Egyptian. One need only think of adjunct clauses with the bare sDm=f/sDm.n=f with which the ambiguity between causal ‘because’, ‘since’, result ‘and so’, temporal ‘while, when, as, after’, conditional ‘if’, continuative ‘and’, final ‘so that’ and even relative ‘which’ -readings can never be wholly removed and whose translation will always be based on the most likely interpretation of the context.101 In stark contrast to this vacuity, interpretation of the geminating sDm=f complements of verbs of attempted manipulation depends on a single variable. Moreover, linguistic items are hardly ever associated with a single self-sufficient semantic-pragmatic value and function. Instead, ‘meaning’ is dependent of use, and contextual deductions are decisive in interpretation. This is particularly the case with modal expressions. For example, without context it is impossible to decide whether the word ‘must’ in “Jack must be at home” indicates deontic or epistemic attitude. An interpretation can be achieved only by adding some further context such as “at six o’clock sharp”, (deontic 100 For example, the deduction that the complement situation in a sentence such as “Jack prevented Jill from coming” remained unrealised requires no particular ‘computing’. 101 GEG § 211 is a pertinent monument to this difficulty.
complementation after notionally non-assertive verbs 169 obligation) or “because I can see that the lights in his room are on” (epistemic deduction) to the original utterance. Also the geminating sDm=f is a multifunctional entity, and speakers rely on their audience’s ability to use context to recover the exact sense that the form employed is intended to communicate. Context is an essential component of meaning of modals, and this general principle applies to Earlier Egyptian as much as to all human languages. Discussing the organisation of object complements after predicates combinable only with irrealis clauses separately from those potentially followed by assertions has methodological advantages, but should not be taken to imply that these groups are wholly separable. Of course, the notionally non-assertive verbs do form their own group by never combining with realis complements. In addition, the rules dictating the use of one or the other of the irrealis patterns appear rather different from those seen in connection with notionally assertive verbs, and the same holds also for the functions of the forms in question. However, these differences are only apparent. The principles of complement selection and the functions of the geminating and non-geminating sDm=f discussed in this chapter can be unified into one generalisation of the use of irrealis in object complements of notionally non-assertive and assertive verbs. In the former group, the variation of the different sorts of sDm=f is a method for providing information of the speaker’s attitude towards and reality-status of the subordinate situation. Signalling the latter explicitly is, as pointed out in various occasions above, clearly conditioned by the relevance of this information. That is, one encounters here, just as after assertive verbs, the same basic roles of grammatical indication of modality: attitude and informational relevance. Further, after the non-assertive verbs, if the complement is, in abstract terms, close to the speaker attitudinally (i.e. is a ‘goal’ of a manipulation-attempt and hence expected, desired and subject to positive orientation) or in terms of deixis, (i.e. is signalled to represent a currently actual situation) a geminating form or some of its functional counterparts is used. If the complement situation is remote attitudinally (is subject to negative or indifferent speaker attitude), or deictically, (is non-actual or not indicated as actual for whatever reason) a nongeminating sDm=f appears. In other words, the bare forms again behave as expressions of proximal and distal irrealis, and these their
170
chapter three
more abstract characteristics are, of course, the same after notionally non-assertive verbs as after verbs of locution, cognition and perception. Thus, ultimately, Earlier Egyptian object complementation can be conceived as one overall semantic and morpho-syntactic continuum of modality. The ‘upper’ end of this is occupied by firm committed assertions of information viewed as optimally relevant and the ‘lower’ by propositions such as the complements of affirmative verbs of preventing, in which various irrealis-inducing factors including unreality and absolute lack of commitment (‘negative implicativity’) are present simultaneously. Sections of this meaning-continuum represent domains of the different formal (grammatical) patterns used to code the various semantic-pragmatic notions: ntt/wnt-clauses and the bare geminating and non-geminating sDm=f forms. The resulting image of the Earlier Egyptian system of object complementation as a spectrum of speaker attitude and informational relevance with superimposed grammatical entities is considerably more intricate than the onesize-fits-all, slot-and-filler-model based on syntactic substitutability of forms and constructions. However, this only reflects the astonishing complexity of the notions that the system is designed to communicate. All users of language are painfully aware of the difficulty of expressing one’s attitudes and emotions verbally. It is thus a small wonder that grammatical organisations serving precisely this purpose should turn out to be complex in an equal measure.
modality in affirmative subject complement clauses
171
PART TWO
MODALITY IN OTHER TYPES OF EARLIER EGYPTIAN COMPLEMENT CLAUSES
172
chapter four
modality in affirmative subject complement clauses
173
CHAPTER FOUR
MODALITY IN AFFIRMATIVE SUBJECT COMPLEMENT CLAUSES 4.1 Introduction If object complementation can justly be said to represent a littleexplored area of Earlier Egyptian grammar, the same characterisation has the flavour of an understatement when subject complements are at issue. Complement clauses of this type are far less common than object complements in all languages and have not attracted nearly as much attention in general linguistics either. Yet, in Egyptology there has been an almost total neglect thereof—save for the regular inclusion of subject complements with sDm=f forms in general descriptions of ‘noun clauses’ and in illustrations of functions of the ‘nominal forms’.1 However, subject complement clauses are most interesting from the perspective of semantic-pragmatics and realis/irrealis modality, inasmuch as their grammar in Earlier Egyptian and elsewhere is, like that of object complements, determined by their status as either asserted or non-asserted. In sentences with clausal objects the modality of the subordinate clause is, as seen, dependent of the attitudes, evaluation and perspective of the real speaker. The same holds also for subject complementation. Also here the individual whose subjective stances regarding the complement proposition decide the grammatical form of the subordinate clause is the speaker, who either asserts it or has some reason to frame it as a non-assertion, i.e. irrealis. The prerequisites for assertion are the same in subject complementation as elsewhere. As a preparatory condition, the speaker has to be capable and willing to express commitment towards the complement proposition:2
1 2
See 0.1.1 for bibliographical details. Examples from Palmer 2000, 123 and Butt & Benjamin 1988, 245; 2000, 253.
174 (li)
chapter four Italian:
Si capisce che sono (IND) arrabiati “It’s clear that they are cross”
Spanish:
Lo increíble era que Pedro no lo sabía (IND) “The incredible thing was that Pedro didn’t know about it” Es imposible que lo dijera/dijese (SUB) “It’s impossible that he said it”
In the Italian and the first Spanish sentence with the indicative, the complement clause situations stated as being or having been ‘clear’ and ‘incredible’ are presented as information towards which epistemic and emotional value judgements are expressed, and in both instances the speakers are committed to the complement state of affairs.3 However, this is of course not the case in the second Spanish sentence above, where the subjunctive is used instead. Non-commitment may be overtly expressed by governing predicates and expressions such as ‘be unlikely, improbable, not the case that’ or, as above, ‘be impossible’.4 Yet, as in object complementation, it may also result from some quite unpredictable ‘situational’ factor; for example:5 (lii)
Spanish:
Parece que lo hace (IND)/haga (SUB) a propósit “It seems like he’s doing/might be doing it on purpose”
3 The use of expletive pronouns such as si/es/it here as non-thematic ‘dummy’ subjects semantically co-indexed with the complement is typical for many modern languages. However, this is a cross-linguistic variant; in Earlier Egyptian the use of co-indexed pre-verbal cataphoric expletives with finite complements does not seem to be licensed; see 4.3 below. 4 Cf. Hooper 1975, 92, 112–14; as expected, there is some variation herein in how ‘borderline’ expressions are treated in different languages. For example, in French the indicative is used in the complement of ‘be probable’, but the subjunctive appears after the less certain ‘be possible’ (Ferrar 1967, 94–95). Yet e.g. in Spanish and Italian no similar gradation is expressed and both these verbs take the subjunctive (Wierzbicka 1988, 148). 5 Lunn 1989b, 689; see also Wierzbicka 1988, 143–44 for an identical situation in French.
modality in affirmative subject complement clauses
175
Here the variation between assertion and non-assertion again expresses greater and lesser speaker confidence respectively, but it is not apparent from the sentence per se what might have motivated the speaker to adopt one or the other of these attitudes. Rather, this depends once again on the (linguistic and extra-linguistic) context in which the utterance is made. Non-commitment is particularly common when reference is to situations purely hypothetical, as in the following examples where the subjunctive/past tense/if in the complement indicate the same lack of certainty; often the governing verb is also a non-assertion:6 (liii)
Spanish:
Lo peor será que no venga (SUB) nadie “The worst thing will be if no-one comes”
French:
Il est important que tu le fasses (SUB) tout de suite “It is important that you do it at once” It would be unfortunate if he came It’s good for him if he does it
But in subject complementation the information value of the complement clause is at least equally, or even more important as commitment in the selection of an appropriate subordinate clause modality. In the examples (li) above, the indicative also serves to signal that the speaker considers the content of the complements to represent information new and relevant to the hearer in the current context of communication. Yet, as in object complementation, this need not be the case everywhere. In fact, more often than not, situations described in subject complement clauses are presupposed, i.e. accepted and assumed as known by the hearer(s), and, because of this presumption of mutual consensus, treated as indisputable background information in no need of asserting.7 Many languages employ non-indicative forms and mood in such instances:8 6
The Spanish and French examples from Butt & Benjamin 2000, 253 and Ferrar 1967, 95. The second English sentence may alternatively be understood as presupposing the complement, although in such a case that would be more appropriate. For the use of past tense as irrealis, see chapter 9.2 below. 7 See 0.1.2 above. 8 Examples from Maiden & Robustelli 2000, 323; Butt & Benjamin 2000, 251 and L’Huillier 1999, 164. However, as everything in modality, also this varies from
176
chapter four
(liv) Italian:
È normale che i ragazzi facciano (SUB) baldoria dopo gli esami “It’s normal for kids to live it up after the exams”
Spanish: Es natural que esté (SUB) alterada “It’s natural for her to be upset” French:
Il est logique qu’il faille (SUB) utiliser le subjonctif dans ce cas “It is logical that one should use the subjunctive in this case”
Note also the English use of the modal should e.g. in the translation of the last sentence above, which similarly signals the concessive status of the complement. The function of the sentences in (liv) is to comment on various ‘facts’—that they are ‘normal’, ‘natural’ or ‘logical’—whose status as existing states of affairs is completely taken for granted by the speakers. But crucially, the speakers also assume the hearers to be aware of them, and consequently, the complement propositions are unasserted. By changing the mood in the complement, speakers may, within limits, indicate whether or not they consider the information presented in the complement to be part of the discourse background assumptions.9 For example, the subjunctive may be used instead of the indicative in (li) if the speaker believes the complement to be information already known to the audience:10 (lv)
Si capisce che siano (SUB) arrabiati “It’s understandable that they should be cross” Lo increíble era que Pedro no lo supiera/supiese (SUB) “It was incredible that Pedro should not know it”
language to language. In these sentences the indicative would be used in e.g. Russian, Swedish and Finnish. 9 When themselves asserted, many governing verbs expressing value judgements (e.g. ‘be odd, strange, interesting, relevant’ etc.) function mostly as presuppositiontriggers. An assertive interpretation of their complements in such circumstances is somewhat difficult, although not excluded; cf. e.g. “it is strange if he really is there” (the speaker is not wholly convinced). 10 See n.2 above; notice also the change of sense in the Italian example.
modality in affirmative subject complement clauses
177
These sentences are intended as expressions of attitude towards supposedly shared information. The speakers’ illocutionary intention is not to present the complement propositions as new and relevant, and as a signal of this they are framed as non-assertions. In many modern languages the choice of irrealis for subject complements is thus based on the by now familiar motives for non-assertion: lack of speaker commitment and redundancy of the information in the current discourse context. In these languages, one of the most common employments of sentences with subject complement is in expression of epistemic judgements with verbs such as ‘be (im-)possible, (im-)probable, (un-)feasible, (not) obvious, (un-)true, (in-)disputable that’ etc., or ‘seem’ as in many of the examples above. In Earlier Egyptian no comparable expressions appear to have existed.11 Instead, in this language finite clauses are found occupying the subject position in the following sentence types, which also include non-verbal constructions:
Adjectival sentences and after governing adjective verbs Sentences with passive main clause predicates Tripartite nominal sentences.
Subject clauses of predicative adjectives and adjective verbs are considerably more common than the other two types. With the second class above there is some question whether clausal complements of suffix conjugation verb forms passivised by means of the elements .t(i), .t(w) and .tw should be subsumed under this heading.12 The most recent analysis of Early Egyptian passives treats the element .t(i) as an Old Egyptian synthetic passive ending attached directly to the verb stem and followed by the grammatical subject.13 Yet, although opinions on this matter diverge, probably sometime in the late Old Kingdom this ending seems to have undergone a reanalysis whereby e.g. the earlier passive sDm.t(i) N “N was heard” was ‘re-bracketed’ 11
In contrast, Coptic is extremely rich in expressions of this sort; see Layton 2000, 391–96. 12 See chapter 2 n.15 above and cf. e.g. Gardiner’s apparent hesitation in GEG §§ 39, 410, end. 13 Reintges 1997, 145–49. For a bibliography of alternative views, see ibid 145. In Old Egyptian the ending is usually written .t(i); full writings are rare (see e.g. PT 204b/ N; PT 1041c/N; 1423a and EAG § 177).
178
chapter four
as [Dd][.t(i)] N, with the erstwhile ending .t(i) interpreted anew as an indefinite pronoun subject ‘one’ and N as an object of an active predication “one heard N” rather than as a subject of a passive.14 However, it is difficult to divide the evidence into ‘before’ and ‘after’ this ‘Aktivierungsprozeß’.15 There are indicators that the reanalysis and the consequent reassignment of semantic roles seems to have been completed probably by the First Intermediate Period or the early XI dynasty, and that this was reflected in the new orthography of .t(i) as .tw,16 but there are various morpho-syntactic problems still associated with this issue.17 Nevertheless, in the present work all complement 14 Loprieno 1984, 92–94; 1986a, 47; Malaise & Winand 1999 § 145; but see Reintges 1997, 145–49; Malaise & Winand, ibid. § 661b. 15 Westendorf 1953, 81. 16 Cf. Loprieno 1984, 93. The ‘activation’ of sDm.t(i) > sDm.tw and the role of .tw as a pronoun rather than inflectional ending of verbs is verified beyond doubt when the latter becomes attached to non-verbal particles and conjuncts such as iw, aHa.n, xr or nty (but see Malaise & Winand 1999 § 661b). Such usages are not attested for certain before the XIII dynasty (e.g. Louvre C11, 17 aHa.n.tw Haw im “then one rejoiced therein”), but already in the First Intermediate Period ‘autobiography’ of Ankhtify one finds an early attestation of .tw in i irtw xr.tw “‘O weakling’, so one would say”, (Moaalla Iα4) where it is attached to xr, a fossilised verb-like expression for parenthetic ‘X says/said’ lacking nearly all inflection. Slightly later, the construal SA.tw grt m rdit pA aqw “One should begin allocating the rations” occurs in Heqanakhte II, rto. 31–32, where the introduction of the infinitive rdit by the preposition m speaks in favour of an active interpretation of SA.tw. Notably, in all these instances the earlier .t(i)/.ti has given way to .tw This new writing would thus seem to have been motivated by a functional rather than a phonetic change i > w as has been postulated (EAG § 142). 17 The reanalysis and ‘pronominalisation’ of .tw apparently did not result in a complete breakaway from its earlier inflectional origin as a passive ending. It has frequently been noted that had personal pronoun patients of .tw-passivised forms come to be seen as grammatical objects, one might have expected the dependent, rather than suffix pronouns to have been employed (e.g. GEG § 39, Obs.; EAG § 463; Reintges 1997, 146; Malaise & Winand 1999 § 661b). Yet, the correct construction in such instances is sDm. tw=f, “he is heard”, not †sDm.tw sw. It would have to be assumed that the suffixes were retained regardless of their new grammatical role as objects. Further, the agent of (at least) .t(w)-passivised forms can be introduced as a postponed agentive phrase by in (or xr), in which case the pronominal interpretation of .tw seems excluded: BM 567, 8 Dd.t(w) n=f ii w(y) m Htp in wrw nw AbDw “May ‘welcome in peace’ be said to him by the great ones of Abydos” (not †may one say… by the great ones of Abydos (cf. also in Late Egyptian with .tw: KRI IV 19, 8 iw.tw Hr waf=f in nsw-bit N “he was captured by king N”, not †one captured him by the king). Further, the writing .t(w), which is widely employed even after the appearance of .tw, seems to be attached to verbs only. It could thus be argued that .t(w), which of course in no way differs from the most common writing of the earlier .t(i), still always represents the earlier passive ending rather than a pronoun and that the two are not the ‘full’ and ‘defective’ writings of one and the same element. Finally, it is clear that even the form .tw never acquired a status of a suffix pronoun: it cannot be used after prepositions or as a possessor in direct genitives (i.e. †n=tw
modality in affirmative subject complement clauses
179
clauses of .tw/t(w)/t(i) -passivised verbs are treated as objects regardless of their synchronic status. Even if in case of Old Egyptian data this may be not strictly correct, the difference is immaterial to the modal analysis of object and subject complements generally. In Earlier Egyptian the grammatical organisation of subject complement clauses involves the same forms and elements with the same functions as in object complementation and parallels exactly the assertion/non-assertion system of Romance languages illustrated above. Asserted clausal subjects are introduced by the realis-operators ntt/wnt whereas non-assertion is signalled by the bare sDm=f, which in case of mutable roots appears either with or without gemination and doubling.18 The irrealis category is again divided up in the expected manner. By and large, the geminating sDm=f and its functional counterparts appear in clauses containing information presupposed to be shared whereas non-geminating forms are reserved for hypothetical, potential and denied states of affairs, as well as for information of maximal lack of discourse relevance. Thus again the grammar reflects the functional divide of irrealis in Earlier Egyptian into distal and proximal types. The three sentence-patterns above represent all environments where subject complementation occurs in Earlier Egyptian; in particular, it should be noted most steadfastly that the negations n sp and nn, often analysed as somehow ‘predicative’ in character, have no place in this syntactic category.19 The same holds also for such thoroughly “to one”; †Hmt=tw “one’s wife”). In all, .tw would seem to represent a ‘semi-pronoun’ that occupies a position halfway between true pronouns and inflection. 18 Since ntt/wnt have already been shown to function as assertion-markers, the examples of irrealis below will be of mutable roots only. There are no examples of the specific distal irrealis moods with the endings -w/-y as subject complements, nor of the bare sDm.n=f (see 7.1 below). 19 The hypothesis of sp in n sp as a sDm=f “of a verb ‘to occur’ related to the noun sp ‘time’, ‘occurrence’” (GEG § 456) with the negated clause as its subject, is erroneous because there is no such verb in Egyptian. For the hypothesis of nn as an adjectival predicate, see Gunn 1924, chapter XVII; GEG §§ 188 Obs, 457; Satzinger 1968, §§ 50, 57; Gilula 1970, 211; Malaise & Winand 1999 § 611 and numerous others. As noted by Jansen-Winkeln, (1996a § 459) nn cannot be analysed as an adjective because it is not used as an attribute or a noun, (the latter feature shows that it is not a noun either) nor is it ever followed by the particle wy, which is common with predicative adjectives. nn-negated existential and adverbial sentences also often function as circumstantial clauses, which adjectival sentences do not. Further, as noted by Collier, (1991c, 17–19) nn-negated existential (a below) and adverbial construals (b) differ in negative scope: (a) Amenemhat I tells that his weakness before his assassins was due to circumstances:
180
chapter four
grammaticalised initial auxiliaries as aHa.n. Nevertheless, as in object complementation, there exists one environment where bare forms and patterns without a specific modal function on their own occur as clausal subjects, namely after the expression xpr.n. The pragmatic and grammatical status of this is, however, of a quite different order from other verbs with subject clauses. Yet, xpr.n seems to cast light on the diachronic history of auxiliaries such as aHa.n and suggests that their development into functional elements may have begun from construals involving subject complement clauses. Besides the modal-typological division into realis and distal and proximal irrealis, Earlier Egyptian subject complements also display the same formal division into marked realis and irrealis (ntt/wntintroduced- and geminating/-w/-y-ending sDm=f-clauses respectively) versus unmarked irrealis (the unmarked sDm=f forms) as object complementation. Finally, the opportunity for greater communicative effect provided by the internal division of irrealis is exploited in subject complementation with the same degree of sophistication and ingenuity as in object clauses. This imaginative use of modality appears characteristic to Earlier Egyptian and speaks eloquently against views nn swt qn grH There is none strong at night. (Amenemhat VIIe) (b) The peasant says to the high steward: iw awn ib=k nn n=k st You are greedy, but it is unbecoming of you. (Peas B1 323) In the first example, the negative nn has scope over the initial noun qn ‘one brave’ and leaves the adverbial grH unscathed. In the second example, however, nn has scope over the entire proposition. An existential translation “It does not exist for you” fails to capture the exact sense. It is clearly not the case that ‘it’, which refers to the high steward’s ‘greed’, does not ‘exist’, but that this is not fitting to the said person. The dative n=k is thus not an optional adjunct and there is a predicative relationship between it and st that the negation invests. Consequently, in negated adverbial sentences, a ‘predicative’ nn should take the entire adverbial clause as its subject, which at least in the ST is forbidden without a ‘nominaliser’ ntt/wnt. However, these never appear in the said constructions. Further, in (b) above the pronominal dative n=k cliticises to nn, which shows that the complex nn n=k st consists of a single clause (Collier, ibid 18–19). These features suggest that at least in case of adverbial negations, nn does not behave ‘predicatively’. Concerning existential negations such as nn qn and the verbal nn sDm=f, the ‘predicative nn’ -hypothesis seems better defensible. However, existential negations of the type nn + bare noun should probably be seen as ellipses for the widely attested pattern nn wn + noun, where a lexical form of the verb wnn ‘exist’ serves as the predicate. No corresponding pattern nn wn sDm=f seems to exist (see below). It thus appears that the existential nn + N and the verbal nn sDm=f are different patterns.
modality in affirmative subject complement clauses
181
of it as ‘Kindersprache’ or as somehow primitive, inflexible and lacking in expressive nuance. 4.2 Asserted versus Non-asserted Subject Complements If complements of verbs passivised by means of .t(i)/t(w)/.tw were to be analysed as syntactic subjects, all such examples quoted above—and particularly Old Egyptian instances such as the following—could have been cited here as illustrations of the use of realis in subject clauses: (186) Weni relates the background of a punitive campaign he orchestrated:
Dd.t(i) wnt btkw nxt m xAstyw pn It was told that there were strong troublemakers among these (Urk I 104, 12) hill-dwellers. But if these are viewed as objects instead, even for no other reason than for the sake of convenience, there remain but two (identical) instances of the sort in this class of complementation.20 Yet, as seen, non-assertion is also the norm in object complementation, instances of subject complementation as a whole are much less frequent than object clauses, and the example(s) mentioned, with the past passive sDm=f, still present a paradigm case of assertion:21 (187) The king overrules previous decrees: ir nf Ddw xr Hm(=i) wnt xtm wDw nw nzw r Sma r irt hA n kAt nt nzw m fAw Sdt kAt nb wDt irt m Sma pn... n rdi.n Hm(=i) iri rmT n nw nTrwy gbtw pr-min fAw Sdt hA nb n kAt nb irrt m Sma pw As for it being said to my majesty that royal decrees have been issued concerning Upper Egypt, namely about performing tasks of royal work, carrying and digging and whatever work ordered to be done in this Upper Egypt… my majesty (regardless) does not allow any people of the temple of Min, Koptos, the V Upper 20 A further possible example is the erased pPurches palimpsest 3–4, which Allen (2002, 73 & pl. 54) reads Dd n=i wnn=k Hr spr, but which may perhaps also read wnt=k Hr spr “I have been told that you are about to arrive”. 21 Similarly Urk I 286, 7. See also 5.3 below for further discussion of this example.
182
chapter four Egyptian nome, to do digging, carrying, or any task of royal work done in this Upper Egypt. (Urk I 282, 15–283, 3)
Here again, the principal information in the first sentence is the content of the reported saying, which is presented as the pivot of the communication and as something to which the speaker is committed. Consequently, the complement clause is introduced by the element wnt. But as seen, there are also clauses serving as subjects of passive sDm=f where the latter in particular is not the case; abundant examples of this sort with Dd as the governing predicate were quoted above in connection with the discussion of CT spells 38–40.22 However, as in other languages, in Earlier Egyptian subject complements non-assertion is most often not strictly a result of attitude and commitment but of the status of the information that the clauses convey as presupposed background. This is the case in all instances after adjectival predicates and adjective verbs; for example:23 (188) Ankhpepy-Meriptah tells of his success in royal service: wn aA Hzz w(i) Hm=f Hr hAbt w(i) Hm=f Hr=s His majesty’s praise of me was always great on account of what (Urk I 221, 4) his majesty used to send me for.24 (189) Rediu-Khnum says in reference to his performance at the service of the queen:
iw Ax wrt irr s Axt n ib=f n nbt=f xntt mnw=f It is very beneficial for a man to do what in his mind seems useful for his mistress, who advances his monuments (in turn). (CGC 20543, 18–19)
22
See 2.1.2 above. A further possible, but badly broken example after wr is Urk I 195, 1 (wr mrr [Hm=f]). The hopelessly obscure CT VI 194c has StA w(y) dgg=k where dgg=k may be a subject geminating sDm=f ‘that you look’ or a relative ‘whom you see’. For further comments on the examples (188)–(189), see n.64 below. 24 The initial wn is probably the same ‘perfective’ auxiliary as e.g. in the Old Egyptian Perfect wn(=f) sDm=f, used before a sDm=f of an adjective verb instead of iw; cf. EAG § 949, ββ; Doret 1986, 111–12; Allen 1986b, 18. 23
modality in affirmative subject complement clauses
183
(190) An appeal to the addressee: mAn=k Hmt=k iw mr rmm=s Tw You should see your wife. Her weeping for you is terrible. (Kemit pl. 9)
(191/4) Redjedet’s travails are described: wa m nn hrw xpr wn.in rd-Ddt Hr Snt=s qsn mss=s One of these days, Redjedet was suffering, for her labour was (pWestcar 9, 21–22) difficult. (192) A medical text describes the symptoms of a fractured temporal bone:
di=f snf m msdty=fy Sr hAA ir(y)25 He (the patient) has nosebleed, but its issuing is meagre. (pEdwin Smith 7, 24)
(193) From a surgical instruction on operating on a swelling: iri.xr=k n=s Dw m swt nt irt Dw-a ir wr dd=f snf SAw.xr=k sw Hr sDt Then you make a rapid incision into it with a knife. If its bleeding (pEbers 109, 14–15) is excessive, you burn it with fire. In all these examples the complement situations are presupposed as actual and not presented as conveying new information. For instance, the sentence in example (188) serves to comment on the character of ‘praising’ whose actuality is the point of departure for the entire utterance and not an issue requiring assertion. Similarly e.g. in example (190), the speaker expresses his attitude towards information assumed to be common ground with the addressee, and informing the latter of it is not the ‘point’ of the communication. In example (193) the governing verb and its complement are parts of a conditional protasis, but unlike in object complementation, this does not affect the status of the subject clause as information or, consequently, its modality. It is not ‘that it bleeds’ which is presented as potential—this remains as a matter of fact—but the excessiveness of this; i.e. the bleeding may be 25 In WGMT § 225, hAA is interpreted as having an omitted subject ‘etwas’, but it is also possible that the subject is ir(y) which has wholly coalesced with the initial ir of the following conditional clause ir di=k Dbaw=k “if you put your fingers…”
184
chapter four
greater or lesser, but its reality in the circumstances described is not in doubt. It is notable that the geminating sDm=f is regularly used in all these presupposed instances, and in view of its role as expressive of proximal irrealis background information elsewhere, this is not surprising. The same holds also for the ease in which the words ‘how’ and ‘the manner which’ are inserted in many of the translations above. Governing adjectives and adjectival verbs in particular describe the manner, quality and degree of their subject states of affairs rather than predicate their ‘reality’ or occurrence. Once again, these properties reflect the more general modal profile of the geminating sDm=f and its association with actual but ‘low relevance’ complement situations.26 Geminating subject complements expressive of presupposed information are also found in tripartite nominal predicate sentences. In the first of the following pair of examples the predicate is an interrogative pronoun, in the second a proper noun or an adjective used as such:27 (194) A sarcastic remark by a man to his adversary amidst a boat-fight: iSst pw hAA=k Hr AHt (Reden und Rufe 58) Why are you going down onto the field? (195) The sage laments that the divine does not accept offerings of men: dd=tn n=f ø Hr m n pH ø sw indw is pw dd=tn n=f28 Why do you offer to him when (things) do not reach him? Your (Admonitions 5, 9) offering to him is a miserable thing. Again, in example (194) the complement situation is quite obvious to both the speaker and, the speaker assumes, the hearer. This is particularly pronounced seeing that the utterance is a wh-question, in which the situation whose motive is questioned is presupposed as a rule.29 Another such question, this time in the guise of a second tense with a geminating sDm=f and an interrogative adverbial phrase, appears 26
See 2.2.2. In (194) the sense is: “I’ve just given you such a blow that you fly off the water altogether”. 28 The isft pw irr=sn in Admonitions 5, 4 is probably an error for irr(t)=sn (so also Faulkner in Simpson 1973, 216 and Parkinson 1997, 176 ). 29 See 10.3. 27
modality in affirmative subject complement clauses
185
the second of the examples above. In this instance the mutually shared status of the complement situation dd=tn in the sentence that follows is most apparent, seeing that the issue of ‘your offering’ was mentioned just before in the question and is most certainly established as shared. The communicative function of the sentences in examples (194) and (195) is hence again to question or express attitude towards non-asserted known ‘facts’, for which Earlier Egyptian uses the geminating sDm=f proximal irrealis. Given the apparent regularity of this, one may also quote some comparable examples of the anomalous verbs iwi and ii, which here perhaps more than elsewhere can be assumed to reveal their ‘geminating’ form—or better, the form which with these verbs serves as the specific proximal irrealis:30 (196) A piece of hate-mail (?) makes mockery of standard salutary clichés: bin wy iy=k aD.ti wDA.ti How very unfortunate that you should come safe and sound. (pUC 32204, 2)
(197) A caption in a fish-spearing scene: [nf]r w(y) iw sxt... Xr rmw How good that the lady of the marshland comes… carrying (Deir el-Gebrawi II, pl. 5, right) fish. (198) The sage says that in the current dire situation foreign goods are scarce; thus:
wr wy iw wHAtw Xr Hbyt=sn How important it is (now) when the oasis-dwellers come carrying (Admonitions 3, 9) their festive-offerings. In example (198) the situation described is not strictly ‘actual’ here and now; rather the attitude is expressed towards a generally occurring state of affairs which is still subjectively more ‘real’ than not. Earlier Egyptian uses the proximal irrealis herein; in Spanish there is a choice between indicative and subjunctive in certain clauses of comparable 30 One may compare here the spelling iy in example (196) and the same writing in pRamesseum I, Bii, 10 noted in chapter 2 n.88, with the form ii in example (97) where the semantic environment is suggestive of distal irrealis. It would again seem that these two modal functions indeed are formally distinct in case of the verb ii.
186
chapter four
sort.31 Also in Spanish, situations towards which emotional reactions are expressed, or which are subject to value judgements, are usually treated as background, and indicative use akin to (li) above is notably rare.32 In Egyptian no comparable asserted examples with ntt/wnt are found at all, and it would seem that in these instances the complement situation is always treated as presupposed (or is not asserted/assertable for some other reason—see below) by speakers.33 Accordingly, in examples (188)–(198) above, the complement clauses describe presupposed actual situations and consequently employ the geminating sDm=f and its functional counterparts. But in contrast, non-geminating sDm=f of ult. inf. roots and its functional counterparts are used instead if the situation described in the subordinate clause is for one reason or other less obviously ‘real’ or if its communicative ‘value’ is even lower than that of the complements in the examples above—i.e., in such instances these forms again function as a more distal irrealis. The first scenario is most prototypically at issue when the complement state of affairs is a mere prospect: (199) The courtiers express their opinion on the king’s grandiose buildingplan:
Twt wrt iri=k mnw=k It would be most fitting if you were to make your monuments. (Berlin leather roll 2, 4)
31
Butt & Benjamin 2000, 253. Cf. Butt & Benjamin 2000, 250–51. 33 There is one example where the context appears to suggest otherwise, but there is no ntt/wnt: (a) The goddesses come to Redjedet’s husband who tells them the reason of his despair: Hnwt=i mtn st pw ntt Hr mn=s qsn ms=s My ladies; look, there is a woman suffering because her labour is difficult. (pWestcar 10, 4) The situation ms=s ‘she is giving birth’ would seem to be presented to an unassuming group of addressees. However, rather than an exception to the rule or an error for mss=s, (cf. example (191/4) above) ms=s is probably not a complement clause at all. More likely, qsn is an adverb qualifying Hr mn=s and ms=s functions as an explanatory adjunct clause: “look, there is a woman suffering badly (qsn) because she is giving birth (ms=s)”. 32
modality in affirmative subject complement clauses
187
(200) The writer asks for help also from her dead addressee’s mother: mk grt in.t(w) tA mnTAt irt mwt=k wDa-mdw r=s nDm w(y) fA=k s(y) Now look, this vessel over which your mother should litigate is brought. It would be most agreeable if you supported her. (Haskell Museum 13945, 3–4)
In example (199) the subject complement situation is potential, just as in the examples (liii) above: the speakers comment on and express attitude towards a hypothetical state of affairs whose ultimate realisation depends on the addressee. Similarly in example (200), the speaker says that a certain response from the addressee’s part would be met with favour—should the latter choose to take the proposed course of action. In neither case can the speakers express commitment towards the complement propositions and the more distal irrealis non-geminating sDm=f appears instead of the geminating form. In the following instance the potential state of affairs appears as a subject of a passive participle mry that functions as an ‘adjectival’ predicate and is followed by a pronominal dative referring to the experiencer:34 (201) Nekhebu asks the functionaries of the necropolis: i Hmw-kA n imAxw in-iw mry n=Tn Hz Tn nsw wnn imAx=Tn xr nbw=Tn itw=Tn m Xrt-nTr O the kA-priests of the blessed dead. Is it desirable to you that the king favour you and that your reverence remain before your (Urk I 217, 15–17) lords and forefathers in the necropolis? In subject complementation the ‘relevance’ motivation for the use of non-geminating forms is most apparent in clauses after the passive verb rdi:
34 Similarly Urk I 218, 16. The wnn-clause appears to be a second subject of mry, referring to a currently actual presupposed state of affairs. Cf. 3.3 and example (180) above. See also GEG §§ 141, 374B and chapter 3 n.87. The subject in this construction is usually non-overt, but overt examples do occur as well; cf. e.g. Florence 2590 4–5 Ax ø n irr ø r irrw n=f ø with Vernus 1976, 140, example 5b Ax st n irr st r irrw n=f st, both meaning “It is more beneficial for him who does it than to the one to whom it is done”. In the latter example the overt subject is pronominal, in example (201), 16, clausal.
188
chapter four
(202) The deceased has assumed the role of the Dawn-god and the appropriate honours therewith:
dy mA N tn Smsw=s It has been granted that this N sees her retinue. (CT VI 348m) (203) The deceased possesses divine knowledge: ink r=i xt mnxt di(w) di.t(w) n=i Tswy pw n DHwty I am an enduring thing; it has been granted that I am given the two spells of Thoth. (CT VII 210f–g) (204) A note at the end of a royal decree:35 iw rdi iwt smr-wat(y) Hmi (zA) intf r=s The unique royal friend Hemi son of Intef has been sent (lit. “made to come”) concerning it. (Urk I 296, 16) The reasons for the use of distal irrealis here are the same as in comparable active environments. The realisation of the complement situations is implied by the nature and time-reference of the main verb and is hence a non-issue.36 Similarly, the only certain instance of a sDm=f from a weak root serving as a subject clause of some other notionally non-assertive verb suggests that the system of form selection after the latter are generally the same here as in object complementation:37 (205/133) The deceased says to the keepers of the mounds of the Field of Reeds:
swab iAt=tn wddt iri=tn pw in wsir n Dt “Purify your mounds!” That is what you have been ordered to do by Osiris forever. (Tb 149, 19–20) Quite interesting for the present argument is also the sole unambiguous example of complementation of a finite clause as the subject in the 35 Similarly Urk I 292, 12; Urk I 298, 7, 15; Urk I 299, 17 (broken); Urk I 301, 10 (broken); Urk I 303, 11 (broken); Urk I 306, 13; Pepi II Dakhla decree, 6. 36 See 3.2 above for discussion. 37 In Urk I 286, 13 one reads “various tasks… wDt irrt m Sma pw ordered to be done in this Upper Egypt”, where irrt is undoubtedly an error for irt (so correctly in Urk I 282, 17).
modality in affirmative subject complement clauses
189
pattern nfr pw X. In this construction, a third person singular masculine participle of the verb nfr is used as an inherently negative predicate in a tripartite nominal sentence expressing the meaning ‘X is one at end/non-existent’ = ‘there is no X’:38 (206) The writer reports a failure in deliveries of goods: swDA-ib pw n nb aws [r ntt Dd.n] bAk im n mr Sna wsr mk di=i n=k pA diw tA bnrt nt imnyt aha.n Dd.n=f nfr pw iri=i st [swDA=ib] pw Hr=s This is to inform the lord l.h.p. that yours truly had said (?) to the overseer of the storehouse User: “Look, I will give you the grain-rations and the confections for (lit. of) the daily offerings”. But then he said: “I will not handle them”. This is to inform of it.39 (pBerlin 10016, 4–6) The complement situation here is again a prospective action, but also one towards which the speaker expresses negative attitude by refusing to carry it out.40 It would seem that also in subject complementation, this latter sort of factors again affect the choice of the subordinate clause form. The evidence of this is here much sparser than in object clauses, but there are some suggestive examples apparently showing that also in subject complementation, speakers could manipulate the system of distal versus proximal irrealis for expression of subtle nuances 38 The predicative nexus of the overall sentence is, of course, affirmative. See Gardiner 1923; GEG §§ 351.2, 456.2; Loprieno 1995, 90; Allen 2000, 190. Satzinger (1968, § 108) analyses this construction as a bipartite nominal sentence with the following noun or verb in apposition (so also Gaskins 1978, 123). In addition to the examples above, pBerlin 10016, 3, pBerlin 10023B, 3 and the palimpsest of the latter have int (infinitive or in.t(w)?). The alleged inn.t(w) in pBerlin 10016, 4 is uncertain. 39 Wente (1990, 75) treats nfr pw iri=i as indirect speech “He said that I didn’t do this” (i.e. say something or give the foodstuffs to the overseer of the storehouse). 40 As the sentence is a direct quote, the relevant speaker is the subject of iri=i. A propos, although opinions concerning this issue vary, it is commonly assumed that negated propositions do not presuppose or imply their affirmative counterpart. However, the negation nfr pw appears to form an at least partial exception to this, seeing that it seems to be a favourite in contexts where the speaker believes, or assumes the addressee to believe, that something should be/have been the case, but, often astonishingly, is/was not. This background assumption seems to be present in all examples of this pattern and appears to be part of its actual pragmatic profile or discourse function.
190
chapter four
of meaning. For instance, in the following example after a passive verb, the choice of the geminating sDm=f for what is apparently a prospective complement situation over a non-geminating form seems to have a rather interesting motivation:41 (207) The deceased says to a divinity that he expects the latter to bring his peer to him; but:
ir in inn=k n=i sw irt Hr r=k mi If your bringing him to me is delayed, the eye of Horus is against (CT II 106b–c) you accordingly. Although the complement situation is unrealised, the speaker does not seem to view it as a mere prospect whose realisation depends on the addressee, but treats it as if its eventual realisation was presupposed or, rather more appropriately, prescribed. The conditional issue is the delay of ‘bringing’, but there is no suggestion that it would not occur; instead, this is clearly taken for granted by the speaker, and a geminating form appears instead. The same effect can be seen also in the following, sadly mutilated example, where the complement situation seems to be similarly a prospect, but one that is again viewed as prescribed and unquestioned by the speaker: (208) Thutmosis I purportedly asks Amun for Queen Hatshepsut to be appointed as his heir:
[sDm]=k n=i nis n sp tp sprwt=i Hr mryt=i [...] Hr HqA.n=f kmt Dsr[t sS]m.n=f idbwy m mAa-xrw ist wr wy irr=k xr Hmt=s Hear (?) me, the call of the first occasion and my petitions on behalf of my beloved one. [Give her (or the like) the throne] of Horus, after he has ruled the Black and the Red Land and led the Two Banks in vindication. How great it is that you should (Urk IV 273, 14–274, 3) act for her majesty! But the reverse seems to hold in the next example where a nongeminating sDm=f occurs in a description of a situation which is 41 PT 1223a has ir wdfi DAA=Tn mXnt n N pn “if your ferrying the barque to this N is delayed…” Tb 89, 7 reads ir wDf rdi=k mA=i bA=i “if you delay in allowing me to see my soul”, but in this instance the non-geminating form rdi is used.
modality in affirmative subject complement clauses
191
treated as actual in the context, but which is also associated with much difficulty and danger of failure; the form seems to be chosen to indicate that the action described occurs ‘only just’: (209) In a description of a certain kind of injury, it is noted of the patient:
iw qsn fA=f a=f Xr=s He has difficulties with lifting up his arm because of it. (pEdwin Smith 16, 19–20)
This differs fundamentally from example (191/4) above, which, although displaying the same governing verb qsn, carries no similar sense. In (191/4) a very ‘real’ situation (state) is difficult for the subject to bear, whereas in example (209) performing (and hence realisation) of an action is difficult and near-impossible for the same participant. The non-geminating sDm=f distal irrealis in the latter has an iconic and almost ‘figurative’ function: the use of modality mimics the physical obstacles and the very concrete ‘force dynamics barriers’ associated with realising the complement situation. The same nuance is discernible also in the next example—assuming that dgA is an idiosyncratic writing of the non-geminating sDm=f of the verb dgi, ‘glance’ as opposed to dgg42—where performing the action described is associated with much difficulty and discomfort: (210) A paraphrase for the expression “he finds no way to look at his breast” in a medical text:
n nDm.n n=f dgA=f n qAbt=f It is not pleasant for him to look at his breast. (pEdwin Smith 1, 26)
Comparable exploiting of the basic meaning of mood for ‘expressive’ purposes is not uncommon in modern languages either, and there is no reason to assume that the ancient Egyptians were any less capable of equally sophisticated language use. It again illustrates well how abstractions such as ‘distal’ find concrete manifestations in actual communication. 42 Cf. the writings dgy=i and dgAy=i in Urk IV 117, 6 and 148, 17; cf. chapter 2 n.177 above.
192
chapter four
The data discussed thus far represents the near-totality of Earlier Egyptian examples of asserted or morphologically revealing nonasserted subject complement clauses of governing lexical verbs. In addition, there is one instance of a bare passive sDm=f complement after the enigmatic verb gmi, itself also a passive sDm=f: (211) In a broken context: gmy sip nn n Smsw n HAty-a N It was found that these retainers had been ascribed to mayor N. (pBerlin 10026) But besides appearing as subjects of lexical verbs, in Earlier Egyptian clausal complements are also attested in the same position after verbs that show signs of beginning or advanced grammaticalisation. Yet, in these instances considerable care is at order. In general, the greater the degree of grammaticalisation of the verb, the less feasible is an analysis that treats its syntax on a par with fully lexical predicates such as those discussed above. A particular case of point here is the verb wnn, which often appears to be followed by complement clauses. Nevertheless, in most such instances this is demonstrably not the case, and wnn has a role as a purely functional tam-modifier instead.43 However, there are a few examples of the negation n-wnt sDm=f, where the sDm=f seems to function as a clausal subject of a “phrase expressing non-existence”:44
43 See Uljas 2004 for a detailed discussion of interrogatives of the type in wnn + S(entence). Cf. also the non-structuralist analysis of ir wnn + S conditionals in Collier 2005 as opposed to particularly those by Satzinger (1993b) and Kruchten (1994). There is one early example of wn that is difficult to analyse as auxiliary, namely Urk I 59, 16 wn irr(=i) mi st-ib Hm=f. Auxiliary wn before a geminating sDm=f would be most exceptional, but the interpretation of irr=i as a complement of lexical wn ‘exist’ would yield a strange meaning [[I acted] existed]. See Uljas 2004, 99 for arguments against Reintges’ (1997, 129) analysis of this sentence as a bi-clausal complex wn=øi irr(=i)i. Reading this example as wn(=i) iri(=i) r(=i) seems rather adventurous. 44 GEG § 188.2. Yet, the alleged example of the corresponding construal nn wn sDm=f (Sin B 197 nn wn mt=k Hr xAst) most probably involves not a sDm=f, but the noun ‘death’ (so Schenkel 1963, 130 contra Gunn 1924, 124 and GEG § 188.2).
modality in affirmative subject complement clauses
193
(212) The gods are demanded to allow the deceased to ascend to heaven; then:
ir nTr nb DA.t(y)[=fy sw] m N pn n-wnt t=f n-wnt t-HD=f n-wnt spd=f n-wnt bA=f n-wnt pr=f r Hwt-Hr irt pt As for any god who will place himself into opposition against this N; his bread shall not exist, his white bread shall not exist, his supplies shall not exist, his soul shall not exist, and it cannot be that he might go to Hathor, who is in heaven. (CT VI 403i–l)
(213) The deceased addresses guardians of the Underworld: i nw n nTrw... Hsqw tpw anxw n-wnt iri=Tn nw r=i O you gods… who decapitate the living; it cannot be that (CT VII 23o–24a) you should do this to me. (214) Queen Hatshepsut stresses her concern for the shrine of Amun: n-wnt qd=i Hr r-pr=f It never occurred that I should have been asleep concerning (Urk IV 363, 12) his temple. These rare instances do not seem to be ‘ordinary’ negations of the occurrences described. Rather, they appear to express forceful denials of the very possibility of the complement state of affairs ever occurring or having occurred by negating their existence, rather like the pattern nfr pw + S(entence). Due to this negative attitude mutable roots appear in the forms pr=f and iri=Tn. However, to analyse wnt as a ‘lexical’ wnn followed by a subject clause seems rather off the mark. Instead, seeing that n-wnt always appears as an inseparable fixed unit, a better approach might be to consider it a grammaticalised (non-isomorphic) complex that as a whole functions as a negative existential predication [not is].45 Thus, although n-wnt does indeed involve the verb wnn, the grammatical status of wnt and, consequently, syntactic relation
45
This analysis resembles that suggested by Antonio Loprieno (PC) for the development of the Coptic affirmative and negative possessive expressions ouNte/mNte. These can be seen to have developed from the earlier wn-m-di and nn-wn-m-di into actual verbs ‘have’ and ‘have not’ (witness the ‘verbal’ manner in which the possessed N is introduced by N after ouNte/mNte).
194
chapter four
to the preceding negation and the following complement clause are different than e.g. those of wn in the similarly existential negation nn wn, where the former is apparently an independent lexical verb and is followed only by noun (rarely pronoun) subjects.46 wnn is not the only verb in the Early Egyptian lexicon with which grammaticalisation needs to be taken into account in the study complementation. By far the most intricate example of this is the verb xpr, to be discussed next. However, a detailed look on this predicate opens unexpected vistas to issues far beyond the relatively restricted scope of complementation. It reveals synchronic and diachronic phenomena and developments that are of immediate relevance to the study of realis and irrealis modality in Earlier Egyptian grammar also in other environments, particularly in initial causes with introducing auxiliaries. 4.3 The Verb xpr47 As seen, Earlier Egyptian subject complement clauses display the same general split into non-asserted/irrealis and asserted/realis modality as object complements. Similarly, they show the division of non-assertions into distal and proximal irrealis according to the informational and ontological status of the situations described and the speaker attitude. However, the verb xpr forms a partial but important exception to this organisation. This verb is used as a lexical predicate ‘come into being/ develop’ with (pro-)nominal and sometimes also clausal subjects: (215) The man insists his soul must remain with him: nn xpr m-a=f rwi=f hr[w qsn] It must not succeed in fleeing on a difficult day.48 (Man and Ba 9–10)
46 There do not seem to be compelling reasons to analyse wn in nn wn as a participial subject ‘Seiend’ followed by an appositive noun/clause (Satzinger 1968 § 56) let alone a stative (Thacker 1949, 32). Rather, the arguments in favour of the time-honoured view of wn as a sDm=f of the lexical wn ‘exist’ (Gunn 1924, 122; GEG § 108) seem more persuasive: see Gilula 1970, 211. 47 The present chapter is a condensed version of Uljas 2006. 48 Literally “that it flees must not come to be with it”; the form is, as could be expected from the negative attitude expressed, non-geminating (cf. Malaise & Winand 1999 § 897).
modality in affirmative subject complement clauses
195
However, xpr appears also in the initial expression xpr.n used to introduce full sentences of various kinds. Middle Egyptian examples of this construal are seldom without ambiguities and archaising texts must often be consulted for less dubious instances:49 (216) The snake says that his avoidance of the fate of his siblings was due to fortune:
xpr.n rs nn wi Hna It happened that I was not with (them).
(Sh.S. 130)
(217) The background-narrative of the prophesy begins: xpr.n swt wn50 Hm n nsw-bit snfrw mAa-xrw m nsw mnx m tA pn r-Dr=f It happened that the majesty of the dual king Sneferu the justified (Neferty Ia–b) was the potent king in this entire land. (218) A new development in a mystery play: xpr.n inw abAwy It happened that two abA-sceptres were brought.51 (pRam Dram 83)
(219) As above: [xpr].n iri sxn-wAx mamaAwy It happened that the sxn-wAx-people made two frames for (pRam Dram 117) ladders. In Late Egyptian xpr is used as an aspectual auxiliary verb, and this use has attracted a fair degree of scholarly interest.52 The Earlier Egyptian 49 Sh.S. 166 has xpr.n rdit wi Hr Xt=i “It happened that I put myself on my belly”, but the analysis of rdit is uncertain (cf. Erman 1906, 22; Junge 1978a, 100; Collier 1987, 2 n.6, 3 n.10; Feder 2004, 74). In Old Egyptian bare xpr often occurs in the same role (e.g. Urk I 182, 11 and 14). Edel (EAG § 901, Anm.) seeks to separate this use from xpr.n. 50 pPetersburg 1116B writes wnn; all other variants have wn. 51 Similarly pRam Dram 64, 72, 76, 87, 89 (in), 91, 101, 104 and 107. pRam Dram 97 and 126 have xpr.n diw N. pRam Dram 15 and 51 have dw N of the verb d(i)/wd(i); 89 has also Dd.t(i). 52 See e.g. Erman 1933 §§ 569–70; Korostovtsev 1973 §§ 218–19; Satzinger 1976, 166–70, 236–39; Kruchten 1982, 30–32 and passim; Collier 1986; Winand 2006, 325– 29. Some commentators have analysed the earlier xpr.n also as an aspect-marker; e.g.
196
chapter four
expression xpr.n + S(entence) has been less extensively discussed, although it has not been neglected either. Erman interpreted xpr.n as an impersonal expression followed by an adjunct S.53 In this he was followed by Polotsky who analysed xpr.n as a second tense with a ø-subject followed by an adjunct S ‘predicate’, i.e. [xpr.n=ø] + [adv. S].54 Neither scholar viewed the (explicitly or implicitly) postulated ø as semantically linked to the following S but rather understood the complex as carrying the meaning “(it) happened while/after (etc.) he heard” and the zero subject referentially as either pleonastic or anaphoric. There are various problems with these analyses. Semantically, it is often difficult to assign an interpretation to the omitted subject of xpr.n and/ or to render S as an adjunct without destroying the sense of the overall situation-description.55 For instance, in example (217) the absence of any preceding discourse excludes an anaphoric interpretation. Also impersonal “it (i.e. something/everything) happened” seems decidedly impossible, as does the reading of S as an adjunct “when N was a king…” Similarly, what is communicated in (218) is clearly not “ø/something happened when/while/after (etc.) two abA-sceptres were brought”.56 Polotsky’s second tense analysis is also syntactically dubious; in the following examples, xpr.n is followed by the very ‘nonadverbial’ geminating sDm=f: (220) A new development in a mystery play: xpr.n hAA ms[w]-nsw [r] wiAwy It happened that the royal children descended into two barks. (pRam Dram 53)
Reintges (1997, 368–71) and Winand (ibid, 326) label it ‘inchoative’, although it clearly does not have this role in most of the examples above (for instance, in example (216) the issue is certainly not the inception of the speaker’s ‘not-being’ somewhere). 53 Erman 1928 § 309 Anm.; see also Junker 1940, 50–54, where the following S is, however, seen as wholly independent. In Erman 1906, 16 the author’s position is still somewhat agnostic; cf. also EAG § 901, Anm. 54 Polotsky 1976, 22–23; this analysis is originally due to Satzinger 1971, 66. Cf. also Junge 1978a, 43. 55 Cf. Collier 1987, 2–3. 56 Analysing S as an independent continuative/paratactic (cf. Malaise & Winand 1999 § 897) does not improve the sense either.
modality in affirmative subject complement clauses
197
(221) As above: xpr.n fAA sxn-wAx it Hr awy=sn It happened that sxn-wAx-people lifted the father onto their (pRam Dram 114) arms. Further, although characteristically found in absolute initial position, xpr.n can also occur after iw, whereas second tenses cannot: (222) Merer tells of his famine-relief activities: iw xpr.n di(=i) di.t(w) it-Smaw n nwt(=i) iw DA.n=i s(y) aSAw spw It happened that I used to secure that Upper Egyptian barley was given to my town. I crossed it numerous times.57 (Krakow MNK-XI-999, 12)
In contrast to these analyses, Sethe assigned S the syntactic function of subject of xpr.n58 and Gardiner similarly interpreted the former as subject ‘virtual noun clauses’.59 More recently, Collier has brought the earlier views together by analysing these constructions as containing a covert expletive as the subject of xpr.n, but one which is semantically coindexed with the following sentence, i.e. xpr.n=øi Si.60 In this hypothesis ø is readily interpretable due to its semantic identity with the following S. Also in the resulting ‘sentence-presentative’ translation of xpr.n S as “it happened that S”, S is standardly analysed as the ‘semantic’ or ‘deep’ subject of happen, i.e. [[S] happened] with the ‘dummy’ expletive it representing a mere ‘syntactic place holder’ for S.61 However, in English it functions as the grammatical subject of happen, and Collier’s analysis assumes the same structure for xpr.n + S. Yet, as seen, such ‘non-nominal’ sentence-patterns as e.g. the past passive sDm=f can function as syntactic complements without being ‘nominalised’, and 57 The reading di(=i) di.t(w) suggested by Mark Collier (PC). The traditional dd.t(w) (e.g. Doret 1986, 126) creates a sense where the speaker is absent from the process of ‘giving’, which fits poorly the general context. The present reading assumes di to function as a past habitual, (or iterative) ‘circumstantial’ sDm=f. 58 Sethe 1928, 90. Earlier (1907, 84), Sethe had expressed uncertainty about the subject- and adjunct analyses of the passage in example (216) above; cf. Edel 1959, 20. 59 GEG § 486 Obs. 1. 60 Collier 1987; 1991b, 38; cf. Collier 1986; Reintges 1997, 369–70. The subscript ‘i’ indicates referentially identical elements. 61 See Haegeman & Guéron 1999, 41–42 and passim.
198
chapter four
there is no need to postulate a zero subject for xpr on these grounds.62 The examples (220)–(221) above also show the ‘expected’ patterning of complement clauses with geminating hAA/fAA. Further, in Egyptian omitted pronouns are, as a rule, anaphoric in character,63 but here the assumed expletive would be referentially cataphoric. Finally, seeing that expletives provide an overt grammatical subject when required by (cross-linguistically variable) syntactic and semantic conditions, it would be highly surprising if an element with such a function were then consistently omitted after xpr.n. One would expect to find instances of an overt expletive after xpr.n if they existed, but there are no examples of e.g. †xpr.n=si Si.64 These notions speak strongly against the presence of an expletive subject of xpr.n in Earlier Egyptian. In Late Egyptian, impersonal co-indexed pronouns do occur with xpr in constructions semantically akin to xpr.n.65 Yet, this need not indicate that a phonetically zero element is gradually becoming ‘visible’. When occurring in its (almost) earlier guise, the ‘sentence-presentative’ xpr appears without expletives also in Late Egyptian.66 However, in Late Egyptian one finds uses such as the following:
62 The following remarks should not be understood as arguments against the phenomenon of pronoun omission in Earlier Egyptian in general. Referential and nonreferential pronouns can be omitted in Earlier Egyptian under agreement and relevance after verbs; see e.g. Urk IV 397, 2–3 xpr.in ø mi wddt r xt nb xpr.n ø n bAw Hmt=s “Then (everything) happened exactly as was ordered, and (it) happened through the manifest power of her majesty”, where first a referentially vague and then a situational pronoun is omitted after lexical xpr.in/xpr.n. 63 Matthias Müller, PC. This holds both for pronouns omitted under agreement and relevance. 64 Cf. Uljas 2004, 99. This argument can be extended to cover various other problem-cases. For example, and contrary to the prior argument of the present author (ibid, 97), the usual postulate of a covert ø after iw in examples such as (188)–(190) above is questionable. There are no overt examples of the type †iw=si mr [rmm=s]i.. Rather than statives, (as in the expletive analysis) (188)–(190) can be analysed simply as regular iw/ wn sDm=f of adjective verbs, (cf. Polotsky 1969, 470) which do not normally occur with anticipated subject pronouns after iw (i.e. as iw=f sDm=f—see 2.4 above). 65 Collier (1986, 16 and 20 n.8) cites pSalt 124 vso. 1, 5 (= KRI IV 413, 4–5) mtw=f xpr i.iri=f Sm Hr-DADA n nA inb “it happened that he went on the top of the walls” and LEM 73, 11–12 iw=f Hr xprw i.iri=f snhA n=i (r) awty m Hwt-nTr R “it happened that he registered me in this document to the temple of R.” As noted by Collier, the pronouns here are not referential; nor does xpr function as an aspectual auxiliary. 66 See e.g. LES 85, 4 xpr swt wn.in tA n kmt m iAdt “It happened that the land of Egypt was in a sad state”. Apart from the spelling xpr, the beginning of the Story of King Neferkara and Sasenet (oDeM 1214, 1/OIC 13539, 1) appears to have been identical to example (217) above.
modality in affirmative subject complement clauses
199
(223) A testimony of a woman accused of possessing stolen goods begins: xr xpr=i Hms.k(w) Hqr.tw Xr nA nhw(t) I just happened to be sitting hungry under the trees… (KRI VI 831, 11–12)
Here xpr is not used to introduce a situation but a participant into a situation.67 The apparent ‘raising’ of the controlling subject of Hms to the subject position of xpr shows that in Late Egyptian the latter is ‘free’ to host such elements. In addition, the paradigm of S after the ‘sentence-presentative’ xpr now consists of inherently non-subordinate ‘initial main clauses’, indicating that there is now clearly no mainsubordinate connection between it and S.68 But with the Earlier Egyptian xpr.n none of this seems to hold. Instead, seeing that its subject position is not available for other elements such as expletives, this is clearly occupied by S, which serves as a subject complement of xpr.n. In other words, the Sethe-Gardiner analysis of the syntax thereof appears to be correct. Yet if this analysis is adopted, the question arises how can S, which often is one of just those bare forms and constructions proposed to be modally unmarked, function as a complement of xpr.n. A further question is how can the latter appear in the same position with the geminating sDm=f. The key to this lies both in the diachronic history of xpr.n and its semantic-pragmatic relationship with its complement. It is notable that the instances with the geminating sDm=f derive from a source that linguistically represents the oldest stratum of Earlier Egyptian,69 whereas the ‘non-nominal’ construals are the only ones found in classical sources. In addition, example (222) above with iw appears to be the sole instance of its kind; otherwise xpr.n S occurs without any introducing element.70 These, and the other properties of xpr.n discussed seem to bear testimony to its gradual grammaticalisation 67
See Collier 1986, 19. Cf. Collier 1986, 17 and e.g. the wn.in-construction in n.66 above. However, there are some exceptions such as RAD 56, 5: xpr iw mn m nA Snwt “whenever it happens that there is nothing in the storehouses…”, with a formally ‘circumstantial’ iwclause. 69 Although the Ramesseum Dramatic Papyrus is of an early (or late) Middle Kingdom date, its language and orthography are highly archaic. 70 Helck 1975, no. 146, 5 reads iw xpr.n iwt Hm n nTr pn “It happened that the majesty of this god came”, but iwt may be an infinitive “There occurred the coming of the majesty of this god”. Cf. also BH I, pl. 8, A, 6 iw xpr.n rnpwt Hqrw “Years of hunger 68
200
chapter four
from an erstwhile ‘normal’ main verb followed by a subject complement into an initial auxiliary. The examples with the geminating sDm=f are likely to represent last vestiges of xpr.n as a lexical verb. In most of Earlier Egyptian this expression represents a quasi-auxiliary whose subject position is still occupied by S, but this no longer displays the ‘expected’ grammar of clausal complements. By Late Egyptian the subject position has become ‘vacated’ and capable of being taken over by e.g. controlling subjects of the following predication, and xpr is followed by unambiguous main clause constructions and functions as a ‘full’ auxiliary. This development has its roots in the meaning associated with the verb xpr plus a clausal complement, i.e. results, as often in grammaticalisation, from a local use of an expression in an environment with favourable semantic-pragmatic conditions. Informally speaking, the sole ‘meaning-function’ of xpr in xpr.n S is indeed to ‘present’ S, but like happen in the corresponding English “it happened that S” it has little independent content. On its own it does not refer to any separate instantiation. Thus, [[that I was not with them] happened] might suffice as a representation of the syntactic structure of example (216) above, but it does not describe the semantic message communicated very well. xpr.n does not introduce any tense- or aspect-modification to the semantic profile of its complement; instead, its contribution can be characterised as modal. It is not a far cry to equate the role of xpr.n “to mark the emergence of a new event the occurrence of something fresh and notable”71 with indication of assertive status. The sole function of xpr.n is to signal that its complement situation most certainly took place, that it carries a high degree of novelty and salience in the current frame of discourse, and that it thus has the illocutionary force and grammatical status of an assertion.72 In time, came about” and UC 14333, 8 iw xpr.n Hap(y) Sr rnpt 45 “A low Nile occurred in year 45”. 71 Capart et al 1936, 175; the Late Egyptian xpr is referred to, but this is also an apt description of the earlier xpr.n; cf. also Junker 1940, 50; Malaise & Winand 1999 § 897 among others. 72 Ransom (1986, 8) labels happen ‘implicative’, i.e. as signalling that the speaker most definitely indicates its complement to be the case. Cf. also “it seems that John left” and “it happened that John left”; both seem and happen add modal qualification to the complement situation but have little semantic function beyond this. Both describe speaker judgement; the first that the complement is considered likely, the second that it most certainly occurred.
modality in affirmative subject complement clauses
201
xpr.n with a clausal subject was re-classified as a ‘true’ auxiliary and its syntactic relationship with its complement S reorganised. Like all chains of grammaticalisation, this development did not take place overnight and in discrete steps with clear cut-off points. Hence, for a while the paradigm of S is a mix of typical and less typical complementpatterns and the drift of S away from the syntactic position of subject is observable only at a relatively late stage. But from early on, xpr. n is not followed by the bare geminating sDm=f nor ntt/wnt: both are incompatible with its functional profile. xpr.n already does the work of ntt/wnt (and has no other role) and the geminating sDm=f is modally irrealis and cannot function as an assertion. Also the nongeminating -w/-y-less sDm=f does not have an irrealis function after xpr.n, as can be seen from examples (219) and (221). The forms with the said endings, of course, do not occur at all. In a sense then, in most of historical Earlier Egyptian xpr.n is a modal marker akin to ntt/wnt: it assigns the following unmarked construal a modal profile and allows it to function as a complement.73 This development undergone by xpr.n might also cast light on the grammar of auxiliation and the pragmatic role of auxiliaries in Earlier Egyptian generally. Regarding the common aHa.n, ‘then’, Gardiner analysed constructions with this element as originally paratactic structures in which “the subject of aHa was the same as that of the following verb, the form aHa.n sDm.n=f… thus meaning ‘he rose up and heard.’”74 Satzinger and Polotsky saw aHa.n, like xpr.n, as having developed out of an impersonal second tense sDm.n=f of aHa, with the following ‘adverbial’ S.75 But alternatively, and by analogy with xpr.n which it greatly resembles formally, aHa.n + S may have originally been just an ordinary case of main verb and a subject complement, with the meaning “(the situation) S arose” (i.e. took place, happened) but later S lost its complement status in lieu with the grammaticalisation of aHa.n as an initial auxiliary. The ability of aHa.n to host anticipatory or controlling subjects of the following predication would thus be a secondary development arising from its new role as a functional element with no superordinate relation to S and an ‘empty’ subject position. 73 74 75
See 2.4 above. GEG § 476; see already Erman 1889b, 36–38. Satzinger 1971, 66; Polotsky 1976, 33; see also Doret 1986, 125.
202
chapter four
The same analysis is also applicable to the much less ubiquitous auxiliaries ii.n and pr.n, both of which clearly mean “(the situation) S came (about)”.76 Something akin to these last two elements might be recognised even in iw. In the light of the considerations above, Gardiner’s suggestion that this auxiliary might have originated in a “specialised use” of the verb iw ‘come’” appears quite cogent,77 particularly as in the PT there are occasional uses and writings of iw seemingly recalling such an etymology.78 The assumed ‘original’ meanings ‘S came about’ and ‘S arose’ of the auxiliary construals above all entail the same semantic component ‘S happened’ which, of course, is the sense of xpr.n.79 Now, since the function of xpr.n is to signal assertion, it may well be that this is shared by other initial auxiliaries as well. As is the case with ntt/wnt, it could be argued that xpr.n is (largely) incompatible with auxiliaries due to their common function of assertion-marking; of course, ntt/wnt-clauses, which are certainly assertion-marked, do not allow auxiliaries either. Many Earlier Egyptian auxiliaries may thus represent former kinetic verbs with subject complement clauses, which gradually acquired a pragmatic role of highlighting the (subjective) ‘actuality’ and discourse-pragmatic relevance of the latter—i.e. its modal realis- and assertion-marking. Again, verifying this beyond doubt presupposes a detailed study of auxiliaries from a modal perspective, but the behaviour of xpr.n is suggestive of this semantic-pragmatically most complex of Earlier Egyptian grammatical categories.
76 Erman (1889b, 36) and GEG § 488 assume a similar origin to these construals as to aHa.n. 77 GEG § 461; cf. Reintges 1997, 76. 78 Cf. particularly PT 377a/W, P; PT 2075a and PT 1480b. However, iw- or aHa.nshows no syntactic trace of such origins in historical Egyptian; cf. Gardiner’s (GEG § 461 Obs.1) denial of the possibility of viewing S after iw as a ‘virtual noun clause’. 79 However, this cannot be extended to all auxiliaries in Earlier Egyptian; certainly such elements as wnn or e.g. sDr.n do not succumb easily to the framework proposed herein. The latter might be a possible candidate for the sort of serial construction (*“I spent the night, I heard”) suggested by Gardiner for aHa.n; this path of auxiliation does indeed exist (Heine 1993, 37–39).
negative complement clauses
203
CHAPTER FIVE
MODALITY IN NEGATIVE COMPLEMENT CLAUSES AFTER GOVERNING PREDICATES1 5.1. Introduction Among the operations carried out on transitive predicates that may determine or affect the modal profile of their complement clauses, negation is one of the most important. As seen, in all languages— including Earlier Egyptian—negation of the matrix verb as well as their inherent negative-like properties such as ‘negative implicativity’ may render the complement non-assertable to the speaker or variously delimit its assertability.2 However, unlike in most other languages, in Earlier Egyptian modality and negation interact extensively also in instances where the governing predicate remains affirmative, but where the complement is negated. The following types of negated complement clauses are attested in Earlier Egyptian:3 iwt- and ntt n-clauses tm-clauses nfr-n-clauses As noted, the employment and mutual distribution of these negative complement clause types has been largely interpreted as being based
1 The present chapter discusses negated complements of governing verbs only. For negated preposition complements, see 6.2 and 6.3 below. 2 See 3.2 above for negative implicative properties of verbs; cf. also example (206) above with the construction nfr pw + subject complement clause. 3 Edel (EAG §§ 1107, 1109) quotes two PT passages where the negative verb imi allegedly functions as an object complement. In the first of these (PT 577b-d) imi=sn is more probably a final clause, seeing that two further clauses separate it from the main verb and the latter is the causative verb rdi after which complements are not negated (see 3.2 above). Also in the second instance, PT 16d sA (i)m=k sfxxw im=f, (i)m=k may not be a complement of sA, as the variant Nt. 67 has the bare imperative m sfxxw instead. No trace of this proposed use is found in later material.
204
chapter five
on syntactic substitution.4 The element iwt, which as a negative word for ‘that not’ is semantically a negative synthetic counterpart of ntt, has been seen to be similar to the latter also syntactically. For Gilula, iwt represented a “nominalization (the subordinated, substantivized form) of the negative word n and possibly nn”5 that subordinates “independent negative sentences”.6 For Satzinger, iwt allowed sentences that could not be ‘nominalised’ by other means to function as noun clauses.7 At the end of the Old Kingdom iwt was superseded by the analytic pattern ntt + n-negation.8 Collier analyses the latter type of construals within the framework of his general system between syntactically ‘specialised’ and ‘non-specialised’ patterns.9 Like all ntt-constructions, ntt + n consists of the ‘connector’ ntt plus a ‘non-specialised’ form—i.e. one that cannot alone function as a complement. By contrast, the negations tm and nfr-n have for long been interpreted as actual ‘nominal’, or ‘specialised’ negations that function as the negative equivalents of the affirmative ‘nominal’ forms, or patterns ‘specialised for nominal use’.10 The use of bare ‘non-nominal/-specialised’ forms and constructions in complementation without ‘nominalisers’ discussed earlier and of tm and nfr-n in fundamentally ‘non-nominal’ environments such as final ‘so that’-clauses are not the only problems marring these hypotheses.11
4 A notable exception here is Loprieno 1991b; see further below. Edel (EAG § 1027) analyses the variation of nfr-n/tm and iwt in Old Egyptian to be based on the futurity versus non-futurity of the complement clause, but elsewhere (ibid. § 1074) it is noted that the former are used in ‘dependent’ clauses. 5 Gilula 1970, 213. 6 Gilula 1971, 17. 7 Satzinger 1968 § 101. 8 EAG §§ 1043, 1054; Satzinger 1968 § 101; Gilula 1970, 213; 1971, 17. For an early example of ntt n after the preposition Dr ‘since’, see EAG § 1080. Gilula (1971, 17) also postulates the construction ntt + the negation nn, but of this there are no actual examples. However, one CT instance (example (226) below) may represent a functional equivalent of this. 9 Collier 1999, 57, particularly n.43. 10 See Satzinger 1968 § 66; Gilula 1971, 17; Vernus 1990, 119; Collier 1999, 57. 11 See 0.1.1 above. A further ‘non-nominal’ use of tm will be discussed in 5.3 below. Callender’s (1975, 107) formulation of tm as generally “used to negate all forms of the verb transformed into other parts of speech” does avoid the difficulty of the use of tm in final clauses, but still cannot explain its use to negate e.g. kA/xr/ix-constructions. Cf. also Doret 1986, 44, who refers to tm and nfr-n as the negations of “the dependent, i.e. subjunctive or nominal, forms of the verb”.
negative complement clauses
205
Even in complementation, the distribution of the said negations and negative elements is once again not explicable on purely syntactic grounds. iwt, like ntt, never ‘nominalises’ such truly ‘non-nominal’ expressions as adverbial phrases or -clauses. Both iwt- and ntt nclauses are also attested only after ‘certain verbs’ (and prepositions),12 which shows that their use has nothing to do with the syntax of the following construal and everything to do with the meaning of the governing entity and the wider semantic-pragmatic co(n)text. Further, once again there does not seem to be any exclusively syntactic reason why e.g. a construal such as ntt n sDm.n=f “that he does not hear” (cf., for instance, example (229/73) below) with the ‘unsuitable’ n sDm.n=f could not have been replaced by tm=f sDm or nfr-n sDm=f allegedly specialised for the purpose.13 Refraining from doing so must indicate that there is some deeper difference in meaning between the two types of clauses. Similarly, if the use of iwt in the construal iwt sp sDm=f (cf. e.g. (224/13)–(225) below) was due to a desire to retain the ‘unsuitable’ negation n sp sDm=f in the complement, there must still have been some semantic-pragmatic motive for this. Nevertheless, with negations the preoccupation with syntax and particularly the early identification of tm as the ‘nominal’ negation par excellence have served as a particularly impenetrable barrier against making full use of the (admittedly infrequent) observations on the meaning of Earlier Egyptian negative complement clauses. For example, Satzinger speaks of tm as the negation of all ‘non-indicative’ or ‘subjunctive’ forms of suffix-conjugation and of iwt as an element introducing ‘indicative’ sentences.14 However, in complementation he treats ‘subjunctive’ and ‘indicative’ as syntactic terms equivalent to ‘nominal’ and ‘adverbial’ respectively.15 Yet, in principle Satzinger’s characterisation in fact captures much of the essence of negative complement clauses in Earlier Egyptian. The modal division of 12 Gilula 1970, 213; Frandsen 1975, 70; cf. Doret 1986, 34 n.263. Vernus (1990, 119) uses the same expression to characterise the use of tm in complementation. See 6.2 and 6.3 below for ntt/wnt-introduced preposition complement clauses. 13 For the formation of tm- and nfr-n-negated clauses, see 5.3 below. 14 Satzinger 1968, §§ 62, 65–66; 45 n.149; 100. 15 Satzinger 1968, 42 n.138 “die subjunktivischen Formen… fungieren als Nomina, wie die infiniten Formen”; on iwt, (and ntt/wnt) ibid. 45 n.149 “Bei diesen wird die Unterordnung durch die Einleitungspartikel vollzogen, was darauf folgt, sind indikativische Zustandkonstruktionen” (emphasis by SU).
206
chapter five
complements into realis and irrealis found in affirmative sentences governs the grammar of these subordinate clauses also in negative environments. Whenever the complement clause constitutes a speaker assertion, it is introduced by the element iwt, or, alternatively, by ntt followed by a clause negated by n. If the complement is not asserted, iwt/ntt n are not used. Instead, and quite unlike in most languages where subjunctive/irrealis form(s) are simply negated, Earlier Egyptian replaces the affirmative sDm=f forms with the negations tm and nfr-n. These negations do not differentiate between distal and proximal irrealis functions. Instead, the said nuances pertain only to the affirmative and are abstracted away in the negations following the cross-linguistic trend of concentrating the highest degree of grammaticalisation of fine-grained tam-oppositions to ‘prototypical’ or ‘cognitively simpler’ active affirmative expressions, which is a ubiquitous feature of Earlier Egyptian grammar.16 This organisation holds for both object- and subject complements of verbs, although, as in the affirmative, there is an overwhelming quantitative bias in favour of clauses functioning as syntactic objects. The parameters for the choice of negative construal show once more that the grammatical organisation of complementation in this language is based on the speaker’s subjective vantage point towards the proposition communicated and the information it conveys. 5.2 Assertion: Complement Clauses with iwt and ntt n Like their affirmative counterparts, asserted negative complement clauses after verbs are in Earlier Egyptian introduced by a specific element which functions as a marker of modal realis. The original morpheme for this purpose is iwt, a particle of obscure etymological origin and a somewhat brief lifespan in historical Egyptian.17 As a complementiser, iwt has the same syntactic role as ntt/wnt, but as an assertion marker it is also their exact semantic-pragmatic equivalent. The characteristic features of and conditions for assertion are, unsurprisingly, the same in negative as in affirmative complements; for example:18 16
See Collier 1994, 76; Uljas 2000, 132 and n.189 of chapter 2 above. For views on the etymology of iwt, see GEG § 202; EAG §§ 345, 1054; Satzinger 1968 § 102; Gilula 1970, 213; 1971, 17. 18 All the examples presented below are object complements. There are no attestations of asserted negative subject complements. 17
negative complement clauses
207
(224/13) The king notes to his envoy concerning a dwarf that the latter is bringing:
Dd.n=k xr Hm(=i) iwt zp in.t(i) mit(y)=f in ky nb You have said to my majesty that never has the like of him (Urk I 129, 2–3) been brought by anyone else. (225) In a dispute over property-rights, one of the participants is reported as denying the authenticity of a document purportedly drawn up by his father:
Dd.n TAw pn iwt zp iri sw it=f m bw nb This Tjaw said that his father had never done anything of the (pBerlin 9010, 3–4) sort. The first of these examples resembles closely the affirmative (18) above, and both are instances of reported indirect speech after a verb of locution. In example (225) one observes again the altering of the pronoun =f in it=f from the presumed ‘original’ it(=i) to correspond to the perspective of the real speaker. The subordinate reported speech is attributed to the subject of the main clauses and the main clause utterances “X said Y” to the real speaker. But again the grammatical marking of the complement does not derive from its relation to the main verb subject but from its content being accepted and communicated as optimally relevant by the real speaker who expresses commitment towards its veracity. As before, this becomes apparent when one examines the information-value of the main clause and the effects of variations in its illocutionary force on the complement. Particularly in example (224), the main communicative function of the utterance in the current discourse is hardly to inform the addressee of his own words, but to express the speaker’s own confidence in them, in this case undoubtedly in view of holding the original speaker accountable of his words. Similarly, in example (225), recalling what the litigant said is certainly the pivot of the sentence rather than that he said something. Thus the parts carrying the highest degree of speaker commitment and informational relevance in the sentences as a whole are the complements, which constitute assertions by the real speaker and, because of this, are introduced by iwt.19 19 Yet, there is one problem involved here with example (225). If the complement is to be seen as an assertion of also the real speaker, this is then tantamount to the latter
208
chapter five
The same can be seen also in the following examples, both of which have the flavour of speakers being anxious to ‘prove’ the point they are making by turning them into complements of perception and cognition verbs:20 (226) In the debate between father and son, the son says he is destined to survive as an heir:
iw grt sDm.n=i mdw i.mr=f m-Xnw iw n anxw m Hr-ib DbAt nt wabw iwt mt=i n=sn n mt sin Moreover, I have heard the word of i.mr=f on the island of the living and amidst the robe-room of the pure-ones, that I did/will not die a sudden death for them. (CT I 170g–i) (227) The deceased says to Anubis that his death is only apparent: iw=k rx.t(i) iwt wnt=i m-m iAtw You know that I am not among the mound-dwellers. (CT II 125f)
In the first of these examples the real speaker reports his own perception, in (227) the addressee is the subject to whom the knowledge is assigned. In the latter it is again rather counterintuitive to speak of the subject ‘asserting’ anything, seeing that the reference is to a mental process, not a speech act. Yet, the complement is clearly presented as something to which the real speaker is strongly committed and which he announces as a relevant and reliable piece of information, i.e. asserts it. There is also at least one negated example where assertion by the main clause subject is excluded because of the unrealised status of the situation, which the governing predicate describes:21
accepting the claim of the litigant concerning his father’s actions or non-actions. Whether this can be the case in a legal texts such as the one from which the example derives is debatable. Admittedly in this instance the assertion would seem to be made primarily by the original speaker. 20 In example (226) there is some question concerning the intended time-reference; both past and future, which in the CT is typically n sDm=f, are possible. 21 Cf. the closely similar affirmative examples (26)–(27) in 2.1.1 above. This passage is followed by the rather obscure iwtt gmt=k N pn (318l).
negative complement clauses
209
(228) The deceased says to a messenger of an unfriendly deity: Dd m n hAb Tw iwt gm=k N pn Pray tell the one who sent you that you did not find this (CT VI 318j–k) N (= me). Here the addressee is again told to say something that he has not yet said and thus cannot have asserted. The assertion is due to the real speaker, but it is quite interesting to note that the complement situation here is not actually ‘true’ in any real sense—the addressee of course did find the speaker. However, it is what the speaker wishes to present as true, and this alone suffices to determine its modal status and grammatical form as an asserted realis. Finally, the modal equivalence of ntt/wnt and iwt as markers of assertion is most obviously shown by the diachronic disappearance of the latter and its subsequent replacement by ntt + n.22 Although this latter construal is very common after prepositions, only the following instance, already quoted as example (73) above and repeated here for convenience, is attested after a governing verb: (229/73) An utterance of a workman to his fellow whilst manufacturing a vessel:
in iw=k Hr mAA ntt n aHa.n pA mnw Do you see that the cup cannot stand up?
(Meir I, pl. 5)
Regarding the set of forms and patterns introduced by iwt, the evidence is rather sparse. Examples (224) and (225) stand for the hypothetical analytic *ntt + n sp sDm=f. The same holds also for the somewhat uncertain instance of an iwt-clause after the verb ib ‘think’ quoted as example (123) above. Examples (226) and (228) correspond to *ntt n sDm=f, which in case of (228) may have a future reference and the example may thus correspond semantically to the unattested *ntt nn sDm=f of the classical language. Example (227) appears to be an instance of a negated existential n wnt, although the suffix subject is most exceptional. By analogy with the affirmatives and the n sDm.
22 There do not seem to be any examples of †wnt n; however, given that wnt was supplanted by ntt at the same time as iwt disappeared from regular use, (see 2.1.1 above) this is hardly surprising.
210
chapter five
n=f after ntt in example (229/73), it seems that the said negations also belong to the modally neutral/unmarked grammatical group. Consequently, there is no need to assume that e.g. n sDm=f or even n sp sDm=f cannot alone be used as complements for syntactic reasons. In negations as elsewhere, the requirement that a complementiser introduce the ‘non-nominal’ forms and constructions has to do with the principles of marking assertions in Earlier Egyptian and the lack of assigned modal profile of the patterns in question. The use of iwt parallels that of affirmative or n-negated complements with ntt/wnt, and all these elements have the same function of marking the clause they introduce as a modally realis assertion. Conversely, also the reasons for their absence turn out to be of familiar sort and reflect a grammatical strategy of expressing non-assertion with negations specialised for this purpose. 5.3 Non-assertion: tm and nfr-n In the surviving textual corpus of Earlier Egyptian there are no examples of negated object- or subject clauses of notionally assertive verbs where the real speaker would be incapable of asserting the complement due to ignorance of its veracity status. In the sole example of a negation after an interrogative main clause, (229/73) above, the scope of the interrogative does not extend over the complement, and the speaker is in no way unaware of the situation it describes. However, non-assertion in negated complement clauses of verbs is not uncommon but actually rather more frequent than assertion, exactly as in the affirmative. Instead of iwt, in such instances Earlier Egyptian employs tm and nfr-n, which due to their status as specific irrealis negations are never introduced by ntt. The etymology and grammatical construal of both tm and nfr-n has been much discussed but remains rather unclear. The use of tm as a negative function word seems to result from grammaticalisation via metaphoric extension of sense of the lexical verb tm meaning ‘finish’ or the like.23 tm is followed by the so-called negatival complement
23 See GEG § 342, end; WGMT § 360 among others. Lexical tm is polysemic and translates variously, and in a rather contradictory manner, as ‘complete’ or ‘fail’. It may be that both these meanings are actually extensions of the basic sense ‘finish’, which can refer both to completion/success and failure/premature cessation.
negative complement clauses
211
(later infinitive) as the immediately following prosodic unit and the carrier of lexical content in the complex, but the syntactic relationship between the two is uncertain.24 The etymology and build-up of the negation nfr-n are similarly obscure, although the most commonly accepted view is that originally it consisted of a predicative adjective ‘be at end’ followed by dative and then the negated clause.25 When finite, the latter takes the form of an affirmative sDm=f for which weak roots always show gemination in preposition complements but both gemination and no gemination after verbs. Given the relatively small number of mutable examples in both environments it is impossible to say whether this split is merely an accident of preservation.26 Although all the uses of nfr-n are perhaps not present in the surviving corpus, the functional overlap of this negation and tm appears to have been complete.27 Perhaps it is because of this that nfr-n was gradually replaced by tm in all but consciously archaising texts and in certain set-phrases of elevated style.28 Yet, whatever the origins, internal composition and diachronic fortunes of these construals, both tm and nfr-n are employed in 24 For a discussion on the syntactic relation between negative verbs and the negatival complement/infinitive, as well as for bibliographical notes, see Schenkel 2000a, 3– 7. 25 See Gardiner 1923, 81; GEG § 351.1; EAG § 1130; Satzinger 1968, 67 n.246. Ogdon’s (1996) suggestion of nfr-n as functionally equivalent to nfr iw (sic) is clearly unfounded. Gaskins (1978, 121) and Reintges (1997, 349) interpret n as a Perfect marker, i.e. understand the complex as a sDm.n=f. However, nfr-n is clearly a fixed (grammaticalised) negation and its internal components do not need to retain their etymological function. The common writing of the element n as ‘negative arms’ and the version nfr A found in the Heqanakhte papers (T. James 1962, 104–05; Allen 2002, 100) shows that nfr-n quickly became subject to phonological reduction and was seen as a nonisomorphic functional unit. This being so, the relationship between n and the following sDm=f is hardly any longer that between e.g. a governing dative and a subordinate clause (Mark Collier, PC). This poses further difficulties to assessing the motives for the geminating/non-geminating variation of the sDm=f following nfr-n between verb- and preposition-complements (see below). 26 See example (230) below for the non-geminating pr N after the verb wD and (241) for a geminating Tzz=sn after mAA (but see n.46). These are the sole instances of weak verbs in complements after verbs. For examples after prepositions, see 6.2. 27 Besides complement clauses, tm and nfr-n both negate the infinitive as well as conditional and final clauses. Urk I 85, 5–6 sk nfr-n irt (or ir.t(i)) mitt n wr xrpw Hmwt nb Hr hAww [nzw nb] might be a concessive ‘setting’ second tense “Now, although nothing of the kind had been done to any great overseer of craftsmen under any king…” Edel’s reluctance to accept the capability of nfr-n to negate participles seems strange in view of the clear examples he cites (EAG § 1137). 28 Doret 1986, 44.
212
chapter five
complementation for modally irrealis clauses. Their principal domain of use in this function is in object clauses of verbs that in the affirmative are never followed by ntt/wnt or, in other languages, by indicative/ realis forms. This is, of course, not a coincidence; the predicates in question are notionally non-assertive and cannot be followed by assertions.29 The most common combination is wD + nfr-n, although all the examples derive from Old Egyptian and often involve weak final radical verbs with the ending -t with which it is impossible to decide whether a .t(i)-passive or infinitive is in question.30 Unambiguous finite instances are e.g.: (230) A royal decree prohibits the seizure of resources from funerary foundations by other branches of the administration:
iw wD.n Hm(=i) nfr-n pr nHs-Htpw nb r irt awA m nwty (i)ptn My majesty has ordered that no ‘pacified Nubian’ may go to make requisitions in these two pyramid-towns. (Urk I 212, 7–8)
(231) A further order: iw wD.n Hm(=i) nfr-n aq rmT nb ntw xr nHs-Htpw (i)pn r wab... m Hwt nt nwty (i)ptn My majesty has ordered that no people associated with these ‘pacified Nubians’ may enter priestly service… in the temple(s) of these two pyramid-towns. (Urk I 212, 9–11) Also tm is found after the same verb: (232) The deceased says that he will not consume faeces: iw wD.n gb it wsir tm=i wnm Hs tm=i swr wsSt Geb the father of Osiris has ordered that I do not need to eat excrement and drink urine.31 (CT III 171j–l)
29 30 31
See chapter 1. Cf. EAG § 1131. The sense here is, in addition, clearly that of negated deontic necessity.
negative complement clauses
213
(233) It is said of the workers who constructed the tomb for Debehni: wD Hm=f tm=sn iT r wnwt nb HAw irt kAt im=f r Htp=f His majesty ordered that they are not to be taken to any labour-service in excess of carrying out the work therein to his satisfaction.32 (Hassan, Giza IV, fig. 118, 5) Of the other non-assertive verbs, only the following example of tm after sAw ‘take care that not’ may be quoted: (234) The deceased is warned against bad preparation on his journey: sAw wdn rk n rdi.tw xsf Hr=k sAw tm=k pr Be careful, be alert, and no-one will obstruct you. But take care that you do not fail to go forth.33 (CT I 71b–d) Both tm and nfr-n are generalised irrealis negations: the fine-grained distal-proximal distinction expressed in the affirmative by variation of the geminating and non-geminating sDm=f are suppressed in the negative, and tm in particular certainly does not “assume the morphological features” of any of the so-called ‘nominal’ sDm=f forms.34 The non-indicative character of these negations in complementation after verbs seems to have passed unnoticed, but this is presumably not only due to the focus on syntax, but also because in many complements the irrealis nature of tm and nfr-n is not necessarily so immediately apparent. For instance, deducing the reasons for the use of nfr-n in the following example again requires meticulous attention to the co(n)text:35
32 The passive sense here has attracted some scholarly interest (see e.g. Doret 1986, 46). It seems that the voice of the negatival complement had not been fixed as active by the time of the V dynasty rather than that here tm itself is passive (cf. EAG § 743). 33 Not “Beware that you do not go forth” (Faulkner 1973–78, vol. 1, 14), which would undoubtedly be expressed by *sAw pr=k. The present sentence is a double negation and an exhortation against failure: “take care that you do not not go forth”. Similarly Merikara E 17 (broken). 34 Depuydt 1993a, 27. 35 Similarly Urk I 286, 7–17.
214
chapter five
(235) The king cancels previous obligations imposed upon certain establishments:
ir nf Ddw xr Hm(=i) wnt xtm wDw nw nzw r Sma r irt hA n kAt nt nzw m fAw Sdt kAt nb wDt irt m Sma pn sk Dd r wDw pf nfr-n iri.t(i) xwt nb m nwtw nt(t) xw.t(i) ntt m Sma pw n rdi.n Hm(=i) iri rmT nb nw nTrwy gbtw pr-min fAw Sdt hA nb n kAt nb irrt m Sma pw As for it being said to my majesty that royal decrees have been issued concerning Upper Egypt, namely about performing tasks of royal work, carrying and digging and whatever work ordered to be done in this Upper Egypt. Although it is said according to those decrees that no exemptions are to be made in towns already exempted in this Upper Egypt, my majesty (regardless) does not allow any people of the temple of Min, Koptos, V Upper Egyptian nome, to do digging, carrying, or any task of royal work done in this Upper Egypt. (Urk I 282, 15–283, 3) This passage, part of which was quoted as example (187) above, does not only provide the sole instance of nfr-n in a subject complement clause (after passive Dd), but it is also a superb example of lack of speaker commitment as a motive for the use of irrealis and non-assertion. Here, the king declares of previous decrees affecting the staff of the temple of Min that although these forbid further exemptions on establishments, this prohibition is hereby overruled. To paraphrase freely, his message is “although those decrees say that X is not to be done, I care little of this and do it regardless”. The king explicitly does not commit himself to the validity of the content of the previous decrees, which he first cites but then dismisses, and for this reason he uses nfr-n. This may be contrasted with the beginning of the passage cited, where a wntclause appears similarly as a subject of the very same verb Dd. There the degree of speaker commitment is strong: the king has indeed been told that there are decrees to the effect noted and there is no reason for non-commitment or non-acceptance. The variation of the two clause types is hence not random or based on syntax, but purely on the subjective attitude of the speaker towards the complement. However, negative examples of this sort are lamentably rare, and against the practise adopted in the present work, the following similar example with tm from the Pyramid Texts may exceptionally be quoted:
negative complement clauses
215
(236) The resurrection of the king is celebrated with the following remark to gods:
mTn nw Dd.n=Tn nTrw tm N wnn m xnt=Tn mTn N mn m xnt=Tn m imnw n smA See this what you said, gods—that the king is not before you; look, the king is firmly before you as a victorious bull. (PT 998)
The sense is clearly: “although you said this, the contrary is the case”, and the speaker’s attitude towards the subordinate proposition is nearly Schadenfreude. Loprieno terms the use of tm here ‘assertive’, which, unfortunately, is precisely what it is not.36 Instead, the negation signals that the speaker entertains no commitment towards the complement proposition, whose veracity has been proved completely void and which is hence coded as a non-assertion. Loprieno also suggests that tm was ‘originally’ a modal negation, for which reason it is used as a negation of e.g. final ‘so that’-clauses, but that in the mature syntactic system of Egyptian it has become a generalised ‘dependent’ negation.37 Yet, there are various indicators that the syntactic dependency of the negated clause is secondary in importance to its pragmatic profile overall. This is best shown by un-introduced, or bare, ‘circumstantial’ adjunct clauses. Here tm is not used to negate the quite clearly ‘dependent’ relative present adjuncts, where one normally uses the bare n sDm=f/ sDm.n=f instead:38 (237) The peasant petitions Nemtynakht about the return of his goods: iri.in sxty pn aHaw 10 r hrw 10 Hr spr n nmty-nxt pn n rdi. n=f mAa=f r=s Then this peasant spent a ten(-day) period of up to ten days petitioning to this Nemtynakht, without him paying (any) attention to it. (Peas B1 62–63) But if the adjunct describes an irrealis situation, tm (or less often 36
Loprieno 1991a, 217. Loprieno 1991a, 218, 223; see also n.73 in the introduction above. 38 See Collier 1999, 56; see ibid, 56 n.38 for the reading n rdi.n=f. In addition, there exists the specific adjunct negation n-is sDm=f/sDm.n=f, whose use is, however, rather limited. 37
216
chapter five
nfr-n) is used. Final ‘so-that’-clauses, which are negated by tm/nfr-n, are of course a prima facie case of this. The same applies also to the so-called ‘virtual clauses of condition’, which alone among non-final bare ‘circumstantials’ are tm-negated:39 (238) A title of a prescription: rdit Ssp Xrd tm=f snqw To cause a child to accept (breastfeeding) if he does not suckle (pRamesseum III, B10–11) (239) As above: kt smAa mwyt tm=s mAaw Another (method) of setting urination in order if it is not in order (pEbers 49, 8) This regularity provides good empiric evidence to the hypothesis that tm is specialised for negating irrealis-modal propositions. In addition, Loprieno has argued elsewhere that tm behaves as a ‘meta-linguistic’ ‘mention-negation’ expressive of negative presupposition.40 This means that, according to Loprieno, tm actually does not ‘perform’ negation per se but instead refers to situations whose negativity has already been co(n)textually established and is taken for granted.41 Although such predetermined negativity is clearly not at issue in the examples after wD above nor in e.g. negative final clauses,42 this analysis seems quite to the point e.g. in case of wh-questions, which are tm-negated.
39
Cf. here the affirmative (119) above. See also GEG § 347.3; WGMT § 364 and Satzinger 1968 § 71. Merikara E 87 has nn mn n=k Hapy tm=f iw “Inundation will not trouble you even if it does not come” and PT 499c n id=f tm=f sDm xrw=k “He is not deaf even if he does not hear your voice”. There are no examples with nfr-n. 40 Loprieno 1986b, 280–81; 1991b, 231–35. 41 For example, in case of a tm-negation such as e.g. Peas B1, 211 sDmw n A sDm. n=k tm=k tr sDm Hr m “Hearer, you really do not hear; why do you not hear?” the WHquestion negative state of affairs [you do not hear] is clearly presupposed. To borrow Loprieno’s terminology, the negation is indeed not actually ‘used’ but only ‘mentioned’; cf. Loprieno 1991b, 233. For remarks on WH-questions and assertion, see 10.3. 42 A final negation such as Ptahh 374 m qA ib=k tm=f dHi “Do not be arrogant of heart lest it be humiliated” hardly carries an “entailment of a modal denial” nor constitutes a ‘negative assertion’ any more than negated final clauses do in other languages (Loprieno 1991b, 234–35). Instead, it describes a hypothetical situation contingent upon the realisation of the preceding order.
negative complement clauses
217
In addition, when it is noted that negative-presupposed states of affairs are in no way different from positive ones in terms of their pragmatic character as fundamentally non-asserted information, Loprieno’s proposal may well be accepted as a characterisation of tm in its role to negate presupposed or more generally low relevance information. There are instances of negated complement clauses after verbs where this hypothesis aptly characterises the use of tm, such as the following celebrated example: (240) The king reminds that men are held accountable of their actions: DADAt wDa sAryw rx.n=k tm=sn sfn hrw pf n wDa mAir The tribunal, which judges the miserable—you know how merciless they are on the day of condemning the wretched. (Merikara E 53–54)
Here a similar ‘manner’-translation seems to suit tm as with the geminating sDm=f after verbs of cognition and perception.43 It was argued above that in the affirmative this semantic nuance is a reflection of the low information-value of the complement situation predicative nexus; the actuality of the latter is part of the discourse-pragmatic background assumptions and what is the centre of interest is merely e.g. the character (degree, manner of unfolding) of the situation.44 The same holds also in the negative but here it is, of course, the falsity of the situation that is taken for granted. In example (240) the assumption [they are not merciful] is already assumed in the context, exactly as formulated by Loprieno, and stands at the background in the discussion centring instead on the inherent properties of this state of affairs. In other words, the use of the irrealis herein has an evaluative motive. As seen, coding propositions as non-assertions is often used by speakers to signal that the information they contain is viewed as less than optimally relevant, not necessarily only because the audience is believed or known to be aware of it, but also because in the current discourse it constitutes something which the speaker does not assume
43 This example is almost always rendered with ‘that’ by translators (the most obvious exception being Borghouts 1985, 37 n.33) which again reflects the assumption that the presence and absence of iwt/ntt n after rx is not semantically conditioned. 44 2.2.2 above.
218
chapter five
to enhance the hearers’ ‘representation of the world’ more generally.45 In such instances it is the task of the audience then to deduce the reasons why exactly did the speaker deem something un-informative enough to warrant its non-assertion, which, as seen, may require rather extensive study of the co(n)text in which the proposition occurs. Yet, for communication to succeed at all, this is normally not overly complicated. For example, in the following passage the reason for the use of nfr-n is recoverable from the preceding text:46 (241) Having outlined the exemptions granted to the temple of Min, the king says:
iw wD.n nzw-bit nfr-kA-ra anx Dt r nHH sxp.t(i) a m wD pn dy r wD n inr r arrwt nt pr-min gbtw r mAA imyw-st-a nw spAt nfr-n Tzz=sn Hmw-nTr nw r-pr pn r kAt nb nt nzw m Awt Dt The majesty of the dual king Neferkara, living eternally and forever, has ordered a copy of this decree to be brought and put on a stela of limestone at the gateway of the temple of Min at Koptos, so that the functionaries of the nome may see that they may not recruit the priests of this temple for any royal (Urk I 286, 1–6) work for all eternity. Here the clause with nfr-n in fact does not convey any new information. Earlier in the decree (and its twin) it was already said that the king does not allow the temple-staff to be put to royal corvèe-work and that any functionary of such-and-such a rank who will recruit (Ts.t(y)=f(y) them into various tasks faces extreme royal disapproval (msDD nsw pw mAa mAa). In other words, the information that the functionaries ‘may not recruit’ the priests has already been passed in the preceding discourse, and here the issue is merely that a public record is to be made of this as a permanent reminder. The complement clause is not intended to present the information [functionaries may not recruit priests] as ‘news’—this is indeed only ‘mentioned’ anew, and for this reason nfr-n is used in the complement.47 The same motive for the use of irrealis appears to be in question also in the next example: 45 46 47
Lunn 1989b, 691. Literally similarly in Urk I 282, 9–14, but with a non-geminating form Tz=sn. In addition, there is a clear sense of denied permission herein.
negative complement clauses
219
(242) The author says a third party has witnessed his bowing to his addressee’s expectations:
iw gr mA.n smr wat(y) mr pr Htp nfr-n aHa sn=k im n Tzt nt mDA wA[wA]t n-mrwt [nfr-n] irr sn=k im msDD[t sS]=k Now, the unique friend and steward Hetep has seen that yours truly does not oppose (?) the troops of Medja and Wawat lest yours truly do what your excellence disapproves of. (pBerlin 8869, 12–13)
The topic of ‘opposing troops’ (if this is the sense of aHa n) is not explicitly mentioned in the letter previous to the present complement, but it is still treated as ‘old news’. In the opening of his letter, the author says that he has “given his utmost attention to the business” about which his correspondent had written earlier and had sent the official Hetep mentioned in the example above.48 He claims to have acted thus “lest yours truly do anything what your excellence may disapprove of”. Given the repetition of this phrase here and in the example above as well as the talk about the messenger, the ‘business’ in question can hardly be anything else but the ‘opposing of troops’. When it is mentioned again, by then the speaker has already said that he has given it his full attention, which must mean that he is ‘not opposing the troops’ and which thus, in the latter instance, is no longer ‘news’. Rather, the new and asserted information is that the messenger of the addressee has verified the author’s claim, to be found in the governing clause. Again the use of irrealis fits the context of discourse and signals low information value of the complement proposition.49 It is thus clear that in complementation after verbs, iwt is no more a ‘nominaliser’ than is ntt before an n-negation or than tm and nfrn are ‘nominal negations’. The use of these opposing strategies of creating negative complement clauses after verbs displays a definable and coherent set of parameters based on the realis or irrealis status
48
Lines 3–4. However, it goes without saying that with letters interpretation is riddled with difficulties precisely because the correspondents rely heavily on mutual background knowledge that is mostly unmentioned and, in the absence of previous correspondence, sometimes quite unrecoverable to outsiders. 49
220
chapter five
of the complement proposition. The existence in Earlier Egyptian of grammaticalised verbs for negation is an unusual feature crosslinguistically, and their employment for a special modal purpose is highly original. Nevertheless, in general terms Egyptian is in the mainstream of languages with respect of its treatment of negativity visà-vis modality. As is usual elsewhere, the marking of a complement as asserted or non-asserted does not follow from it simply being negative, but from the subjective assessment of the speaker as to the reliability, acceptability and ‘news value’ of the information it conveys, exactly as in the affirmative. Put another way, although negative situations are non-actualised, this does not determine their classification as realis or irrealis. In Earlier Egyptian complementation negation has no ‘semantic scope’ over modality and is not one of the factors activating grammatical marking thereof.50 Although the grammaticalisation of modal nuances is less fine-grained in negative than in affirmative complements, the former align themselves in the same continuum of speaker ‘attitudes and opinions’, which forms the foundation of Earlier Egyptian complementation.
50
Cf. the discussion in Mithun 1995, 380–84.
modality in complement clauses after prepositions
221
CHAPTER SIX
MODALITY IN COMPLEMENT CLAUSES AFTER PREPOSITIONS 6.1 Introduction In contradistinction to most ancient and modern languages, Earlier Egyptian did not possess specialised (i.e. grammaticalised) conjuncts for introducing adjunct clauses.1 Instead, and in keeping with the general functional flexibility of this language as well as its tendency to form clausal adjuncts without specialised means, ordinary prepositions were employed for this task.2 The number of elements thus used and of the semantic types of preposition-introduced adjuncts is relatively high. A classification of the latter, disputable in its details, but providing a useful summary of the prepositions employed for the purpose, is due to Gardiner, who distinguishes between3 1 For ‘preposition-ntt’-clauses, see 6.3 below. A partial exception to this are the elements (i)sT/(i)sk, which particularly in Old Egyptian often introduce temporal clauses (see EAG § 855). However, this does not seem to be their sole function even then. 2 As noted by Gardiner, (GEG § 181) there are certain prepositions that are used mainly as conjuncts. However, noun and pronoun complements do occur after these elements as well (see e.g. Urk I 118, 17 and Urk IV 1431, 14 for n-mrwt and n-aAt-n + noun) even if their conjunct use is more common. In Earlier Egyptian prepositions form a relatively closed class, but indicators of elements moving in and out thereof are occasionally discernible. An example of an expression in process of being grammaticalised as a preposition is n-aAt-n. In early examples (Urk I 139, 10; Urk I 216, 5; BM 614, 12; BM 1164, 3) the second n is often still a genitive nt agreeing with the ‘relational noun’ aAt (cf. Hopper & Traugott 1993, 107). Later this feature and even the entire n disappear, (see e.g. Urk IV 1671, 1 for n-aAt mrr=f) which, along with such late writings as n-aAw-n, (e.g. Amarna VI, pl. 21, middle 2; KRI II 260, 6) shows that phonetic contraction of the expression has set in. For an example of a preposition apparently undergoing grammaticalisation away from this functional category, see 6.4 below. 3 GEG § 222, quoted verbatim. However, some of these prepositions, namely tp-a, n-iqr-n and n-wr-n, are attested only with immutable verbs. In case of bare sDm=f, the present discussion concentrates mainly on complements from mutable roots, excluding the forms iwt and int. For the bare sDm.n=f after prepositions, see 7.1 below. To this list may be added the rare elements m-Sw ‘on account of’, (once with irr(=i), see EAG § 804) n-mnx-n ‘through the efficiency of’ (example (257) below) and the negative elements n-msDt and n-snD (6.2 below). A unique example of m-isw-n ‘in exchange of’ + rdi=f is Urk IV 1799, 4; pEbers 99, 8 appears to read m-a wnn wbA n irty “because there
222
chapter six 1. clauses of time. With m ‘when’; r ‘until’; xft ‘when’; Dr ‘since’; m-xt ‘after’;4 r-sA ‘after’; tp-a ‘before’; r-tnw-sp ‘every time that’. 2. clauses of condition.5 With ir ‘if’. Cf. too with m or mi ‘according as’. 3. clauses of asseveration. With m or mi ‘according as’. 4. clauses of concession. With m ‘though’. 5. clauses of purpose.6 With n-mrwt (rarely n-ib-n) ‘in order that’. 6. clauses of result. With r ‘so that’. 7. clauses of cause.7 With n ‘because’; Hr ‘because’; n-iqr-(n) ‘by virtue of the fact that’; n-(or m-)aAt-n(t) ‘inasmuch as’; n-wr-n ‘inasmuch as’. 8. clauses of comparison. With r ‘than’; r ‘according as’; xft ‘according as’; mi ‘as when’; mi ‘according as’. 9. clauses of co-ordination. With Hna ‘and’. 10. clauses of exception.8 With wpw-Hr ‘but’.
As can be seen from the multiple occurrences of various prepositions in the above list, the semantic range of such adjuncts is further expanded by the extensive polysemy displayed by many of these elements in Early Egyptian. As is amply demonstrated by a glance through any standard grammatical work listing meanings of prepositions, some prepositions can express a bewildering array of senses.9 For example, when serving as a conjunct, the preposition m is, depending on context, translatable (at least) as ‘during’, ‘when’, ‘while’ ‘as’, (in a temporal, comparative and concessive sense) ‘though’ and ‘whether’.10 With such elements, is an opening of the eyes”. An example of a non-geminating sDm=f after Xr is Ptahh 246 (Xr iri N) followed immediately (in 247) by geminating Xr dd N. However, the conjunct meaning(s?) of Xr is unclear. 4 For m-xt, see 6.4 below. 5 Conditionals with ir, despite of presumably employing a specialised use of the preposition r, should not be confused with complement clauses. 6 There does not seem to be grounds for delimiting the element n-mrwt to purposeclauses and r only to clauses of result. In Egyptian as in many languages, the demarcation between these types is at best precarious. 7 Gardiner omits the causal sense of Dr, ‘since’. 8 An example of wpw-r ‘except’ + wnn=sn is Urk I 305, 18, cited as part of example (181) above. 9 See e.g. Erman 1928 §§ 444–56; GEG §§ 162–81; EAG §§ 756–817; LGEC §§ 489–534; Callender 1975, 19; Englund 1988, 13–14; Loprieno 1995, 100; Grandet & Mathieu 1997, 49, 59–60, 71; Malaise & Winand 1999 §§ 227–90; Allen 2000, 83–88 among others; cf. also WGMT §§ 335–36. For discussion on the semantics of individual prepositions or generally, see e.g. Smither 1939; Anthes 1969; Junge 1973; Perdu 1978; Collier 1994, 60–67; Zonhoven 1996; 1997a; 1997b. 10 However, the only (probable) attestation of the last-mentioned sense is pBerlin 8869, 7, with mrr N.
modality in complement clauses after prepositions
223
interpretation is partly a matter of contextual deductions and partly based on recognition of idiomatic combinations.11 In Earlier Egyptian, the morpho-syntactic organisation of clausal preposition complements is identical to that found after verbs.12 The forms and constructions used to code the subordinate clauses are the same as in complements of verbs. Besides ntt- and wnt-introduced clauses, the complement often takes the form of a bare sDm=f, for which weak roots show alternatively gemination/doubling or no gemination/doubling. The non-geminating forms occasionally show writings with the endings -w and -y. tm/nfr-n and iwt/ntt n vary in the negative. Yet, the very heterogeneous semantic-pragmatic character of adjuncts introduced by prepositions and of the complement clauses involved therein would seem to have contributed to the perception among Egyptologists of their construal as determined almost solely by syntactic ‘nominality’ and ‘nominalisation’ of the subordinated verb forms and constructions. Many treatises, including the few that have tackled preposition complements from a more semantic perspective, have been characterised by scepticism about the role of ‘meaning’ in their grammar.13 Yet, the similarity between preposition- and verb complements is neither coincidental nor restricted to identical morpho-syntax. Instead, the grammar of both these categories reflects and is conditioned by the same system of expressing realis and irrealis modality. As seen, after verbs the primary modal division of complement clauses is that between assertions introduced by ntt/wnt/iwt and non-assertions without these elements. A further division is made in the domain of affirmative non-assertion, where various root-classes of the Earlier Egyptian verb are capable of distinguishing between what is termed here the distal and proximal types of irrealis modality. Also after prepositions, the un-introduced bare sDm=f-clauses are non-assertive and contrast with complements introduced by the elements ntt/wnt. The most notable difference of preposition complements vis-à-vis complements of verbs is that assessing the modality of clauses serving as complements of 11 See the fine discussion in Junge 1973. This semantic diversity may also occasionally only reflect defects in the translation-language. For example, English captures precisely the interplay of temporal and causal meanings of Dr ‘since’ whereas this is not the case e.g. in German. 12 See 0.1.1 above. 13 See 6.3 below.
224
chapter six
prepositions is largely—but not wholly—free from the complications effected by differences in identity, deictic locus and mental orientation between the original and the real speaker. Nevertheless, exactly as after verbs, the modal status of the former type of clauses is partially determined by the semantics of the governing expression. Like various verbs, certain prepositions systematically disallow ntt/wnt-introduced asserted complements and only occur with bare sDm=f forms. Yet, and again as after verbs, more commonly there is a choice between these alternatives depending on the more general semantic environment. The same holds also for the two types of irrealis, distal and proximal. Even prepositions with a singular sense usually accept both the geminating/doubling and non-geminating/non-doubling sDm=f as their complements. Rather curiously, however, in many instances the un-introduced complements of prepositions seem to be bereft of the key semantic and pragmatic properties that motivate the coding of propositions as irrealis. Instead, often the variance of the forms and constructions does not seem to pertain to modality but rather to temporal differences. However, the precise nature of the modal organisation in Earlier Egyptian preposition complement clauses can be properly understood only by extending the focus of investigation beyond the semantics of the constituent expressions of the phrasal complex [preposition + clause] and, as with verb complements, into the wider discourse context. Once again, this move is, in fact, a necessary one given that modality, which concerns the speaker’s attitudinal and evaluative perspective to propositions, has to do with discourse as a whole, and receives most of its ‘meaning’ thence.14 Finally, and almost as a byproduct, the organisation revealed provides the basis for modal analysis of a domain of Earlier Egyptian grammar that is syntactically quite different from complementation, but whose semantic-pragmatic and grammatical organisation in this language appears to share notable similarities with complement clauses.
14
Cf. Givón 2001 vol.1, 300 and 3.5 above.
modality in complement clauses after prepositions
225
6.2 The Typology of Irrealis in Complement Clauses of Prepositions: The ‘Obvious’ Instances In the Earlier Egyptian textual corpus there are copious examples of complement clauses after prepositions with the bare geminating/ doubling and non-geminating/non-doubling sDm=f forms that can readily be described as irrealis and unasserted. The most obvious case is final ‘so that/in order to’-clauses, which instead of asserting a state of affairs, rather posit it as the motive for the main clause event or speech act.15 For such clauses languages characteristically use irrealis and subjunctive forms, either on their own or ushered in by subordinating elements. In Earlier Egyptian, the latter category comprises complements introduced mostly by the prepositions r and n-mrwt, (literally ‘through love of’), after which both of the said types of bare sDm=f are common:16
15
See Palmer 2000, 128–31 and passim; cf. also Cristofaro 2003, 157–58 Further examples of n-mrwt + non-geminating/-doubling sDm=f: Koptos decree I, x+10 (iri(=i); Hatnub 22, 19 (iri(=i); Peas R 17.7 (in.t(w); Berlin 1199, 8 (wn(=i); see example (402) below); pRamesseum II, rto. 1 (iri=f); Urk IV 366, 15 (wn N); Urk IV 868, 8 (restored: [n-mr]wt iri=f); The occasional writings m mryt (e.g. Urk IV 1260, 11, with iri N) are perhaps variant spellings of n-mrwt. Further mutable examples after nmrwt show almost exclusively the form wnn of the auxiliary wnn; thus Hassan, Giza II, fig. 208A (n mrrw wnn N); Fischer 1994, fig. 1, 3; (wnn N); CT I 78a (n-mryt wnn N); CGC 1641, 5 (wnn N); Turin 1447, middle panel, 9 (wnn N); Turin 1449, 7 (wnn=i); Turin 1534, 7 (wnn=i); Siut VI, 16 (wnn=i); Louvre C34, bottom, 1 (wnn N); Door jamb in Luxor Museum 2 (wnn N); Urk IV 606, 7 (wnn N); Urk IV 853, 13 (wnn[=f]); White Chapel 67 (wnn N). In CT VII 54l one finds n-mrwt mAA N. Further non-geminating/doubling examples after r are: Urk I 126, 4 (wn=f); CT IV 178h (qA=i); BM 614, 10 (iri(=i); pUC 32199, 12 (wn N); pRamesseum IV, Div, 2 (mA=k); pEbers 8, 16 (pr N; see example (393) below); pEbers 88, 19 (hAw N; see example (8) above); Urk IV 1216, 6 (mA N); Urk IV 1798, 17 (gm=tn); Tb 64, 33–34 (mA=i); Tb 149, 75 (mA N). Louvre C14, 9 has a clear -w-form r iw N of the root ii ‘go’. The expression rdi s r wn=f mAa “who treated a man so that he would be soothed” (see Vernus 1999) seems to belong here as well. Further geminating and doubling examples after r are Urk I 202, 7 (mAA N); Urk I 282, 12 (mAA N, broken); Urk I 286, 4 (mAA N, broken); Urk I 306, 12 (mAA N); Kom elKoffar A, 4 (wnn N); Hassan 1975, vol. 2, pl. 4A, 3–4 (mAA N); CT I 391a/M5C, M28C (prr=i); CT II 102b (mAA N); CT III 328a (irr=i); CT IV 75f (wnn=sn); CT V 241c (mAA N); CT VI 73a (prr=f); CT VI 328b (hAA=f); CT VI 336l (dd N); Admonitions 13, 4 (mAA=f); pUC 32036, 1.x+12 (Sdd=k); Tb 149, 98/Nu pl. 86, 107 (mAA=f); pRamesseum III, A19 (prr N); pEbers 40, 14 (ngg=k); pEbers 41, 2 (hAA=f); pEdwin Smith 4, 9 (naa=f); pEdwin Smith 8, 12 (mAA=k); pEdwin Smith 16, 6–7 (hAA N). pHearst 9, 12 has iw N. There are also examples of r mAn=f/N that, given the context, are certainly final clauses rather than the sDm.n=f (e.g. Urk IV 367, 17; Urk IV 1294, 11). 16
226
chapter six
(243) The king instructs the high steward on the treatment of the peasant:
in-mrwt wn=f Hr Dd gr ix in.t(w) n=n ø m Ss sDm=n st In order that he might get into talking, be quiet. Then let (it) be brought to us in writing that we may hear it. (Peas B1 110–11)
(244) The high steward retorts angrily to the peasant’s pledges: Dd.in mr-pr wr mrw sA rnsy in aAt pw n=k-imy Hr ib=k r it.tw Smsw=i Then the high steward Meru’s son Rensi said: “Are your belongings in your opinion of such importance that my retainer (Peas B1 134–35) should be arrested?”17 (245) The king confirms Sinuhe’s status as a royal favourite: iw=f r smr m-m srw rdit=f m-qAb Snyt wDAw tn r aXnwti dwAt r irr.t(w) aHaw=f He shall become a royal friend among officials and installed among the entourage. Proceed now to the audience-hall so that (Sin B 280–83) one may wait for him. (246) Ukhhotep explains his reasons for depicting previous nomarchs in his tomb:
iri.n=i nw n-mrwt wnn rnw=sn mn n Dt I have done this in order that their names might be firm for (Meir III, pl. 11) eternity. Here again the forms appear to be used to articulate degrees of nonassertion in accordance with their distal and proximal irrealis profiles. In example (243) the realisation of the complement event seems less certain and less in control of the speaker, given that obtaining the desired outcome involves a complicated scheme of ‘tricking’ the peasant into petitioning against his will. The situation described is, in addition, clearly not envisaged to occur immediately, but after everything is
17 The traditional reading of tw as a dependent pronoun (cf. “that my retainer should arrest you”, Faulkner in Simpson 1973, 37; Parkinson 1997, 63 “my follower’s seizing you”) is certainly wrong.
modality in complement clauses after prepositions
227
in place for the cunning plan to be put into action.18 In (244) there is a clear element of negative attitude involved: the non-geminating sDm=f complement describes a situation rejected out of hand by the speaker and viewed as unworthy of consideration. Conversely, in examples (245) and (246) various factors appear to conspire to render the complement situation ‘closer’ to the speakers. In (245) they seem to relate to the king’s role as the supreme authority issuing orders, the expectations arising from this as regards compliance, the temporal proximity of the complement situation to the time of speaking and, in (246), even its ontological reality. In this last example, the author clearly is not saying that his actions were motivated by an intention literally to ‘make’ the names eternal, but to perpetuate the existing fame of the forefathers, whose names ‘have been’, ‘are’ and now ‘will’ continue to be renown.19 One may contrast this with (243) above as well as with the following example: (247) Hapdjefa sums up his motives for transferring property to his funerarypriest:
mk smnx.n(=i) Tw m AHwt m rmT m mnmnt m Sw m xt nb mi sr nb n sAwt n-mrwt iri=k n=i xt See, I have provided you with fields, people, cattle, orchards and all things like every official of Assiut, in order that you may (Siut I 271) perform rituals for me. Here again an undefined stretch of time is expected to pass before the performing of the rituals takes place. The contract comes into force and the funerary priest is due to begin his service only after the death of the speaker, who is donating the wherewithal he then wants to be used for the purpose specified. Due to lack of morphologically revealing examples, it is unclear if this sort of formal and semantic-pragmatic variation occurs also after the other prepositions used to introduce final clauses, namely n-ib-n 18 This equals arrangements for food and drink to be provided to the peasant and his family, described just before. The combination r wn=f Hr mdt is apparently a device to combine the non-immediacy brought by wn with the continuous sense of Hr + infinitive. The translation of (243) as “in order that he might get into talking” is an attempt to capture this nuance. See chapter 9 for a full discussion of the temporal features involved here. 19 See 9.2 for further discussion.
228
chapter six
(‘through desire of’) and the extremely rare negative elements n-snD and n-msDt, ‘lest’, literally ‘through fear/hate of’:20 (248) A medical instruction for treating a wounded (and infected?) earlobe:
sf.in=k gs=f wa n-ib-n hAw snf=f Then you cut its one side so that its blood may flow down. (pEbers 91, 15–16)
(249) In a broken context: [...] igrw n-snD prr rxyt […] the silent ones lest the plebs go forth.
(CT VI 368l–m)
In the last instance above the complement clause is affirmative and it is the preposition that is semantically negative. There are relatively few examples of preposition complement clauses where the clause itself is negated. Most examples derive from final clauses after r and n-mrwt after which the negations nfr-n and tm appear regularly:21 (250) Weni boasts of his skilled organising of a military force: ink wn iri n=sn sxr sT iAt(=i) m mr xnty-S pr aA n mt n st(=i) r nfr-n dd wa im m sn-sw=f r nfr-n nHm wa im xAD Tbty m-a Hr(y) wAt r nfr-n iTT wa im dAiw m nwt nb r nfr-n iTT wa im wat nb m-a rmT nb It was me who took care of their organisation—even though my office then was only that of an overseer of royal tenants—due to the exactitude of my disposition, so that not one of them would harm his fellow, so that not one of them would seize a loaf or sandals from a traveller, so that not one of them would steal a cloth from any town and so that not one of them would seize (Urk I 102, 9–16) even a goat from anyone.
20 Gilula 1969; further examples are CT I 174f (n-snD mAA=f); CT VI 143i (n-msDyt wnn=f ) and CT VII 171g (n-msDwt mAA N). 21 Also Urk I 106, 5 (r nfr-n dd N); Urk I 278, 9 (r nfr-n iT[…]); pBerlin 8869, 4 and 13 (n-mrwt nfr-n irr N).
modality in complement clauses after prepositions
229
(251/12) Hapdjefa describes his impeccable conduct in life: ink dr bXbX m qA-sA sgr qA-xrw r tm=f mdw I was one who removed pride from the arrogant and silenced the loud-mouthed one so that he would not speak. (Siut I 229) (252) It is said of Ra and the guardian flame of the underworld: iri.n=f [st] r imw-xt=f n-mr(w)t tm=sn ssn tAw wp-Hr nTr pwy Spsy He made it against those who are in his train so that they could not breathe air—except for this noble god (the deceased). (Tb 149, 56–57)
Besides final clauses, the bare sDm=f forms are seldom found in preposition complements with equally clear irrealis sense. There are some instances that appear to represent deontic injunctions: (253) Nika-ankh establishes rules for his heirs concerning the running of his cult:22
irr=sn Xr a sA(=i) smsw mi irr=sn n xt(=i) Ds(=i) They are to act under the authority of my eldest son just as they are to act with respect of my own religious service. (Urk I 162, 11)
(254) Hapdjefa explains to officials what duties he expects from them and their successors:
ntsn iri=sn n=i pA t Hnqt Hna wnn=Tn m-sA Xnt(=i) nty m S=i Hr Sms=f x[ft wDA=f r pr inpw] It is they who will provide for me the said bread and beer, and you are to be after my statue, which is in my tomb-garden whenever it proceeds to the temple of Anubis. (Siut I 316–17) Rather than simply describing current or eventual states of affairs, the speakers in the examples above seem to be distributing orders and laying directives on the conduct of their addressees. There are also some instances that appear to read as ‘obligative’ statements rather
22
Cf. Urk I 31, 3.
230
chapter six
akin to English ‘should’- or ‘ought to’-clauses,23 but these too are rather infrequent and not absolutely certain at all times:24 (255) Merer describes his care of the people of his district during a foodshortage:
iw xtm.n(=i) AHwt=sn iAwt=sn nbt m nwt m sxt n rdi(=i) mH=sn n ky m iri nDs iqr I sealed off all their fields and their mounds in town and in the countryside, and I did not allow their waters to inundate for someone else, as an astute individual should do. (Krakow MNK-XI-999, 10–11)
However, particularly the bare geminating sDm=f occurs commonly in contexts where the preposition complement proposition clearly consists of low relevance background information. As seen, this form is common as object of cognition and perception verbs as well as in subject clauses of adjectives and adjectival verbs when the centre of interest in the situation described is its degree, quality and character rather than its occurrence and ‘reality’ as such.25 As has been noted by various authors,26 the same nuance is commonplace in complement clauses of prepositions with a ‘correlative’ meaning such as n-aAtn(t) ‘through the extent of/how much’,27 and n-mnx-n ‘through the efficiency of’. Both these prepositions refer directly to the quality and degree of the complement situation and are invariably followed by geminating or doublinig sDm=f:28 23 ‘Obligatives’ involve the crucial ‘element of will’ and are traditionally understood as deontic. But they also express value-judgements and could be characterised as a hybrid between deontic and epistemic, as well as having characteristics not associated with either of these modal categories; see Sweetser 1990, 56, 61, 63; Myhill & Smith 1995; Palmer 2000, 73–74 and passim. 24 Further examples with the spelling m/mi iri and where a similar reading seems suitable are Man and Ba 16; UC 14430, vertical x+7; perhaps also Urk IV 752, 14 (Hr mAn N). A past reading in these instances is excluded, unlike in the common dedicatory formula m iri N “as made by N” with the ‘m of predication’—see n.57 below. 25 See 2.2.2 and 4.2 above. 26 Polotsky 1984, 119; Borghouts 1985, 35; cf. De Cenival 1972, 45. 27 I.e. not simply ‘inasmuch as’; cf. EAG § 777; Reintges 1997, 377–78. 28 The rare counterexamples after n-aAt-n(t) (Sinai 200, x+9; Urk IV 1686, 19, both mr=f) are late and apparently erroneous; e.g. Sinai 200 has correctly mrr=f immediately before (x+1). Further examples of n-aAt-n(t) mrr=f (all with this verb) Roumiantshev Museum 18 17/III 78, 6; Pushkin Museum I.1.a.1137, 5–6; Berlin 1203, 3; BM
modality in complement clauses after prepositions
231
(256) One of the laudatory epithets of Djehutinakht: nHn n=f nwt=f wAH n-aAt-n mrr=s sw One to whom his city prays long life, through how much it (Bersheh II, pl. 7, top, 7–8) loves him (257) Kay says his skills in conducting expeditions were rewarded: rdi.n wi nb=i m sA=sn m rwd=f n aq-ib=f n-mnx-n irr=i wpt n nb=i My lord appointed me as their protection and as his trusted administrator, through the efficiency of how I carried out (Berlin 22820, 6–8) business for my lord. Yet, similar sort of ‘background-ness’ is also discernible after other prepositions, particularly with state- and adjective verbs: (258) Yamu-Nedjeh sings praise to the sun-god: i.nD Hr=k nn gnn HH n sp n{n} wnwt twt n=k iAw mi wrr=k Hail to you, one with no weakness for millions of hours. Adoration to you is fitting in accordance to how great you (Urk IV 943, 4–5) are. (259) The deceased addresses a divinity: i Sp tp=f imy 6 qA=k xft qAA=i Ts=pXr O Blind-of-Face among the six; be exalted as (i.e. in proportion (CT II 156b) to how) I am exalted—and vice versa.29
614, 12; BM 1164, 3; BM 1367, 12–13; Hatnub 11, 4; Hammamat 113, 15; Hammamat 114, 16; Neferhetep Stela, 40; Sinai 196, 14; White Chapel 64; BH I, pl. 25, 45–46; BH I, pl. 25, 74–75; Bersheh II, pl. 21, top 6; MMA 35.7.55, 14–15; Red Chapel 460, 5; Urk IV 28, 14; Urk IV 43, 15; Urk IV 100, 3; Urk IV 141, 7; Urk IV 170, 14; Urk IV 312, 15; Urk IV 322, 12; Urk IV 554, 5; Urk IV 590, 16; Urk IV 843, 14; Urk IV 849, 14; Urk IV 852, 16; Urk IV 859, 4; Urk IV 868, 4; Urk IV 889, 5; Urk IV 1708, 10. Also with the form Hss=f from the verb Hsi, ‘praise’: Urk I 139, 10; Urk I 216, 5; UC 14333, 6; Goyon, Hammamat no. 61, 12; München GL WAF 35, 21–22 (spelled m-aAt-n); Hatnub 28, 5. There are no further examples of mutable verbs after n-mnx-n. 29 With xft, see also Ptahh 76 (Xss=f, pPrisse); Rifeh VII, 31 (mrr=f); Urk IV 272, 10 (irr=i).
232
chapter six
(260) A note concerning the legal treatment of two children: iri aA r aAA=f Sr r Srr=f The older was treated according to how old, and the younger (pBerlin 9010, 3) according to how young he was.30 (261/90) Instruction concerning the use of certain ingredients in preparing a potion:
ps Hr qd r mnx rx=k pss=sn Hr axx mw iry Hr Sww=sn Boil completely and thoroughly. You may tell how cooked they are on the basis of how water evaporates and how dry they (pEdwin Smith 21, 17–18) are.31 Further, in many ‘Appeals to the Living’, the funerary formula discussed in section 3.4 above, the oath serving as a premise for the request for offerings begins with a geminating sDm=f introduced by the preposition m, less often mi:32
30 Cf. also the expressions r mrr=f “just as he (etc.) pleases/pleased” (e.g. CT I 265g; CT II 29b; CT II 47f; CT II 254v; CT III 263e; CT V 123b; CT V 349b; CT VI 73a; CT VI 290f; CT VI 334o; CT VI 344d; CT VII 3p; Hammamat 114, 8; Peas B1 157– 58; Literary Fragments pl. 19, B2, 11; Tomb of Amenemhat pl. 30C; Urk IV 65, 3 and 8; Urk IV 520 1 (erroneously l. 10, r mr=k); Urk IV 617, 9; Urk IV 1064, 6; Urk IV 1236, 1; Urk IV 1298, 9/10; Urk IV 1517, 5/6; Urk IV 1662, 11; Urk IV 1664, 13; Urk IV 1744, 5; Urk IV 1849, 13; Urk IV 1889, 4; Amrah pl. 33, no. 2, 4–5; Tb 17, 105; Tb 69, 6; Tb 110, 15 (r mrr st-ib=f “just as his state of mind desires”); Tb 134, 4; Tb 126/ Nu pl. 70, 14. In Sin B 236 m mrr=f has the same sense. The contrasting expression r msdd=f “according as he hates” (Urk IV 758, 9) is much less ubiquitous. Frequent are also the synonymous expressions r dd ib=k (e.g. Five Theban Tombs pl. 19, 4; Tomb of Amenemhat pl. 30F; Urk IV 1014, 16) and xft dd ib=k (e.g. Urk IV 116, 17; Urk IV 499, 6; Urk IV 1221, 7) “according as your heart directs”. 31 Cf. also pUC 3206, 1.x+14 (example (79) above) with Hr iw N and the expression “I was rewarded (etc.) Hr qnn=i due to how brave I was”, found occasionally in the bellicose ‘autobiographies’ of the early XVIII dynasty (e.g. Urk IV 3, 4; Urk IV 892 12; Urk IV 1371, 16; Berlin 19286, left 11). 32 Alternatively, one may employ the geminating sDm=f without a preposition and sometimes the two appear as variants in the same inscription (e.g. as noted in n.86 of chapter 3 above, BM 1213 vertical 6 has m mrr=Tn wp-wAwt nTr=Tn bnr-mrwt mrr=Tn Sms wsir “As you love Wepwawet, your god sweet of love; and as you love to follow Osiris…”). The omission of the preposition might have been purely graphemic, assuming that the /m/ of the preposition m and of the verb mri were not separated by a vowel (Matthias Müller, PC). Further instances of m + mrr=Tn/=tn are BM 152, 4 (see example (173) above); BM 579, 4; BM 584, 4; CGC 20040, x+10; CGC 20536, 4 & 5; CGC 20538 I, d3; CGC 20538 II, c23; CGC 20683, 2 & 3; Heqaib 52e, 3; Heqaib 67, a6; Heqaib 88, 2; Leiden L.XI.8, 7; MMA 65.120.2, 3; Turin 1447, middle panel, 4;
modality in complement clauses after prepositions
233
(262) Appeal to the Living of Abkau: i anxw tp tA mrrw anx msDD xpt m mrr=Tn wAH tp tA Dd=tn xA t Hnqt... O the living upon earth, who love life and hate death; as you desire to endure upon earth, say: “A thousand bread and (Turin 1534, 10) beer…” (263) Appeal to the Living of Khuy: i anxw tpw tA swA.t(y)=sn Hr iz pn mi mrr=Tn dwA nTr=Tn ib=tn nDm Dd=Tn swA=Tn xA t Hnqt O the living upon earth, who will pass by this tomb; even as you desire to praise your god with your heart(s) joyful, may you say when passing by: “A thousand bread and beer…” (UCLA 3769, 3–5)
The complement situations here are presupposed. The speakers take it for granted that their audience is willing to ‘remain upon earth’ or ‘praise their gods’ etc., and this state of affairs has no information value. The audience is assumed to know and agree upon the content of the propositions, and a plausible alternative translation for examples of this sort would be a concessive “given that you want…”33 The sense is perhaps best discernible in the following famous passage: (264) The king tells the high steward what to do with the peasant: Dd.in Hm=f m mrr=k34 mA=i snb.kw swdf=k sw aA nn wSb r Ddt=f nbt Then his majesty said: “As you desire to see me healthy, retain him here, without answering to whatever he might say”. (Peas B1 109–10)
The king promptly assumes his henchman’s utmost desire to be the well-being of his lord, and rather than informing him of this self-evident matter with zero ‘news value’, he treats it as a mere starting-point for his actual message. In all these instances the adverbial clauses are Tübingen University 458, 8–9; UCLA 97, 1. Further examples of mi + mrr=Tn/=tn are Urk I 268, 13 and Dendereh pl. 2a, bottom left, 3. 33 Cf. GEG § 444.2. 34 In the R variant a bare mrr=k appears instead.
234
chapter six
ante-posed as sentence-initial, which in general is a sign of their status of clausal topics with background informative value.35 Beyond the type of Appeal just discussed, in Earlier Egyptian the phenomenon of true adjunct ante-position is relatively rare, but there are some examples of this with the topic-marker ir:36 (265) King Intef II sings praise to Hathor: Xt(=i) Dd=s spty(=i) wHm=sny iHy wab n Hwt-Hr iHy HHw Hfnw ir Dr mrr=T iHy iw HH n iHy n kA=T My body says, my lips repeat: priestly music for Hathor; music a million myriads! Since you love music, there is music for a (MMA 13.182.3, vertical 4–5) million times for your kA. In the examples above, the complement states of affairs represent the ‘present reality’ of both the speaker and the audience. In instances such as (264) or (265) this is, of course, necessarily the case, but in the Appeal to the Living it need not be. As a reflection of the difference between the ‘here and now’ of the speaker and the audience, a non-geminating sDm=f of the verb mri is often used after the initial prepositions m/mi in this formula instead of a geminating form:37 35
See e.g. Haiman 1978, 573–74; Thompson & Longacre 1985, 229–32; Lambrecht 1994, 125–26 and Givón 2001 vol. 2, 344–45; see also the latter for arguments against the use of the term ‘topic’ for these construals and cf. Ramsey 1987. In Halliday 1985, 44, 48, ante-posed ‘background-establishing’ adjuncts are treated as marked themes. The exact meaning of these terms will be discussed in some more detail in 8.2 and 10.3 below. 36 See also pEdwin Smith 2, 6; 3, 7 and 4, 10, all ir Dr gmm=k “since you find” and Adminitions 12, 12 ir m irr=k “since you act”. For further examples ante-posed with ir, see n.65 and n.110 below. It will be suggested in 10.3 below that the so-called setting second tenses have the same pragmatic function as ante-posed adjuncts stricto sensu. 37 Further examples with m + mr=tn/=tn are CGC 20119, c3–4; CGC 20341, 9; CGC 20458, b2; CGC 20515, 6 (m mr{t}=Tn); CGC 20516, 6; CGC 20540, 2 (m mri=Tn); CGC 20606, 3; CGC 20609, 2 (m mry=Tn, i.e. with the ending -y); CGC 20748, g5; Ny Carlsberg Glyptotek ÆIN 964, centre, 4; Geneve D50, 4; Koptos pl. 12, 2, x+6; Heqaib 20, 2; Heqaib 48, f2–3; Heqaib 49, d3; Berlin 1188, A3 (m mri=Tn); Semna-Kumma R.I.K. 12, 1; Semna-Kumma R.I.K. 51, 3; Felsinschriften no. 400, 1 (see example (185) above); Felsinschriften no. 484, 7; Urk VII 1, 13; BM 805, 4–6; Louvre C181, 1; Oxford Queen’s College 1113, bottom 4; Turin 1541, 4–5. The variant mi + mr is rare, but one may cite e.g. ASAE 55, 240 and CGC 20093, 4, the latter of which has mi Sd=Tn “according as you will recite” at the end of the oath. The geminating and non-geminating versions alternate occasionally after m in the same inscription (e.g. CGC 20683, 2–4 m mrr=Tn Hwt-nTr nt wsir... m mrr=Tn wp-wAwt... m mr=T(n) anx
modality in complement clauses after prepositions
235
(266) Appeal to the Living of Ipi: i anxw imw Abw m mr=Tn mAA r-pa HqA-ib dwAt nt Hb skr Dd=Tn xA t Hnqt... O the living who are on Elephantine; as you will desire to see the noble Heqaib the morning of the festival of Sokar, may you (Heqaib 61f, 3–5) say: “A thousand bread and beer…” (267) Appeal to the Living of Tiy: i anxw tp tA swA.t(y)=sn Hr maHat tn m mr=Tn grg Hwt-nTr tn n(t) wsir xnty-imntw nb AbDw Dd=Tn xA m t Hnqt... O the living upon earth who may pass by this cenotaph; as you will desire this temple of Osiris Khentyamenty, Lord of Abydos, to be well-founded, may you say: “A thousand bread (Florence 2571, 3–6) and beer…” In these variants the speaker, rather than expressing himself generally, seemingly opts to maintain his own current deictic centre and position in time.38 This is different from that of his audience, which from the speaker’s perspective will at some future point come to recite the inscription and whom he addresses in a way appropriate with this respect, i.e. future ‘as you will desire’ in ‘your’ present. That is, e.g. “as you will desire to see N” addresses people who will desire something in the speaker’s future, not their own; the phrasing makes sense only from the speaker’s deictic vantage point. The complement situation is no longer presented as actual from the speaker’s perspective but rather only as epistemically probable—somewhat akin to English “assuming you will desire…” which has a semi-conditional overtone—and as a nsw “As you desire the temple of Osiris to be firm… as you love Wepwawet… as you will desire that the king lives”; Louvre 196, 2–3 m mr=Tn anx m msdd=Tn mt “as you will love life and as you hate death”). Like the geminating sDm=f, bare non-geminating forms appear commonly, sometimes in the same text with m (e.g. Lesestücke 80, 4– 5 mr=Tn anx msDD=T(n) m(t) m mr=Tn xnty-imntw “As you love life and hate death; as you love Khentyamenty”; Pushkin Mus. I.1.b.32/UC 14326, x+6 mr=Tn mAA nfrw HwtHr... m mr=Tn mAA nfrw Hr-smA-tAwy “As you will desire to see the beauty of Hathor… as you will desire to see the beauty of Horus-Unifier-of-the-Two-Lands”). Turin 1628, bottom 2 has two bare -w-forms in a sequence (mrw=Tn anx smxw=Tn mt “as you will recall life and forget death”). 38 See Allen 2000, 364 whose semantic (if not syntactic) analysis of these construals is essentially that presented here.
236
chapter six
‘possibly true’ rather than a ‘true’ situation. The semantic difference achieved by the variation here is, of course, the same as when the form of the complement of mri is altered in the Appeal.39 It allows more latitude for the speakers to express degrees of informational and deictic ‘sameness’ with the audience. The use of non-geminating forms for propositions representing low relevance background information is rare after prepositions. In the following remarkable example, the form in=f seems to have been chosen due to the explicitly past and completed status of the contextually clearly presupposed complement situation:40 (268) Ankhtify reflects upon his mission to Edfu: i[w] in.n w(i) Hr r wTst-Hr n aws r grg=s iri.n(=i) xr wn Hr Hr mrt grg=s Hr in=f w(i) r=s r grg=s Horus brought me to Edfu through l.h.p. to establish it, which I did. Horus must have wanted to establish it because he brought me there to establish it. (Moaalla Iα2) More commonly, non-geminating sDm=f forms occur in instances where the preposition complement clause refers to hypothetical situations: (269) The sailor describes what he found on the island: gm.n=i dbAw iArrt im iAqt nbt Spst kAw im Hna nqwt Sspt mi iri.t(w)=s I found figs and grapes there, and all fine vegetables; there were sycamore figs—ripened ones as well—and cucumbers like it had been (ready-) made.41 (Sh.S. 47–50)
39
See 3.3 above. Compare also this use of non-geminating sDm=f forms with their employment in subject complement clauses with the sense “it is good if you should…” (4.2 above). The non-geminating version is unambiguously distal also temporally, but the difference between m/mi mrr=Tn and mr=Tn is not merely that between ‘present’ and ‘future’. Futurity is already ‘included’, as it were, in the former variant, whose very general meaning ‘in that/just as you desire’ pertains to all thinkable situations. See further chapter 9 below. 40 See 6.3 below for further discussion of this example. The translation of wn Hr Hr mrt as an epistemic conclusion follows Vernus’s (1990, 31) insightful interpretation. 41 So also Simpson in Simpson 1973, 52 “as if they were cultivated”; Junge 1978a, 90, “wie angepflantzt” ; Parkinson 1997, 93 “as if cultivated”.
modality in complement clauses after prepositions
237
Here the completely un-verifiable and -knowable complement state of affairs represents a mere ‘distant possibility’ (note the idiom). It is not, however, imaginary or counterfactual. Yet, the non-geminating sDm=f may also describe counterfactual situations after prepositions. This usage appears to have gone unnoticed in spite of counterfactual being the only semantic area to which the term ‘irrealis’ has been applied in Egyptology.42 The reason for this lies presumably in the rather restricted nature of the example basis. In preposition complements, counterfactual sense occurs for certain only after mi in ‘simulative’ ‘as if’ -clauses:43 (270) Queen Hatshepsut says of the tribute destined for her: inn.tw n=i antw nw pwnt mi Xn Ssw The incense of Punt is brought to me as if grain was flowing. (Urk IV 372, 13–14)
(271) King Kamose narrates the overthrow of a traitor: HD.n tA iw=i Hr=f mi wn bik xpr.n nw n sty-r sAsA=i sw xb.n=i sbty=f smA=i rmT=f di=i hA Hmt=f r mryt mSa=i mi wn mAiw Xr HAqt=sn When morning broke, I set upon him as if it were a falcon. When suppertime came, I was already overthrowing him: I demolished his wall, slew his people and made his wife go down to the riverbank—and my army was as if lions were carrying (Carnavon tablet, 14–15) their prey.44 Cross-linguistically, counterfactual is the meaning most widely associated with the formal categories of subjunctive and irrealis.45 Yet, the 42
See 0.1.2 above. Cf. Chafe 1995, 357 for an example of the use of irrealis in comparable ‘simulative’ utterances in Caddo; cf. also corresponding subjunctive uses in French (Judge & Healey 1983, 132). Further examples where a similar translation seems suitable are Karnak stela of Mentuhetep, 10 (pr N); Urk IV 340, 1 (iri N) as well as Urk IV 1675, 18, surely meaning “I have given you all the land… nsy=k sw mi wn=i m nsw-bit exercise kingship over it as if I were the dual king”. 44 Gardiner 1916, 107 and Smith & Smith 1976, 60 have “as it were a hawk” and “as if I were a falcon” respectively, but “like lions” and “as lions are”. Kamose Stela, x+6 and x+7 have a very similar expression, but the spelling is, inexplicably, wnn. 45 See 0.1.2 above. 43
238
chapter six
term counterfactual is, in fact, somewhat of a misnomer. ‘Factuality’ in the ontological sense is not the primary issue with counterfactual propositions, but the lowest possible speaker commitment and “greatest degree of remoteness from reality” as the speaker understands the latter.46 Against this, the use of the distal irrealis non-geminating sDm=f here is quite expected. Counterfactual situations are known not to obtain, and this total speaker non-commitment is obvious from the fact that, as can be seen also from the examples above, such instantiations are not envisaged to be located anywhere in time. Counterfactual states of affairs are conjured up as ‘antithesis’ to ‘reality’ and represent the nadir of speaker commitment on the scale between assertion and non-assertion. 6.3 ‘Indicative Non-indicatives?’ The Broader System of Preposition Complement Modality in Earlier Egyptian In Earlier Egyptian bare geminating and non-geminating sDm=f forms and their functional counterparts as well as the negations tm and nfr-n thus occur in preposition complement clauses with definitely irrealis senses. The examples presented in the previous section display the same general typological divide of non-assertion into propositions characterised by low speaker commitment to the veracity of the situation described and those with low discourse relevance of the information passed as is found in complements of verbs. Similarly, they once again illustrate the specifically Egyptian division of irrealis into subjectively more distal and proximal subtypes, reflected formally in certain morphological classes of verbs. However, the rest of the data of bare sDm=f forms serving as complements of prepositions at first sight seems to contradict the hypothesis of them as expressive of non-assertive and modally irrealis meaning. In numerous instances clauses of the sort appear to convey no attitudinal nuance or mitigated communicative value of the proposition. The geminating sDm=f and its functional counterparts often describe situations apparently with a simple generic or habitual time-reference:47 46 D. James 1982, 376, emphasis by SU; cf. also Givón 1982, 160 n.24; Roberts 1990, 393–94. 47 As for prepositions other than those exemplified below, for n see Urk I 84, 1 (mrr N, followed by n wn N on line 3); Urk I 204, 10 (mrr=i; see example (169) above);
modality in complement clauses after prepositions (272) The deceased says to the yon ferrymen: in n=i mXnt tw m iww=i Bring me this ferry-boat when(ever) I call.48
239
(CT V 174c–d)
(273) The deceased is served as a god: sar.t(w) n=f ø r-Tnw ibb=f (Something) is lifted up for him every time he thirsts.49 (CT V 11d)
(274) The deceased defies a malicious spirit: n xtm=k r=i Hr HkA imy Xt=i mi irr=k r Axw Hr HkA imy Xt=sn You will not shut up my mouth for the sake of the magic in my belly, as you are wont to do to spirits for the sake of the magic (CT V 322i–j) in their bellies.50
Tomb of Ankhi, false door niche, 4 (mrr=(i); CT VII 177e (wpp=f); Urk IV 1673, 14 (xaa=k). For r-sA, see CT IV 326j (prr=f) and Urk IV 1112, 4 (irr=f, variant iri=f). For Hr, see pBerlin 8869, 6 (mAA N); Hatnub 22, 15 (prr=i); Sin R142 (mAA=f; see example (6) above); pEdwin Smith 22, 4 (prr=f). A possible instance of Hna SAA=sn is Urk I 19, 3. 48 Further examples after m: Berlin 9571C, 2 (xaa=k); MMA 57.95, 5 (dd N, hAA N); Admonitions 12, 12 (msdd=k); Berlin Bowl, 2 (msDD=T); CT IV 187a/T3Be (wnn=i; Tb 17, 3 has wn=f); CT IV 236/37b (itt N, variants either it N, m-xt itt N, m-xt it N or mxt it.n N; Tb 17, 31 has itt N); Tb 72, 12–13 (dd.tw). CGC 20057, d2 has iw=f as does Urk IV 702, 1 and Urk IV 1159, 11 (followed by m Sm=f); Literary Fragments pl. 18, 1 (msdd=sn); pEbers 70, 24 (msdd=f); Urk IV 969, 3 (msdd N). There are also examples after the ‘m of predication’: Semnah desp. 2, 8; 3, 7 and 4, 6 (all m dd N); Urk IV 363, 10; Urk IV 389, 3 and Urk IV 439, 1 (all irr=i/=f); Urk IV 1807, 21 (dd=k); Tb 64, 23 (dd N). 49 pUC 32213, vso. 1–2 has r-Tnw-sp gmm N. Urk I 215, 14 has r-Tnw iw(=i) and Urk IV 1805, 8 tnwt xaa=k. 50 Similar examples after mi: Tomb of Merefnebef pl. 16/32, 3 (wnn N); CT IV 398/99b and Tb 57, 2 (sbb=sn/N); CT VI 54l (mAA=sn); pUC 32157 h/v right 7 (irr N); Sin R 65 (mAA N); Sin B 225 (mAA N); Man and Ba 137 (iw N); Man and Ba 141 (Abb N); Peas B1 273 (iw N); Admonitions 2, 8 (irr N); pPushkin 167, fragment 5, 2 (rmm N); Urk IV 362, 16 (xaa=f); Urk IV 437, 17 (wnn N); Urk IV 687, 13 (xdd N); Urk IV 736, 5 (irr N); Urk IV 1653, 14 (xaa N); Urk IV 1673, 6 (wnn=sn); Urk IV 1919, 12 (wnn=f, followed by mi wn=f on line 13); Tb 1, 24 has mA=Tn, Nu pl. 12, 6 mAA=Tn; Tb 3/Nu pl. 38, 3 (Haa=sn); Tb 72/Nu pl. 20, 14 (wnn=f); Tb 112, 4/Nu pl. 53, 4 (mAA=f). See also n.51 below. Buhen I pl. 62, 15 and Urk IV 809, 10 have “foreigners come bowing to his majesty… mi iri Tsmw like dogs do”, with a presumably erroneous non-geminating form; Urk IV 321, 11 has correctly mi irr Tsmw.
240
chapter six
(275) In the statutes of his funerary cult, Nika-ankh says: in igr msw(=i) ipn wab n Hwt-Hr nbt r-int mi irr(=i) Ds(=i) It is these my children who perform priestly service for Hathor, mistress of Ra-Inet, like I used to do myself. (Urk I 26, 14–15) (276) The writer salutes his addressee in the name of all divinities: iri=sn n=k rnpt HH m aws mi mrr bAk im May they provide you with a million years in l.h.p., according (Meketra letter 4–5) as yours truly wishes.51 (277) Hapdjefa instructs priests on how to deal with certain tapers: dd.t(w) wat im n Hm-kA=i xft prr.t(w) Hr stt tkA im=s(n) n nTr m hrw 5 Hr rnpt grH n wpt-rnpt One of them is to be given to my kA-priest when one goes lighting the lamp for the god with it on the five epagomenal (Siut I 297) days and the night of the New Year’s Eve.52 (278) An inundation-record: r n Hap(y) n rnpt 3 xr Hm n nsw-bity sxm-ra xw-tAwy di anx Dt xft wnn sDAwty-bity mr-mSa rn=i-snb Hr Ts mnnw sxm-xa-kAwra mAa-xrw The extent of the inundation of year 3 under the majesty of the dual king Sekhemra Khutawy, given eternal life, when the royal seal-bearer and general Reniseneb was commanding the fortress (named) ‘Khakaura-True-of-Voice-is-Mighty’53 (Felsinschriften no. 509)
51 Examples of the phrase mi mrr bAk im abound in Middle Kingdom letters and are too numerous to be listed here. Outside epistolary contexts, the expression mi mrr=f/N “just as he/N desires/loves” occurs e.g. in Hassan Giza II, fig 219; Urk VII 4, 8; White Chapel 182; Urk IV 253, 9; Urk IV 278, 10; Urk IV 280, 1; Urk IV 290, 13; Urk IV 304, 4; Urk IV 377, 9; Urk IV 561, 17; Urk IV 567, 17; Urk IV 571, 8 and 12; Urk IV 579, 1; Urk IV 1560, 2; Urk IV 1861, 13; Urk IV 1909, 13; Red Chapel 81 south; Red Chapel 92; Red Chapel 301; Red Chapel 327; Red Chapel 349; Red Chapel 352; Red Chapel 355; Red Chapel 448; Red Chapel 490; Red Chapel 493; Red Chapel 601 and Red Chapel 715). 52 Further examples after xft are CT VI 269k (irr=i); Tb 148, 22 (mAA=f, see example (60) above); Urk IV 1412, 7 (irr=i); Urk IV 1919, 9 (xaa=f); Urk IV 1947, 2 (xaa=f). Urk IV 1023, 5 and 1158, 17 have iy=f. pUC 32197, 2.2 seems to show iw=k; Urk IV 1597, 17 has iw=sn. 53 Similarly Felsinschriften no. 504.
modality in complement clauses after prepositions
241
(279) An additional note to a medical description on how to diagnose a broken skull by the movement of the fractured bone:
xpr nf (A) n nhdhd npApA Xr Dbaw=k Dr ngg Ais n Dnnt=f This quivering and shuddering takes place under your fingers since the cerebral tissue of his skull is broken out.54 (pEdwin Smith 2, 21–22)
Of the non-geminating/-doubling sDm=f forms, examples with past reference are widely attested:55 (280) Idi says he took revenge at his dead father’s enemies: mA.n(=i) nb sfA sw srx n=f nb im=sn m wn=f 56 m pr=f iw sxr.n(=i) sn mi-qd=sn Anyone I saw who had hated him or laid accusations against him among them when he was in his house—I overthrew them (Kom el-Koffar A, 9) all.57
54 Further examples with Dr: Urk I 119, 11 (mAA=Tn); Urk I 218, 2 (Dr wnn mr n=Tn zb(=i) Hr=Tn m Xrt-nTr, with wnn apparently followed by an adjectival sentence: “since it is desirable for you that I protect (?) you in the necropolis”; cf. example (201) above); Urk I 233, 17 (wnn=sn); CT III 158a (mAA=sn/m Dr mAA=sn). CT VII 429d has iw=k. See also Anthes 1969, 12–13 55 Besides prepositions exemplified below, for n, see Urk I 84, 3 (wn N); Urk I 272, 12 (mr N); Kamose stela x+15 (Sw=s); Urk IV 1592, 7 (xa=f); for Hr, see Urk IV 1754, 8 (iri=k) and Tb 17, 34 (wn=s). In the comparative sense ‘more (etc.) than… is/was’, the preposition r is almost invariably followed by the auxiliary wn/wnn; see examples (403), (407) and n.5 in chapter 9 below. Urk I 21, 14 r iri it(=i) may be an exception to this. For r in the sense ‘until’ + sDm=f instead of the sDmt=f-form, see e.g. Florence 6365, 6– 7; pRamesseum III, B5 (both r iw N). 56 The transcript of the publication omits the n in wn, clearly visible in the photograph of the original. 57 Further similar examples after m are CGC 20539, 2. side, 9 (xa=f); Cairo Linen, 2 (wn(=i); Kaw Bowl inside, 2 (in=k); Kaw Bowl outside, 2 (m in N; see example (396) below); CT II 39e (wn=f); CT IV 242/43a (nSn=s; variants m-xt nSn/nSnt=s or m-xt hA=f); CT VII 382c (in=sn; variants int=sn or bare in.n=f); CT VII 487c (nSn N); Louvre C14, 9 (pr=f); Berlin 1204, 5 (iri=k); Tomb of Amenemhat pl. 39, middle band (wn=f); Urk IV 432, 16 (pr=f); Urk IV 897, 13 (wn=k); Urk IV 1409, 3 (wn=f); Hermann 1940, 31*, 14 (pr=f); Hermann 1940, 47*, 10 (xa=f); Tb 17, 16–17 (wn=sn); Tb 17, 34/Nu pl. 4, 38 (wn=s); Tb 17, 31/Nu pl. 4, 35 (wd=f); Tb 78/Nu pl. 39, 16 (wn=sn). The dedication formula m iri N “as N made/made by N” also appears to contain a past non-geminating sDm=f after the ‘m of predication’. See GEG p. 125 n.10 and § 454.4 for examples and also e.g. Urk IV 182, 3; Urk IV 840, 12; Urk IV 1682, 13 and Urk IV 1696, 9. The verb in this formula is almost always iri ‘do, make’, but e.g. in Urk IV 1455, 18 and 1456, 2 one reads m in n=k sxtw “as brought to you by the fields”, in Urk IV 1456, 1 m
242
chapter six
(281) Sinuhe extols the virtues of Senwosret I: iw=f m nsw iT.n=f m swHt iw Hr=f (r=s)58 Dr ms.tw=f He is a king who seized already in the egg. He has been intent (Sin B 68–69) on it ever since he was born.59 (282) The sailor tells the captain to take heed of his words: mA wi r-sA sAH=i tA r-sA mA=i dpt.n=i See me now after I reached land, after I reflected upon what (Sh.S. 179–81) I experienced.60 (283) Khonsuemwase pays homage to the king: DwA=i n Hm=f (n) nt[t Hm]=f mn xaw mi kA-mwt=f mi Hs=f wi m iAt[=i ntt w]i im=s Thus I give praise to his majesty, because his majesty is enduring and appeared in glory like Kamutef, as he favoured me with (Helck 1975 no. 57, 3–4) my office in which I am.61 (284) The writer seeks support to his appeal for the office of his dead father by citing the latter:
iw grt Dd.n n=i pA[y]=i it xft wn=f mr Moreover, my father said to me when he was ill… (a quote (pUC 32055, x+14) follows).62 in n=k Hap(y) “as brought to you by the inundation”. Sometimes too, the preposition m appears to be replaced by mi ‘like’; see n.61 below. 58 Emendation after the R-version. 59 Further examples after Dr: Hassan, Giza II, fig. 219 (wd(=i); CT II, 2c (fA=i); CT VI 347f (wd=Tn); Louvre C202 (pr=f ); Urk IV 157, 7 (wn N); Urk IV 390, 7 (wn N); Tb 17, 45 (hA=f ). Tb 93, 6–7 has the sequence Dr (n)Drr=i (variant Pb: nDr.tw=i) Dr it.tw=i r iAbt Dr iri.tw Hb sbiw im Dr iri.tw Sat nbt im=i Dwt “(Inflammation will occur in the eye of Atum) since I have been taken, since I have been seized to east, since the feast of the rebels has been celebrated there and since I have been subjected to evil harm”. See also Anthes 1969, 11 and Zonhoven 1996, 620 n.25. 60 Further examples after r-sA: Hatnub 8, 3 (pr(=i), see example (394) below); Siut I 298 (iri=f; see example (7) above); Merikara C IV, 5 (iri.tw=ø). 61 Further examples after mi: Ptahh 593 (iri=f ); Kamose Stela x+34 (di=f ; the sign D 36 serves as a ‘determinative’ to D 37 and the word is not to be read dd, but simply di); BM 65340, 7 (di=k, see example (399) below); Urk IV 547, 10 (wn=i); Urk IV 1073, 8 (wn N); Urk IV 1246, 12 (wn N); Urk IV 1676, 15 (pr=n); Urk IV 1729, 13 (mr=f ); Urk IV 1814, 16 (wn=k); Urk IV 1919, 13 (wn=f ); Urk IV 1926, 15 (Hs=sn); Sinai 182, 8 (iri N); Tb 1, 25 (Hms=Tn); Tb 134, 13 (iri=sn). Certain examples with mi appear to be variants of the dedication formula m iri N “as N made/as made by N”; (see n.57 above) for example Urk IV 1269, 8 and Urk IV 1495, 1. 62 A further example after xft is Urk IV 1216, 6 (Ha=f ).
modality in complement clauses after prepositions
243
It is to be noted that these examples differ in whether or not the clauses describe well-delineated situations envisaged as occurring at some particular time or place, i.e. whether they occupy a definite locus in ‘reality’. Some languages mark such distinctions modally; for example, in Spanish the opposing ends of this ‘definiteness’-spectrum are differentiated in temporal ‘when’-clauses:63 (lvi)
Me saludará cuando llega (IND)/llegue (SUB) “She’ll greet me when/whenever she may arrive(s)”
The bare geminating sDm=f complements in examples (272) and (273) are semantically very similar to the subjunctive instance in (lvi) by not referring to any particular occurrence or even series of occurrences located in linear time. However, unlike Spanish, Earlier Egyptian does not indicate such fine-grained distinctions formally; e.g. in (277) the speaker clearly has very definite occasions in his mind, but the grammar remains the same as in examples (272)–(273). This is also the case with the past wnn in (278), which refers to one continuous stretch of time in verifiable ‘history’. As for the non-geminating sDm=f and its functional counterparts, the past situations in examples (280)–(284) do not seem to display the hallmark properties of irrealis but appear to be viewed simply in retrospect and portrayed as unambiguously completed, both of which are properties most commonly associated with modal realis.64 Also regarding the evaluation of discourse-pragmatic relevance, it is difficult to say whether e.g. the causal situation Dr ngg N “since N is broken out” in example (279) carries a lower or higher informationand assertion-value than e.g. the ante-posed ir Dr mrr=T “since you love” in (265), and in any case in both the grammatical form of the subordinate verb is the same. This holds also for past instances such as the m wn=f m pr=f “when he was in his house” in (280) above versus e.g. such past ante-posed circumstantials as the following:65
63
Givón 2001 vol.1, 324. Givón 1994, 270. For the issue of ‘completion’, see 9.2 below. 65 See also Hatnub 22, 1 and 22, 6 (both ir m wn=i); CT II 344b (example (406) below). pEbers 41, 16 has ir r-sA iw=f and pBerlin 10025, 9 ir rf r-sA Htp.n=i. 64
244
chapter six
(285) A retrospect to the early days of the future king Amenhotep II: ist ir m wn=f m inpw mr=f ssmwt=f Ha=f im=s Now, when he was a stripling, he loved his horses, rejoicing in (Urk IV 1281, 8–10) them. Also the few examples of tm after prepositions outside final clauses with r or n-mrwt seem to lack discernible traits of non-assertion:66 (286) It is said of the yon gatekeeper ‘He-who-lives-on-maggots’ and the unfortunate dead:
ir xpr m fnTw wnm.xr=f st Dr tm=f rx r n swA Hr=f As for him who ends up amidst maggots, he has to eat them since he does not know the spell for passing over him (the (CT VII 438c–d) gatekeeper). To solve this problem, one could argue that such suspiciously ‘indicative’ preposition complements as those above involve different forms from the clearly non-indicative ones. This solution is obviously excluded in case of the geminating sDm=f and the negations tm/nfrn. However, it has actually been applied to the past non-geminating sDm=f complements—although not to deal with their modal, but temporal characteristics. As seen, the opinio communis in Egyptological linguistics is that non-geminating sDm=f forms conceal one or several ‘prospective’ forms.67 The past instances after prepositions naturally posit a problem for this analysis. Consequently, mention of the latter has either been omitted altogether,68 or the examples have been argued to involve some other form than a ‘prospective’. For Vernus, rather than the ‘prospective sDm=f’, this was the ‘prospective sDmw=f’, which after prepositions was “not bound to convey future meaning”,69 for Schenkel the Old Egyptian ‘perfective indicative’ sDm=f as opposed to
66
The only other example is Urk IV 150, 2 (n tm=f msbb). See 0.2. 68 E.g. Doret (1986, 41) subsumes the use of his ‘subjunctive sDm=f’ after prepositions under the generalisation of it as “used for actions in the future relative to the time of the main clause” (ibid, 39). But when discussing the form wn after the preposition n in Urk I, 84, 3, the author in a rather contradictory fashion labels it a ‘subjunctive’ and translates it past (ibid, 51). 69 Vernus 1990, 30–31. 67
modality in complement clauses after prepositions
245
the “abstrakt-relativisches Futur” sDm=f.70 Yet, diachronic or morphosyntactic evidence does not seem to support these assumptions.71 Loprieno’s apparent suggestion that in the past the “gewöhnliche” sDm=f is used as a counterpart to the ‘prospective sDmw=f’ fares rather better by correctly predicting the virtual non-appearance of past examples with forms showing the ending -w.72 However, it still entails the assumption that in past preposition complements the form employed is different from the one(s) used for future. Nevertheless, in view of the fact that the writing of the sDm=f in the past examples such as (280)–(284) above is the same as in all other -w- and -y-less non-geminating examples after prepositions, it is clear that one is dealing with a single form in all these instances. This apparent semantic diversity of preposition complements has led to doubts as to whether their grammatical organisation is actually based on any ‘meaningful’ principle. For example, in his study of Middle Egyptian modality Vernus speaks of the ‘objective dependency’ of syntactic subordination as the most salient facet in the grammar of complementation generally.73 In preposition complements in particular, 70 Schenkel 2005, 200–01. In Schenkel 1980, 90 the first form is called “das sog. perfektische” sDm=f, in Schenkel 2005, “Das Perfekt” sDm=f. However, for Schenkel the difference is not purely that of time-relation. He suggests that Vernus’s ‘sDmw=f’ may be used after prepositions when ‘eventuality’ is referred to, whereas when a ‘fact’ is in question, the complement pattern is the ‘perfective’ (= ‘indicative’) sDm=f (Schenkel 1992, 379). 71 There are rather few past preposition complements with the ending -w to support Vernus’s hypothesis, (see n.72 below) whereas past examples displaying the morphology of his ‘prospective sDm=f’ (Allen’s ‘subjunctive’) such as int, iwt and mA are rife. The same holds also for the form wn, but for Vernus the past instances involve a separate, yet again morphologically indistinguishable ‘past converter’ (Vernus, ibid 51). See 9 n.35 below for further comments on this issue. Moreover, Allen states the PT evidence for the sDmw=f after prepositions to be inconclusive (1984, § 251). As for Schenkel’s ‘indicative’ sDm=f, this form disappeared after the Old Kingdom (save for the bound negative n sDm=f) and was even then used only of transitive verbs in initial main clauses and with nominal subjects, (Doret 1986, 24–27) none of which holds for preposition complements. The use of the ‘indicative’ sDm=f in Classical Egyptian in general and/or preposition complements in particular is explicitly denied by Kammerzell (1988). 72 Loprieno 1986a, 37 n.36, 53, 53 n.86. For the issue of -w-ending forms and past reference, see 9.1 below. Examples to the contrary are rare and only CT VII 308c Dr irw=k “since you did” and pWestcar 3,2 (example (307) below) m-xt hAw N “after N has descended” can be cited. Past use of forms with the ending -y is similarly most infrequent: Urk IV 1279, 18 has xft TAy=f “when he has taken” and Urk IV 1675, 7 m pry=i “when he went”. 73 Vernus 1990, 16–17 and the diagram on p.19, column 1.
246
chapter six
he argues the ‘less subjective’ sDmw=f to be the form primarily used and omits a mention of the possible modal value(s) of the geminating sDm=f in the same syntactic environment.74 Kammerzell expresses doubt as to whether there existed any particular ‘system’ behind the use of the different sDm=f forms after prepositions at all, and Junge denies this possibility altogether.75 Nevertheless, rather than representing a chaotic non-system of freely varying forms, the grammar of Earlier Egyptian preposition complement clauses is indeed organised in a principled manner and, furthermore, is based on the differentiation between realis and irrealis modality. The most obvious feature suggesting this is the already noted identical formal organisation of preposition- and verb-complements. As seen, instead of the bare forms of sDm=f, the complementisers ntt/wnt may be used to introduce the complement clause after a definable class of verbs. The same elements appear also after a number of prepositions. Here ntt is considerably more common than wnt; the attested combinations with the former are illustrated below:76 (287) Hapdjefa instructs those in charge of carrying out his will: mTn pXr nA n hrw n qnbt nbt nt Hwt-nTr xpr.t(y)=s(y) m-a ntt ntsn iri=sn n=i pA t Hnqt Pass on these (temple-) days to every temple-council that there shall be, because it is them who shall provide me this bread (Siut I 288–89) and beer. (288) As above: pXr grt nA n hrw 3 nw Hwt-nTr n SnDty nb xpr.t(y)=f(y) Hr ntt prr n=f nA n gmHwt Pass on these three temple-days to any future wardrobe-keeper, (Siut I 301) because these tapers revert to him. 74 Vernus 1990, 30; however, e.g. the deontic nuances of the geminating form in initial environments are duly noted (ibid 42–43). 75 Kammerzell 1988, 51; Junge 1978a, 102, 104 and 109; see 0.1.1 above. 76 Examples of most of the combinations abound in all texts, but m-a ntt is rarely attested (e.g. Meir I pl. 5; pUC 32197, 2.9; pEbers 100, 21; pEdwin Smith 5, 22). n ntt seems to have a graphemic variant in ntt in the Kahun material (e.g. pUC 32203, 6; pUC 32205, 11). Of mi ntt there is only one further instance besides example (289), pBerlin 9010, 5.
modality in complement clauses after prepositions
247
(289) The author tries to manoeuvre his addressee into supporting his case against an adversary:
sbAqq.n swt sS=k n sn=k im m wsxt nt Hr mi ntt wnn is sS=k sn=k im m xt wa nfr-n wAh HAty-a pn awA iri.n=f r tA But your excellence’s having cleared yours truly in the Court of Horus is like our being (now) in total agreement, lest this towngovernor manage to brush aside the crime he has committed. (pBerlin 8869, 9–11)
(290) Having endured Sinuhe’s eulogies of Senwosret I, Ammunenshi says:
xr Hm kmt nfr.t(i) (n) ntt s(y) rx.t(i) rd=f Then Egypt must be happy, because it knows how flourishing (Sin B 75–76) he is. (291) Mery bequeaths his office in his will: iw=i Hr rdit pAy=i mty-n-sA n sA=i mry sA intf Ddw n=f iwsnb r mdw-iAw xft ntt wi tn.kw I am (hereby) giving my (office of) controller of a phyle to my son Mery’s son Intef called Yuw-seneb as a ‘staff of old age’, in view of the fact that I have become old. (pUC 32037, rto. 3–5)
(292) Neni begins his letter with the standard formula: swDA-ib pw n nb aws r ntt hAw nb n nb aws aD wDA This is to inform the lord l.h.p. of the fact that all the affairs (pUC 32199, 1–2) of the lord l.h.p. are safe and sound. (293) Khentyemsmyt appeals to visitors to his cenotaph: rmT imA-ibw n maHat=i DAt n srx=i Dr ntt n iri=i DAt xnm.n=i nTr m mAat People, be kind to my monument and light-handed with my memorial, since I did no wrong, but gladdened the god through (BM 574, 18–20) righteousness. wnt is attested but rarely, but there are some examples after the preposition Dr ‘since’ in its causal sense:77 77
See also CT I 141e.
248
chapter six
(294) Kagemni tells of his strong position when faced by accusations by an adversary (?)
[n Dd.]n=f is xt r(=i) Dw n nzw m grg Dr wnt iti rx(w) qd(=i) Against me he could not [say] any bad thing to the king falsely, (Urk I 195, 9–10) since the sovereign knew my character. In one instance Hr + the negative element iwt appear (spelled with an extra t and a superfluous f after sXm):78 (295) Tefibi protests his incorruptibility: n sXm{f}(=i) r nDs Hr iwtt xpr=f r(=i) m sprti in inw I was not impetuous towards a commoner because he did not appear before me as a petitioner who has brought gifts. (Siut III 11)
These combinations have been analysed as consisting of governing prepositions followed by complement clauses with introducing elements,79 but also as clauses ushered in by fixed (non-isomorphic) true conjuncts of the type ‘preposition-ntt connectors’.80 However, various indicators show that the correct syntactic division here is [preposition [ntt + clause]] rather than [preposition-ntt [clause]] and that the complements governed by the prepositions are marked as modally realis and asserted. The ntt following the prepositions above is clearly the same element as that found after verbs, and the same applies to wnt and iwt, which are unlikely to represent parts of bound ‘preposition-wnt’- and ‘preposition-iwt connectors’ in (294) and (295) above. The paradigm following these elements appears to be the same in both instances, although after prepositions the evidence is more abundant than after verbs. Thus, alongside future cleft sentences, second tenses, sentences with subject + stative and the negation n sDm=f exemplified above, after ntt/wnt/iwt following prepositions one finds also the sDm.n=f and the passive sDm=f, participial cleft sentences, nominal and adverbial sentences, (the latter both affirmative and nn-
78 Also in PT 809a–b one finds n iwt itw=k m rmT n iwt mwt=k m rmT it=k smA wr mwt=k Hnwt “because your fathers were not human and because your mother was not a human; your father is the Great Wild Bull and your mother is the Maiden”. 79 Erman 1928, § 532b; GEG § 223; Gilula 1971, 16; Allen 1986b, 25; 2000, 137. 80 Collier 1991a, 29–30; 1999, 54; Loprieno 1995, 100.
modality in complement clauses after prepositions
249
negated), pseudo-verbal sentences with subject + Hr/r + infinitive, existential sentences, (negative only) n sp-negations, n sDm.n=f negations and sentences with elements ante-posed with ir.81 In addition, after prepositions, ntt also occurs with following conditional and adjectival sentences as well as the negations nfr pw and nn sDm=f.82 However, the bare geminating sDm=f is again absent from the paradigm—except when it is part of a larger second tense complex.83 Similarly, none of the examples argued to represent a ‘prospective sDm=f’ after ntt show the endings -w and -y,84 exactly as after verbs. Such instances are also very rare and mostly rather dubious.85 Instead, although examples here are also most infrequent, in expressions of futurity the pseudo-verbal subject r + infinitive appears to be used:86 (296) In an obscure context: in A sA=i irr r rxt-Hw n ntt=f r mt It is hardly my son who can fight against the rxt-Hw, because (CT III 355a–b/S1Ca) he is going to die. Also, the negations tm and nfr-n are never found following ntt after prepositions. In other words, the restrictions pertaining to the paradigm of forms following ntt/wnt/iwt are exactly the same here as after verbs and suggest strongly that they have the same motives. In the conjunct-hypothesis the ‘preposition-ntt connectors’ are seen to ‘convert’ the same syntactically ‘non-specialised’ or ‘unconverted’ forms and constructions that do not substitute for nominal parts of 81
Cf. 2.4 above. Examples of most of these construals are too numerous to be noted here, but mention should be made of the more rare passive sDm=f (e.g. pBerlin 10025, 2; pUC 32202, 7), adjectival sentences with pronominal subjects (pUC 32158, 2.5), nn sDm=f (CT III 47k; Moaalla IIα2) and conditional sentences (pUC 32291; Moaalla IIβ1–2, (with is) as well as the negations n sp, (Rifeh VII, 13) and nfr pw, (pBerlin 10016, 3; pBerlin 10023B, 3). For subject + r + infinitive, see below. 83 See for instance example (288) above. 84 Johnson 1984, 81; Collier 1991a, 29. 85 In Siut I 296–97 one reads Hr ntt pr N which in the identically phrased Siut I 301 (example (288) above) is written Hr ntt prr N. Similarly, Siut I 311 with Hr ntt rdi=sn repeats the identical Siut I 282, where the writing Hr ntt dd=sn appears. pUC 32210, 17– 18 writes r ntt rdi N in what might be future reference and pBerlin 10066, 1 r ntt iri.tw N in what is certainly so. 86 The only other instance attested, Urk IV 656, 3 shows r ntt iw.tw r THn, with extensive Late Egyptian influence. 82
250
chapter six
speech but which can neither be directly subordinated as adjuncts.87 The ‘connectors’ do not govern the ‘unconverted’ construals, for which reason the bare geminating sDm=f, a ‘specialised’ nominal conversion of the verb, is not found after them and instead appears in adjuncts where it occupies a properly ‘nominal’ position, i.e. after prepositions. Yet, just as ntt cannot appear after all verbs, also the ‘connectors’ are formed only of some prepositions or correspond only to some senses of more polysemic elements—and these restrictions demonstrably stem from modality. For example, combinations such as †n-mrwt ntt, †n-ib-n ntt or †n-aAt-n ntt do not occur because clauses with n-mrwt and n-ib-n ‘in order that’ as well as n-aAt-n, which directs the attention to the character of the complement situation, are always irrealis and non-asserted. Similarly r ntt never means ‘so that’, although bare r can, as seen, express this sense. In fact, the combination r ntt is of particular interest here seeing that in the vast majority of examples it is not used to create adjuncts at all, even if this function forms the semantic basis of the conjunct-hypothesis.88 Instead, in the epistolary formula swDA-ib pw n N r ntt “This is to inform N of the fact that” exemplified in example (292) above, it is used precisely to introduce the main communicative content of the correspondence—i.e the assertion—just as in the alternative phrase Dd=i di(=i) rx=k r ntt, “I am writing to inform that”, where it introduces an asserted complement of the verb rx, ‘know’.89 Diachronically, the writing r ntt instead of the bare ntt/wnt in the latter environment appears later than the epistolary swDA-ib pw -formula. This orthographic shift seems to be a direct analogy motivated by the use of r ntt in similar contexts and for the same function.90 Further signals of the asserted status of ntt-introduced preposition complements come from their use alone in various specific contexts such as the ‘question and answer’ sequences found in mortuary texts:
87
Collier 1991a, 29–30; 1999, 54. Cf. GEG § 223; Collier 1991a, 29n.36. 89 See 2.2.1 and n.74 in chapter 2 above as well as GEG § 225; Collier 1991a, 29 n.36. This is in spite of the assertion often consisting of little more than stereotyped clichés in both instances. 90 Thus r ntt is not merely a ‘Doppelpunkt’ (Scharff 1924, 38) but partakes in the modal organisation of the clause by marking it as ‘indicative’; cf. Luft 1984, 107. 88
modality in complement clauses after prepositions
251
(297) Gods try to force the deceased to eat faeces; a dialogue ensues: wnm irk in=sn r=i n wnm=i n=Tn Hr iSst in=sn r=i Hr ntt wi Tb.kw m Tbty nty skr “Eat”, they say to me; “I will not eat for you”. “Why”, they say to me; “Because I am shod with the sandals of Sokar”. (CT III 48e–49a)
(298) As above: Hr iSst ir(f) tm=k wnm Hs swr wsSt n Swt Hr wDa n ntt w(i) iri.kw r xtw Hr wxAt tw wrt nt wsir Hr gs imnt(y) n pt “Why do you not eat excrement and drink urine for the emptiness (?) of Horus and the Outcast”? “Because I was made for the offerings on that great altar of (CT III 202i–j/B1L) Osiris at the western side of heaven”. In the examples above and their like, the main clause proposition [I will not X] is presupposed to the extent that it has been omitted altogether and only the proposition that explains it, the Hr/n ntt-clause, is present. The latter is thus clearly the focal point of information as a whole, i.e. the assertion. It is notable, moreover, that bare Hr/n- etc. -clauses without ntt never occur in similar environments. On the other hand, preposition + ntt clauses are in turn very seldom found as the ‘emphatic’ part in second tenses, a property that has hitherto gone unnoticed and is rather surprising if the function of these constructions is to create syntactically ‘specialised’ adjuncts. Yet, ‘emphasised’ prepositional adjuncts without ntt occasionally consist of pragmatically presupposed, i.e. non-asserted information: (299) Words addressed to Min in a magical incantation: imi n(=i) irty=i mA=i im=sny dd=k n=i irty=i n mAA=i im=sny Give me my eyes that I may see through them. The reason you should give me my eyes is because I see through them. (pTurin 54003, vso. 10–11)
252
chapter six
This does not, of course, prove conclusively that ntt-introduced preposition complements are asserted, but it is suggestive of the reasons why prepositional adjuncts with ntt hardly ever occur as ‘emphatic’ vedettes and in any case adds to the cumulative evidence in favour of the hypothesis of their realis status. Another notable property of preposition + ntt-clauses is that, with the exception of r ntt, they are used to create causal adjuncts. Yet, these mostly do not express ‘real-world’ causality but rather explain epistemic conclusions expressed in the main clause or alternatively speech acts performed.91 For example, in (290) above the n ntt-clause seems to furnish an explanation to the speaker’s conclusion “Egypt must be happy” rather than simply states that the happiness is caused by something. In (287)–(288) the explanatory clauses clearly give reasons for the speaker’s orders and in example (293) for his requesting something. The speech acts thus explicated are most commonly deontic orders and requests as well as promises, but e.g. in the clearly performative (291) above, the making of a declaration is explained. Further cases of such epistemic and ‘speech act’ causals with preposition + ntt are: (300) Senwosret III sums up the basic qualities required when dealing with the Nubians:
qnt pw Ad Xst pw Hm-xt Hm pw mAa Arw Hr tAS=f Dr ntt sDm nHs r xr n r Aggression is bravery; retreat is timidity, and the real coward is one who is driven away from his border, since the Nubian (Berlin 1157, 10–11) listens only to hard talk. (301) The peasant exhorts the high peasant to carry out his duty as an official:
Dd mAat iri mAat Dr ntt wr s(y)92 aA s(y) wAH s(y) Speak righteousness and perform righteousness, since it is great; (Peas B1 351–52) it is mighty; it is enduring!
91 See the illuminating discussion in Sweetser 1990, 76–86 and Rutherford 1970; cf. also Kac 1972 and Thompson & Longacre 1985, 203–04. French employs different conjunctions to differentiate between ‘real’ and ‘speech act/epistemic causals’ as a rule: the former are introduced by parce que, the latter by puisque (Sweetser ibid, 82). 92 Or should wr=s, aA=s and wAH=s be read?
modality in complement clauses after prepositions
253
(302) The writer expresses his opinion why he should be rewarded: kA Hsw bAk im... xft ntt ir pA bit Hnw 1 rdy n bAk im gm.n bAk im swr.n sw pA aAm Thus yours truly should be rewarded… seeing that, as for the 1 hin of honey given to yours truly, yours truly found that the (pUC 32124, fragment ii, 4–8) Asiatic had drunk it. In example (300) a reading in which the situation after Dr ntt ‘causes’ someone’s being a ‘real coward’ is impossible; the cause must refer to the king’s saying what he says. Similarly in (301) the peasant explains his reasons for exhorting something, whereas in (302) the speaker employs the causal clause to substantiate an opinion expressed in the main clause. There is a strong correlation between this sort of linkage and ntt-introduced causals, and the reason for this seems to be the same as before. Explaining one’s speech acts, promises and epistemic reasoning aims at increasing the likelihood of positive response from hearers: belief in case of declarations, promises and conclusions, obedience in case of deontic acts. The possibility for success in this is highest if the explanatory material is presented in committed terms. For example, the injunction “you must not enter, because the animals may be dangerous” where the reason given for the denial of permission is presented as a mere possibility runs a greater risk of being ignored than “you must not enter, because the animals are dangerous” with more uncompromising ‘back-up’ thereof. Similarly, the speaker of a sentence such as “Jack must be at home, because that’s his VW in front of his house” can be more certain that his reasoning will be accepted than someone uttering “Jack must be at home, because that may be his VW in front of his house” where the reason for concluding something about ‘Jack’s’ whereabouts is presented in weaker terms than in the first instance.93 None of these alternatives is, of course, excluded as such, but given the pragmatic function that the explanations are intended to serve the stronger asserted options are clearly preferable. Hence the correlation between ‘speech act’ and epistemic causality and preposition + ntt-clauses: 93 Indeed, in the first case a possible challenge from hearers would most naturally target the acceptability of the claim presented in the subordinate clause rather than the conclusion drawn from it, whereas in the second instance the reasoning would be the most obvious first line of attack.
254
chapter six
the latter represent assertions and communicate strongest speaker commitment. The information presented is also new, although it is of course possible to give a reason to e.g. an epistemic conclusion that consists of pragmatically old information. Yet, ntt is not found in such instances, and contrasts such as the following occur: (303) After telling how he entered his tomb, Intefoqer tells of his expecting eternal life:
ix mA=i wp-wAwt m Hbw=f nbw m nmtwt=f nbt Hr ntt ink mry nb=f Then I shall see Wepwawet at all his festivals and processions, because I was one beloved of his lord. (Leiden AP 7, 5–6) (304/268) Ankhtify reflects upon his mission to Edfu: i[w] in.n w(i) Hr r wTst-Hr n aws r grg=s iri.n(=i) xr wn Hr Hr mrt grg=s Hr in=f w(i) r=s r grg=s Horus brought me to Edfu through l.h.p. to establish it, which I did. Horus must have wanted to establish it because he brought (Moaalla Iα2) me there to establish it. In example (303) the explanation consists of new, in (304/268) of old information. The first is construed with Hr ntt, the latter with a bare Hr + complement clause. In Earlier Egyptian preposition complements the same formal distinction between un-introduced sDm=f complements and ntt/wnt/iwtintroduced clauses is thus made as after verbs. The bare forms and the construals ushered in by extra elements are the same in both instances. From whichever perspective the latter are examined, they consistently turn out to represent propositions that are subject to highest speaker commitment and discourse relevance, i.e. they are clearly asserted. Equally, in every turn they can be contrasted with examples that either lack positive speaker commitment regarding the veracity of the situations described or are low in information value, and where introducing elements are absent as a rule—i.e. that bear the hallmarks of non-assertion and irrealis. But as seen, there are many examples of bare sDm=f complements of prepositions that do not seem to share these latter characteristics. Nevertheless, it is
modality in complement clauses after prepositions
255
suggested here that these ‘indicative’ instances do not represent an anomaly in the modal organisation, but that in Earlier Egyptian all un-introduced preposition complements are irrealis and non-asserted without exception. The key issue here is that the grammatical status of such clauses does not always stem directly from the nature of the situation that the complement verb describes as counterfactual etc., but from the character of the entire complex [preposition + clause]. Adjuncts thus formed are, to borrow the terminology coined by Collier, indeed ‘specialised’, because their semantic-pragmatic function is explicitly marked by the introducing preposition as causal, temporal etc. The same holds also for their syntactic status, which is similarly unambiguously indicated by the preposition. Adjuncts of this sort contrast grammatically with ‘unspecialised’, directly embedded ‘circumstantial’ clauses, whose syntactic and semantic-pragmatic (including modal) status is—courtesy of the unmarked character of the forms and constructions thus used— not specifically indicated.94 ‘Specialised’ or marked circumstantial clauses, on the other hand, are also marked for modality, and in most cases this is irrealis simply because they represent adjuncts. The linkage between main- and subordinate-, and particularly adjunct clauses, and the gestalt-psychological notions of background and foreground has for some time been recognised among (especially cognitive-) linguists.95 Irrealis modality, which besides signalling lower speaker commitment is also a reflection of mitigated relevance of the information communicated can be said to belong to the discourse background. Adjuncts, in turn serve one global role that sets them ‘close’ to modal irrealis. Regardless of their more precise semantic value, in terms of discourse-pragmatics and information structuring on a sentential level, most adjuncts present exactly the sort of supporting or amplifying ‘commentary’ to the main clause, which is frequently characterised as ‘less important’, ‘background setting’ and even ‘presupposed’ material that e.g. in narration “does not 94
However, see below for negated ‘circumstantials’. See e.g. Talmy 1978; 2000, 320–29; Reinhart 1984, particularly 796–97; Tomlin 1985; Givón 1987; Thompson 1987; Croft 2001, 328–35; Cristofaro 2003, 26–27 These notions are often treated as synonymous to ‘figure’ and ‘ground’. Although related, these two pairs of terms should not be equated given the spatial differentiation implied by ‘background’ and ‘foreground’, the latter of which is “necessarily that which is less distant” (Waugh & Monville-Burston 1986, 855; cf. also Cristofaro 2003, 28) See 9.2 below for further discussion. 95
256
chapter six
constitute the assertion of events in the story line.”96 Also experimental evidence, all of which concerns adjuncts, shows that these clauses are systematically comprehended as secondary information against which the main clause is projected and which is, akin to visual background, processed and interpreted less immediately.97 In Earlier Egyptian, whenever an adjunct is of a ‘specialised’ type, i.e. is introduced by a preposition whose function is to define its specific syntactic and semantic-pragmatic character and relation to the co-text, the proposition itself, i.e. the complement clause of the preposition, receives modal marking as irrealis or realis. Irrealis is, once again, signalled by the use of bare sDm=f forms and is the ‘default case’ due to the generally ‘circumstantial’, i.e. less-than-optimally relevant background character of the information provided and its status as something less worthy of full attention and concern of the speech participants. However, the information can be also presented as belonging to the discourse-pragmatic foreground and asserted, in which case it is marked as modally realis by ntt (or, less often, wnt and iwt). The use of the latter strategy is, as after verbs, a subjective speaker choice, but, once again, not a universally available option. For example, final ‘so-that’-clauses after n-mrwt, n-ib-n and r cannot be realis-marked inasmuch as they are irrealis also by virtue of their more specific semantic value as descriptions of speaker assumptions and expectations. Similarly, the sole function of temporal adjuncts is to provide ‘out of sequence’, ‘circumstantial’ information about the prior or simultaneous conditions which furnish the ‘frame’, ‘setting’ or background to the main clause situation.98 Consequently, just as there is not †n-mrwt ntt, neither is there a combination such as †m ntt, and e.g. xft ntt never has a temporal reading. The Earlier Egyptian typological cum modal organisation of affirmative adjunct clauses may thus be summarised as follows:
96 Hopper 1981, 215–16; see also Talmy 1978a; 2000, 320–29; Reinhart 1984, 796; Lambrecht 1994, 67–69, 125–26; Croft 2001, 330–35. 97 E.g. Townsend & Bever 1977, 7–15. 98 In Egyptology this point is most explicitly made by Ritter (1995, 69–70, 81–82).
modality in complement clauses after prepositions specialised/introduced Realis Preposition + ntt Irrealis Preposition + bare sDm=f modality marked syntactic status marked
257
non-specialised/un-introduced Bare verb-forms used ‘circumstantially’
modality unmarked syntactic status unmarked
As can be seen, this system is basically the same tripartite one as in complement clauses, which are similarly divisible into marked realis and irrealis versus modally unmarked types. 99 These terms refer again to the same discourse functions as before: realis is equivalent to strong commitment and high relevance, irrealis signals diminished acceptance and diminished relevance. The concepts of ‘markedness’ and ‘unmarkedness’ pertain simultaneously to syntactic and semanticpragmatic function. Marked (introduced) adjuncts are marked for their modality as well as their structural and semantic link with the discourse. Unmarked (un-introduced) adjuncts are unmarked modally and regarding their exact discourse- and syntactic status. However, as seen, in bare negative circumstantial clauses the same difference as in preposition-introduced affirmative construals is made.100 Adjuncts that are not semantically irrealis take the form of simple n-negations, whereas those that are, namely final- and ‘virtual conditional’ adjuncts, are negated by the irrealis negations tm and nfr-n. What conditions the choice between the ‘specialised’ (i.e. preposition-introduced) and ‘unspecialised’ adjuncts is another matter. Suffice it to say that the choices seem to correlate, at least partly, with particular textual genres and registers.101 This analysis is less audacious than would appear at first sight. Although very little research has been conducted on this matter, many languages 99 See the diagram in 2.4, end. There are certain differences, however. With the exception of the modally unmarked active suffix-conjugation forms, which can assume an irrealis role, in complementation the other ‘circumstantially’ used forms function only as realis and require realis-marking by ntt/wnt. 100 See 5.3 above. 101 For example, ‘unspecialised’ adjuncts predominate in literary works, whose linguistic idiom has a semi-artificial flavour whereas e.g. the contracts of Hapdjefa, a ‘legal’ text, are notable of their extensive use of preposition-introduced ‘circumstantials’.
258
chapter six
display what can only be described as a ‘tendency’ towards similar generalisation of irrealis into adjunction, resulting e.g. in various ‘inexplicable’ uses thereof in ‘circumstantial’ clauses seemingly illqualified as non-assertions. For example, in Latin one occasionally encounters the subjunctive in adjuncts that describe realised situations not subject to speaker doubt:102 (lvii)
Pugnatum… incerto Marte, donec proelium nox dirimeret (SUB)
“The fight went on indecisively until night broke it off” In Fula, (Atlantic Niger-Congo) subjunctive is systematically used in temporal adjuncts introduced by conjuncts ‘before’ and ‘until’, again irrespective of speaker attitudes and regardless whether the situation described is realised or not:103 (lviii)
(SUB) “They continue farming until they finish” ooko e-njottoo, (SUB) o- yami gorko on, o-wii “Before they arrived, he asked the man, and said…”
In Mangarayi, (Australian isolate) irrealis appears to be almost a general marker of subordinate status:104 (lix)
(IRR) “He buried the dog when it died/ that died”
There seems to be no apparent reason for irrealis here, save for the general background-status of the information, which the clauses convey. In Bemba, verbs receive explicit marking if they are asserted. This marking is absent in such clearly presuppositional environments as e.g. cleft sentences and wh-questions, but it is neither found in adverbial
102 103 104
Palmer 2000, 142. Arnott 1970, 310–11. Palmer 2000, 143.
modality in complement clauses after prepositions
259
clauses.105 Also in Biblical Hebrew explanatory and final adjuncts display increasing diachronic tendency of being introduced by the element asher, which is otherwise used in relative clauses and with concessive adjuncts.106 This development appears to reflect a new organisation of all these clauses within the same category of irrealis due to their background- or attitudinal character. In view of these cross-linguistic phenomena, it is not overly bold to suggest that Earlier Egyptian might simply have been more systematic in its grammatical recognition of the overall ‘background-ness’ and lessthan-optimal relevance of adjunct-information.107 However, it is not the case that all adjuncts are treated as irrealis in this language: some are specifically indicated as realis, and there is also a parallel method of creating them that stands wholly outside the system of marking for modality, namely the use of unmarked ‘circumstantial’ clauses without introducing prepositions. Furthermore, organising grammar on basis of discourse-pragmatic relevance is, after all, information structuring, for which, as seen, all languages employ modality. Its extension also to adjunct-clauses in Earlier Egyptian is actually quite in keeping with its general tendency to grammaticalise information structuring to a considerably greater extent than is usual in languages overall. As another example of this, one may consider the marking of negative scope. Here most languages tolerate a notably high degree of ambiguity, whereas Earlier Egyptian employs a set of specific negations n/nn, tm and n…is which mark the scope explicitly.108 Such thoroughness in signalling the exact informative contour is highly unusual crosslinguistically, but is a systemic feature of Earlier Egyptian negations. Similarly, the modal organisation of adjuncts—in which complement clauses play an integral part—may be cross-linguistically unusual as regards its extent, but as an innerägyptisches phenomenon it is far from being unique.
105 See Givón 1982, 139. Interestingly, it is also not used in negative or relative constructions, nor, most decisively, in environments where, according to Givón, the ‘scope of assertion’ does not include the verb; see 10.3 below. 106 Givón 1991, 281–86; 296–97. 107 In fact, the hypothesis presented here is considerably less radical than many recent linguistic analyses of subordination that treat all subordinate clauses as ‘nonasserted’, regardless of their formal and functional differences (so Croft 2001, 323; Cristofaro 2003, 28, 30). 108 See Loprieno 1991b.
260
chapter six 6.4 Clauses with m-xt
Before moving on to consider Earlier Egyptian verb- and preposition complement patterns that supplement the system outlined thus far and ones that are wholly independent of it, some remarks must be made of the element m-xt. m-xt is traditionally classified as a compound preposition that in conjunct-use carries the temporal meaning ‘after’ (although ‘when’ is sometimes just as feasible, or even better, rendering109). Yet, it has various syntactic and syntagmatic properties that differ markedly from those of other prepositions, suggesting that it has been, or is being developed, into a ‘true’ conjunct. m-xt is commonly found followed by geminating/doubling and non-geminating/non-doubling sDm=f forms:110 (305) From an instruction on preparing an eye-ointment: iri m iwSS rdi Sw=f XAw im=f m-xt Sww=f Make into a mixture; allow drying, and separate part of it after (pEbers 56, 20–21) it has dried up. (306) A contract by Hapdjefa obliges a wab-priest to offer oblations: ...m-xt pr=f Hr irt xt m Hwt-nTr ra nb …after he has set forth from performing the daily ritual in the (Siut I 308–09) temple.111 In at least one instance, a non-geminating form appears with the ending -w:112
109
Cf. GEG § 178 (p. 133). For examples with the sDm.n=f, see 7.1 below. Further examples with the geminating sDm=f: CT V 333p (prr=f); pRamesseum IV, C16 (hAA=f); pEbers 68, 3 (qbb=s); pEbers 78, 4 (ir m-xt irr=k, variant pHearst 11, 18 has ir m-xt iri.xr=k); pEbers 91, 7 (irr=k); pHearst 11, 2 (mrr=k, variant pEbers 88, 14 mr=k); Tb 86, 11–12 (prr=f, some variants pr=f); Tb 112, 10/Nu pl. 53, 9 (wnn N). 111 Further examples with a non-geminating sDm=f: CT IV 242/43a (nSn=f, variants nSnt=f); CT IV 243b/BH1Br (hA=f); Heqaib 9, 5 (wn=f); Pushkin Museum 1695, vso. 2 (wn=f); Literary Fragments, pl. 17, 5 (hA N); pEbers 97, 3 (bS=s); Tb 154, 4 (sy=f); Tb 144/Nu pl. 76, 48 (iri.tw). pWestcar 11, 26 has iw=f, MMA 35.7.55, 10 rdi=i and Urk IV 220, 2 iw=f (variant Urk IV 1714, 13 iy=f). Siut I 298 has rdi N; CT IV 236/37b varies between m-xt itt=f, m-xt it=f, m itt=f and m it=f. 112 In CT VII 31k a strange spelling m-xt Xnmw=f, but given that Xnm is a strong root, this instance is dubious (cf. Schenkel 2000b, 90). 110
modality in complement clauses after prepositions
261
(307) The magician instructs his butler on what to do with a magical waxcrocodile and an adulterer:
i[r m]-xt hAw nDs r p[A] S mi nt-a=f nt ra nb kA=k xAa=k [pA m]sH [Hr mw] r-sA=f After the ‘little man’ has gone down to the pool according to his daily habit, then you throw this crocodile after him into the (pWestcar 3, 2–4) water. But, most surprisingly, m-xt is also found before stative and passive sDm=f clauses:113 (308) Isis says to the hands of Horus, equating them with his eyes: iw=Tny r wDaty Hr m-xt=Tny gm.twny You will be the two severed parts of Horus after you have been (CT II 350a) found.114 (309) Amenhotep describes the conclusion of a royal construction-work: xr m-xt snfrw kAt tn aHa.n wAH.n Hm=f Htp-nTr m mA m Xrthrw nt ra nb After these works had been perfected, then his majesty re(Urk IV 1795, 18–19) dedicated daily offerings.115 In this m-xt differs from all other prepositions.116 These unusual combinatory properties seem to reflect a gradual change in the grammatical profile of this element. Very few examples of m-xt as a conjunct introducing finite clauses are forthcoming from Old Egyptian, and none at all are followed by the stative or the
113
Edel (EAG § 906cc) analyses m-xt as originally a final adverb of the main clause rather than a subordinating preposition, or, alternatively when followed by stative, as governing only the head noun of the latter. 114 Further examples with the stative are CGC 20001, 6; BM 1671, 6–7; Moaalla IV24; pRamesseum I, A11; pWestcar 3, 10; pWestcar 3, 17; pWestcar 7, 11; pWestcar 12, 8–9; pLythgoe rto. 7–8; Urk IV 1308, 2; pEbers 53, 10–11. 115 Further examples with the passive sDm=f are Urk IV 978, 15 and Urk IV 1282, 13. 116 In Tb 167, 6 one reads m-xt wA sy “after it (fem.) had fallen into rage”, with m-xt being followed by what appears to be an adjectival sentence wA sy “it (fem.) was (one) fallen”. However, the correctness of this may be doubted inasmuch as immediately before (line 5) there appears wn.in=sy nSny.ti “Then it was enraged”, where the sy clearly stands for the suffix =s.
262
chapter six
passive sDm=f.117 Examples of the former appear during the First Intermediate Period and the latter is found only in post-classical sources. m-xt seems to have been in an extremely slow cline of development from a preposition into a ‘real’ conjunct or a ‘connector’ with everloosening grammatical relationship with the following expression and, in particular, loss of its governing force. Just as with the expression xpr.n, its ‘mixed’ paradigm appears to bear testimony to the incompleteness of this grammaticalisation-process.118 In other words, m-xt seems to represent a ‘semi-conjunct’ that occupies a position mid-way between the functional categories of prepositions and true ‘adverbialising’ clausal/sentential connectors. It appears to have undertaken a further step away from the former in one particular syntagmatic environment. m-xt-clauses are often ante-posed as sentence-initial by the particles ir/xr, as in examples (307) and (309) above.119 Mostly these uses differ little from ‘ordinary’ m-xt-‘circumstantials’ after main clauses and, as seen, other prepositions can also be similarly ante-posed. However, the latter show no differences to their usual paradigms, whereas in later texts the set of construals following ir/xr m-xt becomes ever more unlike that of prepositions. For instance, in early XVIII dynasty medical papyri there are examples of even the ‘contingent’ patterns sDm.xr=f, sDm.in=f and xr=f sDm=f:120 (310) A prescription on treating a swelling with various ingredients concludes:
ir m-xt iri.in=k n=f mrHt 1 Afterwards, you then prepare for it 1 (measure of) oil. (pEbers 56, 3)
(311) A prescription on treating a burn with various sort of bandaging concludes:
ir m-xt xr=k wt=k sw m ftt n dbyt Afterwards, you then bandage it with reeds of the dbyt-plant. (pEbers 70, 16–17) 117
EAG §§ 906cc, 1032b. See 4.3 above. 119 Sometimes ir/xr is absent—see e.g. Urk IV 836, 6; pEbers 66, 3; pEbers 66, 4; pEbers 92, 9—but in these instances m-xt is not followed by a finite verbal clause. 120 Instances of ir m-xt sDm.xr=f are found throughout pEbers, pHearst and pEdwin Smith, whereas the other two construals above seem to be limited to the examples quoted. 118
modality in complement clauses after prepositions
263
Here m-xt together with ir apparently functions as a connecting expression ‘afterwards’.121 m-xt most certainly does not govern the following clause/sentence and the combination ir m-xt can even be preceded by initial particles such as kA ‘then’.122 Nevertheless, the ‘de-categorisation’ of m-xt clearly never reached its conclusion. With or without ir/xr, m-xt is found before both the geminating and nongeminating sDm=f until the onset of Late Egyptian after which it all but disappeared.123 Its demise coincides with the complete overhaul of the Earlier Egyptian system of forming adjunct clauses with prepositions used as conjuncts.124
121
Cf. WGMT 244 n.3, where the combination is characterised as an ‘adverbial’. See the badly damaged pUC 32215, 4. 123 1erný & Groll (1993, chapter 35) analyse ‘m-xt-stp.f’ as a single fossilised unit. In WGMT 158 n.3 examples of geminating sDm=f after ir m-xt are argued to represent second tenses. 124 See 10.2 below. 122
264
chapter seven
CHAPTER SEVEN
EARLIER EGYPTIAN SUPPLEMENTARY PATTERNS OF COMPLEMENTATION AFTER VERBS AND PREPOSITIONS The system of expressing realis and irrealis modality in Earlier Egyptian affirmative complement clauses after governing verbs and prepositions is founded, first and foremost, on the variation between clauses introduced by the operators ntt/wnt and bare sDm=f forms. Yet, there are also various other forms and constructions that occur in the same syntactic positions, but whose status vis-à-vis the organisation outlined thus far is either complementary or quite extraneous. The former class comprises clauses with one particular verb form employed on its own as a functional counterpart to bare sDm=f forms in past environments. The latter group consists of clauses subordinated by means of additional elements that, unlike ntt/wnt, do not form part of the modal system of complementation. Yet, they shed light on its diachronic position in Ancient Egyptian more generally. 7.1 The sDm.n=f The sDm.n=f, one of the few suffix conjugation forms demonstrably formed of almost every verb in the Earlier Egyptian lexicon, is relatively common in complement clauses after verbs and prepositions both with and without an introducing ntt and wnt. Abundant examples of sDm. n=f object complements ushered in by these elements have been cited in previous chapters. No corresponding subject clauses seem to be attested, but given their general rarity, this is perhaps not surprising. When ntt/wnt are absent, there are differences in frequency of the use of sDm.n=f between verb- and preposition complements as well as between different verbs and prepositions. After verbs, there again do not appear to be any examples of subject clauses with the bare sDm.n=f—unless passives in .t(i)/.t(w)/.tw be considered such.1 Also 1
See 4.1 above.
egyptian supplementary patterns of complementation 265 as an object, un-introduced sDm.n=f most frequently appears after the exceptional verb gmi, ‘find, discover’:2 (312) The deceased says to gods: iri.n=i gmt.n=i iri.n=Tn I have done what I found that you had done. (CT VII 232m) (313) A letter describes the fate of some honey entrusted to the writer: gm.n bAk im swr.n sw pA aAm Yours truly discovered that the Asiatic had drunk it. (pUC 32124, fragment ii, 7–8)
(314) A report of a dispute concerning an official appointment notes: wHm rw Hna Dd-xnw r=s m xA TAty gm.n.tw iri.n mr nwt TAty iy imt-pr m pA HAty-a n nxbw n sA=f The statements and complaints concerning it were revised in the vizier’s office. It was found that the town-governor and vizier Iy had made a will concerning the (office of) town-governorship (Stèle Juridique 24) of El-Kab in benefit of his son. Otherwise, examples of un-introduced sDm.n=f object clauses of verbs are rare:3 (315) The deceased proclaims his free access to heaven: nwA.n4 irw pt sDm.n irw tA iwa.n=i wAwt=s The keepers of the sky have seen and the keepers of the earth (CT III, 332e–g) have heard that I have inherited its roads. (316) The author reminds her addressee of a past threat by a third party: i(w)=k rx.t(i) Dd.n=f n(=i) ink smy im=T Hna Xrdw=T You know he said to me: “I will personally report against you (Louvre Bowl 17–18) and your children”. In contrast, bare sDm.n=f-complements of prepositions are more
2 3 4
See 2.4 above as well as examples (330), (335) and (336) below. See also example (331) after mAA below. So the damaged variant S2C; the better-preserved S1C has nwAw.
266
chapter seven
common, particularly in early XVIII dynasty inscriptions,5 but the most frequent combination by a wide margin is with mi ‘like, according as’:6 (317) The annals of Thutmosis III tell how the king secured the obedience of vanquished Syrian towns:
in Hm=f msw=sn m sqr-anxw r dmi n wAst r mH Sna n (i)t=f imn m ipt-swt m wDyt=f tpt nt nxt mi wD.n (i)t=f imn smS sw r wAwt nfr(w)t His majesty brought away their children as living captives to the town of Thebes to fill the production facilities of his father Amun at Karnak, from his first campaign of victory, according as his father Amun, who follows him to the goodly roads, ordered. (Urk IV 781, 1–4)
(318) The king reveals his grandiose building-plan: iry=i kAt m Hwt-aA n it(=i) tmw di=f wsx=f mi rdi.n=f iT=i I will undertake work in the great temple for my father Atum; he will cause himself to be rich, even as he allowed me to (Berlin Leather roll 1, 15) conquer. With other prepositions the sDm.n=f is less ubiquitous, although examples—sometimes no more than one—occur also after the elements wpw-Hr ‘except’, m in the sense ‘when’, xft in the sense ‘in accordance with/according as’, r in the sense ‘until’, r-DAwt ‘in return of’, Hr ‘because’ and Dr ‘since’; for example:7 5
Sometimes in these texts one suspects that the apparent sDm.n=f stands for a neuter relative sDm(t).n=f. The examples cited below are ones where this possibility seems the remotest. 6 Excluding the uncertain spellings [mAn] (mA.n or mAn?), further examples after mi are Helck 1975, no. 26, 2 (iri.n N); pHearst 11, 13 (Sni.n N); Sinai 244, 6 (rdi.n=s); Urk IV 224, 1 (wD.n N); Urk IV 624, 5 (read mi rdi.n=i n=k); Urk IV 795, 14 (wD.n N); Urk IV 1008, 6 (wD.n N); Urk IV 1349, 18 (wD.n(=i); Urk IV 1529, 10 (sDm.n=i); Urk IV 1652, 8 (rdi.n=f); Urk IV 1656, 4 (rdi.n=i); Urk IV 1689, 9 (rdi{t}.n=f). 7 For a morphologically rather peculiar example after r-nfr-n ‘through the goodness of’, see example (329) below. For further examples after xft, see (332) below as well as Berlin Leather Roll 2, 13 (SA.n=k) and Urk IV 1675, 10 (wD.n N). For the sole instance after r-DAwt, see example (333) below. After Hr the only other example besides (323) is Urk IV 1845, 20 dm rn=i Hr iri.n=i “Say aloud my name on account of how I acted”. After Dr one finds the damaged CT VII 353a/B4Bo, (xr.n=s, other variants simply xr=s/=f) as well as MFA 04.2059, vso. 3 (mr.n=k); Cairo JE 52000, col. x+3, line
egyptian supplementary patterns of complementation 267 (319) Sebekemhat says he made his career entirely in the service of his nomarch:
iri.n=i sSm(=i) m-a=f m-Xnw pr=f n rdi.t(w) mA=i Sw n st nbt wpw-Hr hAb.n=f wi r Hwt-nbw I carried out my business beside him in his house and was not made to see the shadow of another place, except when he sent (Hatnub 22, 12–14) me to Hatnub. (320) From a gynaecological handbook: SsAw s[t Hr] mn warty=sy rdi.xr=k [r]=s stpw n Hatw txb m antw [ir...]st [...] nDm m iri.n=s xt nbt snb pw Examination of a woman aching in her calves; you apply to it strips of fine linen soaked in myrrh. If… (a long gap, probably describing what the patient must do) sweet/pleasant when she has done everything—this means health. (pUC 32057, 2.7–9) (321) Thutmosis III is characterised: Hr nbw hr-Hr Hw HqAw xAswt pHw sw xft wD.n n=f it=f ra nxtw r tA nb dmD The Golden Horus ‘Lenient-of-Face’ who smites foreign chieftains who attacked him, according as his father Ra decreed for him (Urk IV 593, 10–11) victories over all lands in one. (322) It is said that the mother of the nomarch acted as a vice-regent: ...r xpr.n sA=s m nxt-a (Siut V 29) …until her son had grown up. (323) The deceased says to a divinity: ink rx r=f ii.n=i m [...] Hr rx.n=i I am one who knows his spell; I have come from […] because (CT III 316h–i) I have gained knowledge. x+6 (iri.n N) and Urk IV 1543, 15 (rx.n=f). A broken example after r-sA ‘after’ is pBerlin 10025, 9 ir rf r-sA Htp.n=i. Of wpw-Hr there are no further examples. Besides pUC 32057, 2.16 m iri.n=s, most examples after m ‘when’ are rather uncertain. In CT II 40h/B2L one reads m xpr.n=i “when I came into being”, but this seems to be a corruption for m xpr rn=i “when my name comes/came into being” of the other variants. In CT VI 173r m sr.n=i appears in an obscure context. Tb 98/Nu pl. 25, 6 has m swA. n=i.
268
chapter seven
(324) The ground at the gates of the underworld addresses the deceased: Dd n=i rn=i in sAtw Dr xnd.n=k Hr=i “Tell me my name”, says the ground, “—since you trod on me”. (Tb 99, 34/Nu pl. 64, 30)
Examples after the ‘semi-conjunct’ m-xt are more common:8 (325) The duel between Sinuhe and the ‘Goliath of Retenu’ approaches its climax:
aHa.n ikm=f mnb=f Hpt=f n nsywt xr m-xt spr.n=i xaw=f rdi. n=i swA Hr=i aHAw=f sp n iwtt wa Hr-Xn wa Then his shield, axe and his armful of spears fell (to me), after I had escaped his weapons and made them pass by me, with his arrows having met with nothing, one after one.9 (Sin B 134–36)
In ntt/wnt-introduced complements with the sDm.n=f the question often arises whether the form might be part of a larger second tense construction. Morphological evidence is of little aid here as examples of intransitive vom or sDm.n.tw=f-passives after ntt/wnt are extremely rare.10 Consequently, one must rely primarily on semantic interpretation. After verbs, the sDm.n=f following ntt/wnt usually seems to have no focalising or setting role.11 Conversely, instances of preposition complements similarly introduced but where the form certainly does not function as a second tense are less abundant. Sometimes, though, the semantics of the situation described and/or the absence of adjuncts clearly subject to ‘emphasis’ exclude such an interpretation:12
8 Further examples (with and without ante-position with ir/xr): Qurneh pl. 3, 2, 3 (m-xt aHA.n(=i); pUC 32158, 2.6 (ir m-xt rdi.n=f); pBerlin 10066, 4 (ir m-xt gm.n=f); CGC 20541, 10 (xr m-xt sAq.n=f); pEbers 91, 16–17 (ir m-xt rx.n=k); pEbers 96, 21 (ir m-xt rdi.n=f); Urk IV 3, 2 (xr m-xt grg.n=i); Urk IV 5, 4 (xr m-xt smA.n N); Urk IV 1842, 5 (m-xt Dd.n N). 9 Sinuhe’s adversary appears to have furiously thrown all his close combat weaponry against him after having wasted his arrows. 10 See example (103) and n.139 in chapter 2 above. ntt-introduced examples of sDm. n=f intransitive VOM with the element is are more common; see 7.2 below. 11 See examples (18), (19), (21), (22), (24), (50), (52), (55) and (61) above. 12 See also Collier 1991a, 30 and his example (23).
egyptian supplementary patterns of complementation 269 (326) Merishenet writes to his superior: swDA-ib pw n nb aws r ntt sDm.n bAk im mdt nt sSw pn iny n bAk im r-Dd... This is to inform the lord l.h.p. that yours truly has taken note of the matter of this document brought to yours truly, quote: (pUC 32212, 1–3) “…” (a quote follows) However, the virtual impossibility of distinguishing formally between second tenses and ‘non-second tenses’ here is part of the wider question of whether or not the ‘nominal’ versus ‘non-nominal’ functions of the sDm.n=f should actually be allocated to different forms. Ever since Polotsky’s classification of sDm=f forms as syntactically nominal, adverbial or adjectival, a corresponding split has been sought in the sDm.n=f, but no secure morphological evidence has been found to support the assumption that it be divisible into more than one form.13 Polotsky himself remained rather vague on this question, contenting himself with the often-made case that at least the initial passive sDm. n.tw=f and the sDm.n=f of intransitive vom are ‘emphatic’ or ‘nominal’, and that syntax (paradigmatic similarity with nouns, complementary distribution with the past passive sDm=f and the stative, absence of iw in initial environments) was here more important than morphology.14 Yet, the term form occurs everywhere in ST discussions, and the later work by Polotsky (and his followers) seems to be characterised by belief that the ‘nominal’ and ‘non-nominal’ uses of the sDm.n=f indeed hide the required “two different grammatical forms, indistinguishable in writing”.15 But insofar as the issue is writing, as it unavoidably is with Ancient Egyptian, then if some two writings are indistinguishable, they 13
See already Polotsky 1944 § 29 and Gardiner’s (1947, 100) severe criticism there-
of. 14 Polotsky 1957, 110; 1965, 17; 1969, 475; 1976, 2.6, 3.2. As noted also by Hannig, (1991, 156) this somewhat evasive mode of discussion characterises most (particularly earlier) studies among this school of thought (e.g. Frandsen 1975, 35–43; Junge 1978a, 17–18, 105; Silverman 1985, 270 n.7; Satzinger 1993, 198). This seems to reflect unease with the apparent violation of the fundamental ST equation of form and function. Explicit reference to the ‘nominal/emphatic’ sDm.n=f as a morphological entity is seldom explicitly made (see e.g. Johnson 1980, 69 n.5) but denials of the ‘emphatic’ form are equally rare in works influenced by Polotsky (e.g. De Cenival 1977, 26 n.12; Schenkel 1975, 55–56) as are statements to the effect that the sDm.n=f is one form with more than one function (e.g. Doret 1979a, 16; 1986, 67–68). 15 See Polotsky 1984, 116–18 and e.g. Greig 1990, 314; Hannig 1991, 151–52, 274; Allen 2000, 383. The quote is from Allen 1991, 6.
270
chapter seven
must be treated as the same form. Even the sDm.n=f of intransitive verbs, particular of motion, is hardly a morphologically separate form, but merely a sDm.n=f of such verbs. The same applies also to the sDm. n.tw=f, of which there are some examples after prepositions: (327) The author says he provided shelter for refugees in times of trouble:
iw iri.n=i pr=i m rwt n ii nb snD hrw n HAayt wn=i m mnat Hr Aty n ii nb inDw r ssnb.n.t(w)=f I made my house a door for everyone who came frightened on the day of the strife. I was a nurse and caretaker for everyone who came afflicted, until he was healed. (Hatnub 16, 9–10) (328) Singers describe Amenhotep III with solar imagery at his Sedfestival:
mH.n=k tAw m nfrw=k mi pt st.ti m tHnt mi ms.n.tw=k m itn m pt You have filled the lands with your beauty, like heaven when it glitters as faience, even as you have been born as the sun-disc in the heaven. (Urk IV 1861, 20–1862, 2) Rather than any ‘nominal sDm.n=f’ form, this is clearly just a .twpassivised sDm.n=f. Both this and the sDm.n=f of intransitive verbs may have different functions from e.g. an active sDm.n=f of a transitive verb after iw, but this does not make them separate forms—provided that the latter is understood correctly as a morphological rather than a syntactic term. Yet, the issue of separating the ‘predicative’ (‘non-nominal’) and ‘non-predicative’ (relative and ‘nominal’) sDm.n=f formally has recently been discussed anew by Schenkel, according to whom these classes are indeed morphologically distinct.16 The decisive feature is the morphology of 2ae gem. roots, which sometimes show doubling of the last consonant. According to Schenkel, such writings do not occur after
16 Schenkel 2006. For Schenkel the ‘non-predicative’ sDm.n=f doubles as a ‘concrete’ and ‘abstract’ relative form, the latter being tantamount to the form in its nonrelative ‘nominal’ uses such as complements and second tenses (see n.25 below).
egyptian supplementary patterns of complementation 271 iw (or isT or wn) but do occur as second tenses.17 Spellings such as mAA. n=f are thus a signal of the ‘nominal’ status and there are, accordingly, (at least) two separate forms of sDm.n=f. Leaving aside the question of how is it possible to accurately differentiate between bare sDm.n=f’s functioning as second tenses and continuative or adjunct clauses, this hypothesis faces various difficulties. In general, it is strange that in the sDm.n=f the morphological signal for ‘nominal’ function should be doubling of 2ae gem. rather than gemination of ult. inf. roots as in the sDm=f.18 Further, also the infinitive and the imperative show occasional variation between doubling and non-doubling writing of 2ae gem. roots (at least mAA) that is certainly not due to any syntactic difference.19 Schenkel also argues that after the negative n, the sDm. n=f of doubling roots does not show the said morphological feature.20 In fact, this argument does not hold beyond the corpus of CT, outside of which there are instances of doubling 2ae gem. roots after n and even the auxiliary iw.21 But even following the assumption of the lack of doubling after the negative n, Schenkel finds it difficult to correlate the form used there with either of his two sDm.n=f forms. As he notes, intransitive vom appear regularly after n and the passive of n sDm.n=f is n sDm.n.tw=f—i.e. the construction appears to involve a form that does not occur after iw either and seems to be identical with Polotsky’s original ‘nominal’ sDm.n=f.22 Consequently, Schenkel speculates whether this might be yet another, third form of sDm.n=f.23 Unfortunately, instances of sDm.n=f exist whose morpho-syntactic characteristics do not fit even these assumed three forms. For instance, there are examples showing gemination of ult. inf. roots that occur equally 17
Schenkel 2006, 46, 47. Schenkel is clearly aware of this problem as well (2006, 59). 19 Matthias Müller, PC; see GEG §§ 299, 336. 20 Schenkel 2006, 48. 21 For examples after n, see for instance Literary Fragments, pl. 2, 2.7 (n mAA.n=f ); Merikara E 33 (n tkk.n N); Merikara E 68 (n qbb.n N); Urk IV 367, 12 (n ann.n=i). Tb 136B, 10 and 11 as well as Literary Fragments pl. 8, 2.2 have have iw mAA.n=i. 22 Polotsky 1957, 116. For discussion and research-historical remarks on the issue of n sDm.n.tw=f and the ‘nominal’ sDm.n=f, see Zonhoven 1993–94, 40–43, 49–53. Zonhoven’s solution to the problem is rather similar to Schenkel’s. Since the form after n is not semantically ‘emphatic’ nor syntactically adverbial, it must again be a yet third form of sDm.n=f, this time dubbed ‘indicative’ (ibid 50). 23 Also in Schenkel 1990, 190 it is noted that the form in the negation n sDm.n=f cannot be analysed as ‘nominal’ without ease and the pattern should therefore be considered as non-isomorphic, with perhaps some ‘secondary’ form. 18
272
chapter seven
after the negative n, after iw and once also after the preposition nnfr-n ‘through the goodness of’:24 (329) Henu says his departure as a blessed dead was due to the following reason:
...n-nfr-n irr.n(=i) tp tA …through the goodness of how I acted upon earth. (CGC 20011, a4)
These writings could be argued to constitute evidence for a fourth sDm.n=f form. Further, as noted, there are at least three examples of iw mAA.n=i, which, given the presence of the auxiliary, surely cannot involve the ‘nominal’ sDm.n=f. On basis of these instances, one might venture to suggest that the form after iw is neither a unit, but perhaps hides two separate ‘predicative’ forms, raising the number of sDm. n=f’s to an unprecedented five. Now, of course, some of the instances just noted are likely to be erroneous—surely no-one would wish to take such rare writings as evidence of entire paradigms. But here exactly lies the core of the problem. There is no non-arbitrary point at which one may stop generalising rare writings into full paradigms pace ST, and this holds with the sDm.n=f as much as it does with the sDm=f. Schenkel’s ‘nominal’ sDm.n=f, like its predecessors, not only appears to lack a clear syntactic or syntagmatic function. There are also various other more or less plausible ways in which the sDm.n=f may be split into further forms. In view of these considerations it seems best to consider the sDm.n=f a unitary form with multiple syntactic and semantic-pragmatic functions and to treat the exceptional writings such as mAA.n=f or irr.n=f as just that—exceptions.25 This has various repercussions to the analysis of complement clauses with this form. If the same form can be used with and without ntt/ wnt, with the sDm.n=f the said elements certainly do not behave as ‘nominalisers’ or ‘converters’ any more than they do elsewhere. Instead, their function must consequently be semantic-pragmatic. On basis of 24 The examples after iw are from the same stela CGC 20011, a4 and b1 (both iw irr.n(=i), pRamesseum II, vso. 5 has n gAA.n N “N does not lack”, with the ult. -w root gAw; pWestcar 5, 19 has n Xnn.n=t. 25 Given Schenkel’s equation of his ‘nominal’ sDm.n=f with the relative sDm.n=f, (the ‘concrete’ relative form—see n.16) this of course entails the disappearance of also the latter as a separate form. See 10. 3 for some discussion of this issue.
egyptian supplementary patterns of complementation 273 the numerous examples quoted thus far, there is little question of what this function might be, and of the nature of sDm.n=f-complements introduced by these elements. Unless part of a larger second tense complex, after ntt/wnt the form clearly corresponds to a past assertion both in complements of verbs, and, insofar as examples exist, of prepositions. In un-introduced clauses of both sorts, the semanticpragmatic function of the sDm.n=f is less apparent. However, it is argued here that these examples are expressive of irrealis modality. There are various instances where this property of the bare complement sDm.n=f is more clearly seen than in the examples cited thus far. In the following passage the un-introduced form occurs as object complements of conditional main verbs, where, as seen, the subordinate clause describes a mere open possibility:26 (330) A medical handbook describes a procedure for diagnosing an intestinal blockage:
ir xA=k s Hr mn r-ib=f rdi.in=k drt=k Hr=f ir gm=k Ts.n=f Hr gs=f imnty Dd.xr=k iw TA.n ø iri.n ø dp If you examine a man with stomach pains, you should put your hand on him. If you find that it (i.e. excrement) has solidified on his right side, you say consequently: “(It) has accumulated (pEbers 40, 18–20) and created a lump”. In the next instance, partly cited as example (99) above, the sDm.n=f is used as a variant for the immutable xpr=f in an environment where the complement consists of co(n)textually wholly shared information:27 (331/99) The deceased says that aspects of his rebirth are a mystery even to the divine:
mA.n wi nw xpr.k(w) n rx=f bw xpr.n=i im n mA=f S1C, M3C, M5C, M18C, M20C, M.Ann: B1C, B2L, M4C:
xpr=i m Hr=f xpr.n=i m Hr=f
26 See 2.1.1 above. Further examples are pEbers 39, 13 (DbA.n=f DA.n=f) and pEdwin Smith 15, 10 (sqd.n=sn), but neither occurs directly after the conditional ir ‘if’. 27 Cf. Malaise & Winand 1999 § 574. For the other versions varying between xprw=i and xpr rn=i, see chapter 2 n.109 above. B3C and B1Bo have n mA.n=f xpr=i “he could not see that I had come into existence”.
274
chapter seven Nu has seen me after I came into existence, but he does not know where I came to be because he did not see with his own eyes that I had come into existence. (CT I 334/335a–c)
Typologically, the irrealis expressed by the bare sDm.n=f overlaps with that associated with both the geminating and the non-geminating sDm=f of mutable verbs. In many instances the semantics of the unintroduced complement sDm.n=f parallels the by now familiar individual characteristics of each of the said forms. For example, often the word ‘how’ seems highly appropriate in translation, which, as seen, is typical with geminating sDm=f complements and reflects the non-optimal discourse relevance assigned to the situation predicative nexus:28 (332) Khakheperraseneb says his words can be trusted: Dd.n=i nn xft mA.n=i I have said this in accordance to how I saw it. (Khakheperraseneb rto. 6)
(333) Ramose addresses those bearing his remembrance: iri=tn n=i m ib mrr=i r-DAw(t) iri.n=i tp tA Act for me as I wish in return of how I acted upon earth. (Urk IV 1777, 17)
The functional overlap with the bare non-geminating sDm=f forms is apparent in the use of the sDm.n=f for counterfactual complements, as in the following example: (334) The deceased, as Horus, says to his father Osiris: di(=i) DA=k ptrty DA=k S nmi=k wAD-wr Tb.t(i) mi iri.n=k tp tA I will let you cross the two Sky-Windows and the lake. You shall traverse the sea in sandals, as if you were on earth.29 (CT I 267a–b/B10Cc)
28
See 2.2.2, 4.2 and 6.2 above. Certainly not “as you did upon earth”, which would mean that when alive, the person addressed practised walking on water! 29
egyptian supplementary patterns of complementation 275 The sDm.n=f has a comparable function sometimes also after the negated verb gmi, where it may describe situations that did not occur: (335) Harkhuf stresses the uniqueness of his achievements abroad: n zp gmy(=i) iri.n s(y) smr mr aw nb pr r iAm tp-aw I never found that any royal friend or overseer of barbarians who had previously gone to Yam would have done such. (Urk I 125, 10–11)
(336) The deceased refers to his survival of death: n gm tr pn iri.n=f r=i That time did not accomplish its work against me.30 (Tb 42, 22/Nu pl. 17, 21)
Also outside complementation, the sDm.n=f is used similarly for unfulfilled wishes and conditionals after the elements HA ‘would that’ and ir ‘if’: (337) The sage laments his own inaction at the time of crisis: HA rf iri.n=i xrw=i m tAy At Would that I had made my voice (heard) at that moment. (Admonitions 6, 5)
(338) The sage admits to god that people tend to beseech the divine only when faced with the worst imaginable disaster:
ir snm.n.tw=n n gm.n=i tw If we had (only) been saddened, I would not have found (Admonitions 12, 6) you.31 Nevertheless, not all conditionals with the sDm.n=f carry this meaning,32 and the same holds also for clauses after gmi, as can be seen from 30
Literally “…did not find that it could have acted against me”. For the interpretation of this difficult passage, see Parkinson 1997, 197 n.97. For another similar example, see chapter 2 n.79 above. In Amenemhat VIIc, pSallier II has ir sSp.n=i As=i and B3019 ir sSp=i Ast “If I had taken (my) haste”. According to Lopez, (1963, 32 n.II, 3a) the lost pMillingen had ir sSp=i A st, “If I had received it”, which seems to involve a sDm=f + the particle A (so Allen 2000, 230). 32 See WGMT § 242, 1. 31
276
chapter seven
examples (312)–(314) above. All these disparate uses, along with the appearance of the sDm.n=f after ntt/wnt (as well as e.g. after iw) in what can hardly be anything but assertions would seem to indicate that the sDm.n=f itself is unmarked as to its pragmatic function and on its own merely represents a neutral reference to past states of affairs. In complementation, then, together with the unmarked sDm=f forms, the sDm.n=f constitutes a second exception to the rule posited that no form or construction without a specific modal profile may be used as a complement without being preceded by ntt/wnt.33 This is why the fourth principle of construing complement clauses in Earlier Egyptian was formulated as follows: 4) …unmarked active suffix-conjugation forms, i.e. ones that do not show the formal markers of irrealis… may also be used alone in non-asserted complements. Accordingly, like the unmarked sDm=f without the morphological indicators of irrealis mood, (gemination, the endings -w/-y) also the sDm.n=f could be used alone in non-asserted complements. Similarly, and again like the unmarked forms of sDm=f, it could be introduced by the realis operators ntt/wnt to assign it an assertive value. The factor facilitating the bi-functional behaviour of the sDm.n=f was the non-availability of modally marked forms in this verbal formation. As seen, most verbal root classes did not have the irrealis marked forms of sDm=f with gemination and the endings -w/-y (the latter of which were also in process of being lost altogether) and for these the unmarked forms were used as functional counterparts. The sDm.n=f, by contrast, was a morphological unity, and there existed no separate irrealis form(s) at all. Thus, the same unmarked form was used for asserted and non-asserted complements, and these functions were differentiated by the presence or absence of the elements ntt and wnt respectively. There remains the question of what might have motivated the use of the un-introduced sDm.n=f in complementation instead of bare sDm=f forms. After prepositions the sDm.n=f seems to occur primarily in such
33
See 2.4 above.
egyptian supplementary patterns of complementation 277 instances where the governing preposition does not have, or at least does not express, temporal sense.34 As seen, bare sDm=f forms could be used equally for past, ‘present’ and future and the correct sense can usually be recovered from the co(n)text and/or the meaning of the preposition. The same holds also for complements of verbs. But it is notable that also the rare instances of bare verb complement sDm. n=f all involve matrix verbs that do not impose restrictions on the time-reference of their complement clauses, which could potentially be located anywhere in time. It seems thus that the use of the sDm. n=f was particularly called for in instances where the context and the semantics of the governing expression could not be relied upon to clearly reveal the intended time reference. Put another way, the sDm.n=f appears to have been favoured when it was deemed necessary to show in no unambiguous terms that the time reference of the complement situation was past rather than aorist or future.35 The sDm.n=f could do this because there is one clear property that sets it apart from both the modally marked and unmarked constructions. Unlike the latter, the sDm.n=f is specifically marked for past tense. This latter feature also answers the question how the sDm.n=f could be used as a modal counterpart for both distal and proximal irrealis sDm=f forms. It will be seen below that the irrealis functions of the bare mutable sDm=f and particularly the geminating and -w/y-ending sDm=f used in complementation are intertwined with their temporal properties, the latter of which differ radically from those pertaining to the sDm.n=f.36 More particularly, the sDm.n=f does not display the fundamental aspectual characteristics inherent to the temporal-modal functional profiles of the former. As a consequence, the sDm.n=f is not bound to express any one of the two types of Earlier Egyptian irrealis, but can instead cover both functions. However, the interrelation of tempus and modality in all these forms can be fully examined only after the remaining patterns of complementation in Earlier Egyptian have been discussed. 34
Wolfgang Schenkel, PC. Vernus (1987, 168–74) and Allen (2000, 312) similarly postulate an analogous diachronic process of sDm.n=f replacing the sDmt=f after prepositions motivated by strive towards more precise time-expression. In later texts there is also a tendency to replace the non-geminating sDm=f by the sDm.n=f after the genitival n(y) (e.g. Urk IV 807, 8, 12; Urk IV 812, 9, 13) which might relate to the same phenomenon. 36 Chapter 9 below. 35
278
chapter seven 7.2 The Particle is
This clitic particle is occasionally found introducing complement clauses in Earlier Egyptian, but its distribution and functions are simultaneously more restricted and more general than that of ntt and wnt. In the CT and particularly in the PT, is is relatively common in complement clauses, but elsewhere its use in these constructions is marginal. After verbs is is hardly ever found beyond object clauses of rx ‘know’, and even there it mostly combines with nominal, adjectival and cleft sentences:37 (339) Osiris is asked to announce the divine status of the deceased: di=k rx imnt nfrt sA=k is pw May you let the Beautiful West know that he is your son. (CT I 104/105d–e/B1P)
(340) The deceased declares that the gods know his status: is(T) rx.n=sn nnk is tm sp-sn They know that All Entirety belongs to me. (CT VI 348d–e) (341) The deceased extols the triumph of Horus to gods: wsr w(y) Hr di(=i) rx=sn wr is nrw=f spd is Hnwt=f r stX How mighty is Horus! I will let them know that the dread of him is great and that his horn is sharp against Seth. (CT IV 84h–i)
(342) The deceased says to a group of gods concerning one of their kind: di=i rx=f ink is ms Tn I will let him know that it was me who fashioned you. (CT VII 492h)
37 This is what motivated Gilula’s (1971, 16; 1972, 59; 1978, 47) formulation of is as being used to subordinate “non-verbal sentences without iw” as complements. Further examples after rx: Urk I 222, 12 (mrrt nTr is); CT II 24c (ink/in is N pn qmA/iri Tn); CT II 214b (nb=Tn is pw); CT II 219a (iwa wsir is pw). There are also some instances after Dd: besides examples (344), (351) and (353), see CT I 155g (mr is mrt). CT III 181b–182a has Dd=f sA=f is pw iwa=f is pw, apparently “He says that he is his son and that he is his heir”. The variants M22C and pGard. IIa continue this in CT IV 182b–c with Hr is pw ink is pw, which, given the deixis in the initial Dd=f “he says” seems unlikely to be further complements “that he is Horus and that this is me”.
egyptian supplementary patterns of complementation 279 In the following unique example, the complement is a conditional sentence: (343) Niankhsekhmet explains the significance of the king’s blessing of him:
sk sw rx Hna Smsw r-Dr=f ir is pry xt nb m r n Hm=f xpr ø Hr-aw He and his entire entourage knew that should something issue from the mouth of his majesty, (it) would be realised at once. (Urk I 39, 12–14)
The most interesting use of is after verbs and apparently occasionally also after prepositions, is its employment in complements consisting of entire second tenses:38 (344) It is said of the couriers and messengers of the deceased: Dd=sn n ra Dsr rmn m iAbt pr.n=k is m nTr sb.n=k is m nTr hA.n=k is m nTr They say to Ra, who holds arm aloft in the east, that you have gone, passed, and descended as a god. (CT I 278c–f) (345) Pepiankh says he overcame his accusers: [i]r x[t] nb Dd[d]t r(=i) m-bAH srw pr.n(=i) Hr=s m Htp sk ø xr Hr Ddw Dr bAk(=i) im m-bAH srw Dr Dd=sn is r(=i) m sDw As for anything said against me before officials, I came forth cleared and (it) fell upon those who had brought the charges— since I was innocent before the officials, and since they spoke against me only through ill-will. (Urk I 223, 12–16)
38 As noted in n.139 of chapter 2 above, Allen (1979, 8) and Silverman (1985, 272; 1986b, 317) argue that the presence of is is actually required for second tenses to be able to appear as object complements (cf. also Allen 1984 § 238). However, as seen in connection with the discussion on preposition complements introduced by ntt, (6.3 above) this does not seem to hold. Silverman’s further claim that is allows the embedded second tense to maintain its initial status (1986b, 317) is most incongruous; how can a patently subordinate complement proposition be ‘initial’ or a subordinator function simultaneously as a marker of ‘initiality’?
280
chapter seven
By this token is could be characterised as a complementiser, were it not that the same element is also found as an indicator of pragmatic focus in the restricted-scope negation n…is/n-is and in affirmative sentences.39 In the PT is occurs also as a post-positive element expressing similes,40 and overall in these texts it appears to have been a wholly general marker of dependency, marking adjunction and complementation of all kinds of construals.41 Certain later examples seem to suggest that it retained its ability to be used to subordinate nominal sentences and second tenses as adjuncts:42 (346) Queen Hatshepsut claims to have obeyed Amun’s will at all time: iw Hmt=i rx.ti nTrr=f iri.n=i is (s)t Xr wD=f ntf sSm wi My majesty knows just how divine he is, because I did it under (Urk IV 363, 6–8) his command; it was he who guided me. Further, rather than alone, in complement clauses is occurs often in connection with the elements ntt/wnt, which is almost a rule with prepositions:43 (347) The passage quoted in example (340) continues: isT rx.n=sn ntt N tn is HqA iaHw They know that this N is (= I am) the ruler of the farmlands. (CT VI 348f)
(348) The writer asks his dead addressee for help: idr n(=i) grt nqmt nbt ntt r Hmt(=i) i(w)=k rx.t(i) ntt irr(=i) is DArw im Destroy for me every ill directed against my wife; you know that it is this where I have a need.44 (Haskell Museum 13945, 6–7) 39 Borghouts 1986, 66–67; Loprieno 1991b, 218–22; 1995, 154–55. Cf. also Depuydt 1993b, 16–17 where is-constructions are, however, viewed as syntactic main clauses. 40 See EAG § 828. 41 Loprieno 1995, 154–55. 42 Cf. Allen 1979, 9, but see Allen 1986b, 17. 43 Further examples abound in the CT. The last two examples below are also fine instances of subordinated second tenses. 44 Wente (1990, 213) translates “my wife, whom you know I have need of”, but there is no resumptive im=s, and ‘my wife’ is not an antecedent sufficiently indefinite for i(w)=k rx.t(i) to function as a ‘virtual relative’ clause.
egyptian supplementary patterns of complementation 281 (349) An ‘ascension-text’ says to opposing spirits: imi sn Tswy bqsw bAw xnt pt n ntt pA.n=sn is r pt m bikw iw N pn Hr DnHw=sn Do not kiss the vertebra of souls who are foremost to heaven— because when they flew to heaven as vultures, this N was on (CT VI 283e–g) their feathers. Syntactically and syntagmatically, is thus has a character markedly different from ntt/wnt, and the same seems to hold also for its pragmatic profile. Loprieno speaks of its use to “transform ‘categorical’ into ‘thetic’ sentence”, which, among other things, explains its use in “explanatory clauses representing the object of verbs of perception such as Dd ‘to say’, sDm ‘to hear’, rx ‘to know’ or the like”.45 In view of the fact that is is often followed by cleft sentences and adjunct focus second tenses, this is unlikely to be correct.46 Loprieno suggests further that is represents a generic marker of focus as well as dependency, the latter covering both subordination and other more ‘semantic’ dependency-relations.47 This seems to capture the essence of the use of is rather well, but with the modal organisation of assertion/non-assertion outlined in the present work this element appears to have little to do. Not only are its syntactic functions apparently much wider than those of the realis operators ntt and wnt. In addition, is-complements can neither be divided exclusively as assertions or non-assertions. In the examples cited above the complement situations with is do not appear to be subject to speaker doubt or non-acceptance, nor does their discourse relevance seem to be mitigated. On the other hand, in various other complements with this element the speaker’s commitment to the state of affairs described is clearly lacking. In the following instance, which constitutes part of example (100) above, is occurs in a complement whose propositional content is subject to speaker doubt and which is under an interrogative scope:
45
Loprieno 1995, 153; cf. Loprieno 1991a, 214. See the discussion in 2.1.1 above and 8.2 below. 47 Loprieno 1995, 154, 165; cf. also Borghouts 1986, 66. Loprieno uses the label ‘hypotactic’ for the latter, which, however, in the present work is used in a rather different sense—see chapter 1 above. 46
282
chapter seven
(350) Isis has claimed that her unborn child is ‘Osiris’ seed’ and divine; Ra-Atum replies:
ihi i.n ra-tm sA ib=T Hmt i.rx Tn rf mi iSst nTr is pw nb iwa psDt... “Well”, said Ra-Atum; “Be prudent, woman! How do you know if he is a god, lord and heir of the Ennead?” (CT II 215b–216a)
In the following example is occurs after a negative imperative of Dd in a sentence forbidding the addressee from saying the complement proposition:48 (351) The ritualist disclaims responsibility of his address to the deceased: imi=k Dd ink is Dd n=k nw in gb Dd n=k nw Hna wsir Do not say that it is me who says this to you. It is Geb and Osiris who say this to you. (CT I 302d–f/B10C) Here the speaker shows clear non-commitment to the complement proposition, whose falsity he goes on to substantiate in the following sentence. These characteristics of is-complements, the overall distribution of this element, and the fact that is was apparently eclipsed by ntt/wnt in complements where the latter can be used suggest that it represents an echo from a diachronically earlier grammatical system in which complementation was not organised around modality.49 The redundant appearance of is with ntt is also a tell-tale sign of this,
48 The change of deixis shows that one is not dealing here with direct speech. The variants T9C and Sq3C have negative imi=k Dd n ink is Dd n=k nw, which fits the context rather poorly. T2C has imi=k Dd ink Dd n=k nw, seemingly omitting is. 49 Gilula (1972, 59; 1978, 47; 1986, 161) similarly maintains that is belongs to an older language-stratum. Here a note should be made of the peculiar passage CT VII 475i–j, which Loprieno (1995, 155) and Allen (1979, 15 n.25) cite as an example of an iw-sentence subordinated by is without any structural change whatsoever: Dd.n=sn iw=i is rx.kw sn m sSmw=sn “They said that I know them in their behaviour”. However, the context of this unique instance is obscure (it is not clear e.g. who ‘they’ are) and thus there are also various other equally possible readings. For instance, one may follow Faulkner (1973–78, vol. 3, 170) and translate “I speak to them, for I know them by their images” (i.e. Dd(=i) followed by an adjunct). Dd may also be read as an imperative “say to them: ‘I know them by their images’”.
egyptian supplementary patterns of complementation 283 as is the occurrence of these elements after verbs in different variants of the same passage, examples of which abound in the CT:50 (352) It is said of the deified deceased and gods: CT VI 348r–s
isT rx.n=sn N tn is HD-tA CT VI 348l
isT rx.n=sn ntt N tn HD-tA CT VI 349m
rx ra ntt N tn is HD-tA They/Ra knows that this N is the Dawn-god. (353) A messenger of a demon is told: i sxm is Dd=k n hAb Tw Ax is r n N pn r ds=f ntt Ax is r n wsir N r ds=f O mighty one, go and tell the one who sent you that the spell of this N is more powerful than his (the sender’s) knife. B6C, B2Bo, B4Bo, B1Y, S2C: B4C:
(CT V 49b–c)
After the stage of Earlier Egyptian represented by the CT, is all but disappeared from complementation. However, it experienced a brief revival in this role in the royal and private monumental inscriptions of the early XVIII dynasty. In these texts it appears to have been employed as a conscious archaism with a particularly elevated flavour, and, it seems, not always correctly:51 (354) Thutmosis III explains his motives for increasing the daily offerings in Karnak:
rx.n=i is nHH pw wAst Dt pw imn I knew that Thebes is an eternal place and Amun is forever. (Urk IV 164, 5–6)
50
Cf. Gilula 1971, 17. One would have expected nHH is pw wAst and Hr ntt ink is iwa. A further XVIII dynasty example is Urk IV 260, 6 (sAt nTr is pw). There are also contemporary examples of is as a post-position in similes, (e.g. Urk IV 367, 7) a use long since obsolete. 51
284
chapter seven
(355) Amenhotep Son of Hapu addresses his audience: iriw n=i iri.tw n=tn Hr ntt is ink iwa grg nwt=f dr twA=s m st nbt Act for me and one will act for you, because I was an heir who established his town and drove away its evil from every place. (Urk IV 1824, 10–11)
This development coincided with another diachronic shift in the fortunes of the Earlier Egyptian system of complementation to be considered next. 7.3 The Element r-Dd Unlike is, the element r-Dd is a latecomer in Ancient Egyptian. A precursor of the Coptic element je used to introduce complement, final and result clauses,52 the roots of r-Dd lie in an Earlier Egyptian expression employed to introduce direct quotes.53 This was later grammaticalised as a complementiser following a widespread crosslinguistic tendency.54 Sporadic examples of its use to introduce object complements of the verbs rx ‘know’ and sDm ‘hear’ in particular occur in early XVIII dynasty sources:55 (356) Thutmosis III explains why he has embellished the temple of Amun: [iw]=i rx.kw r-Dd xnw[=f] pw (Urk IV 736, 16) I know that it is his resting-place. (357) The background to Amenhotep II’s assault on the Syrian city of Ugarit is narrated:
ist sDm.n Hm=f r-Dd nhy[.n] nA.n stw nty m dmi n ikA(r)ty Hr ngmgm r irt sxr n xAa tA iwayt n Hm=f [r-b]l m pA dmi 52
For details, see Layton 2000, chapter 24, 265–67, 415–16. See e.g. (326) above. 54 A representative sample of languages is provided in Heine & Kuteva 2002, 261– 65. See also Hopper & Traugott 1993, 14–16, 180–81 as well as Peust 1996, 39–40, Peust 2004, 378–79 and Kammerzell & Peust 2002, 316–17 for Egyptian in particular. 55 GEG § 224; see also e.g. Helck 1975, no. 113, 3–4 sDm.tw r-Dd aq TArw “One heard that (the town of) Tjaru had been entered”. 53
egyptian supplementary patterns of complementation 285 r pna Hr pA [wr n ikA(r)Ty] nty Hr mw n Hm=f Now, his majesty heard that some of the Asiatics who were in the town of Uga(r)it were plotting in order to concoct a plan to expel the garrison of his majesty from the town and to bring about a change of mind in the [chief of Ugarit] who was loyal to his majesty. (Urk IV 1312, 7–11) (358) The sphinx of Giza says to the future king Thutmosis IV: sin.n=i r rdit iri=k ntt wn m ib=i rx.kw r-Dd ntk sA=i nDty=i I have been waiting to let you do what has been in my mind, for I know that you are my son and my protector. (Urk IV 1543, 9–10)
In Late Egyptian r-Dd is used quite differently from the earlier ntt/wnt. It combines freely with the verb gmi, is employed to introduce sDm=f functioning as final (i.e. irrealis) clauses in second tenses56 and occurs also in other types of unasserted object clauses: (359) Mai-Seteh scolds certain military chiefs for interfering with the religious administration:
ir sDm(=i) r-Dd th=tn nA.n rmT n pA nTr nty (m) ww=tn gr mk iw=i r aHA aA Hna=tn If I hear that you have harmed the god’s people who are in your districts, I shall reprimand you severely. (KRI I 322, 7–9) (360) Maani-nakhtef asks his correspondent to send a reply: mtw=k dit in.tw=f amA=i r-Dd tw=k Aty.tw mi-qd And you should send it so that I will know whether you are also in need. (KRI VI 671, 8) This latter property already characterises the earliest attestation of r-Dd as a complementiser ‘that’:
56
See Sweeney 1986, 340; 1erný-Groll 1993, 384.
286
chapter seven
(361) The author voices his exasperation with the complaints of his household:
in-wnn ib=i mH r-Dd rdi.n=i aqw n prw=i iw hAb.t(w) n=i rDd nn wn aqw How could I be confident that I have given rations to my household when people write to me saying: “there are no rations”? (pBM 10549, vso. 1–2)
The idea of this rhetorical question is that, in the circumstances described, the speaker feels (or would feel) anything but confident of having sufficiently seen to the well-being of his household.57 However, this has no effect on the use of r-Dd.58 In the few pre-Amarna examples of its use as a complementiser, this element represents a harbinger of a new emerging linguistic system that gradually replaced the Earlier Egyptian idiom, including its modal organisation of complement clauses.59
57
See Uljas 2004, 101 for discussion. Cf. Luft 1984, 107, where r-Dd is similarly seen to leave open the modality of the following proposition. 59 For further remarks to this effect, see 10.2 below. 58
predicate complement clauses
287
CHAPTER EIGHT
PREDICATE COMPLEMENT CLAUSES1 A well-established and -recognised type of Earlier Egyptian nominal sentences, or ‘sentences with nominal predicate’, is constructions in which complete verbal, pseudo-verbal or adverbial clauses are embedded as predicate complements. When the subordinate clause is verbal, it usually takes the form of a bare sDm=f for which ult. inf. roots show gemination and doubling roots doubling;2 the negation is with tm:3 (362) A gloss explaining an expression in a medical text: ir ib=f mH mhh ib=f pw mi nty Hr sxAt kt mdt As for ‘his heart is flooded’; this means that his (the patient’s) heart is oblivious, like one who is thinking of something else. (pEbers 102, 15–16)
(363) As above: ir wbnw m Srt=f isdb gnn spty wbnw pw As for ‘a wound in his nostril is weakened’; this means that (pEdwin Smith 6, 12–13) the edges of his wound are soft.
1 The research for this chapter was carried out in conjunction with the project Basel Diachronic Grammar of Egyptian, led by Prof. A. Loprieno. 2 In addition to passages cited below, further examples of geminating sDm=f are pEbers 99, 16 (dd ø N pw) and pEbers 100, 16 (wgg N pw). See also example (10) above. A further example of a doubling form is pEdwin Smith 9, 12 (gnn N pw). According to Loprieno (1995, 263 n.27) also ‘prospective’ forms can occur in this construction, but his only example in support of this argument, pRamesseum IV, C18, has mwt=f pw, with the immutable root mwt. There are no examples showing the endings -w/-y. In CT VI 158k one reads hA nsw=Tn pw, (with a non-geminating for hA of the root hAi, ‘descend’) which might mean “this means that your king will come down”, but the co-text and the reference of ‘you’ are obscure. Yet, some examples of the form wn of the auxiliary wnn are forthcoming—see below. 3 Further examples are pEbers 100, 14 (tm mdt N pw) and pEdwin Smith 14, 14 (tm sSp N pXrt pw). See also example (14) above.
288
chapter eight
(364) As above: ir n abA.n irtw=sn tm nbibiw inm=sn pw ir nn tms Hr=s tm wn xt pw Hr=s dSr As for ‘their colouring does not sparkle’; this means that their skin does not shine. As for ‘there is no reddening on it’; this means that there is no red matter on it. (pEdwin Smith 16, 14–16)
Of the anomalous roots, only iwi ‘come’ is attested, but then in great numbers in the literary colophon of the following form:4 (365) The end of the dialogue between the man and his soul: iw=f pw HAt=f r pH(wy)=fy mi gmyt m sS This is how it comes from its beginning to its end, in accordance (Man and Ba 154–55) with what was found in writing. Rarely, a verbal predicate may also be a bare sDm.n=f. In the following example this is passivised with .tw:5 (366) A gloss explains an expression in a mythological text: ir sAwt sbiw Htm.n.tw smAyt swty pw As for ‘guarding the rebels’; this means that the cronies of Seth (Tb 18, 5–6/Nu pl. 9, 5) were annihilated. Pseudo-verbal and adverbial complements are ushered in by the auxiliary wnn, which appears both as wnn and, less often, wn:6 4 See also e.g. Sh.S. 186; Sin B 311; Kagemni II, 9; Ptahh 645; Merikara E 144. For examples in the CT, see Doret 1990, 48 n.89. 5 A further example is Tb 18, 15–16 (Dd.n N pw). 6 Additional instances with wnn + stative are pEbers 102, 1 (wnn N pw tA); pEbers 102, 10 (wnn N pw gAw); pEbers 102, 12–13 (wnn N pw wrd); pEdwin Smith 3, 19 (wnn N pw fd); pEdwin Smith 8, 17 (wnn=f pw gr); pEdwin Smith 13, 10 (wnn N pw sS); pEdwin Smith 14, 8–9 (wnn N pw wDfA) and pEdwin Smith 14, 10–11 (wnn N pw Sfw). Further examples of wnn + Hr + infinitive are pEdwin Smith 8, 15–16 (wnn N pw Hr iit) and Tb 17, 56–57 (wnn N pw Hr irt). Further examples with wnn + adverbial predicate are pEbers 101, 13–14 (wnn N pw m Swt=f); pEbers 101, 16 (wnn pw {in} N m gs=f iAbty); pEdwin Smith 3, 21–4, 1 (wnn N pw mi wsSt n awt) and pEdwin Smith 15, 17 (wnn N
predicate complement clauses
289
(367) As above: ir ar=sn n=i and im=sn wnn smAyt swty pw Hr tkn im=s As for ‘few of them ascend to me’; this means that the cronies of Seth were approaching her. (Tb 17, 108–09/Nu pl. 9, 118) (368) A gloss explaining an expression in a medical text: ir mtw nw nHbt=f dwny wnn mtw pw nw nHbt=f dwn nxt m-a ih=f As for ‘the muscles of his neck are stretched’; this means that the muscles of his neck are stretched (i.e. cramped) and hard (pEdwin Smith 3, 19–20) due to his pain. (369) As above: ir ib=f sS=f wnn mtw pw n HAty Xr Hs As for ‘his heart spreads out’; this means that the veins of heart (pEbers 100, 17–18) carry (lit. “are under”) excrement. (370) A gloss explaining an expression in a mythological text: ir psx[=i Sny=i] r Hr=i txtx r wpt=i wn7 Ast pw Hr StA As for ‘I let my hair on my face and on my brow in disorder’; (Tb 17, 107–08) this means that Isis was hiding. The grammar of these construals appears to fulfil the syntactic expectations of the ST ‘nominal hypothesis’ to a fault. Only the ‘nominal’ sDm=f and sDm.n=f forms are used in the complement ‘slot’, tm appears in the negative and the adverbial and pseudo-verbal patterns are ‘converted’ by wn(n) (but not, strangely, by the main ‘converters’ or ‘nominalisers’ ntt and wnt). Unfortunately for this analysis, the patterns above do not exhaust the inventory of Earlier pw Hr qAbt=f). Besides example (370) below, of the form wn one may cite only Tb 17, 57, wn N pw Hr mAa, for which Nu has wnn. pEdwin Smith 16, 13 may be an instance of an existential sentence wnn xt pw Sfw wr “this means that there is a greatly swollen thing”. In Tb 17, 34 there appears the strange construction wnn irt=f pw mr sy, “This means that his eye was in pain”. The papyri of Nebseni and Nu are in agreement of the writing mr sy, whereas other variants have either mr=s or mr.ti. 7 So Nebseni; the papyrus of Nu has wnn.
290
chapter eight
Egyptian nominal sentences with clausal predicates. In addition to constructions such as those just cited—which one may label ‘type A’—there exists another class of bipartite pw-sentences where clausal embedding takes place—namely ones of the following kind: (371) Khakheperraseneb adds a self-perceptive note amidst his laments: ink pw Hr nkAy m xprt I am meditating upon what has happened. (Khakheperraseneb rto. 10)
(372) Neferty launches upon the pivotal part of his prophesy of a new order that will dispel chaos:
nsw pw r iyt n rsy A king will come from the south.
(Neferty XIII a)
(373) The king tells of his spoiled boat-trip; one of his female rowers had suddenly stopped rowing:
aHa.n Dd.n=i n=s tm=t Xn Hr m aHa.n Dd.n=s n=i nxAw pw n mfkAt mAt xr Hr mw Then I said to her: “Why don’t you row?” And she said to me: “A pendant of new malachite just fell into the water”. (pWestcar 6, 4–6)
(374) A new turn in a mythological narrative describing the wounding of the Eye of Horus:
ra pw Dd.n=f n Hr imi mA=i irt=k Then Ra said to Horus: “Let me see your eye”. (CT II 334/35b–c)
Over the years, sentences of this particular sort have been subject to considerable debate, and not all Egyptological linguists would view them as involving clausal predicates at all. However, as will be seen below, syntactic and semantic considerations show that this analysis is indeed the correct one, but constructions of this kind, termed here ‘type B’, also differ considerably from bipartite pwsentences of ‘type A’ exemplified in (362)–(370) above. Syntactically, they involve complementation of sentences with ante-posed elements
predicate complement clauses
291
whose predicate may be the sDm.n=f as in example (374) above, but also some of the by now familiar ‘non-nominal’, ‘non-specialised’ or ‘unconverted’ forms and construals such as the pseudo-verbal Hr/r + infinitive in (371) and (372) or the stative in (373).8 All these construals are embedded directly without any ‘converters’ or ‘nominalisers’. As with the subject clauses of the semi-auxiliary xpr.n, this exceptional structural phenomenon is a reflection of the special semantic-pragmatic value that the ‘non-nominal’ patterns assume in construction with the element pw in this sentence type.9 It will be seen that type B nominal sentences are in fact used to express realis assertions with a particularly marked information-structure and with a referentially ‘bleached’ pw that has lost its semantic ‘subjecthood’. In both respects they differ from constructions of the type A in which pw retains its subject properties, the complement clauses remain non-asserted and irrealis and are coded with (some of the) forms/constructs seen to be associated with these functions also elsewhere in complementation. In addition, a closer scrutiny of both these nominal sentence types, and particularly type B, reveals rather striking facets of the Earlier Egyptian nominal sentences as a wider grammatical category. These constructions represent perhaps the most problematic of the Earlier Egyptian non-verbal sentence patterns. Particularly the precise mechanism(s) of identifying the ‘subject’ and ‘predicate’ in these constructions continues to divide opinions.10 Yet, nominal sentences seem to have at least one redeeming feature, namely their apparently easily definable and well-delimited character as a sentence type. Prototypically, a ‘nominal sentence’ is a non-verbal construction whose semantic predicate is a ‘nominal (phrase)’, i.e. an expression that can be used as the head of a term as e.g. in “I saw a/the man”, 8 For other patterns found occurring or argued to occur in this construction, see further below. In addition to the examples cited here and further below, further instances with the sDm.n=f are pRam Dram 48 (N pw w[D].n=f); CT IV 384b (N pw wD.n=f); pPushkin 167, fragment 2, 2 (N pw mn.n=f); pPushkin 167, fragment 2, 3 (N pw xpr. n=s) and pEbers 103, 9 (N pw sSp.n=sn—so also pEbers 103, 6). See also example (11) above. Additional instances with the stative are pRam Dram 8 (N pw Spt); pRam Dram 21 (N pw dy); pRam Dram 46 ([N] pw sxm); CT I 269i (N pw xr) and pRamesseum III, B23 (ink pw pr.kw). A further example of Hr + infinitive is pEbers 36, 14–15 (N pw Hr mxr). Of r + infinitive there are no further attestations. 9 For xpr.n-constructions, see 4.3 above. 10 This topic is not relevant to the present hypothesis and will not be discussed here.
292
chapter eight
“the man killed the goat” or “what the man killed was a goat” etc.11 Also semantically, Earlier Egyptian nominal sentences seem to have their own definable typology that differs from adjectival, adverbial and verbal predication.12 However, besides demonstrating the modal differentiation between irrealis and realis clausal predicates, the present discussion also seeks to show that rather than being a ‘closed’ grammatical category, the Earlier Egyptian nominal sentence is more correctly described as a semantic-pragmatic and morphosyntactic continuum with verbal, pseudo-verbal and even adverbial sentences. The core feature in this merging of these seemingly distinct sentence classes is modality, but with an additional twist of pragmatic information structuring. There is no clear line of demarcation between nominal- and the latter types of sentences in either meaning or form. Instead, nominal sentences of the type B above represent a zone of ‘cross-categorical’ merging with the said sentence types. Constructions of the type A are also part of this continuum, being the ‘next step down’ towards more prototypical nominal sentences/predication. This amalgamation is observable first in the meaning expressed by construals with clausal predicates and, the greater the merging, also in their morpho-syntax. Nominal sentences are thus not a semantically or syntactically coherent or self-sufficient category of uniform ‘sentences with nominal predicate’. Instead, they are part of an interconnected fabric of meaning and form that is the linguistic system of Earlier Egyptian. 8.1 The Syntax of Type B Nominal Sentences with Clausal Predicates Below is a selection of further examples of ‘type B’ nominal sentences: (375) A beginning of a mythological narrative: S2P +13 other variants: ra pw Hr mdt Hna imy-whm=f b S1C : ra pw mdw.n=f Hna imy-whm=f Ra was speaking/spoke to the imy-whm=f-snake. (CT II 274/75c–276/77a)
11 12
See Hengeveld 1992, 58 and passim. See 8.2 below.
predicate complement clauses
293
(376) The sailor begins his narrative to the snake: ink pw hA.kw r biA m wpwt ity I went down to the mine-country on a royal mission. (Sh.S. 89–91)
(377) In a broken context: ink pw hAb.kw r s ff I have been sent against the man of Fef (?) (pTurin 54003, rto. 5)
(378) A new development in a mythological narrative: ra pw xp.n sw imy-ins=f Ra was encountered by He Who Wears His Bright-red Cloth. (CT II 280/81d)
(379) The deceased (the ‘magician’) is asked beside a boat: in iw tr sxm=k m tm in n=k HkAy pw dpt tn pw n wnt mDAbt=s Do you have power over what has not been brought to you, O magician? This boat does not have its bailer.13 (CT V 109h–j) (380) Amun says to Queen Hatshepsut: rn(=i) pw xnt nTrw [r]n=T pw xnt anxw nbw Dt My name is before gods; your name is eternally before all the (Urk IV 347, 2–3) living. There are at present two (main) alternative analyses of constructions of the above sort. Both view the sentences as bipartite nominal sentences, but opinions diverge as to what element or set of elements actually represents the predicate. In some analyses this is understood to be only the initial noun/pronoun, whereas what follows the subject pw is interpreted as an attributive adjunct clause outside the sentence proper.14 Alternatively, the predicate has been seen to include also 13 The same construction occurs also in CT V 112f. The same holds also for the variant T1C in CT V 110g, but here the version T1Be writes dpt pw n apr=s m AxAxw=s (M2C omits the negation) “this boat has not been equipped with its spars”, apparently with n + an archaic passive sDm=f with a suffix pronoun subject. 14 Polotsky 1976, 3.10.3; Doret 1990, 48; Malaise & Winand 1999 § 469; cf also WGMT 1962 § 405. Vernus (1994, 339) brackets what follows pw together with the
294
chapter eight
the ‘circumstantial’ and together with the initial noun/pronoun, to form a full verbal, pseudo-verbal or adverbial sentence complement clause subordinated to pw.15 In defence of the first of these analyses, it could be argued that inasmuch as e.g. pseudo-verbal constructions such as ink hA.kw with independent pronoun subjects are extremely rare in Earlier Egyptian generally,16 they can scarcely occur widely ‘embedded’ as predicates of pw. Yet, as will be seen shortly, this is in fact not a serious objection to the clausal complement analysis, which is supported by two strong arguments. Firstly, the clauses and phrases following pw are not free ‘circumstantials’, but display obligatory subject- or, in (377) and (378), object/patient control. They cannot be ‘substituted’ with adjuncts lacking this referential linkage with the initial noun/pronoun antecedent:17 rai pw mdw.n=fi
†
ra pw mdw.n rmT
Secondly, were the alleged ‘circumstantials’ after pw mere adjuncts, they should behave accordingly also as regards their semantic value. Thus, e.g. in ra pw mdw.n=f the sDm.n=f should be relative past † “This/he/it is Ra after he spoke…” which is clearly impossible.18 The conclusion that what follows pw in type B sentences are not adjuncts, but instead constitute part of the predicate seems inescapable. But then the question arises, what is the role of the initial noun/pronoun and why does it appear in the position that it does. A convincing answer to this has been suggested by Loprieno, who analyses constructions of type B as bipartite pw-sentences with predicate complement clauses, whose subject (or, rarely, object) is topicalised. In case of patterns with the order vso, the latter is resumed after pw as a co-referential pronoun:19 initial noun/pronoun + pw, but still refers to the former as “forme verbale en fonction adverbiale”. 15 Sethe 1928, 90; GEG §§ 190.1, 325, 332; LGEC §§ 614–15; Junge 1978a, 60–62; Roccati 1979, 40; Westendorf 1981, 14; Callender 1984, 141–42; Loprieno 1988b, 37; Allen 2000, 210; Schenkel 2005, 305–06. Cf. also Sethe 1916 §§ 102, 104. 16 The rare instances of ‘rhematised’ (focalised) pronouns are an exception; see Allen 1994, 6–8. 17 The subscript ‘i’ indicates referentially identical elements. 18 Loprieno 1995, 112; cf. also Junge 1978a, 61. Of course, the alleged adjunct can hardly function as a ‘virtual’ relative clause either, although this seems to be the assumption behind Simpson’s (1972, 53) translation of (376) above as “It is I (myself) who have gone down to the mines”. 19 Adapted from Loprieno 1995, 111–12; see also Borghouts 1986, 54–55 and cf. Junge 1978a, 61.
predicate complement clauses ra pw mdw.n=f Hna imy-whm=f
295
=
[[rai]TOPIC[mdw.n=fi Hna imy-whm=f]] [pw] In instances with pseudo-verbal patterns with svo order, one could postulate an analogous covert pronoun in the structure:20 ra pw Hr mdt Hna imy-whm=f
=
[[rai]TOPIC[øi Hr mdt Hna imy-whm=f]] [pw] Yet, as noted by Loprieno,21 the complements here can be more straightforwardly viewed as the basic constructions with the initial subject cleft apart from the rest of the sentence by the syntactic ‘gap’ created by the clitic pw, i.e. simply as [ra Hr mdt][pw] → [raTOPIC [pw] Hr mdt] without ‘movement’ of the initial noun/pronoun and its equideletion after pw. This also renders the use of independent pronouns in the corresponding sentences with topicalised pronoun subjects/objects quite explicable and removes the above-mentioned objection against the complement clause analysis. Independent pronouns appear in all the subordinate clauses above, including pseudo-verbal ones, simply because these construals are embedded in bipartite pw-sentences, where their topicalised pronominal subjects come to occupy the initial position and thus cannot take the form of dependent or suffix pronouns.22 However, this then means that such constructions are not strictly isomorphic, and the same holds even more strongly with sentences such as ra pw mdw.n=f. There does exist a pattern ‘noun + sDm.n=f’,
20
Cf. Borghouts 1986, 55 Loprieno 1995, 111. 22 Cf. Schenkel 2005, 306. Jansen-Winkeln (1996b, 47) views type B nominal sentences as paratactic complexes with ‘directly quoted’ ‘Zitierte Hauptsätze’. This hypothesis entails an assumption of universal isomorphism of these constructions and the existence of independent main clauses (Hauptsätze) such as ink sDm.n=i. Yet, as noted, although such sentences do exist, they are rare to the extreme and, were JansenWinkeln’s analysis to be accepted, they would mostly occur as ‘quotes’ with pw rather than were used independently. However, this is not to say that such construals do not exist at all with pw. On the contrary, there are numerous examples of paratactic pw-sentences where the apparent ‘predicate’ is in fact an independent (non-subordinate) construal (e.g. an imperative, a nn-negated verbal or adverbial sentence etc.). See Vernus 1994, 345–47 for discussion and examples. 21
296
chapter eight
but this is also nowadays commonly analysed as a topicalisationconstruction,23 and it would be strange to find it here. Moreover, in the said construal the initial topicalised element is always the subject, which is not the case in type B nominal sentences (see example (378) above). ‘Deconstructing’ sentences such as (379) above, where what follows pw is a negative existential sentence, is no longer feasible at all. This shows that type B nominal sentences are not derived from simpler ‘atomic’ structures, but represent a well-formed and independent category of marked topic sentences. The clausal analysis of type B nominal sentences thus seems to be best in accord with the syntactic and semantic evidence. In favour of Loprieno’s topicalisation-hypothesis speaks the fact that the first element in these constructions is most often a 1st person subject, which, as noted by Loprieno, conforms well with the recognised hierarchy of discourse topicality where 1st person ranks higher than other persons and subjects are more topical than other grammatical roles.24 However, this analysis naturally entails yet another violation of the ban on subordination of adverbial and pseudo-verbal sentences as complements without a ‘converter’ or ‘nominaliser’ such as ntt or wnt. Yet, as seen, this ban is far less absolute than has been argued. It is precisely the sort of ‘exceptions’ as these that show the clear need for an alternative analysis of the grammatical phenomena studied. 8.2 Assertion and Non-assertion in Bipartite Nominal Sentences with Clausal Predicates and the Limits of ‘Nominal’ Predication One may now turn to consider what factors determine the use of the bare ‘non-nominal’ forms and constructions as predicate complements of type B nominal sentences and how do these differ from the parameters demanding the employment of the ‘nominal’ forms in constructions of the type A above. This is a rather complex matter that can only be clarified by studying the semantics of the predication expressed in and by the sentences with clausal predicates—how these differ from other nominal sentences as well as between the two classes A and B. In the latter instance it is necessary to consider the character of both the syntactically ‘higher’ level predication between the element pw 23 24
Loprieno 1988b, 43–44; 1995, 187–88; cf. also Malaise & Winand § 1050. Loprieno 1988b, 42; 1995, 112.
predicate complement clauses
297
and the clausal complement, and the ‘lower’ level predication between the subject and predicate of the embedded construal. Both types of bipartite nominal sentences with clausal predicates deviate from what one may term the semantic ‘prototype’ of this sentence pattern. Nominal sentence predication is in principle concerned with the identification, classification or specification of the subject.25 Classifying nominal sentences most often predicate their subject as a member of some class—real or imaginary—whose constituents are less defined and more referentially opaque than the subject itself. They define one or more of the semantic properties of the subject without, however, providing an exhaustive characterisation of it. For instance, in the following sentence the referent of the subject pw is profiled as a ‘member’ of the class of ‘peasants of the Wadi el-Natrun’: (381) The peasant is introduced in the Tale: sxty pw n sxt-HmAt He was a peasant of the Wadi el-Natrun.
(Peas R 1, 1)
Specifying nominal sentences establish a relation of complete or ‘coextensive’ semantic equity between the subject and the predicate. They provide, as it were, a total coverage of the subject’s semantic scope by indicating that it overlaps wholly with that of the predicate so that the subject and predicate actually have the same referent: (382) The deceased proclaims his transformation into the eye of Horus: ink pw s(y) stt pw (w)i Ts-pXr (CT VII 157c) I am it and it is me—and vice versa. Finally, identifying nominal sentences represent a semantic subclass of specifying sentences. The subject of such propositions is portrayed as the sole referent to which the predicate applies and becomes a recipient of additional pragmatic prominence and focus in the overall proposition:
25 For what follows, see Dik 1980, 102–04; Hengeveld 1992, 80–89; Loprieno 1995, 104–09; Winand 2006, 154–56.
298
chapter eight
(383) The author singles out himself as the one responsible for catering for a workforce:
ink dd n=sn aqw r Tnw rnpt It is me who gives them rations every year.
(pBerlin 10033, 5)
Yet, this taxonomy of nominal sentence semantics is in fact not applicable to the predication between pw and the complement in bipartite pw-sentences with clausal predicates. Concerning type A first, consider the following typical instance of this category of sentences: (384) A gloss in a medical text explaining a term: ir rwt nt HAty rww=f sw pw Hr mnd=f iAbty As for ‘palpitation of heart’; this means that it parts its self from (pEbers 101, 11–12) his left breast. Here the subject pw functions as an anaphoric ‘index’ for the expression rwt nt HAty ‘palpitation of heart’ ante-posed by ir and is thus fully referential.26 There also seems to be a predicative nexus between pw and rww=f, seeing that the sentence clearly serves to establish some kind of semantic relation between the referent of the subject and the complement state of affairs. Yet, the subject is clearly not identified, specified or classified as the clausal predicate ‘it moves its self’, but rather somehow merely assigned the same denotation or significance. Loprieno has given this intuition a rather more precise formulation by analysing constructions of type A as thetic sentences.27 In the present work this interpretation has already been rejected twice in connection with other complement structures,28 but now Loprieno’s hypothesis seems to capture the semantic relation involved precisely. Here it is necessary to briefly recall and elaborate on the meaning of the term ‘thetic’.29 Non-thetic ‘categorical’ sentences predicate something of a subject that also serves as the topic of the proposition. The topical status of the subject is equivalent for it being the ‘centre of interest’ and 26 See Vernus 1994, 333 for pw as an index for the ‘real’ semantic subject; cf. also Doret 1990, 48. 27 Loprieno 1995, 110–11. 28 See 2.1.1 and 7.2. 29 For what follows, see Sasse 1987; Lambrecht 1994, 118–19, 124, 137–46 as well as 2.1.1 above and 10.3 below.
predicate complement clauses
299
what the sentence is ‘about’. The predicate adds relevant information about the subject/topic and enhances the hearer’s understanding or mental image of it. By contrast, the subject of ‘informatively level’ thetic sentences is not the topic. Propositions of the sort are not, freely speaking, ‘about’ their subjects and do not provide information about them. For example, a sentence such as “my car broke down” may be understood as categorical, i.e. as telling something about the speaker’s car, which constitutes the topic and is what the sentence is about. But in a particular context, such as when uttered by a man carrying a large bag in a crowded bus and in response to angry looks from his fellow passengers, the sentence receives a thetic interpretation. It provides an explanation for the speaker’s presence and is not actually about his car, which consequently is not the topic of the utterance. This is, rather, what the sentence explains. Hence, a thetic proposition such as this may be seen as pragmatically equivalent to the more complex “the reason for my being a nuisance is [my car broke down]”, where the sentence in brackets is merely ‘posed’ as a whole rather than its predication elaborating a topical subject. It seems that this is exactly what is at issue in Earlier Egyptian nominal sentences of type A above. As can be seen from the examples cited, constructions of this sort appear almost exclusively in glosses and explanatory or ‘aetiological’ contexts where they clarify the meaning, symbolic significance, etymology etc. of various sorts of expressions ante-posed by ir.30 The topic of these sentences is not the subject of the subordinate proposition, but the referent of pw, which in fact usually appears topicalised (ir-introduced) just before the sentence itself. Consequently, a sentence such as “this means it departs its self” is thetic in that it is not about something ‘departing’, but rather about what the demonstrative ‘this’ in the translation—and pw in the original—refers to. Against this state of affairs it is relatively obvious why Earlier Egyptian should use forms and constructions argued here to be expressive of irrealis modality. In type A nominal sentences with predicate complement clauses the subordinate propositions do not 30 Cf. GEG § 189; Loprieno 1988b, 38; 1995, 110. There are, of course, contexts such as the literary colophon in example (365) above where this is not the case. However, for the analysis proposed here it is of some interest that e.g. Gardiner should have chosen to describe these instances as ones “where pw must be rendered ‘that is how’” (GEG § 189.1). It has been seen at various occasions in the present work what the appropriateness of this translation actually reflects.
300
chapter eight
actually offer information about their subjects to be accepted or denied by the audience, but they are rather simply presented an sich. This is almost tantamount to saying that such sentences assert nothing of their subjects. Hence, the embedded proposition is modally irrealis with the geminating sDm=f and negated by tm. Yet, there is no question as to the situation it describes being subject to non-acceptance, doubt or any other distal-inducing traits of attitudinal nuance, for which reason in the affirmative the geminating sDm=f and its functional counterparts are used. Accordingly, Earlier Egyptian once again behaves in a principled manner in its grammatical treatment of propositions that do not explicitly affirm or deny something of a subject and present this as optimally relevant information. Such expressions are regularly assigned to the domain of irrealis modality and non-assertion, and the complements in type A nominal sentences with clausal predicates are no exception to this. Turning now to sentences of the type B, in these constructions the status of the subordinate predication is quite different from type A, but the same pertains also to the relationship between the complement clause and the element pw. In type B sentences this level of predication actually disappears altogether. For example: (385) A new stage in the battle between Horus and Seth: stX pw iri.n=f xprw r=f m SAi km Seth then transformed himself into a black pig against him. (CT II 342b)
Sentences of type B have for long been noted to be “specially appropriate to the beginning of narratives and the answers to questions”.31 As in the ‘literal’ or ‘traditional’ translation of (385) as “it was so that Seth transformed himself…” the sentence after the complementiser that is not classified, specified or identified as anything else. But in addition, the element pw no longer semantically indexes for anything. In “it was so that Seth transformed himself…” the pronoun it functions as an expletive and has no semantic content independent of the thatclause but is rather co-indexed with the latter. Similarly, pw has no
31
GEG § 190.1; see also Polotsky 1976, 3.10.2.
predicate complement clauses
301
anaphoric or cataphoric referent beyond the subordinate construction itself.32 This is true also when type B constructions occur in answers to questions (example (386) below) and in seemingly ‘explanatory’ contexts such as that in (387): (386) A dialogue between the queen and king: pty spr r HAty=k nsw Ds=f Dd=f xr=s ink pw sxA.n=i mwt mwt=i mwt it(=i) “What has come to your mind?” The king himself answered to her: “I have recalled the mother of my mother and the (Urk IV 27, 12–14) mother of my father”. (387) A new development in a mystery play: xpr.n dw nwH r Dd stX pw Sa It happened that a rope was attached to a Djed-pillar. Thus (pRam Dram 51) Seth was slain. Yet, in (386) the ‘referent’ of pw is clearly not equivalent to “what has come to my mind”, which it would need to be for the sentence to be a ‘direct’ reply. At most, it could be said to be some highly general ‘the situation’ or ‘the case’ (is). As for ‘explanatory’ instances akin to (387), Loprieno correctly observes that since in type B constructions the subject or rarely the object of the embedded clause is established as a topic, the sentences cannot be characterised as thetic.33 But in addition, a close reading of instances such as (387) above shows that pw does not actually refer back to the situation mentioned just before.34 The referentially zero value of pw is most apparent from 32 Cf. Vernus 1994, 340, who maintains that the referent of pw is actually just the ‘enunciation itself’. 33 Loprieno 1995, 111. 34 Reading type B sentences e.g. in the Ramesseum Dramatic Papyrus as explanatory “this means that X” assumes them to represent ‘commentary’ for the ritual acts performed that explains the symbolic and mythological meaning of the latter. However, as seen, this is what is achieved by type A constructions instead. It seems rather that the type B is employed in a manner suggested by the translation of (387) above. The act of e.g. attaching a rope to a Djed-pillar magically brings into effect the annihilation of Seth as the play is enacted. That is, rather than it being simply clarified what the ritual act with the rope means, the xpr.n-sentence tells what happened concretely and the sentence stX pw Sa what took place symbolically as a result. Accordingly, e.g. attaching a rope to the
302
chapter eight
the again often-noted difficulty of finding a fixed translation for type B constructions and their use in absolute initial environments:35 (388) The Tale of the Eloquent Peasant begins: s pw wn xw-n-inpw rn=f (Once upon a time) there was a man named Khuenanupe. (Peas R 1, 1)
pw has completely lost what may be called its ‘semantic subjecthood’; i.e. its status as a referential and semantically definable/identifiable ‘landmark’ of which something is predicated. As a consequence, in type B constructions the only predication in the complex is actually the embedded proposition. Put another way, although e.g. the clause [s wn] in example (388) above is indeed subordinate to pw, it is not its predicate. Whereas in type A nominal sentences there are still two predications in the complex, in s pw wn there is only one, namely that between the subordinate clause subject s and its predicate wn. Yet, in type B the overall structure, as it were, ‘lags behind’ semantics. Syntactically these constructions are, of course, perfectly well formed bipartite nominal sentences with pw, but semantically they are no longer ‘sentences with nominal predicate’. But even if pw here appears relatively ‘meaningless’, it has a crucial syntactic and semantic-pragmatic role. The syntactic function was already hinted at above where it was suggested that constructions of type B involve topicalisation of one of the situation participants. Now, it could be argued that in the pseudo-verbal and adverbial patterns with the order svo the subject already occupies the most topical position attainable. Yet, this ordering is not marked: the ‘topicality’ of the subject is a mere default setting and is not pragmatically pronounced.36 A priori, one might expect there to be some way of marking the subject of these construals explicitly as the topic, and besides simple antepositing, this is indeed what is achieved by embedding the pseudoverbal or adverbial sentence as a complement of pw and cleaving off
Djed-pillar in the play does not ‘mean’ that Seth was slain, but it has the latter as its outcome. There is thus no antecedent-resumptive relation between the two and pw does not index for the xpr.n-sentence or a part thereof. 35 See e.g. Junge 1978a, 62; Loprieno 1995, 111. 36 Cf. Loprieno 1988, 44; 1995, 188.
predicate complement clauses
303
the subject from the rest of the construal.37 In verbal sentences with the order vso, the communicative outcome of this procedure is the same, but its syntactic effects are more clearly apparent due to the initial position of the element topicalised and its pronominal resumption later in the sentence. Of course, participants could be topicalised also by means of ante-position either alone or after auxiliaries/initial particles and the element ir, but not as generally as in type B nominal sentences.38 For example, topicalisation of pronominal subjects and objects cannot be carried out by simple ante-position or with the element ir; only nominal elements of both sorts can be so treated. Concerning the sDm=f, in Earlier Egyptian the phenomenon of subject anticipation after auxiliaries and initial particles (mk sw/iw=f sDm=f) as well as the operators ntt/wnt occasionally still seems to function as a vehicle for topicalising the subject, even if its pragmatic ‘markedness’ is greatly diluted.39 It is hardly surprising, then, that no examples of type B sentences with the sDm=f are forthcoming.40 However, the same construal is only marginally attested with the sDm.n=f and topicalisation of objects in a similar manner is excluded in principle. Accordingly, type B nominal sentences provide a complementary method of topicalising of participants of verbal, pseudo-verbal and adverbial sentences aside ante-position by other means. In sum, the syntactic function of pw in the constructions above is that of a topic-marker and the overall syntactic raison d’être of type B construals is to function as marked-topic alternative for the various verbal, pseudo-verbal and adverbial patterns. However, pw also has a semantic-pragmatic role in type B nominal sentences. The loss of its semantic ‘subjecthood’ is what ultimately licenses the occurrence of the ‘non-nominal’ forms and construals in the complement syntactic ‘slot’. 37 For pseudo-verbal and adverbial sentences with participants topicalised by simple ante-position, see below. 38 For examples and discussion of the first to methods, see Collier 1991a, 37–43. 39 See Collier 1991a, 37–38, 41–43; Eyre 1994, 126. See also the discussion in 2.4 above on ntt=f/wnt=f sDm=f. 40 CT V 39a, sometimes thought to be an instance of this, is hardly a convincing counterexample; cf. Polotsky 1976, 3.10.7. In the Ramesseum Dramatic Papyrus there are multiple occurrences of the past ‘indicative’ sDm=f (with the highly archaic use with pronoun subjects) in the type B nominal sentences. Yet, iw was never used with the ‘indicative’ sDm=f and the option of topicalising the subject as iw=f sDm=f was not open. It is thus quite according to expectations that type B nominal sentences should have been used instead.
304
chapter eight
The disappearance of predicative relationship between the two main constituent parts of the sentence leaves the subordinate predication as the only information-bearing part in the overall construction. The latter are, of course, definitely ‘about’ their own topicalised subjects (less often objects), but the salience of the information they provide seems particularly, or one might say, exceptionally, pronounced. The communicative role of type B nominal sentences has quite correctly been noted to be ‘presentative-apocritic’ ‘Sachverhalts-Aktualisierung’ and ‘exposition’ of ‘new developments’ in discourse and narrative.41 The messages they convey are clearly specially highlighted as being of pivotal communicative importance in the context in which they occur. This their pronounced ‘importance’ is due to the element pw and as a phenomenon is, once again, modal in character. Informally speaking, the semantic contribution of pw in type B nominal sentences is not to signal that situation p ‘is something’, but only that situation p ‘is’. This in fact resembles greatly the role of the semi-auxiliary xpr.n discussed earlier on.42 Just like xpr.n, pw in type B nominal sentences lays stress on the actuality and occurrence of the situation described in what in both instances is syntactically a complement clause, and signals its high degree of relevance and novelty as information. The only real difference is that the proposition in case of type B sentences is also marked for topicality of one of its grammatical roles. Put another way, like xpr.n, and in addition to its topicalising function, pw merely ‘presents’ the situation and has no independent semantic value beyond this. Consequently, pw does exactly the same job as xpr.n—and, indeed, ntt/wnt—in that it marks the unmarked subordinate pattern as a modally realis assertion and allows it to function as a complement. As in the case of xpr.n, this explains why only the bare ‘non-nominal’ forms and constructions are acceptable in type B sentences. There are no ntt/wnt-introduced versions of this sentence pattern, because pw carries out the assertion marking in the construction and no ntt/wnt are needed.43 Of course, all irrealis-marked forms are categorically 41
See GEG § 190; Polotsky 1976, 3.10.2; Junge 1978a, 60; Shisha-Halevy 1987, 174; Vernus 1994, 340–41; Malaise & Winand 1999 § 469 among others. 42 See 4.3 above. 43 An apparent counterexample to this is pEbers 99, 5 nt pw mdw=f xnt mtw nw at nbt, which Loprieno (1995, 109) analyses as a nt(t)-clause nt(t) mdw=f… subordinated to pw, i.e. as a ntt-introduced counterpart of type A sentences (“this means that it (the heart) speaks through the veins of every member”). However, alternatively one may
predicate complement clauses
305
barred from occurring in type B sentences precisely because of this function of pw. The ‘meaningless’ pw of type B nominal sentences with clausal predicate complements thus has a decisive functional role. Besides its topicalising role, it represents a modal operator just like ntt/wnt. 8.3 Interim Summary Bipartite pw-constructions with clausal predicates form their own peculiar subclass of Earlier Egyptian nominal sentences in that modality plays a significant part in their grammar. As noted earlier, in this language bare non-verbal sentences generally stand outside the organisation of tam. However, this is apparently not an absolute ‘rule’; witness e.g. the introduction of adverbial sentences with initial auxiliaries, which seem to have a semantic-pragmatic function of some sort. Similarly, from the moment that verbal, pseudo-verbal and even adverbial constructions begin to appear as constituents of nominal sentences, modality steps in. This holds both for the type A and B of the present discussion. In the coding of the subordinate verbal/ pseudo-verbal/adverbial construal as realis and irrealis, the shift in the semantic and pragmatic properties of the element pw is decisive. As one metaphorically ‘moves’ from type A to type B sentences, pw relinquishes its characteristic subject-indexing role and the predication it entertains with the complement first begins to deviate from the nominal sentence prototype and eventually disappears. Yet, pw does not actually ‘lose out’ in this process, but rather acquires new functions. Syntactically, it becomes a topic-marker in type B sentences. Semantic-pragmatically, it acquires the role of a modal operator, signalling that the proposition with which it is associated—and which syntactically is still a complement clause—has the illocutionary force and grammatical status of a realis assertion. This its latter role read this example as a predicate-subject-order tripartite sentence whose subject is the clause mdw=f xnt mtw nw at nbt and the predicate a lexical nt(t): [nt(t)]P [pw] [mdw=f]S, i.e. roughly “It is a fact that the heart speaks through every vein” (cf. Sin B 115–16 in nt(t) pw wn.n=i sA pr=f “Is it a fact that I have trespassed his privy quarters?”). The same analysis is, of course, not applicable to the only other example of this construction, pEdwin Smith 1, 7 nt(t) pw mtw=f m HA m st-ib, where the subject clause would have to be a bare adverbial sentence mtw=f m HA m st-ib “its veins are at the back of the heart’s place”. Yet, it is notable that the variant pEbers 99, 4 writes Hr ntt mtw=f n at=f nbt “because its veins pertain to his every member”.
306
chapter eight
is highly interesting not only because it apparently represents a third instance of grammaticalisation of the original demonstrative pronoun pw in Earlier Egyptian, (the other two being its adoption as a subject element in bipartite sentences and as a copula in tripartite ones44) but also because it seems to resurface again much later in Coptic. It has recently been suggested by Reintges that the enigmatic and seemingly superfluous pe often found in the imperfect conjugation nefswtM “places emphasis on the truth of the reported event or state of affairs and thus contributes a modal force of certainty or assertion to the clause it modifies”.45 Against the discussion of Earlier Egyptian above, this seems more than a mere possibility and would represent an interesting case of re-grammaticalisation of a long-dead functional device. The constructions labelled A and B can be seen to provide a link that unifies nominal, verbal and adverbial sentences into one syntactic and semantic spectrum. Functionally, type B ‘nominal sentences’ that are not ‘nominal predications’ are in fact marked-topic equivalents of ‘ordinary’ asserted verbal, pseudo-verbal and adverbial sentences. Their counterparts without topicalisation are, of course, no longer ‘nominal sentences’ even formally. That is, non-topicalised assertion of, say, sDm.n=f is not †ntt pw sDm.n=f or pw anything, but probably a normal verbal main clause preceded by an auxiliary, i.e. e.g. iw sDm.n=f. Both ‘present a new situation’, and although in one of them an element is assigned additional pragmatic marking, their discourse functions are quite akin. Structurally, however, they are of a very different sort: one is a verbal, the other a syntactically ‘nominal’ sentence. Interestingly, like topicalisation, pragmatic focalisation of participants of verbs displays exactly the same juxtaposition between verbal and nominal syntax.46 A subject-focus counterpart of a past and present verbal proposition such as mA.n=f sS “he saw the scribe” is a cleft sentence ‘participial statement’ ntf mA sS “it was he who saw the scribe”—syntactically a bipartite nominal sentence. Object-focus is achieved by means of a pseudo-cleft sS pw mA.n=f “the scribe is
44
See (380)–(382) above for examples. Reintges 2004, 247. Emphasis by SU. 46 An exception here is the rare instances of ‘pronominal rhematisation’ referred to in n.16 above and discussed in Allen 1994. However, this phenomenon seems to be restricted to adverbial and pseudo-verbal predications. 45
predicate complement clauses
307
the one whom he saw”, which is syntactically a tripartite nominal sentence. Although of these pragmatically marked patterns only the latter two are properly ‘sentences with nominal predicate’, they are all structurally ‘nominal sentences’. It seems thus that assigning pragmatic prominence to arguments of verbal states of affairs involves a shift from ‘verbal’ to ‘nominal’ syntax more generally. The two bipartite pw-sentence types discussed, and particularly type B, represent what might be called the ‘limit’ of the Earlier Egyptian nominal sentence; a semantic-pragmatic and ultimately also syntactic ‘transit area’ of mediation between “nominal”, verbal, pseudo-verbal and adverbial sentences. Such ‘form-meaning continuums’ are not marginal ‘exceptions’. The present writer has recently attempted to describe a comparable linkage between nominal and adjectival sentences.47 Also the merging between nominal and verbal meaning/ syntax noted in cleft constructions actually goes further than noted above. The future equivalent of the past and present cleft sentence is the pattern ‘independent pronoun/in + noun + sDm=f’, where the agent of the verbal clause is ante-posed exactly as in the ‘participial statement’, but clearly without any ‘nominal predicate’ nexus between the two. Semantically the sentence †“It is me that I will hear” is just as impossible in Egyptian as in English. Although the focalised element entertains an antecedent-resumptive relation with the sDm=f, it stands outside the predicative frame, which consists merely of the verb and its subject. Semantically, and in fact also syntactically, the future cleft sentence is hence not a non-verbal ‘nominal’, but a verbal construal. It too can be conceived as standing at the border of verbal and nominal sentences: formally, it resembles the latter to a notable extent and enters the same semantic class as the ‘participial statement’, but the only predication expressed in the complex is that between the sDm=f and its suffixed subject. Like the future cleft sentence, bipartite nominal sentences with clausal complements of pw represent concrete manifestations of the phenomenon of merging of Earlier Egyptian grammatical categories, and particularly of nominal and other sentence types.
47
Uljas, in press.
308
chapter nine
CHAPTER NINE
AT THE CROSSROADS OF TEMPUS AND MODUS: THE ASPECTUAL-MODAL CORRESPONDENCE AND THE CONCEPTUAL FOUNDATIONS OF IRREALIS MODALITY IN EARLIER EGYPTIAN COMPLEMENTATION The Earlier Egyptian system of affirmative complement clause modality has been seen to be based on the variation of ntt/wnt-introduced construals and bare active suffix-conjugation forms. In general, the absence of a complementiser signals non-assertion, which is a reflection of diminished speaker commitment to or diminished relevance of what is said. However, unlike immutable forms of the active sDm=f and the sDm.n=f, the different forms of sDm=f of mutable classes divide the domain of irrealis further. Geminating/doubling forms of the sDm=f are expressive of the so-called proximal, non-geminating/-doubling ones of distal irrealis modality. The former is tantamount to lesser, the latter to greater degree of speaker non-commitment or information redundancy. Some of the forms used to express irrealis appear to be marked for this function, i.e. at least in complementation they have no other role, and some are not. This characterisation might suffice for the purposes of grammatical description, but explaining the use of the bare suffix-conjugation forms for irrealis modality requires a more holistic inspection of their semanticpragmatics in complementation and even beyond. On the whole, it seems that the modal characteristics of the bare complement suffixconjugation forms and their thus far largely overlooked temporal profiles are connected. The fixed modal function displayed by the forms analysed as being modally marked is paralleled by a fixed temporal and, more specifically, aspectual function both in complementation and elsewhere. By contrast, forms argued to be unmarked for modality and capable of being used in realis and irrealis propositions show corresponding variation also as regards their aspect value. The modal and temporal functions of the marked forms are also connected in a
tempus and modus
309
principled and non-arbitrary manner. The correspondences observed are reflective of underlying conceptual representations shared by the modal and temporal values associated with the forms. These essentially metaphorical schemata are not mere abstractions, but actually reveal the more general or ‘basic’ meanings of the Earlier Egyptian marked irrealis patterns. They are also recognisable in almost all other languages and represent some of the most basic mental imagery with which language users structure their experience for the purpose of linguistic communication. 9.1 The Interrelations between Temporal and Modal Functions of the Active Suffix-Conjugation Forms in Complementation and Beyond Besides their role as expressive of irrealis modality in Earlier Egyptian complementation, the bare active suffix-conjugation forms are also associated with various temporal values both in these and other environments. As will be seen shortly, rather than conveying distinctions of tense, or the locus of the situation in linear time, with the exception of the sDm.n=f the said forms make reference to the internal temporal composition of the situation described, i.e. aspect.1 However, there are important differences between the various forms in whether or not their temporal characteristics remain the same in all instances where they appear. These differences in turn seem to correspond to whether or not the forms have a fixed modal function. Forms analysed in the present work as marked for a particular (irrealis) modal function have the same aspectual function in all their uses, whereas forms without modal marking—the modally unmarked suffix-conjugation forms—are bereft of a fixed aspectual value. Below is a list of the various bare suffix-conjugation forms attested in complement clauses of Earlier Egyptian, divided according to their argued modal markedness and unmarkedness, but omitting the forms iwt=f, int=f and iit=f of the anomalous roots iwi, ini and ii, whose status vis-à-vis the latter characteristics will be discussed further below:2 1 Comrie 1976b remains the most accessible general work on aspect; e.g. Binnick 1991 is more thorough and up-to-date. The most thorough Egyptological discussion to date is Winand 2006. 2 See 9.3. For the division between modally marked and unmarked forms, see 2.4, end.
310
chapter nine Modally Unmarked
1) The immutable sDm=f of strong roots (e.g. stp=f) 2) The doubling sDm=f of doubling roots (e.g. mAA=f) 3) The non-geminating sDm=f of ult. inf. roots (e.g. hA=f) 4) The form iw=f of the anomalous root iwi 5) The form in=f of the anomalous root ini 6) The form i(i)=f of the anomalous root ii 7) The form (r)di=f of the anomalous root rdi 8) The form wn=f of the anomalous root wnn 9) The sDm.n=f
Modally Marked 1) Proximal: The geminating sDm=f of ult. inf. and the anomalous roots ini and rdi (hAA=f, inn=f, dd=f); the forms iy=f and wnn=f of the anomalous roots ii and wnn 2) Distal: The non-geminating sDm=f of ult. inf. and the anomalous root ii with endings -w/-y (hAw=f/hAy=f, iw=f); the short form of doubling roots (at least mA=f)
This division is, of course, essentially that seen in the introduction of the present study differentiating between specific modal forms and their functional counterparts.3 With one exception,4 each of the modally unmarked forms in the left-hand column above is also associated with more than one aspectual value either within or without complementation, or, in the one instance of the past tense sDm.n=f, is associated with a single temporal function that typologically differs from all the other forms. By contrast, each of the modally marked forms in the right-hand column is associated with one and only one aspectual value, seemingly regardless of where it occurs. Put another way, active suffix-conjugation forms with no fixed modal function that in complementation can thus be used to express both realis (when introduced by ntt/wnt) and irrealis (when appearing alone) neither have a fixed aspect function. Forms that at least in complementation are used exclusively for irrealis mood display the 3 4
See 0.2 above. See 9.3 below.
tempus and modus
311
same stabile character also as regards their aspectual characteristics. This division becomes clear when the temporal values of the above forms in complement clauses are compared with the ones manifest in their use(s) in other syntactic environments. For the unmarked class a good point of comparison are main clauses introduced by the auxiliary iw. After this element the unmarked forms, which, it should be borne in mind, are used for non-assertion in complementation, are not only used to make statements that look suspiciously as assertions. In addition, their temporal function there are as a rule the exact opposite to the one they display in complement clauses. Accordingly, the invariable sDm=f of immutable roots has a continuous, habitual or, as in the following example, generic meaning after iw, all characteristics of imperfective aspect (the question of exact definition of the aspectual values involved will be addressed shortly; the following characterisations are preliminary): (389) A diagnosis in a medical text: ir xA=k s Hr mn r ib=f iw=f mn=f gAb=f mnd=f gs n r ib=f iw Dd.tw r=f wAD pw If you examine a man suffering in his stomach and in his arm, his breast and the side of his stomach, one says concerning it: (pEbers 37, 10–11) “it is the wAD-disease”. In complement clauses the same form can have the same temporal profile, (example (390) below) but it may also express past completed terminative or non-immediate prospective meaning, both of which can be understood as subcategories of perfective aspect (391)–(392): (390) Hapuseneb ends his ‘autobiography’ with a plea for offerings from future generations:
iri=sn n=i xft Dd=i xpr n=sn mitt May they act for me according as I say, so that alike may happen (Urk IV 485, 4–5) to them. (391) The party of herdsmen follows the advice of their colleague and departs from the marshes:
HD.n rf tA dwA dwA iw iri mi Dd=f At dawn, very early, there was done according as he had said. (Herdsman 22–23)
312
chapter nine
(392) Queen Hatshepsut says concerning her role as the organiser of an expedition to the land of Punt:
iw=i r rdit Dd.tw n m-xt nfr wy xprt nn Hr=s I am going to make people say for the posterity: “How goodly is she through whom this happened”. (Urk IV 350, 8–9) Leaving aside for the moment the doubling sDm=f of roots such as mAA, the non-geminating form of ult. inf. roots has in complementation a consistently non-immediate future or past completed perfective meaning, ((393)–(394) below) whereas after iw (395) the same form has a generic, habitual or continuous imperfective sense: (393) In a medical text describing the preparation and use of an emetic potion:
wSa sam m Hnqt r pr ntt nbt m Xt=f To be chewed and made to swallow in beer so that everything that is in his (the patient’s) belly will come out. (pEbers 8, 15–16)
(394) Seri tells of his expedition to the quarrying-region: pr.n(=i) gr r Hwt-nbw tn r-sA pr(=i) r bhks n gm(=i) ø iry in ky mit(y=i) After I had gone to Behkes, I came here to Hatnub, and I have not found such having been done by someone else of my kind. (Hatnub 8, 2–4)
(395) A concluding note in a mortuary spell says of anyone who knows its content:
iw=f pr=f m xrw He comes forth by day.
(Tb 68/Nu pl. 18, 16)
Exactly the same holds also for the forms in=f, (396)–(398) rdi/di=f (399)–(401) and wn=f (402)–(404) of the anomalous roots ini, rdi and wnn:
tempus and modus
313
(396) A Letter to Dead gives background to a following plea: Tnw-r nw Dd=T n sA=T im int=k n(=i) pAawt wnm(=i) s(y) m in n=T sA=T im pAawt 7 This is a reminder of your saying to me, your son: “Please bring me some quails that I may eat them”, and of when I, your son, brought you seven quails. (Kaw Bowl, outside, 2–3) (397/166) The author tells her dead addressee that she would rather welcome death for his son than subordination to an adversary:
mr(=i) in=k n=k wn aAy r-gs=k r mAA zA=k xr zA izzy I would rather that you brought away to you the one who was here by your side than saw your son subordinated to Izezy’s son. (Cairo Linen 8–9) (398) Ptahhetep notes the importance of having friends: iw in.tw aqw wn Aq Close friends are brought when there is trouble.
(Ptahh 349)
(399) Sennetjer thanks Harakhti for his success in life: sqA=i tw mi di=k (wi) m-HAt rxyt I exalt you, according as you placed (me) before the plebs. (BM 65340, 7–8)
(400/148) Coffin Texts Spell 859 incorporates excerpts from the Pyramid Texts offering ritual:
wsir N pn mi n=k a n nbt-Hwt xw di=s sw r=sn Osiris N; take the arm of Nepthys. Prevent her from using it against them. (CT VII 62r–s) (401) The father of king Merikara stresses the importance of good works, because:
iw hrw wa di=f n nHH wnwt smnx=s n m-xt (Even) a single day contributes to eternity, and an hour provides for the future. (Merikara E 66–67)
314
chapter nine
(402) Ikuded reveals his motivation for setting up a monument in Abydos:
aHa.n iri.n(=i) maHat tn r rwd n nTr aA n-mr(w)t wn(=i) m Smsw=f Then I constructed this cenotaph at the Terrace of the Great God, so that I might enter among his followers. (Berlin 1199, 8–9)
(403) Nebamun tells of a royal act of favour: wHm.n n=i Hm=f Hswt nsw-bit(y) mn-xpr-ra di anx saA.n=f wi r wn=i r-HAt My lord, the dual king Menkheperra, given life, repeated favours for me and made me greater than I was before. 5 (Urk IV 150, 14–15)
(404) Prince Hardjedef informs his father of a famous magician: iw wn nD[s] Ddi rn=f There is a commoner whose name is Djedi. (pWestcar 6, 26–7, 1)
There appear to be no examples of the verb iwi ‘come’ after auxiliaries, no certain examples of it in adjunct clauses and even the second tense use is very rare.6 The form iw=f occurs in complementation, but its aspectual character there is extremely difficult to assess.7 Of the verb ii ‘come’ there are similarly no examples after auxiliaries or in complement clauses expressing past meaning. Yet, in the following example, cited as (122) above, the form iy=f occurs as an object of mAA n ‘see, look at’ and contrasts with the immediately following ii describing a generally occurring state of affairs:
5
Similarly Urk IV 879, 4. See, however, e.g. CT I 142f. 7 For instance, after verbs iw=f regularly refers to relative present situations or to states of affairs that seem to be indicated as currently realised. However, after prepositions there is much uncertainty: in past instances iw=f often clearly describes a completed state of affairs (see e.g. chapter 6 n.55 above for instances after r, ‘until…came’); cf. also instances such as e.g. Kaw Bowl, inside, 2 Tnw-r nw... m in=k xpS n kA m iw sA=k im “This is a reminder… of when you brought a foreleg of an ox and when your son came there”, with the forms m in=k and m iw N in close proximity and both describing past completed states of affairs. 6
tempus and modus
315
(405/122) The peasant, believing he is doomed, utters in despair: xsfw n ib m mw DAt-r n Xrd n sbnt m irtt ntf mwt nHy mA=f n iy=f ii wdf mwt=f r=f The approach of a thirsty man for water, the reaching of the mouth of an infant for milk—death is their lot. But as for one who prays that he may see it coming, tardily comes his death. (Peas B2, 119–22)
In contrast to these changing aspectual values of the modally unmarked sDm=f forms, which like their modal characteristics are clearly determined by use, the temporal properties of the modally marked irrealis forms remain the same in all their occurrences. From the examples quoted thus far it can be seen that after governing predicates the proximal irrealis forms of the sDm=f mostly have a relative present or generic time-reference, and although they may also refer to relative future, in such instances the situation is treated as actual at (i.e. temporally simultaneous with) the time of speaking. Similarly after prepositions, these forms mostly express continuous ‘general present’, but the situation may also equally well be located in the past or future. A past proximal preposition complement profiles the situation as ongoing or habitual, and in the future, depending on the inherent Aktionsart of the situation, as continuing unbroken from the ‘present’ or as more or less immediately adjacent to it. For illustration, one may compare the meaning of the marked form wnn of the auxiliary verb wnn, whose only function is tam-specification. wnn-complements may equally well be located in the past, ‘present’ or future: (406) The myth describes why Seth’s transformation into a pig made it detestable to gods:
xr m wnn8 Hr m Xrdt=f xpr xrwt=f m SA iw n mrt irt=f When Horus was (still) in his youth, his sacrificial animal came to be the pig, but (then) his eye had not yet suffered injury. (CT II 344b–345b)
8 So B2Bo, B4Bo, B2P, B4La, B4Lb, B1L, B17C and B1C; the variants S2P, S3P and S1C have m wn=f and B9C bare wnn=f.
316
chapter nine
(407) A song of palanquin-bearers: hr Xrw xwdt nfr s(y) m mH r wnn=s Sw.t(i) Happy are the ones carrying the chair! It is better full than (Deir el-Gebrawi II, pl. 8, top register, left) when it is empty.9 (408/246) Ukhhotep explains his reasons for depicting previous nomarchs in his tomb:
iri.n=i nw n-mrwt wnn rnw=sn mn n Dt I have done this in order that their names might be firm for (Meir III, pl. 11) eternity. In the following instance, rather than completion versus continuity, the variation between past wn and wnn appears to be motivated by the singularity/plurality of the subject and the resulting coherence versus ‘distributiveness’ of the situation:10 (409) The deceased says: ink iri.n nb wa... m hAb=f wat=f m wn=f way m prt m r=f m wnn HH=f n kA m sAH wnDwt=f I am one whom the Sole lord created… when he sent forth his only eye, when he was alone with what was to issue from his mouth, and when his million kA’s were the protection of his (CT III 382e–383e) companions. That is, the proximal forms carry distinct characteristics of imperfective aspect in complement environments. In contrast, the distally marked forms with the ending s -w/-y are well-nigh consistently future in relation to the governing verb and after prepositions. In such instances there appears to be always a notion of temporal separation and lack of immediacy involved between ‘now’, or the time of reference and the state of affairs described in the subordinate clause. Past reference of these forms is not attested after
9 Further identical examples are Reden und Rufe 52; CGC 1419; CGC 1536 and Meir V, pl. 31; see Altenmüller 1984–85. 10 See here the most insightful discussion in Schenkel 1965; cf. also Allen 1984, §§ 608–09; Meltzer 1991, 232–33; Jansen-Winkeln 1997.
tempus and modus
317
verbs at all and is extremely infrequent after prepositions as well.11 However, even if rare, it does not seem to be wholly excluded, and in these rare occurrences the forms describe the situation as completed and having reached its conclusion. Put another way, besides their distal irrealis character, in complement clauses the distal forms clearly behave consistently as expressive of perfective aspect. The rarity of the past use seems to suggest that these forms were seen as having a mostly prospective function. It is suggested here that the sort of ‘detached’ prospective seen with the -w/-y-ending sDm=f can be analysed as a subcategory of perfectivity.12 If these uses are compared with the other forms of employment of the irrealis-marked patterns, there is no discernible difference in their aspectual functions. For example, in second tenses, the most important class of employment of the proximal forms, the latter appear to retain their imperfective character: (410) The father of king Merikara describes the Sun God’s special care for men:
irr=f Ssp n ibw=sn sqdd=f r [m]AA st For their delight he brings forth dawn; he sails only to see them. (Merikara E 134) (411) The deceased reveals his knowledge of divine secrets: iw=i rx.kw Dw pw n bAXw nty pt rhn=s Hr=f wnn=f m THn(?)... wnn sbk nb bAXw m iAbt Dw pn wnn Hwt-nTr=f m Hrst wnn HfAw Hr wpt Dw pf I know this mountain of Bakhu upon which the heavens rest. It is of crystal (?)… East of this mountain is Sobek, lord of Bakhu. His temple is of carnelian and at the summit of that mountain there is a serpent.13 (CT II 375c–377c)
11
See n.72 in chapter 6 above. See 9.2 below. 13 So most versions; the variant B2Bo has iw sbk nb bAXw… instead of wnn. The variants B2P, B17c and B1C have iw wn HfAw instead of wnn HfAw Hr wpt Dw pf. 12
318
chapter nine
(412) The deceased is transformed into a divine falcon: xaa ra ra nb r sdm14 mdw N pn Ra appears every day only to hear the words of this N. (CT IV 59l)
(413) The king tells Sinuhe that his wandering in exile was not a result of royal wrath: dd tw xAst n xAst Hr sH n ib=k Ds=k15 One country used to pass you to another under the counsel of (Sin B 182–83) your very own heart. The situations described in (410)–(412) are clearly imperfective generic, continuous or iterative in character; (413) is past habitual. Here a few specific remarks are in order concerning the form wnn of the auxiliary thus written and of ‘injunctive’ second tenses. Concerning the first of these issues, it seems that all wnn-introduced initial adverbial and pseudo-verbal sentences can be analysed as technically ‘second tenses’ and temporally imperfective in character. When initiating adverbial sentences, wnn sometimes appears to be ‘emphatic’ and sometimes not, but this arises from the status of wnn as a lexical-semantically empty auxiliary. As will be discussed later on in more detail,16 in second tenses the initial verb appears to represent background, less than optimally relevant information, which leaves something else (usually an adjunct) as the carrier of the highest informative focus. Yet wnn represents merely a highly schematic situation with nil semantic content that never carries a degree of informative salience sufficient for it to become the central locus of interest of a ‘first tense’ expression. wnn-introduced adverbial or pseudo-verbal sentences are thus, paradoxically, both ‘always’, and ‘never’, ‘emphatic’ and ‘second tenses’. Further, although the function of wnn in such instances also often seems to be ‘future conversion’, given the similarly common non-future uses, (such as in (411) above) this is hardly its true function. This is not the time or place to enter into a detailed discussion of the aspectual complexities of second tenses; suffice it to say that the ‘futurity’ involved in many instances of wnn-introduced adverbial/pseudo-verbal sentences is 14 15 16
Sic, not sDm. The Ds=k appears in the Ashmolean Ostracon; the B-version has ib=k n=k. See section 10.3 in the conclusion.
tempus and modus
319
often more correctly understood as involving a continuous unbroken development from and including the time of speaking/reference: (414) A decree in the last will of Ankhreni: ir grt nA.n awt qd.n n=i sn=i... wnn tAy=i Hmt im Now, concerning the living quarters, which my brother… built for me, my wife shall be/remain there. (pUC 32058, rto. 13) (415) Ammunenshi says to Sinuhe: mk tw aA wnn=k Hna=i Look, you are here, and you will stay with me.
(Sin B 77)
It has also been recognised for long that some second tenses with 2nd person subjects appear to be used to make requests and commands for which the characterisation as imperfective seems unsuitable.17 The following example could be said to represent something of a bald king of France in Egyptological linguistics: (416) A request in a letter delivered to a temple scribe: imi in.t(w) mskA n iwA mskA n awt r-pw dd=k sw n Tbw ptHwr-rn Let there be brought a hide of an ox or a hide of a goat. You (are to) give it only to the sandal-maker Ptahwerren. (Lesestücke 97, 6–7)
Yet, it is questionable whether injunctive (or any deontic) meaning really is an inherent property of the geminating sDm=f and the other proximal forms.18 At the face of it, in terms of its grammatical tempus the verb in the above example seems simply to refer to a generic 17 This is, of course, the old Gardiner-Polotsky chestnut, admitted by the former as being a problem to his analysis of the geminating sDm=f as imperfective and expressive of ‘repetition and continuity’ (Gardiner 1947, 99; GEG § 440.5; cf. Polotsky 1964, 278. See also n.27 and section 10.3 below). 18 Second tenses with subjects other than the 2nd person may also be rendered e.g. as deontic obligations (Man and Ba 117–18 snw bin inn.tw m DrDrw r mtt nt ib “Brothers are evil and one must have recourse to strangers for honesty”) or epistemic possibilities (Sin B 152 rww s tA=f n HAyt “A man may forsake his country because of nakedness”) etc. It is also notable that negated second tenses, even in the 2nd person, never translate as injunctions; i.e. a sentence such as *tm=k rdi sw n Tbw cannot be rendered as “You are not to/may you not give it to the sandal-maker”.
320
chapter nine
occurrence ‘you give it to the sandal-maker’, but this need not be how the proposition is to be interpreted as a speech act. Given a suitable linguistic and extra-linguistic context such as that above, where the recipient is well aware that the speaker cannot be referring to a habitual or generally occurring situation, an injunctive interpretation of the sentence is the most obvious one. Provided that the context of the utterance allows this, the same holds, of course, also to the English translation “you give it to the sandal-maker”. In other words, in the particular context of its appearance, the sentence is to be understood as an indirect speech act from the speaker’s part.19 The imperfective character of the proximal forms thus seems fixed throughout their uses, and the same phenomenon is observable also with the distal forms and perfective aspect. As noted, in complementation the use of the -w and -y-ending sDm=f for past reference is of extreme rarity; in initial environments such use is completely non-existent, again seemingly bearing witness to the mostly prospective function of these forms.20 Instead, they are used to make promises, wishes, expressions of deontic necessity and other obviously modal statements referring to non-immediate future states of affairs: (417) A truncated Appeal to the Living of Ipy near the Nubian fortress of Semna:
iri=Tn ixt n itw=Tn sHtpw=Tn nTrw=Tn swDw=Tn iAwt=Tn n imw-xt xdw=Tn m Htp mi Dd=Tn Htp-di-nsw... You will offer to your fathers, you will satisfy your gods, you will hand over your offices to (your) successors and you will sail north (i.e. home) in peace, just as you say: “An offering which (Felsinschriften no. 530, 1–4) the king gives…”21 (418) Senbu begins his letter with a salutation: iry n=k Hr(y)-S=f nb nn-nsw ib=k May Herishef, lord of Nennisu make your mood for you! (pUC 32206, 2–3)
19 See Junge 1978a, 122–23 for a very similar view as that expressed here; cf. also Winand 2006, 284 and 10.3 below. 20 However, forms with -y are common in the non-prospective negation n sp. 21 See also Felsinschriften no. 531.
tempus and modus
321
(419) Words of a man who has just seen his wife and children perish on (or by) a lake:
n rm=i n tfA mst... mHy=i Hr msw=s sdw m swHt mAw Hr n xnty n-anxt=sn I will22 not weep for that child-bearer… but I must share a thought for her children who were crushed in the egg and saw the face of the crocodile-god before they had even lived. (Man and Ba 76–80)
These temporal-modal correspondences with the marked and unmarked suffix-conjugation forms have major effects to the present analysis. But specifically the interrelation between modal and aspectual function shown by the specialised proximal and distal irrealis forms warrants a closer inspection seeing that it turns out to be systematic also in terms of modal/aspectual typology. It is not coincidental that proximal irrealis modality and imperfective aspect and distal irrealis modality and perfective aspect should be paired. On the contrary, this reflects a strategy of grammatical coding driven by conceptual similarities between the relevant aspectual and modal meanings that results in them being associated with the same morphological patterns. This also explains why the modally bi-functional unmarked forms show no similar match between tempus and modus. The character of the said correspondences follows a number of simple iconic principles the like of which can be found not only in Earlier Egyptian, but also in languages generally. 9.2 Too Far or too Close? The Modal-Aspectual Continuum of the Earlier Egyptian Irrealis-Marked sDm=f Forms The characterisation of the irrealis-marked sDm=f forms as temporally perfective and imperfective opens up a particularly untidy can of worms. Aspect and the categories perfective and imperfective have historically been among the most disputed issues in Egyptological language studies.23 Prior to the advent of the ST, the general theory 22
Future reference seems to be called here, but the writing is clearly n, not nn. For historical and theoretical summaries of the research on aspect in Earlier Egyptian, particularly in respect with the sDm=f formation, see e.g. Thacker 1954, 190–93; Depuydt 1983, 19–25 and passim; Hannig 1987; Junge 1989, 30–34; Polotsky 1990; Schenkel 1990, 103–04, 149–58; Ritter 1995, 45–60; Winand 2006, 12–18. 23
322
chapter nine
of suffix-conjugation forms was firmly based, in a Semitist tradition, on the assumption of a fundamentally aspectual character of the various forms, which culminated in Gardiner’s analysis of the geminating and doubling forms of the active sDm=f as expressive of imperfective, and those without these morphological characteristics, of perfective aspect.24 Polotsky’s discontent with the views of Gardiner et al on the nature of the sDm=f forms and especially of their ‘nominal’ use was partly a reaction against the obvious shortcomings of the ‘aspectual hypothesis’ as an all-embracing explanatory model. For example, it is inexplicable why ultimae infirmae roots should show gemination e.g. in complement clauses but not e.g. in iw-sentences if both simply express imperfective aspect.25 The solution of the ST was to draw a clear morphological line between these forms and then to introduce the concept of ‘nominal forms’ based on the alleged possibility of generalising forms out of syntactic position.26 However, particularly in accounting for the mutual distribution of the ‘nominal forms’, the geminating/imperfective versus non-geminating/perfective dichotomy maintained its status in discussions also after the ‘Polotskyan revolution’, even if there has been no shortage of revisions of its terminological and definitional basis or internal hierarchy.27 24
See GEG §§ 295, 438–39, 447–48. See e.g. GEG §§ 462 & 463 where these roots are said to employ the ‘perfective sDm=f’ after iw, but where its meaning is nonetheless defined as imperfective. This obvious discrepancy is then argued to result from the presence of iw rather than from the form itself. However, if the sDm=f aspectual profile can be thus changed by an extraneous element, this nullifies the semantic basis of Gardiner’s ‘perfective sDm=f’ rather completely. 26 See 0.1.1 and 0.2 above. 27 Gardiner in particular has received much scorn for his famous characterisation of imperfectivity as ‘repetition and continuity’ (GEG § 295). Various authors (Assmann 1974, 63; Depuydt 1983, 24; Loprieno 1984, 88 n.5; Satzinger 1986, 300) have seen in this confusion between aspect and Aktionsart, others (Vernus 1986, 378-79) between iterative (several repeated actions) and durative (continuous, unbroken). For a summary of attempts to align the imperfective/perfective opposition with tense and Aktionsart or to replace/integrate it with the concepts ‘accompli’ and ‘inaccompli’, see Ritter 1995, 45–60; Winand 2006 provides a full synthesis thereof. Also the usefulness of linguistic analogues, particularly Slavic aspect, has been variously asserted (Hannig 1984, 65; Loprieno 1984, 90; 1986b, 263; Eyre 1989, 58–59; Reintges 1997, 104) and denied (Satzinger 1986, 300–05). The imperfective profile of the geminating sDm=f has seldom been disputed (but see Polotsky 1944 § 31, Appendix; 1964, 275–84 versus 1976, 2.3.1; cf. Junge 1989, 31 n.18). For Winand (2006, 281, 283 and passim) the form is more particularly ‘inaccompli global’, and overarches other imperfective subcategories. The assumed ‘prospective’ form(s) have been characterised as perfective e.g. by Eyre, (1985, 61; cf. 1987, 40) Loprieno (see below) Satzinger (1989, 220) and Winand (2006, 189), 25
tempus and modus
323
Yet, other things being equal, there is some consensus among scholars over what imperfective and perfective ‘internal temporal constituency’ denote more broadly. When describing the phenomenon of aspect, linguistic analyses recurrently refer to a ‘point of perspective’ from which the situation aspectual profile is assessed. This ‘viewpoint’, or ‘aspect locus’,28 is best understood as the ‘location’ where the speaker is metaphorically “situated” in relation to the instantiation and from which the latter is viewed. Perfective aspect is said to imply that this viewpoint is external to the duration of the event described.29 In contrast, the characteristic profile of imperfective states of affairs involves an internal perspective. This definition for imperfective and perfective aspect is widely accepted among Egyptologists.30 A further (and somewhat better) metaphor for the same opposition is that externally-viewed perfective situations seem bounded or closed in character, whereas the internal vantage point to imperfective states of affairs gives them an open appearance. In Egyptology, this characterisation has been masterly adapted by Loprieno, who views perfective and imperfective aspect in general as presence or absence of an abstract property of ‘closure’ of the event described.31 Perfective situations are indicated as ‘closed’ whereas the ‘openness’ of imperfective aspect is tantamount to lack of such a feature. Yet, the ‘location’ of the ‘closure-mark’ (Grenzsymbol) is not predetermined and may lie at the ‘end’ [___#] or at the ‘beginning’ [#___] of the situation duration. In the first instance the result is a past completed, in the latter a non-immediate,
but see Roccati 1979, 46 for a contrary view regarding the ‘sDmw=f’. For Hannig (1982, 47) the ‘prospective’ occupies a position between imperfective and perfective; also for Vernus (1986, 376) such “mere modal forms” are aspectually neutral. 28 Timberlake 1982, 310. 29 This characterisation of perfective, and that of imperfective below, derives from the writings of early Latin grammarians and was adopted in 19th century studies on Slavic aspect; see Binnick 1991 135–36, 157; Winand 2006, 180, 184. Its most succinct formulation in general linguistics is Comrie 1976, 4. 30 See e.g. Assmann 1974, 62; Hannig 1982, 44, 52; Borghouts 1985, 30; Ritter 1995, 66, and in fact already Gunn 1924, 110. But see Winand 2006, 180–81, 184–85 and the next two notes below. 31 Loprieno 1984, 88–89; 1986a, 19, 22; 1986b, 263. For counter-arguments, see Satzinger 1987, 619; 1989, 207 and reply in Loprieno 1991a, 210 n.41. Loprieno’s analysis is modified by Winand (2006, 181–82, 185) according to whom perfective aspect involves singling out part of the situation as particularly salient and imperfective focus on some internal point of the situation and either ignoring its termination (‘global’ imperfective) or taking the limits thereof into account (progressive).
324
chapter nine
‘detached’ prospective event.32 However, regardless of this difference, in both the situation appears ‘closed off’ from the speaker’s ‘point of perspective’. Loprieno’s hypothesis has elements that are open to doubt and in contradiction with the views of the present work,33 but somewhat adapted, it accounts admirably for the perceived temporal characteristics of the irrealis-marked sDm=f forms in complementation and elsewhere. If it is assumed that the proximal forms contain an inherent feature of ‘lack of closure’ in their functional profile, this allows them to describe situations located throughout linear time, whence their ‘eternal’, habitual, continuous and ‘plural’ imperfective profile. In contrast, an assumed situation-initial or -final ‘closure-mark’ inherent to the -w/-y-ending sDm=f accounts for their non-immediate prospective reference and explains also their apparent potential for being used also for past completed reference, even if this seems to have been but rarely exploited in complementation and not at all elsewhere.34 The idea of perfective aspect as ‘closure’ also removes the last semantic motivation for postulates of non-geminating forms without the said endings as concealing various indistinguishable ‘not necessarily’ future ‘prospective’ etc. sDm=f and sDmw=f forms. That is, since these forms can assume a perfective value as functional counterparts of the marked forms with -w/-y, nothing prevents them then from being used equally 32
However, just as the latter cannot be seen as ‘completed’ situation, it can neither be interpreted as complete (cf. Winand 2006, 180). Yet, the concept ‘closed’ does not mean enclosed. The situation is closed vis-à-vis the conceptualiser, but it need not have (at least ‘visible’) end-point. 33 The ‘closure’ is, according to Loprieno, an inherent marking-property of certain mutations of the verbal root, including the ‘prospective sDmw=f’. At least in the classical language, this is not found in all verb-classes, which instead use other forms, e.g. the ‘prospective sDm=f’ (=Allen’s ‘subjunctive’) and the ‘gewöhnliche’ sDm=f as its functional counterparts (Loprieno 1984, 95, 97; 1986a, 33 and n.20–22; 38 n.39). Except for the ‘prospective sDm=f’, this theorem resembles the view adopted in the present work, but Loprieno introduces another semantic level of these forms, namely +/-realisation and labels the ‘prospective’ perfective/-realised and the sDmt=f form (including iwt=f/int=f/iit=f—see below) perfective/+realised (1986, 32, 34). The geminating sDm=f is seen only as ‘continuative’ (a subcategory of imperfective) and hence also ‘unrealised’ (ibid, 23) whereas the ‘gewöhnliche’ sDm=f is ‘neutral’ in this respect (ibid, 27). The definitional basis of these purely ontological labels is questionable—in what sense is a continuative situation ‘unrealised’?—and they are not justified by the evidence: e.g. habitual sense of the geminating sDm=f is commonplace (see e.g. (413) above) and the forms iwt=f/int=f/iit=f are certainly used for ‘unrealised’ in complementation. Loprieno also suggests that his system pertains only to the earliest stratum of the language, giving way subsequently to the ‘transposition’-system of the ST (1986a, 83–87, and passim; see also n.73 in the introduction). 34 For a speculation on the possible reasons for this, see 9.3 below.
tempus and modus
325
for both future and past reference.35 As seen from their behaviour in preposition complements, the correct interpretation is usually derived from the context without difficulty. Yet, the implications of these notions go far beyond morphological and tempus-related issues and extend also to the modal properties of the said sDm=f forms in complementation and possibly beyond.36 The above metaphors used to describe the aspectual oppositions can be used to explain why the particular temporal values of the marked irrealis forms ‘go together’ so naturally with their distinct modal characteristics. Cross-linguistically, the various categories of tempus are not limited to their basic referential role of describing time, but they also have various modal functions. For example, past tense in its different degrees is commonly used for a very specific purpose.37 In the following English sentences the tense variation marks hierarchies of tentativeness, hypotheticality, doubt and deontic force: (lx)
35
It may/might be true If I have/had time, I will/would come You may/might stay at home
Here a special mention should be made of the form wn of the auxiliary wnn (see chapter 6 n.71 above). This writing has been argued to represent separate (but, of course, once again morphologically indistinguishable) future and past ‘converters’ (Vernus 1990, 49–51; Ritter 1995, 189) or has been interpreted as an idiosyncratic spelling for ‘wn.n=f’ (EAG § 533, 2; Doret 1980, 40–41—in Doret 1979b n wnn=f is analysed as ‘n sDm.n=f’!—and cf. already GEG § 414.1, Add. to § 413). If the separate ‘converters’ truly existed, either wnn would have to be the only verb in the entire lexicon exhibiting such a split, or all past and future non-geminating sDm=f’s would also have to represent different forms. Vernus chose the latter alternative with his prospective sDmw=f that was “not bound to convey future meaning” (see 6.3 and n.71 therein). Instances of a relative sDm.n=f of the auxiliary wnn occur in Old Egyptian, (e.g. Urk I 217, 6 wnw.n=sn DAs(=i) Hna=sn “with whom I had an argument”). However, in non-relative environments the form is consistently wn. Of course, the assumption that in the latter instance no vowel fell between the two n:s, resulting in a singular writing is quite plausible, but it may be asked is it necessary seeing that the analysis proposed above allows wn, like wnn, to be seen as a unit both morphologically and semantically. For doubt of the possible sDm.n=f of wnn, see Gunn 1924, 104 n.1; T. James 1962, 103 and Roccati 1979, 47. 36 The following discussion pertains to complementation only; see below for the possibility of extending the analysis outside this grammatical category. 37 D. James 1982; Wallace 1982, 202–03; Fleischman 1989; Bybee 1995; Palmer 2000, 203-21. Similarly, in “I can/could lift 200 lb”, the use of past could implies conditional/lesser commitment to success. Cf. also (xlvi)–(xlvii) in 3.3 above for the variation of past and present after the verbs wish and hope.
326
chapter nine
The following instance is particularly illuminating:38 (lxi)
The ancients thought that the sun moved round the earth; they did not know that it is the earth that moves round the sun.
The variation of tense here obviously has nothing to do with timelocation, but indicates the extent to which the speaker accepts the complement proposition—i.e. its assertivity. These phenomena are based on a metaphorical link drawn by language users between temporal and modal ‘distance’ that allows the former to be used to indicate degrees of the latter.39 This is not a prerogative of tense, however; the following remarks apply quite generally:40 Whatever is happening ‘now’ (and ‘here’), i.e. whatever is proximate to the time and place of the speech event, can normally be vouched for by the speaker, who experiences it as actual and real. However, a situation which takes place ‘not-now’ and ‘not-here’, i.e. one which is distant from the time and place of utterance, cannot be vouched for by the speaker in the same way: it is not experienced as actuality and reality… It is thus reasonable to see the move from spatial and temporal proximity to a broader and more abstract conceptual and cognitive proximity (actuality/reality), and similarly, from spatial and temporal distance to a more abstract conceptual and cognitive distance (non-actuality/non-reality).
The metaphor of ‘distance’ is a cognitive ‘primitive’ associated with, and used to express irrealis modality in great many languages.41 The same principle underlies also the use of the irrealis-marked sDm=f forms in Earlier Egyptian complementation as well as their peculiar distal and proximal characteristics there (and, one suspects, elsewhere). As regards the expression of speaker commitment, the characteristic greater ‘distance’ of the distal forms is a result of their generally ‘external’ and detached vantage point to the situation described. 38
See Palmer 1986, 167. Fleischman 1989, 2–3; Sweetser 1990, 55–56; of course, ‘temporal distance’ is in itself already a metaphor based on the perceived similarity between time and space (Fleischman ibid, 39 n.3). 40 Fleischman 1989, 2; emphasis in the original. 41 Cf. Steele 1975, 216–17 and Slobin & Aksu 1982, 196–98. Vidal & Klein 1998 present an interesting case of a language in process of grammaticalising irrealis from a general marker of spatial distance; cf. also Hardy & Gordon 1980, 193. 39
tempus and modus
327
This is manifest as ‘closed’ perfective aspect in the domain of temporal reference, whereas in the domain of modality the ‘distance’ functions as a force-dynamics barrier that sets the instantiation described more beyond manipulation, certainty and commitment.42 In contrast, the greater ‘proximity’ of the proximal forms of the sDm=f is a result of the generally less ‘disassociated’ speaker perspective to the situation described that is manifest both aspectually and modally as absence of a temporal and force-dynamics barrier. Although part of the overall category of non-assertion, a proximal situation is less strongly irrealis and higher on the scale of subjective speaker commitment than its distal counterpart. The connection between the concepts near/far and temporal/ modal function also accounts for the respective profiles of the irrealismarked sDm=f forms with regards the discourse relevance of the information conveyed. Because irrealis modality can be a signal of reduced relevance, as a modal category it belongs to the discourse background.43 Yet, as with commitment, there are again degrees involved in this, and in Earlier Egyptian this finds expression grammatically. With the proximal forms, the ‘situation-internal’ vantage point, which translates into an imperfective aspect profile, can be likened to standing on a circumscribed geographical feature, e.g. a field. From such an angle the features in the immediate vicinity are likely to attract most attention, whereas the ‘boundaries’ separating the field from the surrounding landscape are not primarily focussed upon and may be wholly invisible. Lunn describes this with a visual parable of ‘aspectual lens’.44 By adopting an imperfective ‘internal’ vantage point, the speaker may be metaphorically ‘too close’ to obtain a focussed view to the totality of the situation, which appears as lacking in coherence and ‘out of focus’.45 Imperfective profiling is thus eminently suitable for concentrating on the internal development and character of a situation, but the result of this is that the state of affairs itself tends to fade away from “the realm of precision and concrete reality into that of vagueness and unreality”46 and to receive mere ‘reduced 42
A comparable system is found e.g. in Itzaj Maya (Hofling 1998). See 0.1.2 above. 44 Lunn 1985. 45 Lunn 1985, 57. 46 Wallace 1982, 204; negated perfective is actually referred to, but the description seems particularly apt for imperfective; see also D. James 1982, 399–400. 43
328
chapter nine
assertion’.47 This is why the proximal irrealis forms express not only imperfective aspect but also signal the situations described as being of mitigated discourse relevance. They remain relatively ‘close’ to the speaker, but appear ‘out of focus’ and are thus associated with lesser relevance than more ‘clearly’ conceptualised realis states of affairs. Further manifestations of this same phenomenon are e.g. the grammatical links in many languages between habitual and irrealis noted earlier,48 or the use of imperfective in Russian to indicate that the verb and its complements are not the focus of the sentence, but e.g. an optional adverbial adjunct:49 (lxii)
V etoy porternoy ya obdumyval svoyu dissertatsiyu i napisal (PERF) pervoe lyubovnoe pis’mo k Vere. Pisal (IMPF) karandashom. “In this tavern I pondered my thesis and wrote my first love letter to Vera. I wrote it in pencil”
Another example is the often noted use of imperfective in narrative to create background or setting for more focal situations that form the ‘backbone’ of the ‘story’ and carry it forward.50 Not coincidentally, the latter are often perfective in character. The external viewpoint of this aspectual category renders the totality of the situation as the primary focus rather than its internal detail or some extraneous landmark. Perfective situations thus appear as sharply individuated and defined ‘figures’ “isolable from their contexts and easy to perceive as wholes”.51 In Earlier Egyptian, the generally distal profile of forms so labelled results in a perfective temporal profile of exactly this sort, but in terms 47
Hopper 1981, 238. See 0.1.2. In Italian and French, imperfective aspect is also common in descriptions of dreams, hallucinations and other semi-conscious states in which the speaker’s ability to focus on the events narrated is impaired; see Lunn 1985; Fleischman 1995, 530–32. 49 Hopper 1981, 218–19. 50 E.g. Hopper 1981, 213–15; Wallace 1982, 208–09 and numerous others. 51 Lunn 1985, 52. The differences between the inherent figure/ground and focus/ de-focus-properties of perfective and imperfective aspect have been widely commented upon in linguistics. Besides the works noted in notes 42–48 above, fundamental discussions on this topic are also e.g. Hopper 1979; Hopper & Thompson 1980, 281–88; Reinhart 1984 791–75; Fleischman 1985; Chvany 1985 and Waugh & MonvilleBurston 1986; see also Binnick 1991, 378–83. 48
tempus and modus
329
of discourse relevance the outcome of this is not to profile the situation as ‘most important’ information, as one might expect from its status as a ‘figure’.52 On the contrary, the associated ‘distance’ means that the state of affairs described is metaphorically so far away that it is not only difficult to ascertain and control (i.e. to treat in committed terms) but also is almost completely irrelevant. This is intuitively obvious: whatever is far removed from one’s primary sphere of action and influence tends to be equally removed from one’s sphere of interest.53 Consequently, whereas in Earlier Egyptian, complement clauses with proximal irrealis still ‘locate’ the situation close or within the speaker’s mental and deictic scope, the distal forms profile them as outside the speaker’s deictic sphere as well as beyond verification, belief and as notably insignificant in the current context of communication. 9.3 Interim Summary The above remarks on the temporal properties of the Earlier Egyptian active sDm=f forms have been kept deliberately brief for the simple reason that the present work is first and foremost a study on modality rather than aspect. Further research on the interweaving on these two domains of meaning is certainly called for. Nevertheless, the manner in which the said forms behave in terms of their modal and aspectual functions in complementation and elsewhere appears to show clear correspondence between these two semantic-pragmatic areas both grammatically and conceptually. Forms that in complementation display a fixed irrealis modal function also have a fixed aspectual function across their uses, whereas forms lacking a fixed modal function in complementation do not. There is one apparent exception to this. The doubling form of doubling roots, analysed as belonging to the latter group, has an imperfective function in complementation, in second tenses as well as e.g. after iw:
52 See here Reinhart 1984, 787 and Croft 2001, 332–33, who similarly point out that figure (their ‘foreground’) does not equal focus of information. 53 As an illustration of this point, one may consider e.g. the differences in people’s usual response to humanitarian disasters closer at home versus in more distant lands.
330
chapter nine
(420) The deceased says to certain divinities: stkn=Tn bA=i Hna=Tn r pr wsir sDm=f mi sDm=Tn mAA=f mi mAA=Tn Bring my soul with you to the house of Osiris and it will hear like you hear and see like you see. (Tb 1, 23–24/Nu pl. 12, 6) (421) The deceased identifies himself with the divine: mAA=i mAA Sw (CT V 107f/T1C, M2C) Whenever I see, Shu sees. (422) It is said of the successful dead who has entered the hereafter: ir s nb nt(y) im iw=f mAA=f wsir ra nb As for anyone who is there, he sees Osiris every day. (CT VII 365g–366a)
No explanation for this phenomenon seems readily available. It may simply be an ‘exception’—no linguistic system is ever 100% regular. Nevertheless, the distinction still holds with all other root-classes. This analysis has far-reaching repercussions to the present hypothesis, to the wider study of active sDm=f forms as well as the other suffix-conjugation patterns, and the syntactic constructions in which they occur. Firstly, the correspondence between tempus and modus displayed by the specialised irrealis forms sets them further apart from their non-specialised counterparts. Although the latter may assume the same temporal and modal functions as the marked forms, their employment in and particularly outside complementation shows that they are not specialised in either respect. Secondly, the correspondences between a fixed aspectual and irrealis modal function give reasons to suspect that forms displaying this interrelation are not used for non-assertion only in complementation, but that they are so used in all their occurrences. That is, forms with varying aspectual values have varying modal values, depending on where they are used, but forms displaying a fixed aspectual value, which one may thus view as marked in this respect, may well also have a fixed irrealis modal value and be universally marked for irrealis and non-assertion. This of course has considerable consequences to the analysis of e.g. second tenses, where the geminating sDm=f is ubiquitous, as well as e.g. on how the sDm.n=f and the forms iwt=f,
tempus and modus
331
int=f and iit=f of the anomalous roots iwi, ini and ii should be viewed. Concerning the sDm.n=f, this form, as noted before, is marked not for aspect but tense, which explains its belonging to the modally unmarked class. As for iwt=f, int=f and iit=f, it has been argued here that these forms do not belong to the sDm=f formation, but are rather sDmt=f forms.54 This hypothesis derives from Loprieno, who defines them and the sDmt=f in general as a specific marked perfective form which these verbs use in complementation.55 Given the lack of other convincing explanations for the presence of the element -t, the fact that the sDmt=f is certainly used as a complement of (at least) the prepositions r and Dr, and that in complementation the use of iwt/int/iit=f parallels exactly that of the distal (=perfective) irrealis sDm=f forms with the endings -w/-y, this suggestion has great appeal. But it is these very same properties that also seem to suggest that the sDmt=f should be analysed as a form marked for irrealis modality. Again, also this matter requires further study, but one may note a yet further phenomenon seemingly pointing towards this same conclusion, namely the negation of the sDmt=f with tm:56 (423) Ptahhetep recommends finding out the motives of the opinions of one’s friend:
iri sp Hna=f waw r tmt=k mn xrt=f Deal with him privately until you are no longer vexed by his (Ptahh 465–66) stance. Thirdly, the tight correspondence between modus and aspect shows that the assignment of temporal functions on the various types of sDm=f is anything but unprincipled. The price of the hypothesis proposed in the present work has from the start been abandonment of the assumed one-to-one match between suffix-conjugation form and both modal and temporal function. However, this holds only for the unmarked forms, and even with them the variation is rule-governed. As revealed in this study, at least in complementation and possibly elsewhere, the marked irrealis modal forms correspond to the ideal of ‘one form, one 54
See 0.2 above. Loprieno 1984, 92, 99; 1986a, 34–35, 45. Yet, according to Winand (1991, 378– 79) and Zonhoven, (1996, 621) the sDmt=f is not formed of the root iwi. 56 See also Naga ed-Deir Letter to Dead, vso. (r tmt=f mAA). 55
332
chapter nine
function’ both in terms of tempus and modus. Modally unmarked suffixconjugation forms show systematic binary variation of both pragmatic and temporal function. There is nothing chaotic or arbitrary in this organisation. Assuming that a stabile or marked aspectual function truly is a signal of a marked modal function generally, the hypothetical temporal-modal system of Earlier Egyptian active suffix-conjugation forms may be summarised as follows:57 modality marked Proximal, Imperfective Other: sDm=f : hAA=f inn=f, iy=f, dd=f, wnn=f
–
modality unmarked
Distal, Perfective Other: sDm=f: hAw=f, hAy=f mA=f
aspect marked
Other:
sDm=f:
i w t = f , stp=f i n t = f , hA=f iit=f in=f, ii=f, wn=f
sDm.n=f
aspect unmarked
The unmarked forms in the two right-hand side columns display counterpart-relations with the marked forms at the left in various ways. Here a specific note is in order concerning the apparent prospective marking of the distal forms. Whether or not their prospective value is a result of a diachronic specialisation whereby the earlier ‘symmetric’ prospective cum past completed perfective forms relinquished the latter function (or situation-final ‘closure mark’) cannot be said with certainty, but this remains a possibility.58 In most historical Earlier Egyptian, in past completed complements the marked distal forms are replaced primarily by their unmarked (-w/-y-less) sDm=f counterparts and secondarily by the sDm.n=f past tense. In initial environments the sDm.n=f is the only form used for this counterpart function. However, it may well be that this organisation does not represent an original state of affairs. 57 The problematic doubling sDm=f of doubling roots has been omitted from the diagram. 58 Yet, the possible role played in this postulated development by the moribund ‘indicative’ sDm=f occurring in Old Egyptian as a past completed of transitive verbs with noun subjects is wholly unclear. Cf. here n.98 in the introduction.
tempus and modus
333
Fourthly and finally, the correspondence between temporal and modal function seen with the irrealis-marked sDm=f forms in Earlier Egyptian complementation reflects and results from the metaphorical ‘viewpoints’ to the situations described. This interrelationship need not be viewed as derivative; i.e., for example the proximal irrealis character of geminating sDm=f is not necessarily a corollary of its aspectual properties.59 Instead, their co-occurrence is based on the perceived similarity of the said aspectual and modal functions with the more abstract concepts distant/close, and this is, in turn, used as a guiding principle in coding semantic-pragmatic content onto grammar.60 Imagination is capable of conjuring up an unlimited array of thoughts, ideas and ‘meanings’ that one may wish to express using language, but there does not and cannot exist a unique and specific expression for each and every meaning. Thus there arises a need to ‘map’ meaning onto e.g. verb forms in a manner that is both economical and ‘makes sense’ intuitively. There are various ways for achieving this, but all reflect the same principle of associating the same expressions with such meanings that are viewed to share some general ‘family resemblance’ or ‘ideational’ similarity that makes them ‘natural’ associates. The organisation and expression of irrealis in Early Egyptian complementation provides a glimpse into the inherent ‘logic’ of this associating of semantic-pragmatics with grammatical expressions which is both elegant and approachable. In fact, results of similar processes are found throughout this language. For example, the grammaticalisation of the combination [preposition Hr/m/r + infinitive] as a pseudo-verbal tense-aspect pattern results from a metaphorical application of spatial concepts for temporal description.61 Such a ‘mental leap’ is possible because the imagery used to cognise time and space is essentially the same. Thus, the mechanisms of organising temporal-modal meaning onto the specialised irrealis sDm=f forms in 59 Cf. e.g. Thacker (1954, 212) who claims that the ‘energic’ sense of the geminating sDm=f is derived from its earlier imperfective meaning. Borghouts (1985) argues similarly for the ‘howness’ of this form. However, the possibility of derivation cannot be excluded out of hand; studies on the development of TAM-systems suggest that diachronically temporal functions of grammatical patterns precede possible modal ones (see Givón 1982 for discussion). 60 Cf. Waugh & Monville-Burston 1986, 853, who note on what they call ‘detachment’, a synonym to ‘distance’, that it is “more general than any of its contextualizations… it simply means separation or dissociation within a universe”. 61 See the fine discussion in Collier 1994, 60–67.
334
chapter nine
Earlier Egyptian complementation do not represent a curiosity of this grammatical domain alone, nor are they restricted to this language. The more abstract characteristics and functional overlaps of the Earlier Egyptian organisation of irrealis modality in the constructions studied emphasise its deeper similarities with other modal systems. They also allow it to be viewed as a part of a larger non-atomic ‘meaningcontinuum’, which ultimately covers and extends over all the categories of tam.
conclusion: retrospect and prospect
335
CHAPTER TEN
CONCLUSION: RETROSPECT AND PROSPECT 10.1 Summary The modal system revealed by Earlier Egyptian complementation is unparalleled among languages of the ancient Middle East both in its extent of grammaticalisation and richness of expressive force. To recapitulate; in Earlier Egyptian the grammar of non-infinitival complement clauses is generally based on the expression of modality. The primary division is between realis assertion and irrealis nonassertion. Complement propositions, towards which the speaker entertains a high degree of commitment and acceptance and which are presented as new and high in information value and relevance, are asserted and marked as realis by the specific realis operators ntt/ wnt/iwt. Clauses describing states of affairs low or lacking in one or more of these characteristics are not asserted, but are treated as grammatically irrealis. Irrealis is characterised by the absence of the elements ntt/wnt/iwt. In affirmative complements, irrealis is typically expressed by employing bare sDm=f forms, less often the sDm.n=f; in the negative the negations tm and nfr-n appear. Affirmative irrealis also displays an internal division into two subcategories termed proximal and distal that is reflected in the use of different sorts of un-introduced sDm=f forms. Yet, these differences are not made formally in all root-classes of the Earlier Egyptian verb, but pertain only to mutable roots. Bare sDm=f complement clauses employing the latter can be divided into binary categories on the basis of Modal typology: proximal versus distal irrealis Morphology: gemination/doubling versus no gemination/ doubling Form-function: marked versus unmarked The first two of the above divisions stand in mutual one-to-one
336
chapter ten
correspondence. Proximal irrealis is expressed by the geminating and doubling, distal by the non-geminating/-doubling sDm=f forms. However, the form-function divide cuts across the two former divisions. The geminating sDm=f of ultimae infirmae and certain anomalous roots (ini, rdi), various individual forms of other anomalous roots (iy/wnn=f) as well as forms with the endings -w and -y represent specialised, or marked, irrealis moods with only this function, which renders them incompatible with the elements ntt and wnt. With the possible exception of the form mA=f of the doubling root mAA, all other active sDm=f forms are modally unmarked. In complementation they may be used as functional counterparts of the marked forms and obtain their irrealis value through negative marking by not being introduced by ntt/wnt. The same forms preceded by these elements function as realis. The unmarked forms of the sDm=f show differences in the way in which they function as counterparts of the marked irrealis forms. Immutable roots do not distinguish between proximal and distal irrealis. Of the mutable roots, doubling forms of doubling roots have a proximal function; non-geminating forms of ult. inf. roots without the endings -w/-y function as distal. Anomalous roots display various irregularities, some (iwi, ini, ii) even borrowing forms for the distal function from the sDmt=f paradigm. The varying characteristics displayed by the different root-classes in the mapping of the modal roles stresses the importance of paying attention to individual Formenbildung in each class. The variation of the different bare sDm=f forms of mutable roots indicates the degree to which the irrealis proposition is viewed as nonassertable. It communicates variation in the speaker approximation and evaluation of the reliability, acceptability and discourse relevance of the information it contains. In principle, situations viewed in committed terms that often are also realised at the time of speaking, but that are deemed as somehow defective as regards their information value, are typically coded as proximal irrealis. Conversely, distal is used for states of affairs towards which little or no commitment is expressed and which are typically non-realised or whose information value is wholly negligible. However, the dividing line between these, and, indeed, all modal-typological categories is scalar rather than absolute. Overall, realis and irrealis form one continuous scale of modal meaning that extends from the expression of information known, accepted and viewed as newsworthy, to unknown, rejected and completely lacking
conclusion: retrospect and prospect
337
in relevance. The various grammatical forms and construals divide up this continuum and are, as it were, ‘located’ at its different points. Yet, the lines of demarcation on this scale are quite ill-defined. Already the basic categories of realis and irrealis are, in Earlier Egyptian as elsewhere, inseparable. This is because treating propositions in one way or the other is fundamentally a speaker choice and in most cases reflects his subjective perception of likelihood of the situation described and the reliability and ‘newsworthiness’ of the information passed. This is poignantly the case also with the typically Earlier Egyptian phenomenon of grammaticalised distal and proximal irrealis. For example, even in case of notionally non-assertive governing verbs after which irrealis is the only option available, with some of these verbs the speaker may, by employing the proximal, indicate that the situation is currently realised. Nevertheless, distal may also be used if signalling this is deemed superfluous for whatever reason.1 The approximate domains of use of construals signalling assertion and the different types non-assertion are represented in the diagram below: Most assertable
Most relevance/ commitment Less relevance/ commitment Least relevance/ commitment Least assertable
1
See 3.1 and 3.3 for details.
ntt/wnt
proximal function sDm=f (hAA=f, mAA=f) distal function sDm=f (hAy=f, hAw=f, hA=f)
Immutable sDm=f, sDm.n=f tm/nfr-n
338
chapter ten
The relative mass of the above pyramid also reflects the quantita tive differences of actual attestations between the forms and constructions. The fundamentals of this system are most apparent in affirmative object complement clauses of verbs of locution, cognition and perception, after which ntt/wnt-introduced and bare sDm=f complements are found in abundance (chapter 2). From the start, the use of these construals after the said predicates shows the intricacy and contextsensitivity of the speaker’s modal choices and their dependence of the surrounding discourse. An examination of the variation between the different complement types shows that ntt/wnt-clauses do not simply substitute for the bare sDm=f, but the two are used under very different conditions. All the characteristic uses of the two types of non-assertion for propositions variously lacking in speaker commitment or informativity are already present in the example corpus of object complements of these verbs. From this it may be seen how the morphologically different bare sDm=f forms divide up the domain of irrealis into proximal and distal types. A close look at the morphological and paradigmatic properties of complements of the verbs of locution, cognition and perception shows that ntt/wntintroduced clauses represent analytic construals consisting of modally neutral forms and constructions ushered in by functional elements that assign them the grammatical role of realis. Of the neutral patterns, the active suffix-conjugation forms may also function as irrealis by not being introduced by ntt/wnt. In contradistinction to verbs of locution, cognition and perception, the modal profile of affirmative object complements of deontic, implicative and volitive verbs of e.g. ordering, preventing and desiring, is primarily determined by the notional properties of these verbs (chapter 3). In a manner similar to other languages, they are semantically non-assertive and thus systematically incompatible with asserted complements and the realis-markers ntt/wnt. However, also here the attitude and deictic locus of the speaker may, depending on the verb in question, be variously brought to fore in the complement.2 Thus e.g. with verbs of attempted manipulation the use of the proximal may indicate success
2
See 3.1 and 3.3 for details.
conclusion: retrospect and prospect
339
of reported manipulation or, alternatively, indicate that the speaker’s illocutionary intention is manipulative. The interplay between asserted ntt/wnt-introduced- and nonasserted bare sDm=f clauses as the verbs of locution, cognition and perception is also found in affirmative subject complements (chapter 4). Here too, the possibilities for increased expressiveness facilitated by the variation of proximal and distal is employed with the same degree of sophistication. In the guise of the initial expression xpr. n, subject complementation also seems to provide a rare example of grammaticalisation of an auxiliary in progress. This in turn is suggestive of the semantic-pragmatic function of these elements and provides further support for the hypothesis of ntt/wnt as modal markers and the forms and construals introduced by them as modally unmarked. Assertion and non-assertion condition the grammar of subject- and object-clauses of verbs also in the negative (chapter 5). Here iwt/ntt nintroduced assertions and tm/nfr-n-non-assertions appear in variation, and the same integrity of the factors motivating their use can be observed as in the affirmative, even though the differentiation between proximal and distal irrealis is not made in the negative. It seems also that tm and nfr-n may be analysable as specialised irrealis negations used to negate non-assertions more generally, e.g. in adjunction. Clauses introduced with ntt and wnt as well as bare sDm=f forms occur also after prepositions used as conjuncts (chapter 6). Here, in a fashion analogous to complements of governing predicates, the subordinate clause is assigned a status as realis and irrealis partly on basis of the speaker’s attitudinal stance towards the situation described, but also, and, in fact primarily, of his evaluation of its discourse relevance and value as information. There are various environments where the use of irrealis in preposition complements parallels that of other languages, e.g. in case of counterfactual and final ‘so that’clauses. However, owing to their role as components of adjunct clauses, which by default provide ‘circumstantial’ background-information, irrealis tends to be generally used for complements of prepositions in Earlier Egyptian. Yet, when needed and possible, speakers could also profile the same clauses as realis. As a third alternative, the modality might be left unmarked by employing modally ‘unspecialised’ adjuncts without prepositional conjuncts. The element m-xt seems to provide an example of a preposition partly grammaticalised as a ‘true’ adjunct subordinator.
340
chapter ten
In addition to ntt/wnt-clauses and the use of bare sDm=f forms, there existed various alternative methods of complementation in Earlier Egyptian (chapter 7). The most notable of these is the use of the bare sDm.n=f, which was employed as a supplementary form in irrealis complements. Its use appears to have been motivated by a desire for clear indication of past time-reference in environments where this could not be reliably deduced from the context.3 Along with sDm=f forms without gemination or the endings -w/-y, the sDm.n=f is also a member of the class of modally neutral unmarked active suffixconjugation forms and could similarly be used alone in non-asserted, and after ntt/wnt, in asserted complement clauses. Of the other more marginal types of Earlier Egyptian complementation, clauses with the element is appear to represent remnants from a diachronically earlier system of creating complement clauses, whereas ones ushered in with the expression r-Dd represent first examples of the type of complementation characteristic to Late Egyptian. The pragmatic character of both these types of clauses is not entirely clear, but their use does not seem to have been based on differentiation between realis and irrealis.4 The differentiation between irrealis and realis typical to Earlier Egyptian complementation finds a rather surprising manifestation in clauses standardly analysed as serving as predicate complements (chapter 8). Here, non-assertions occur in bipartite nominal sentences with the element pw, which serves as a referential subject to a clausal predicate. In sentences of this sort, the subordinate clause is thetic in character, i.e. it does not serve to provide information ‘about’ its own subject and does not assert anything about it. By contrast, in pwsentences where the embedded clause is from the modally unmarked class of patterns, pw does not serve as a referential subject. Instead, it has a double function as a syntactic marker of topicalisation of the sentence-initial element and a modal marker of assertion of the subordinate clause. Finally, irrealis modality interacts in Earlier Egyptian with the semantic domain of tempus, particularly aspect (chapter 9). The modal and aspectual characteristics of the irrealis-marked sDm=f forms showing gemination and the endings -w and -y remain the same across their 3 4
See 7.1 for details. See 10.2 below for further remarks on is and r-Dd.
conclusion: retrospect and prospect
341
uses, whereas the modal versatility of the unmarked forms is paralleled by shifting aspectual values. These correspondences may be used to hypothesise that the ‘markedness’ of the irrealis forms is a universal property thereof and that they are exclusively used for non-assertion. The assignment of the modal and temporal values on these forms also follows systematic principles. The prototypical characteristics of proximal irrealis share conceptual similarities with imperfective aspect whereas distal displays ‘family resemblance’ with perfectivity, and these similarities are exploited in mapping the temporal and modal functions onto the forms specialised for expressing irrealis modality in complementation and possibly beyond. 10.2 The Diachronic Status of the System The organisation of the grammar of complement clauses around modality is a characteristically Earlier Egyptian phenomenon. It is clearly functional already in the earliest non-religious royal and private inscriptions of the Old Kingdom, and it may be that the same holds also for the Pyramid Texts. In these inscriptions the elements ntt/ wnt occur and are absent after the same verbs and prepositions and probably under the same conditions as later on.5 Indeed, as seen, some of the finest examples of non-assertion arising from lack of acceptance after verbs are to be found in these texts.6 Similarly, in the PT verbs and prepositions introduce bare geminating, non-geminating and -w/-y-ending sDm=f forms, whose difference may well be the same as elsewhere in Earlier Egyptian. However, the Pyramid Texts nevertheless both warrant and apparently require an independent investigation into their complement modality. For example, it may be that in these texts or in parts thereof, non-geminating sDm=f forms with and without the ending -w and/or -y (still?) carry differences in meaning. For example, the variation between the -w-ending and w-less forms in the following instance may be intentional given the apparent difference in the main verb tense:7
5 6 7
See Allen 1984 §§ 218, 228–39, 242–51, 272, 410–11, 573; Allen 1982, 24–25. See e.g. chapter 2 n.47 and the use of the negation tm in example (236) above. James Allen, PC.
342
chapter ten
(424) The king’s burial ground is addressed: sDm nn Dd.n nTrw P/M:
Dd ra sAxw=f N pn N: Dd.n ra sAx=f N pn Listen to this what the gods have said: “Ra says/has said that (PT 795a–b) he will/would transfigure this king N”.
The possibility that here one has an instance of a ‘simple’ and perhaps asserted future ‘sDmw=f’ versus a modal “possible, contingent or desirable” non-asserted ‘subjunctive’8 is well worth taking into serious consideration. Yet, as seen, here it is not merely the original speaker’s perspective that has to be taken into account, but also the reporter’s. In the above example the editors of the different PT-variants might have understood the speakers’ attitude towards the reported proposition or its communicative relevance differently; after all, this is also what the variation of will and would (if this is the correct translation) signals in English. Nevertheless, if the form-variation here and elsewhere in PT complement clauses is based on modality, this is certain to reveal new insight into the development and origins of the system described in the present work. Indeed, the PT are likely to hold an answer to e.g. such puzzles as the exact semantic-pragmatic and diachronic position of the element is in the grammar and modal organisation of complement clauses in Earlier Egyptian.9 Further, the PT are characterised by the widespread use of sDm=f forms displaying the i-prefix, which sometimes occur also in complementation, but whose semantic-pragmatic character is similarly unclear.10 It is also noteworthy that in Old Egyptian the bare sDm.n=f is apparently not used after prepositions, but the employment of this construct increases in time and is at its most prolific in early XVIII dynasty texts.11 Yet, some examples are forthcoming from the PT, although, as noted by Allen, they all derive from a very limited sequence of spells.12 These phenomena represent further targets for subsequent research on modality in Earlier Egyptian complementation, a topic which 8
Allen 1984 § 365; see also ibid §§ 267, 364, 705–06 and Allen 1982, 22. See 7.2 above. 10 See EAG §§ 12, 449–54; Allen 1984 §§ 14, 214, 259, 264, 395; in complements, EAG § 481; Allen ibid §§ 218, 229, 232, 235. See also example (38) and n.43 in chapter 2 above. 11 Cf. EAG § 534; Junge 1978a, 105. 12 Allen 1984 § 411. 9
conclusion: retrospect and prospect
343
the present work does not by any means claim to have exhausted. Even if the modal system of realis and irrealis discussed would appear to characterise Earlier Egyptian as a whole, it did undergo changes during the almost one and a half millennia of language history discussed here. In addition, it did not survive the diachronic transition to Late Egyptian, i.e. the linguistic idiom of non-literary texts of the New Kingdom and particularly the Ramesside era. As seen, the differentiation between assertions and non-assertions in indirect speech was gradually phased out already in the transition from Old to Middle Egyptian after which true indirect speech in general is a signal of non-assertion as opposed to neutral direct speech.13 At around the same time, the use of the element wnt was in decline and the negative iwt passed into oblivion, being replaced by the analytic ntt n. 14 Nevertheless, after this matters seem to have remained unchanged until major changes begin to appear during the first half of the XVIII dynasty. First signals of collapse of the system are at first isolated ‘abnormal’ instances, such as the following example of an object complement of the notionally non-assertive predicate mri introduced by the initial auxiliary iw: (425) Amenemhab boasts of his camaraderie with the king: iw Sms.n(=i) nb=i r nmtwt=f Hr xAst mHtt rsyt mr=f iw=i m iry-rdwy=f ti sw Hr priw nxtw=f I followed my lord to his journeys in northern and southern lands, for he wanted me to be his closest companion when he (Urk IV 890, 10–12) was on the battlefields of his victories. However, it is the appearance and spread of the expression r-Dd that signals the definitive beginning of the end for the organisation outlined.15 In mid-XVIII dynasty monumental texts the element ntt is still widely used, but is increasingly confused with the expression r ntt and apparently has become subject to phonetic erosion:16
13
See 2.1. See 2.1.1, 5.1 and 5.2. 15 See 7.3. 16 See (70) and chapter 2 n.74 for further instances of r ntt instead of ntt after verbs other than Dd in XVIII dynasty sources. 14
344
chapter ten
(426) A caption in a scene depicting Amenhotep III with a deity: di=f rx psDt aAt r nty tw xa.ti m nsw-bity Hr st Hr [nt anxw] He shall let the Great Ennead know that you have appeared as a dual king on the Horus’ throne of the living. (Re-used block T/I.R 319 in the mortuary temple of Merenptah, l. 3)
In Late Egyptian ntt/wnt have disappeared completely and r-Dd occurs also in other environments where ntt/wnt did (or could) not.17 However, as noted, r-Dd does not seem to have become grammaticalised as a wholly generic complementiser. Besides its use in final clauses, in complementation r-Dd-introduced clauses vary with ones without it after the same verbs: (427) The writer of the letter assures his addressee: xr m di HAty=k m-sA pAy=k it sDm(=i) a=f m sS sp-sn hAb n=i pA mr pr n imn wab N r-Dd m di hAty=t m-sA=f sw m sS sw snb Now, do not worry about your father; I have heard that he is fine. The steward of the house of Amun and wab-priest N wrote to me, saying: “Do not worry about him; he is fine, he (LRL 59, 1–3) is healthy”. (428) The writer of a letter tells of his arrangements to reach his addressee:
xr ir tw=i m xnty iw=i (Hr) sDm r-Dd tw=k xd.tw iw=i (Hr) wD pAy sDmw ink Now, as I was travelling south, I heard that you had departed north and I dispatched this messenger of mine. (KRI VI 68, 1–2)
The parameters conditioning this variation are unknown and require further study.18
17
See 7.3. In example (427), it may be that the information in the complement clause of sDm represents not wholly reliable ‘hearsay’, whereas this is clearly not the case in (428). However, whether or not this is the true motivation for the variation of r-Dd and no r-Dd here is uncertain. 18
conclusion: retrospect and prospect
345
Besides the rise of r-Dd and the disappearance of ntt, the most significant factor in the collapse of the Earlier Egyptian system of expressing modality in complementation was the general demise of the suffix-conjugation. In Late Egyptian sDm=f complements of governing verbs are largely restricted to causative constructions with rdi, whereas they are scarcely found after fully lexical predicates.19 Save for such semi-grammaticalised compounds as e.g. m-Dr sDm=f, sDm=f is neither used after prepositions, where it is replaced by the infinitive (for present and future) or the relative form (for past) preceded by the possessive pre-formative pAy=f:20 (429) The author quotes his recipient from an earlier correspondence: Dd=k n=i gr tw m iri mdwt iw=i r dit in.tw n=k pA kAr m pAy=i spr You said to me: “Be quiet, say nothing, and I will send you the (KRI III 255, 5–7) boat when I arrive”. (430) A question put to a suspected tomb-robber: ix pA sxr n Smi i.iri=k irm N iw=k (Hr) pH tAy st aAt iw=k (Hr) in pAy HD im r-bl (m)-sA pA Sm i.iri nA iTAw What was the business of your going with N and entering this Great Place and bringing this silver out of it after the thieves (KRI VI 776, 5–7) had left? This development took place, in turn, parallel to the major remodelling in Late Egyptian of the methods of forming adjunct clauses. The unmarked ‘circumstantial’ clauses disappeared altogether and the language developed a generic adjunct subordinator from the auxiliary iw. These changes are likely to be connected and to reflect a global demotion in Ancient Egyptian of the role of modality from its prior status as the dominant factor in the grammar of complement- and adjunct clauses. However, as the example of indirect speech shows, similar shifts away from modally based grammatical organisation occasionally took place also during the diachronic stage of Egyptian 19
Cf. Junge 2001, 214, 216–17. See 1erný-Groll 1993, chapter 51.9; Junge 2001, 102–04, 229–31. But see Layton 2000, 363 for very ‘old-fashioned’ Coptic uses of second tenses directly after the prepositions mNNsa ‘after’ and jin ‘since’. 20
346
chapter ten
studied in the present work, well before the onset of Late Egyptian. Moreover, as seen, non-assertion continued to be specifically indicated by ‘real’ indirect speech, even after the language had reverted to using modally neutral direct speech in instances where assertions had once occurred. In addition, although in Late Egyptian no difference appears to have been made between asserted and non-asserted indirect speech, it may be that, rather like the use of the ‘modal’ pw/pe,21 this division was once again re-introduced much later in Coptic, and in a rather familiar grammatical form. As shown by Peust, the person deixis of Late Egyptian and Coptic indirect speech follows a set of complex parameters that seem to have little to do with modality.22 Typically, the shifts of deixis affect only one of the participants referred to, whereas there are only a handful of (certain) instances of indirect speech where the deixis is completely adjusted to the perspective of the real speaker, and all these derive from Coptic. Yet, interestingly, in these examples a clear modal component is observable:23 (431) The writer of a letter chastises his correspondent: a iwannhs ei nai efjw mmos je ntakti xnsouo mn xniwt enouf ne ebol xwris dikaion... loipon sto petntaf erof John came to me saying that you have sold some of his wheat and barley without permission… Now, give back to him what (Revillout 1914, 26 no. 74) is his! (432) As above: a2ax aplw ei fol efmemve Nmok je akfi pefkahl[iths]... auw ešwpe ce te taei kakws kakws The scribe Aplo came complaining of you that you have taken his camel driver… Now, if so, it is very, very bad. (Worrell 1942, 203, 1–3)
In the first example above the speaker quite clearly does not accept the situation described, in the second he is not certain about the reliability 21
See 8.3 above. Peust 1996; 2005. 23 Example (431) as cited by Peust 2005, 89; see also ibid, 90–91 for less clear instances, including one in Demotic. 22
conclusion: retrospect and prospect
347
of the information—both ‘classic’ environments of non-assertion. It is thus possible that the old differentiation between asserted and nonasserted indirect speech was partially re-grammaticalised late in the history of Egyptian. On the other hand, the fortunes of the element is in Earlier Egyptian give reason to assume that the system of realis and irrealis described represents a relatively recent development in historical Egyptian.24 Hence, it may be that, in a fashion analogous to e.g. the difference between generic versus progressive present grammaticalised in Middle Egyptian, Demotic and Coptic, but not in the PT and Late Egyptian,25 the indication of irrealis and realis modality in complementation (or some of its sub-categories) and the organisation of the grammar of these construals accordingly fluctuated through the diachronic stages of Ancient Egyptian. Further research on complementation might clarify this question and should, in any case, adopt an increasingly historical perspective to the topic. 10.3 Extending the Hypothesis The principles and categories of modality established in the present work would also appear, as noted repeatedly in the course of discussion, to have application in the study of the grammar and semantic-pragmatic character of Earlier Egyptian beyond complementation. Research on negations with tm/nfr-n, 26 adjunct clauses, 27 the constructions following the operators ntt/wnt,28 the peculiarities of the verb gmi,29 the genesis of auxiliaries,30 and e.g. relative clauses/constructions and second tenses would also seem to benefit from the hypothesis proposed in the present work. Although a full discussion of these topics belongs to the domain of future investigation, certain very brief preliminary remarks on the last two types of construals will perhaps not be out of place and will serve as an illustration of how the analysis proposed here for complementation might perhaps be extended into other areas of Earlier Egyptian grammar.
24 25 26 27 28 29 30
See 7.2. Cf. Eyre 1994, 119. See 5.3. See 5.3, 6.3. See 2.4. See 2.4. For this, see Uljas 2006.
348
chapter ten
In the present work the issue of relative forms and -constructions has been deliberately avoided in response to the belief that “there appears to be nothing relevant that can be said” concerning modality and relative clauses.31 However, this contention might be somewhat hasty. Both in the original Polotskyan analysis and its later developments the ‘nominal’ geminating sDm=f is seen as an ‘abstract relative form’, and it is true that the morphological parallelism between the form of bare sDm=f of ultimae infirmae roots appearing in relative and complement clauses can hardly be coincidental. The same applies also to the striking similarity of the complementiser ntt and the relative adjective nty/ntt/ ntw, the possibility of creating a relative sDm.n=f from intransitive vom and the tm-negation of participial and the bare suffix-conjugation relative clauses versus the use of the negative iwty or nty n as negative equivalents of nty. Now, seeing that most—but not all32—relative clauses do not assert but pragmatically presuppose the situation described, it is conceivable that in Earlier Egyptian relativisation might have been part of the same modal organisation as complementation. It is similarly quite remarkable that relative construals seem to display the same tripartite split between bare tm-negated participial, ‘relative’ sDm=f- and sDm.n=f-clauses, iwty/nty n-negated nty-clauses, and the n/nn-negated wholly ‘unspecialised’ or ‘virtual relative’ types, the latter of which are formed of the same ‘unmarked’ patterns as are found in complementation after the elements ntt/wnt. Notably, there is nothing inherently ‘relative’ in the latter set of constructions per se; they are clearly simply the same patterns used also in a number of other environments (initially after particles, ‘circumstantially’ as adjuncts and in complements after ntt/wnt).33 That is, their function seems to be determined by use throughout. This poses an interesting question whether this might not be the case also with the element nty and even the relative forms. The suggestion made here and earlier in the present work that the relative adjective nty and the complementiser ntt may be simply the same element differently used is not new.34 But there appears to be neither any particular reason to presume the
31
Palmer 1986, 128. This is particularly the case with non-restrictive relative clauses; see Cristofaro 2003, 195. 33 Cf. Collier 1991b. 34 See e.g. GEG § 237; Gilula 1970, 213 and 2.1.1 above. 32
conclusion: retrospect and prospect
349
existence of separate ‘abstract’ and ‘concrete’ relative forms simply because the latter are used as attributes of nominal heads and show gender/agreement properties. The morphological characteristics of these assumed separate forms are practically identical, and thus it may well be that the gender marking and agreement are actually features that can be added to the bare geminating and non-geminating sDm=f and the sDm.n=f when they are used ‘relatively’.35 This would mean that in fact the same modally marked and unmarked un-introduced forms used as irrealis in complementation can also be used as unintroduced relative clauses, and this parallelism suggests that they may share the same modal profile in both instances just as they share the same negation and features such as .tw-passivisation/vom use of the sDm.n=f as well as seem to contrast with nty-clauses. However, there are various unsolved questions involved herein,36 the clarification of which requires nothing less than a full reanalysis of Earlier Egyptian relative clauses from a semantic-pragmatic and functional perspective. Nevertheless, carrying out such an undertaking from the perspective of the present analysis may yield interesting results. Outside complementation, the most promising and also much wider area for applying the modal analysis developed herein is second tenses, which have been touched upon at various points of the present work. For the past six decades there has hardly been a topic more intensely debated among scholars of the Ancient Egyptian language.37 As a 35
See here n.25 in chapter 7. For example, it is not obvious where the -w/y-ending sDm=f would fit in the above schema, and the -w/-y-less non-geminating sDm=f and the sDm.n=f do not normally occur after nty as they do after ntt. On the whole, the syntactic explanation that (excluding attributive relative clauses of indefinite antecedents) nty is used whenever the relativised construal is non-verbal or has no initial verb works rather well in relative clauses. Nevertheless, it does have certain anomalies. For example, the non-verbal sentences could be ‘verbalised’ by using the auxiliary wnn (participial examples are commonest; a relative form appears e.g. in Sin B 44–45 nTr pf mnx wnnw snD=f xt xAswt “that potent god, the fear of whom is throughout foreign lands”). The conditions under which this was preferred over nty are unclear, but appear to relate to features of TAM. Similarly, at least after the negative iwty the use of the sDm.n=f (e.g. BM 159, 12–13 iwty sDr.n rmT Sp.t(i) r=f “one with whom no-one stayed anger overnight”) and the sDm=f, for which ult. inf. roots show gemination (e.g. BM 334, bottom 1–2 mnx-ib iwty bAgg=f “One strong of heart who did not (use to?) tire”) is commonplace. On the other hand, from the standpoint of the present analysis, this last property is, again, rather unexpected. 37 The bibliography on the topic is of unparalleled magnitude in Egyptological linguistics and references here are reduced to bare minimum; Depuydt 1983 provides a 36
350
chapter ten
reminder, if such were needed, second tenses fall into two principal semantic classes. 38 They can consist of a setting followed by a subsequent main clause: (433) The king says men are well-tended by the sun-god: rmm=sn iw=f Hr sDm (Merikara C V, 7) Whenever they cry, he is listening. Alternatively, and more commonly, the initial verb may function as a main clause in the translation, but its role is then to highlight or ‘emphasise’ some subsequent expression, typically (but not exclusively) an adjunct: (434) The benevolence of the king towards his followers is extolled: dd=f kAw n ntw m Sms=f To those in his train he gives sustenance. (Loyaliste § 3, 7/St) Both these types are negated by tm; below is an example of a negated ‘emphatic’ sentence: (435) Thutmosis III explains his motives for abbreviating a list of plundered goods:
tm.tw rdit rxt=sn Hr wD pn r tm saSA mdwt Their number is not put on this record merely in order not to (Urk IV 693, 12–13) multiply words. Except when referring to past situations, where a bare sDm.n=f appears, (also in .tw-passive and of intransitive vom) second tenses make use of un-introduced sDm=f forms for which ult. inf. roots show gemination, as above.39 Although they can be preceded by initial particles such as mk, second tenses never occur after iw or other auxiliaries. The original ST analysis of these constructions was firmly rooted in the
historical summary of the argument up to the early 1980s; see also Schenkel 1990, 145– 58. 38 The so-called ‘balanced sentences’ are a third and more marginal category, but can also be integrated to the hypothesis sketched below. 39 Particularly in earlier texts, apparently similarly functioning non-geminating forms, sometimes with the ending -w, are occasionally encountered.
conclusion: retrospect and prospect
351
‘nominal hypothesis’ and actually marked its beginning.40 The initial verb form was argued to represent a nominal subject to the following ‘emphasised’ adverbial ‘predicate’, with the following structure and ‘literal’ sense: [dd=f kAw]Nom subj. [n ntw m Sms=f]Adv. pred. “That he gives sustenance (is) to those in his train” However, this orthodoxy has been largely abandoned due to various problems, some of which have already been noted.41 Most crucially, it proposes a very strange analysis of ‘predication’ and is fundamentally incompatible with the concept of argument-structure.42 In e.g. (434) above, the initial verb rdi ‘give’ projects the correct number and type43 of arguments and hence clearly behaves as the predicate in the complex. The arguments are not simply ‘satellites’ of which some may be cleft apart from the verb, but indispensable elements fulfilling the semantic roles required by the inherent lexical meaning of the verb ‘give’, which demands the presence of an agent (=f ), a theme (kAw) and a recipient (ntw m Sms=f ). Moreover, the original Polotskyan analysis does not seem very plausible with the setting second tenses; e.g. in (433) what follows the initial verb is not an ‘adverbial’ but a syntactically fully independent main clause with iw that can hardly be treated as the ‘predicate’ of rmm=sn. Because of these and other problems with Polotsky’s analysis, an alternative view of the nature of second tenses has recently gained popularity among Egyptologists, either as an augment or replacement of the ST hypothesis. In much recent research the initial verb form of second tenses is argued to be of a nature, or to having undergone a process that sets it into the information background, where it constitutes a topic or theme of the overall proposition. There is no consensus as to exactly how these terms should be interpreted or used in analysis, but this is largely because the same holds also for general linguistic theory.44 40
Polotsky 1944. See the discussion on ‘embedded second tenses’ in 3.1. 42 See Collier 1992 for a detailed discussion. 43 This is clear from the introduction of the datival recipient ntw m Sms=f with the preposition n. 44 Here one may note e.g. Levinson’s (1983, x) bemoaning of the “terminological profusion and confusion, and underlying conceptual vagueness (which) plague the rel41
352
chapter ten
It suffices here to note the most extensive and recent discussions by Junge and Loprieno.45 For Junge, who seeks to defend the ST analysis of second tenses, ‘theme’ is synonymous with ‘topic’ and represents what the sentence is ‘about’; an item ‘given’ to both the speaker and the hearer.46 The ‘predication’ in the sentence is, as in the traditional ST analysis, between this and the following ‘rheme’ and both the ‘emphatic’ and ‘setting’ second tenses are, according to Junge, uses of backgrounded and ‘nominal’ clausal themes only differently translated.47 Loprieno’s point of view is slightly different. He does not subscribe unequivocally to the ST theory on predication, but considers the head verb in ‘emphatic’ second tenses similarly as a theme, whereas in setting second tenses the initial clause provides the sentence topic, “a clause nominalization functioning as pragmatically ‘given’ within the communicative flow of discourse”.48 These and the other approaches concerned with the organisation of theme/rheme and topic/comment in second tenses are laudable in their shifting of the focus of discussion into the domain of information structuring and pragmatics, which certainly hold the key to the correct understanding of these construals. However, even leaving aside the question of predication, there remain various problems and open questions. In general, it is not the case that topic represents ‘given’, ‘old’ or presupposed information.49 For example, in the ‘traditional’ ‘second tense translation’ of (434) above as “It is to those in his train that he gives sustenance” the open proposition “he gives sustenance to X” can indeed be characterised as what is ‘given’ in the sentence—although this is more a result of rendering of the construal as an English cleft sentence than follows from the Egyptian original. Yet, even if correct, it cannot be the topic inasmuch as it does not constitute a full proposition, does not correspond to any syntactic constituent nor has any referent that the sentence could be said to be ‘about’.50 Moreover, terms such evant literature to the point where little may be salvageable”. Cf. also Collier 1992, 23 n.16. 45 Junge 1989; Loprieno 1995. For critique of Junge’s analysis, see Collier 1992. 46 Junge 1989, 43–45; as seen, ‘topic’ was defined as what the sentence is about also in the present work (8.2 above). Junge is careful not to equate ‘theme’ with initial position (pace Halliday 1985, 37). 47 Junge 1989, 53. 48 Loprieno 1995, 107, 192; for ‘theme’ no explicit definition is given. 49 See Lambrecht 1994, 122, 129, 151–52. 50 For the ‘referentiality’ of topics, see below.
conclusion: retrospect and prospect
353
as ‘given’ pertain to what is shared with or ‘given’ to the speaker and the hearer mutually, whereas ‘topic’ is something that is chosen as the starting point for communication by the speaker only and need not be anything but identifiable to the hearer. Further, topic expressions are typically referential (usually noun phrases) rather than propositional. True, propositional topics do exist, but these are not predicates but either adjuncts (as the ante-posed clause in “When Jill had left, John arrived”) or, it is sometimes argued, arguments of predicates (such as the complement clause in “It is nice that Jill left”).51 As the original ST theory of predication in second tenses is clearly incorrect, this excludes the applicability of the second of these alternatives. The first, adjunct-analysis is conceivable with setting second tenses,52 but fits ill with the idea of the initial verb form as a nominal expression. In addition, it is not clear how the ‘thematisation’ or ‘topicalisation’ of the second tense head verb supposedly takes place. Indeed, it is nowhere made clear whether the verb forms involved should be seen as thematised and/or topicalised, or whether they are inherently ‘thematic/ topical’. In the first instance, if they have acquired their topical/ thematic character through e.g. some derivative process, there must be some source from which they are derived, but no suggestion as to what this might be has so far been made. In the second instance, the inherent ‘thematicity/topicality’ of the forms is open to considerable doubt seeing that they do not function thus everywhere, e.g. in object complements of governing verbs and after prepositions. Some other ‘basic’ character thereof seems thus to be called for. From the incompatibility of second tenses and initial auxiliaries it could be conjectured that, besides gemination in the root-classes showing this feature, the absence of these elements is a signal of ‘thematic’ or ‘topical’ value of the initial verb. Yet, what then constitutes the ‘theme’ or ‘topic’ in auxiliary-introduced sentences and why? If it is now not the verb together with its subject but e.g. only the subject, how does the auxiliary ‘direct’ the ‘thematicity’ to the right address? Is it the auxiliary that somehow signals that the following construal has a
51 Cf. Lambrecht 1994, 151 and 6.2 above for clausal topics introduced by ir + preposition. Yet, there are some problems with the definition of topics as what the sentence is ‘about’ and such clausal topics, particularly in case of the argument-analysis. Discussion of these must be left to another occasion (but see n.56 below). 52 Cf. Vernus 1981, 74–75, 77, 79; see also further below.
354
chapter ten
different theme/topic-rheme/comment-structure than second tenses, is this perhaps signalled by the forms and constructions themselves or should the auxiliary itself be seen as a ‘theme/topic’? It is impossible here to delve further into the complexities of these issues. As noted, the discussions attempting to clarify second tenses on basis of the organisation of theme/rheme and topic/comment are undoubtedly correct in focussing on information structuring and in taking their pragmatic characteristics more into account. Yet, in spite of the considerable advances made, it would seem that they still do not represent the final word on this most problematic area of Earlier Egyptian grammar. There is, however, a further possible approach to second tenses that makes use of the findings of the present work. Study of complement clauses after governing verbs shows the geminating sDm=f to have a fixed non-assertive irrealis function and the same holds also for the negation tm. The sDm.n=f, the doubling sDm=f of doubling roots and the unitary sDm=f of immutable roots may also serve an irrealis function in complements without introducing elements, and they are also used alone in second tenses.53 These formal and syntagmatic similarities already suggest that second tenses are analysable as constructions headed by non-assertions, and this proposal seems quite appropriate also from a semantic-pragmatic perspective. The underlying motive for such a modal status is perhaps clearest in second tenses used to express wh-questions: (436) The peasant argues that the high steward’s inaction is against his own good and asks:
irr=k r=k irf r m Why do you act against yourself?
(Peas B1 146–47)
In wh-questions the verb stands outside the scope of the interrogative and is presupposed information—in this instance the presupposition is [you act against yourself].54 The verbal head is thus most certainly
53 For an instance of the sDm.n=f and a sDm=f of an immutable root used as a second tense, see examples (437) and (438). For the distal forms and their functional counterparts without gemination, see below. 54 See Levinson 1983, 184; Lambrecht 1994, 244, 282–86; Givón 2001, vol. 1, 312; vol. 2, 232–33, 300–01 inter alia.
conclusion: retrospect and prospect
355
un-asserted here. In setting second tenses such as in example (433) above, it is no accident that the initial situation-descriptions translate best as ante-posed adverbial clauses, inasmuch as this actually seems to be their pragmatic function.55 As noted earlier on, the role of such clauses is to establish a background for the following main clause, and may perhaps be termed ‘clausal topics’.56 Such settings may refer to situations that are not in any way implied to be or having been ‘unreal’, but which nevertheless do not represent the high point of salience in the flow of information: (437) Sinuhe tells of the effects that receiving a letter from the king had on him:
spr.n wD pn r=i aHa.kw m Hr-ib wHwt=i Sd.n.t(w)=f n=i di.n(=i) wi Hr Xt=i When this decree reached me, I was standing in the midst of my tribe. But when it was read out to me, I placed myself on (Sin B 199–200) my belly. Yet, this may not be the case everywhere. In example (433) above, the word ‘whenever’ in the translation captures rather well the nonreferring, indeterminate character assigned to the state of affairs described. In (433) it is not said or, indeed, asserted, that ‘they’ ‘cry’ even as a matter of habit or custom. Instead, the situation is generic to the extent that its reality or non-reality is largely immaterial. Rather than to communicate information about someone ‘crying’ in a given moment of time or generally, the sole purpose of the initial clause is to provide background to some more salient state of affairs. Setting second tenses thus establish a real or potential ‘frame’ where the latter can then be focussed upon, but they themselves remain non-asserted. The same principle is also applicable to ‘emphatic’ second tenses. As in wh-questions, where the verb has the status of presupposition and the interrogative by default carries the highest pragmatic salience, in these construals a similar background/low relevance character and
55 Cf. Vernus 1981. Once again, appreciating this fully requires giving up the idea of the ‘nominal’ character of the forms used once and for all. 56 See chapter 6.2 and n.35 therein for references. Givón 2001 vol. 2, 344–45 argues against use of the term ‘topic’ here on the grounds that clausal topics are not referential nor cataphorically accessible.
356
chapter ten
reduced ‘assertivity’ of the main verb predicative nexus is observable generally, with the result that the informational salience of some subsequent adjunct is elevated. The latter is not so much focalised as it is focal owing to the downgrading of the initial verbal situationdescription into background. As recognised by Junge, the setting- and ‘emphatic’ second tenses indeed appear to be two sides of the same coin. Although what exactly is left to stand in relief varies, both involve the same demoting of the initial verb into irrealis and non-assertion. Also Junge speaks of ‘weakening’ of the situation ‘information weight’, although his reference is more properly to the translation of second tenses and contains no suggestion of non-assertivity of the initial verb.57 Yet, taking this step in the analysis was certainly contemplated by Gardiner, who, whilst seeking to challenge Polotsky’s analysis of the ‘emphatic’ second tenses with the geminating sDm=f, remarked the following:58 This result would naturally be best achieved by removing the stress59 from the verb-form in the sentence. It was thus important for the Egyptian to avoid saying positively that such and such an action happened or would actually happen. Now this avoidance of direct assertion may be effected by giving the verb-form a general or non-committal character.
This characterisation captures a great deal of what seems decisive herein. Second tense verbal heads constitute non-assertions which not only signal that the situation they describe is not the centre of interest in the proposition overall and that this lies rather in what follows. Their employment is also fundamentally a modal speaker choice and a device for shaping the discourse. It is to be noted that Gardiner was actually speaking only of the geminating sDm=f and his notion of ‘general or non-committal character’ was a reference to imperfective aspect rather than modality. Yet, as seen, the domains of aspect and of non-assertion are indeed very close to each other and particularly in the case of the geminating sDm=f they merge into one proximal functional continuum. This hypothesis, sketched in the most preliminary, tentative and 57 58 59
Junge 1989, 61. Gardiner 1947, 100; emphasis by SU. Gardiner seems to be referring to communicative, rather than phonetic stress.
conclusion: retrospect and prospect
357
in for mal terms, does not pretend to represent anything but the mere beginnings of a possible theory on second tenses; numerous further problems and questions remain. For instance, it is notable that neither the bare marked distal sDm=f forms nor their designated counterparts are (at least standardly) used in second tenses.60 This is probably because when used as irrealis, these forms are reserved for non-assertions that are (subjectively and objectively) most removed from reality and/or whose communicative salience is lower still.61 Yet, this preliminary hypothesis does not assume any ad hoc-theory of ‘predication’ nor ignores the relationship between the second tense verb and its associate arguments. It corresponds well to the views on the relationship between information structuring by means of modality outside Egyptology—theories linking lack of ‘relevance’ and irrealis have been noted frequently in course of the discussion—as well as brings the different types of second tenses under one explanatory umbrella of irrealis modality without simply replacing ‘nominality’ with ‘modality’. Thus it also integrates the analysis of the verb forms employed with their other domains of use, e.g. complementation. However, it is not incompatible with earlier observations on second tenses, but is able to accommodate and enhance many of them. For example, if it is ever possible to reach consensus on the meaning of these terms, the description of the initial verb in setting second tenses as ‘topical’ may be quite appropriate semantically.62 The hypothesis presented here proposes that this is a result of the inherent modal profile or an assigned function of the form employed. Similarly, the aspectual observations concerning second tenses may be integrated into the modal theory with ease, as can e.g. such details as the occasional deontic/optative translation of second tenses with 2nd person subjects discussed above.63 It was argued then that deontic value is not inherent to these construals, but a contextual implication by the audience, compatible with the overall discourse context in which the second tense appears. However, it is also one that is compatible with the non-assertive and non-indicative character/function of the head verb. 60
But see n.39 above for second tenses with -w-forms. That is, in initial environments these forms are rather used to make promises, wishes, epistemic deductions etc. just as in complements they create propositions with least commitment and informational relevance. 62 Cf. 6.2 and the literature cited in n.35 therein. 63 See section 9.1. 61
358
chapter ten
It has further been recognised for some time that in negated second tenses tm has only the initial verb in its scope, whereas when the scope is the highlighted adjunct, Earlier Egyptian uses the negation n…is: (438) Pepiankh assures his audience of the reliability of his words: Dd(=i) m mA[a] n Dd(=i) (i)s m aA-r (Urk I 224, 18) I speak truthfully; I speak not in boasting. Here the proposition [I speak], whose status is hardly more than a co-textual presupposition, remains outside the negation scope. But such a separate strategy also makes sense from the modal perspective. tm only has scope over the non-assertive verb because the adverbial does not represent part of the irrealis background and tm, as an irrealis negation, cannot negate such information. In addition, if the initial second-tense verb is non-asserting, the absence of iw and other auxiliaries must surely indicate that the role of these elements, as suggested at various points in the present work, is to do with assertion and realis-marking. Although their paradigms are not identical,64 many of the forms and constructions following auxiliaries are the same ‘unmarked’ ones found after ntt/wnt, which alone do not function as initial clauses any more than as complements. Active forms of sDm=f without gemination and the endings -w/-y as well as the sDm.n=f do, but in initial environments a bare un-introduced sDm=f of the said sort is always non-asserting—exactly as in complementation. It seems that these forms can be seen as members of a similar system of auxiliary-introduced realis-, and un-introduced irrealis initial clauses, and that a modally-based account of iw and auxiliary use in Earlier Egyptian is possible. True, iw has been labelled ‘assertion particle’ before, but it has not been suggested that it might be this what renders it incompatible with second tenses. Finally, previous analyses of second tenses have seldom tried to substantiate their postulates with cross-linguistic parallels.65 In contrast, uses of irrealis modality for very similar effect as in Early Egyptian
64
Most notably, nominal and adjectival sentences do not occur after auxiliaries. An exemplary sample of African languages with comparable phenomena is provided in Peust 2004, 382–90. 65
conclusion: retrospect and prospect
359
second tenses occur also in other languages.66 For example, in New Testament Greek the subjunctive is employed for ‘scene-setting’ anteposed adjunct clauses similar to ‘setting’ second tenses:67 (lxiii)
όταν δοχήν ποιής (SUB) κάλει πτωχούς “Whenever you prepare a feast, invite the poor” (Luke 14:13)
In Bemba habitual and certain past tenses, verbal inflection indicates whether the verb is included in the scope of the assertion or not. In the latter case it lacks a signal of inclusion and is marked as presupposed (i.e. non-asserted/background) information, with the result that something else in the sentence, e.g. an adjunct, receives additional focus:68 (lxiv)
Context: Response:
What did they do? ba-à-lí (INCL) -boomba “They worked”
Context: Response:
Where did they work? ba-à-boomba mu-mushi “They worked in the village”
Context: Response:
When did they work? ba-à-boomba bulya bushiku “They worked the day before yesterday”
The similarity of this and the ‘emphatic’ second tenses is obvious. Thus neither the phenomenon nor the proposed mechanism of second tenses represents an idiosyncrasy of Earlier Egyptian. Instead, they appear to be innovative uses of irrealis modality in need of a full inquiry.
66 Cf. also chapter 7, example (lxii) above for the use of imperfective aspect marking of verbs in Russian that can also have very similar effects as ‘emphatic’ second tenses. 67 Perschbacher 1995, 325. 68 Givón 1972, 178–79, 212; 1975b; 1982, 139; 2001 vol. 1, 343–45.
360
chapter ten 10.4 Final Remarks
Egyptology is as much study of inscriptions as it is analysis of material remains. The necessity of solid grammatical theory in the former is as self-evident as the requirement that archaeological fieldwork be based on tested and universally valid principles. Notwithstanding the largely theoretical focus of the present work, the hypothesis proposed has also been intended to serve a more utilitarian purpose. Hitherto there have been no guidelines beyond personal taste and educated guesses for translating complements. Clauses with geminating/nongeminating sDm=f and the negations tm/nfr-n have all been translated mechanically with that, and ntt/wnt-clauses with both that and whether without further ado, as if the distinct grammatical types did not differ in meaning beyond some nebulous shades of temporal nuance. Yet, the variation is meaningful and affects both translation and interpretation of texts. The appearance of ntt/wnt/iwt is a signal that the clause is an assertion and is to be translated as such. Except for a number of special cases such as gmi, xpr.n and the construction ink pw hA.kw, whenever these elements are absent, the clause is unasserted, and non-indicative may be appropriate also in the translation-language once the exact type of the non-assertion has been established from the co(n)text.69 Similarly, e.g. different types of non-asserted complements of verbs incompatible with assertions may signal what is the discourse motive of the utterance or whether the reference is to a realised or merely hypothetical situation. Differences of this sort can have dramatic effects on the understanding of the text studied and its potential value as a source of information. For example, in translating historical inscriptions, the difference between knowing that, whether or how something is the case, or whether or not a situation said to have been ordered was realised at the time of speaking, may be of considerable importance. Further, the analysis of the use of different complement types in Earlier Egyptian as based on realis and irrealis modality can also be both explained and learnt in a way that allows teachers, practising 69 Whether or not there is full correspondence in the expression of modality between Earlier Egyptian complement clauses and the language of translation of course depends greatly of the methods of expressing modality in the latter. For example, English that is a highly generalised complementiser that does not necessarily signal the status of the complement as an assertion or non-assertion and can be used far more extensively in translation of Egyptian complements than e.g. the Romance indicative.
conclusion: retrospect and prospect
361
Egyptologists and students alike to mobilise their knowledge of other languages. For example, those with knowledge of Romance languages in particular can often make direct comparisons between the use of modality in the latter and Egyptian, but, as seen, parallels can be drawn with a formidable number of different languages. Calling attention to these similarities makes the grammar of Earlier Egyptian complementation easier both to present and to comprehend. Finally, the study of language is, albeit in a rather oblique manner, study of the mental processes that lie at its root. Although pessimism as to whether grammatical research can actually make any substantial contact with the actual thoughts of language users seems to be once again gaining some foothold,70 at the very least, it allows access to the more abstract reasoning behind linguistic expression, including the use of metaphors and iconic representations that seem to be shared by all languages and are of notable anthropological and psychological interest. Egyptologists interested in language are no strangers to such views; the Anglo-Saxon research-tradition of 1920s and 1930s to Egyptian made conscious attempts to understand the ‘ancient thought’ and the Egyptian ‘mind’ through grammatical and philological examination.71 Subsequent to the ‘Polotskyan revolution’ such aspirations were hardly expressed, but they arguably maintain their credibility. The present work has also attempted to show how Earlier Egyptian modality can be successfully analysed and described with analytical models fundamentally metaphorical in character. These spatial-temporal-modal explanatory and descriptive metaphors are based on the same mental representations shared by language users both ancient and modern and are not products of western speculative thought. For all individuals, what is distant or behind a barrier is progressively less controllable and verifiable than what is closer and more at hand. Similarly, it does not strike one as strange or curious to treat something that is ‘out of focus’ or generally distant as lacking in relevance and as less part of the ‘real world’; this is apparent in all expression. People tend to regard clear ideas and concepts that are easy to grasp and have readily apparent and immediate use as outstanding and ‘the real thing’, and the rest as opaque speculation that is difficult 70
See e.g. Sperber & Wilson 1986, 10–11. See GEG §§ 3 and 211 in particular; more recently the same goal has been pursued, although from a more non-linguistic perspective, e.g. by Kemp (1989). 71
362
chapter ten
to fathom and may be totally far out. That ‘distance’ and ‘relevance’ should be common Leitmotivs in modal systems across languages is an iconic reflection of this rather basic truth. The ‘avenues of the mind’ involved are open irrespective of the passage of time, and the ‘ancient thought’ lives on in communication and linguistic expression. It may at times be difficult to recognise, but this is not because it is alien and incomprehensible, but because it is so deeply entrenched in human mentality as to go unnoticed.
bibliography
363
BIBLIOGRAPHY Akatsuka, N. 1986 “Conditionals are discourse-bound”, in E. Traugott, A. Ter Meulen, J. Reilly & C. Ferguson (eds.) On Conditionals. Cambridge: Cambridge University Press, 333–51. Allan, K. 1986 Linguistic Meaning vol. 2. London/NewYork: Kegan Paul/Routledge. ——— 1998 “Speech Act classification and definition”, in Mey, J. (ed.), Concise Encyclopaedia of Pragmatics. Amsterdam: Elsevier, 922–27. Allen, J. 1979 “Is the ‘emphatic’ sentence an adverbial-predicate construction?” GM 32, 7–15. ——— 1982 “Synthetic and analytic tenses in the Pyramid Texts”, in L’Égyptologie en 1979. Axes Prioritaires de Recherches, vol. 1. Paris: Centre National de la Recherche Sqientifique, 19–27. ——— 1984 The Inflection of the Verb in the Pyramid Texts. Malibu: Undena; Bibliotheca Aegyptia II. ——— 1986a “Tense in Classical Egyptian”, in Simpson, W. (ed.) Essays on Egyptian Grammar. New Haven: Yale Egyptological Seminar; Yale Egyptological Studies 1, 1–21. ——— 1986b “Features of nonverbal predicates in Old Egyptian”, in Englund & Frandsen 1986, 9–44. ——— 1991 “Form, function and meaning in the early Egyptian verb”, LingAeg 1, 1–32. ——— 1994 “Pronominal rhematization”, in Silverman, D. (ed.), For His Ka: Essays Offered in Memory of Klaus Baer. Chicago: The Oriental Institute; Studies in Ancient oriental Civilizations 55, 1–13. ——— 2000 Middle Egyptian. Cambridge: Cambridge University Press. ——— 2002 The Heqanakht Papyri. New York: The Metropolitan Museum of Art. Altenmüller, H. 1984–85 “Das ‘Sänftenlied’ des Alten Reiches”, BSEG 9–10, 15– 30. Anthes, R. 1969 “Zur Übersetzung der Präpositionen und Konjuktionen m und Dr”, in: Kadish, G. (ed.), Studies in Honor of John A. Wilson. Chicago: The University of Chicago Press; SAOC 35, 1–13. Arnott, D. 1970 The Nominal and Verbal Systems of Fula. Oxford: Clarendon Press. Assmann, J. 1974 “Ägyptologie und Linguistik”, GM 11, 59–76. Bach, K. & Harnish, R. 1979 Linguistic Communication and Speech Acts. Cambridge, MA: MIT Press. Barta, W. 1986 “Subjunktivische Konjunktionen als Einleitung von Umstandssätzen”, GM 90, 7–10. Bell, A 1980 “Mood in Spanish: a discussion of some recent proposals”, Hispania 63, 377–90. Bendix, E. 1998 “Irrealis as category, meaning, or reference”, Anthropological Linguistics 40:2, 245–56. Bergen, J. 1978 “One rule for the Spanish subjunctive”, Hispania 61:2, 218–34. Binnick, R. 1991 Time and the Verb. A Guide to Tense and Aspect. Oxford: Oxford University Press. Borghouts, L. 1979 “A new approach to the Late Egyptian conjunctive”, ZÄS 106, 14–24.
364
bibliography
——— 1985 “Aspectual values of the second tenses in Middle Egyptian”, in Schoske, S. (ed.), Akten des Vierten Internationalen Ägyptologen Kongresses vol. 3. Hamburg: Helmut Buske, 29–42. ——— 1986 “Prominence constructions and pragmatic functions”, in Englund & Frandsen 1986, 45–70. Breasted, J. 1905–06 Ancient Records of Egypt. Historical Documents from the Earliest Times to the Persian Conquest. 5 vols. Chicago: The University of Chicago Press. Brovarski, E. 2001 The Senedjemib Complex, Part I. The Mastabas of Senedjemib Inti (G 2570), Khnumenti (G 2374), and Senedjemib Mehi (G 2378). Boston: Museum of Fine Arts; Giza Mastabas vol. 7. Butt, J. & Benjamin, C. 1988 A New Reference Grammar of Modern Spanish. London: Arnold. ——— 2000 A New Reference Grammar of Modern Spanish (3rd ed.). London: Arnold. Bybee, J. 1995 “The semantic development of past tense modals in English”, in Bybee & Fleischman 1995, 503–17. ——— 1998 “‘Irrealis’ as a grammatical category”, Anthropological Linguistics 40:2, 257–71. Bybee, J. & Fleischman, S. (eds.) 1995 Modality in Grammar and Discourse. Amsterdam/ Philadelpia: Benjamins; Typological Studies in Language 32. Bybee, J., Pagliuca, W. & Perkins, R. 1994 The Evolution of Grammar: Tense, Aspect and Modality in the Languages of the World. Chicago/London: University of Chicago Press. Callender, J. 1975 Middle Egyptian. Malibu: Undena Publications; Afroasiatic Dialects 2. ——— 1977 Review of Schenkel 1975, BiOr 34, 305–07. ——— 1983 “Sentence initial position in Egyptian”, CdE 58, 83–96. ——— 1984 Studies in the Nominal Sentence in Egyptian and Coptic. Berkeley/Los Angeles/ London: University of California Press; Near Eastern Studies vol. 24. ——— 1986 “Problems of tense and aspect in Egyptian”, ZÄS 113, 8–18. Capart, J., Gardiner, A. & van de Walle, B. 1932 “New light on the Ramesside tomb-robberies”, JEA 22, 169–93. 1erný, J. & Groll, S. 1993 A Late Egyptian Grammar (fourth edition). Rome: Pontificio Instituto Biblico; Studia Pohl: Series Major 4. Chafe, W. 1995 “The realis-irrealis distinction in Caddo, the Northern Iroquoian languages, and English”, in Bybee & Fleischman 1995, 349–65. Chvany, C. 1985 “Backgrounded perfectives and plot line imperfectives: toward a theory of grounding in text”, in M. Flier & A. Timberlake (eds.), The Scope of Slavic Aspect. Columbus (Ohio): Slavica; UCLA Slavic Studies 12, 247–73. Claudi, U. & Mendel, D. 1991 “Noun/verb distinction in Egyptian-Coptic and Mande: a grammaticalization perspective”, in D. Mendel & U. Glaudi (eds.), Ägypten im Afroasiatischen Kontext. Aufsätze zur Archäologie, Geschichte und Sprache eines unbegrenzten Raumes. Gedenkschrift Peter Behrens. Cologne: University of Cologne; Afrikanistische Arbeitspapiere, special issue, 31–53. Collier, M.A. 1986 “A note on the syntax of xpr and omitted impersonal subjects in Late Egyptian”, Wepwawet 2, 15–22. ——— 1987 “xpr and the ‘raising’ paradigm in Middle Egyptian”, Wepwawet 3, 1–10. ——— 1990a “The circumstantial sDm(.f)/sDm.n(.f) as verbal verb-forms in Middle Egyptian”, JEA 76, 73–85.
bibliography
365
——— 1990b Review of Schenkel, W. 1987 Materialen zur Vorlesung. Einführung in die klassisch-ägyptische Sprache und Schrift, Wintersemester 1987/88, DE 16, 81–88. ——— 1990c Review of Schenkel, W. 1989 Materialen zur Vorlesung. Einführung in die klassisch-ägyptische Sprache und Schrift, Wintersemester 1989/90. DE 18, 83–90. ——— 1991a “‘Circumstantially adverbial’? The circumstantial sDm(=f)/sDm.n(=f) reconsidered”, in S. Quirke (ed.), Middle Kingdom Studies. New Malden: SIA Publishing, 21–50. ——— 1991b “The relative clause and the verb in Middle Egyptian”, JEA 77, 23–42. ——— 1991c “Constructions with HA revisited”, GM 120, 13–32. ——— 1992 “Predication and the circumstantial sDm(=f)/sDm.n(=f)”, LingAeg 2, 17–65. ——— 1994 “Grounding, cognition and metaphor in the grammar of Middle Egyptian”, LingAeg 4, 57–87. ——— 1999 “Of verbs and times past”, in A. Leahy and J. Tait (eds.), Studies on Ancient Egypt in Honour of H.S. Smith. London: Egypt Exploration Society Occasional Publications, 49–58. ——— 2005 “Reading ir wnn conditionals in the Heqanakhte documents”, LingAeg 13, 1–29. Comrie, B. 1976a “The syntax of causative constructions”, in Shibatani, M. (ed.), The Grammar of Causative Constructions. New York/San Francisco/London: Academic Press; Syntax and Semantics vol. 6, 261–312. ——— 1976b Aspect. Cambridge: Cambridge University Press; Cambridge Textbooks in Linguistics. ——— 1985 Tense. Cambridge: Cambridge University Press; Cambridge Textbooks in Linguistics. Coulmas, F. 1986 “Reported speech: some general issues”, in Coulmas, F. 1986 (ed.), Direct and Indirect Speech. Berlin/New York/Amsterdam: Mouton de Gruyter; Trends in Linguistics, Studies and Monographs 31, 1–22. Cristofaro, S. 2003 Subordination. Oxford: Oxford University Press; Oxford Studies in Typology and Linguistic Theory. Croft, W. 2001 Radical Construction Grammar. Syntactic Theory in Typological Perspective. Oxford: Oxford University Press DeBuck, A. 1952 Grammaire Élémentaire de Moyen Égyptien. Leiden: E.J. Brill. De Cenival, J. 1972 “Sur la forme sDm.f a redoublement ou mrr.f”, RdE 24, 40– 45. ——— 1977 “Les écritures du verbe mAA ‘voir’ dans les Textes des Sarcophages leurs consequences théoriques”, RdE 29, 21–37. De Jong, A. 1994 “Coffin Texts spell 38: the case of the father and son”, SAK 21, 141–57. De Morgan, J., Bouriant, U., Legrain, G., Jéqueir, G. & Barsanti, A. 1894 Catalogue des Monuments et Inscriptions de L’Égypte Antique 1: De la Frontière de Nubie a Kom Ombos. Vienna: Adolphe Holzhausen Depuydt, L. 1983 “The Standard Theory and the ‘emphatic’ forms in Classical (Middle) Egyptian: a historical survey”, OLP 14, 13–54. ——— 1993a “A history of research on the prospective sDm.f forms in Middle Egyptian”, JARCE 30, 11–31. ——— 1993b “Zur Bedeutung der Partikeln jsk und js”, GM 136, 11–24. ——— 1995 “Sentence patterns and verb form: Egyptian grammar since Polotsky”, Le Muséon 108, 39–48.
366
bibliography
Dik, S. 1980 Studies in Functional Grammar. London: Academic Press. Dixon, R. 2000 “A typology of causatives: form, syntax and meaning”, in R. Dixon & Aikhenwald, A. (eds.), Changing Valency: Case Studies in Transitivity. Cambridge: Cambridge University Press. Doret, E. 1979a “La forme sDm.n.f narrative”, BSEG 2, 13–22. ——— 1979b “The reading of the negation ”, JEA 65, 161–63. ——— 1980 “A note on the Egyptian construction Noun + sDm.f”, JNES 39, 37– 45. ——— 1986 The Narrative Verbal System of Old and Middle Egyptian. Geneva: Patrick Cramer; Cahiers d’Orientalisme XII. ——— 1990 “Phrase nominale identité et substitution dans les Textes Sarcophages, deuxième partie”, RdE 41, 39–56. Edel, E. 1955–64 Altägyptische Grammatik. Rome: Pontificium Institutum Biblicum; Analecta Orientalia 34/39. ——— 1959 “Die Herkunft des neuägyptisch-koptischen Personalsuffixes der 3. Person -w”, ZÄS 84, 17–38. Edmondson, W. 1981 Spoken Discourse: a Model for Analysis. London/New York: Longman. Elanskaya, A. 1981 “The T-causativa in Coptic”, in Young, D. (ed.), Studies Presented to Hans Jakob Polotsky. East Gloucester, MA: Pirtle & Polson, 80–130. Englund, G. 1988 Middle Egyptian, an Introduction. Uppsala: Balder. Englund, G & Frandsen, J. (eds.) 1986 Crossroad. Chaos or the Beginning of a New Paradigm. Copenhagen: CNI Publications. Erman, A. 1884 “Spuren eines alten Subjunctivs im Koptischen”, ZÄS 22, 28–37. ——— 1889a Die Sprache des Papyrus Westcar. Göttingen: Mittheilungen der Königlichen Gesellschaft der Wissenschaften zu Göttingen. ——— 1889b “ und die ihm analogen Formen“, ZÄS 27, 29–39. ——— 1906 “Die Geschichte des Schiffbrüchigen”, ZÄS 43, 1–26. ——— 1928 Ägyptische Grammatik (4th ed.). Berlin: von Reuther & Reichard. ——— 1933 Neuägyptische Grammatik (2nd edition). Leipzig: Wilhelm Engelmann. Erman, A. & Grapow, H. 1926–31 Wörterbuch der Aegyptischen Sprache, 5 vols. & Belegstellen. Leipzig: Hinrichs’sche Buchhandlung. Eyre, C. 1986 “Approaches to the analysis of Egyptian sentences: syntax and pragmatics”, in Englund & Frandsen 1986, 119–43. ——— 1985 “Tense or aspect in Middle Egyptian?”, in Schoske, S. (ed.), Akten des vierten internationalen Ägyptologen Kongresses vol. 3. Hamburg: Helmut Buske; SAK Beihefte, 51–65. ——— 1987 “Speculations on the structure of Middle Egyptian”, in Ray, J. (ed.), Lingua Sapientissima. A Seminar in Honour of H.J. Polotsky. Cambridge: Faculty of Oriental Studies, 22–46. ——— 1991 “Was Ancient Egyptian really a primitive language?”, LingAeg 1, 97– 123. ——— 1994 “Word order hierarchies and word order change in the history of Egyptian”, LingAeg 4, 117–38. Faulkner, R. 1962 “Spells 38–40 of the Coffin Texts”, JEA 48, 36–44. ——— 1968 “The pregnancy of Isis”, JEA 54, 40–44. ——— 1973–78 The Ancient Egyptian Coffin Texts, 3 vols. Oxford: Aris & Phillips. ——— 1979 “The prospective sDm.f in the Coffin Texts”, in J. Ruffle, G. Gaballa & K. Kitchen (eds.), Orbis Aegyptiorum Speculum : Glimpses of Ancient Egypt. Studies in Honour of H.W. Fairman. Oxford: Aris & Phillips, 1–3.
bibliography
367
——— 1996 A Concise Dictionary of Middle Egyptian (7th reprint). Oxford: Griffith Institute. Feder, F. 2004 “Das Ende des ‘narrativen’ Infinitivs im Sinuhe”, LingAeg 12, 53– 74. Ferrar, H. 1967 A French Reference Grammar, Second Edition. Oxford: Oxford University Press. Fleischman, S. 1985 “Discourse functions of tense-aspect oppositions in narrative: toward a theory of grounding”, Linguistics 23:6, 851–82. ——— 1989 “Temporal distance, a basic linguistic metaphor”, Studies in Language 13:1, 1–50. ——— 1995 “Imperfective and irrealis”, in Bybee & Fleischman 1995, 519–51. Foley, W. & Van Valin, R. 1984 Functional Syntax and Universal Grammar. Cambridge: Cambridge University Press. Frajzyngier, Z. 1995 “A functional theory of complementizers”, in Bybee & Fleischman 1995, 473–502. Frandsen, P. 1975 At-former. Copenhagen: Ægyptologisk Institut; Ægyptologiske Arbejdshæfter I. Gardiner, A.H. 1916 “The defeat of the Hyksos by KamÙse: the Carnavon Tablet, no.1”, JEA 3, 95–110. ——— 1920 “On certain participial formations in Egyptian”, Revue Égyptologique Nouvelle Série, vol. II, 42–55. ——— 1923 “A hitherto unnoticed negative in Middle Egyptian”, RecTrav 40, 79– 82. ——— 1947 Review of Polotsky 1944, JEA 33, 95–101. ——— 1957 Egyptian Grammar (3rd ed.). Oxford: Griffith Institute. Gardiner, A. & Sethe, K. 1928 Egyptian Letters to the Dead. London: Egypt Exploration Fund. Gaskins, L. 1978 Notes on Middle Egyptian Syntax. Berkeley: University of California. Gilula, M. 1969 “Two new conjunctions in Middle Egyptian”, JNES 28, 122–23. ——— 1970 Review of Satzinger 1968, JEA 56, 205–14. ——— 1971 “Coffin Texts spell 148”, JEA 57, 14–19. ——— 1972 “Enclitic particles in Middle Egyptian”, GM 2, 53–59. ——— 1978 “Pyr. 604C–D and Westcar 7/17–19”, JEA 64 45–51. ——— 1985 “The ‘emphatic’ form of the verb iw ‘to come’”, in J. Osing & G. Dreyer (eds.), Form und Mass. Beiträge zur Literatur, Sprache und Kunst des Alten Ägypten, Festschrift für Gerhard Fecht. Wiesbaden: Harrassowitz; Ägypten und altes Testament 12, 137–41. Givón, T. 1972 “Studies in ChiBemba and Bantu grammar”, Studies in African Linguistics 3, Supplement 3. ——— 1975a “Cause and control: on the semantics of interpersonal manipulation”, in Kimball, J.P. Syntax and Semantics. New York/San Francisco/London: Academic Press; Syntax and Semantics vol. 4, 59–89. ——— 1975b “Focus and the scope of assertion: some Bantu evidence”, Studies in African Linguistics 6:2, 185–205. ——— 1980 “The binding hierarchy and the typology of complements”, Studies in Language 4, 333–77. ——— 1982 “Tense-Aspect-Modality: the Creole prototype and beyond”, in Hopper, P. (ed.), Tense-Aspect: between Semantics and Pragmatics. Amsterdam/Philadelphia: Benjamins; Typological Studies in Language 1, 115–63.
368
bibliography
——— 1987 “Beyond foreground and background”, in Tomlin, R. (ed.), Coherence and Grounding in Discourse. Amsterdam/Philadelphia: Benjamins; Typological Studies in Language 11, 175–88. ——— 1991 “The evolution of dependent clause morpho-syntax in Biblical Hebrew”, in E. Traugott & B. Heine (eds.), Approaches to Grammaticalization vol. 2: Focus on Types of Grammatical Markers. Amsterdam/Philadelphia: Benjamins; Typological Studies in Language 19, 257–310. ——— 1994 “Irrealis and the subjunctive”, Studies in Language 18:2, 265–337. ——— 2001 Syntax: an Introduction, 2 vols. Amsterdam/Philadelphia: Benjamins. Givón, T. & Young, P. 2002 “Cooperation and interpersonal manipulation in the society of intimates”, in Shibatani, M. (ed.), The Grammar of Causation and Interpersonal Manipulation. Amsterdam/Philadelphia: Benjamins; Typological Studies in Language 48, 23–56. Goedicke, H. 1955 “Quotations in Old Kingdom inscriptions”, in Firchow, O. (ed.), Ägyptologische Studien: Festschrift Hermann Grapow. Berlin: Akademie Verlag, 93–106. ——— 1967 Königliche Dokumente aus dem Alten Reich. Wiesbaden: Harrassowitz; Ägyptologische Abhandlungen 14. Grandet, P. & Mathieu, B. 1997 Cours d’Égyptien Hiéroglyphique. Paris: Khéops. Greenberg, M. 1965 Introduction to Hebrew. New Jersey: Prentice-Hall. Greig, G. 1990 “The sDm=f and sDm.n=f in the Story of Sinuhe and the theory of the nominal (emphatic) verbs”, in Israelit-Groll, S. (ed.), Studies in Egyptology Presented to Miriam Lichtheim vol. 1. Jerusalem: Magnes Press, 264–348. Grieshammer, R. 1970 Das Jenseitsgericht in den Sargtexten. Wiesbaden: Harrassowitz; Ägyptologische Abhandlungen 20. ——— 1975/76 “Zur Formgeschichte der Sprüche 38–41 der Sargtexte”, OLP 6/7, 231–35. Griffiths, J. 1960 The Conflict of Horus and Seth from Ancient and Classical Sources. A study in Ancient Mythology. Liverpool: Liverpool University Press; Liverpool Monographs in Archaeology and Oriental Studies. Groll, S. 1969 “iw sDm.f in Late Egyptian”, JNES 28, 184–91. Gunn, B. 1924 Studies in Egyptian Syntax. Paris: Geuthner. ——— 1930 Review of Gardiner, A. & Sethe, K. 1928. JEA 16, 147–55. ——— 1949 “A special use of the cDm.f and cDm.n.f forms”, JEA 35, 21–24. Haegeman, L. & Guéron, J. 1999 English Grammar, a Generative Perspective. Oxford: Blackwell; Blackwell Textbooks in Linguistics. Haiman, J. 1978 “Conditionals are topics”, Language 54:3, 564–89. ——— 1985 Natural Syntax: Iconicity and Erosion. Cambridge: Cambridge University Press. ——— 1995 “Moods and metamessages: alienation as a mood”, in Bybee & Fleischman 1995, 329–45. Halliday, M. 1985 An Introduction to Functional Grammar, Second Edition. London: Arnold. Hannig, R. 1982 “Zum mittelägyptischen Tempussystem”, GM 56, 35–52. ——— 1987 “Modelle zum mittelägyptischen Tempussystem. Ein Überblick 1928– 1982”, Journal of Ancient Civilizations 2, 1–46. ——— 1991 Pseudopartizip und sDm.n=f. Der Kernbereich des mittelägyptischen Verbalsystems II. Hildesheim: Gerstenberg Verlag; Hildesheimer Ägyptologische Beiträge 32. ——— 2003 Ägyptisches Wörterbuch I Altes Reich und Erste Zwischenzeit. Mainz am Rhein: Philipp von Zabern.
bibliography
369
Hardy, H. & Gordon, L. 1980 “Types of adverbial and modal constructions in Tolkapaya”, International Journal of American Linguistics 46:3, 183–96. Hayes, W. 1946 “Royal decrees from the temple of Min at Coptus”, JEA 32, 3– 23. Heine, B. 1993 Auxiliaries. Cognitive Forces and Grammaticalization. Oxford: Oxford University Press. Heine, B. & Kuteva, T. 2002 World Lexicon of Grammaticalization. Cambridge: Cambridge University Press. Helck, W. 1975 Historisch-Biographische Texte der 2. Zwischenzeit und Neue Texte der 18. Dynastie. Wiesbaden: Harrassowitz; Kleine Ägyptische Texte. Hengeveld, K. 1992 Non-verbal Predication. Theory, Typology, Diachrony. Berlin/New York: Mouton de Gruyter; Functional Grammar Series 15. Hofling, C. 1998 “Irrealis and perfect in Itzaj Maya”, Anthropological Linguistics 40:2, 214–27. Hooper, J. 1975 “On assertive predicates”, in Kimball, J.P. Syntax and Semantics. New York/San Francisco/London: Academic Press; Syntax and Semantics vol. 4, 91–124. Hooper, J. & Terrell, T. 1974 “A semantically based analysis of mood in Spanish”, Hispania 57, 484–94. Hopper, P. 1979 “Some observations on the typology of focus and aspect in narrative language”, Studies in Language 3:1, 37–64. ——— 1981 “Aspect and Foregrounding in Discourse”, in Givón, T. (ed.), Discourse and Syntax. New York/San Francisco/London: Academic Press; Syntax and Semantics vol. 12, 213–241. Hopper, P. & Thompson, S. 1980 “Transitivity in grammar and discourse”, Language 56:2, 251–99. Hopper P. & Traugott, E. 1993 Grammaticalization. Cambridge: Cambridge University Press; Cambridge Textbooks in Linguistics. Horie, K. & Comrie, B. 2000 An introductory essay in Horie, K. (ed.), Complementation Cognitive and Functional Perspectives. Amsterdam/Philadelphia: Benjamins, 1–10. James, D. 1982 “Past tense and the hypothetical, a cross-linguistic study”, Studies in Language 6:3, 375–403. James, T. 1962 The ¡eqanakhte Papers and other Early Middle Kingdom Documents. New York: The Metropolitan Museum of Art Egyptian Expedition. Jansen-Winkeln, K. 1995 “Finalsatz und Subjunktiv”, GM 146, 37–60. ——— 1996a Spätmittelägyptische Grammatik der Texte der 3. Zwischenzeit. Wiesbaden: Harrassowitz; Ägypten und Altes Testament 34. ——— 1996b “Zitierform und Kontextform”, GM 154, 45–48. ——— 1997 “Intensivformen und ‘verbale Pluralität’ im Ägyptischen”, LingAeg 5, 123–36. ——— 2002 “Zur Bedeutung von Hzj und mrj”, GM 190, 47–52. Jenni, H. 2002 “‘Sagen zu’ im Ägyptischen”, LingAeg 10, 239–59. ——— 2004 “Sätze zum Ausdruck von Zugehörigkeit und Besitz im Ägyptischen”, LingAeg 12, 123–31. Johnson, J. 1980 “NIMS in Middle Egyptian”, Serapis 6, 69–73. ——— 1984 “The use of the particle mk in Middle Kingdom letters”, in Junge 1984, 71–85. Judge, A. & Healey, F. 1983 A Reference Grammar of Modern French. London: Edward Arnold.
370
bibliography
Junge, F. 1973 “Satz und Feld”, GM 6, 61–90. ——— 1978a Syntax der Mittelägyptischen Literatursprache. Grundlagen einer Strukturtheorie. Mainz: Philipp von Zabern. ——— 1978b Review of Callender 1975, BiOr 35, 31–35. ——— 1979 “Adverbialsatz und emphatische Formen, Nominalsatz und Negation. Eine ‘Gegendarstellung’”, GM 33, 69–88. ——— 1983 “Form und Funktion ägyptischer Satzfragen”, BiOr 40, 545–59. ——— 1984 (ed.) Studien zu Sprache und Religion Ägyptens zu Ehren von Wolfhart Westendorf überreicht von seinen Freunden und Schülern vol. 1: Sprache. Göttingen: Hubert & Co. ——— 1989 “Emphasis” and Sentential Meaning in Middle Egyptian. Göttingen: Harrassowitz; Göttinger Orientforschungen IV/20. ——— 1991 “How to study Egyptian grammar and to what purpose, a summary of sorts”, LingAeg 1, 389–426. ——— 2001 Late Egyptian Grammar: an Introduction (Translated by D. Warburton). Oxford: Griffith Institute. Junker, H. 1940 Die Götterlehre von Memphis (Schabaka-Inschrift). Berlin: Verlag der Akademie der Wissenschaften. Kac, M. 1972 “Clauses of saying and the interpretation of because”, Language 48, 626–32. Kammerzell, F. 1988 “Zweifel and der Existenz einer morphologisch-syntaktischen Kategorie “indikativ-perfektivishes sDm=f“ im Mäg.”, GM 102, 41–57. Kammerzell, F. & Peust, C. 2002 “Reported speech in Egyptian. Forms, types and history”, in T. Güldemann & M. von Roncador (eds.), Reported Discourse. A Meeting Ground for Different Linguistic Domains. Amsterdam/Philadelphia: Benjamins; Typological Studies in Language 2, 289–322. Karttunen, L. 1971 “Implicative verbs”, Language 47, 340–58. Kemp, B. 1989 Ancient Egypt: Anatomy of a Civilization. London/New York: Routledge. Kiparsky, P. & Kiparsky, C. 1971 “Fact”, in D. Steinberg & L. Jakobovits (eds.), Semantics: an Interdisciplinary Reader in Philosophy, Linguistics and Psychology. Cambridge: Cambridge University Press, 345–69. Klein, F. 1975 “Pragmatic constraints on distribution: the Spanish subjunctive”, in R. Grossman, L. San & T. Vance (eds.), Papers from the Eleventh Regional Meeting, Chicago Linguistic Society. Chicago: Chicago Linguistic Society, 353–65. Korostovtsev, M. 1973 Grammaire du Neo-Egyptien. Moscow: Departement de la Litterature Orientale; Edition Naouka. Kruchten, J. 1982 Etudes de Syntaxe Néo-égyptienne. Les Verbes aHa, Hmsi et sDr en Néo-égyptien. Emploi et Significations. Brussels: Université de Bruxelles; Annuaire de l’Institut de Philologie et d’Histoire Orientales et Slaves, Supplement 1. ——— 1994 “Ir wnn cDm.f (cDm.n.f) et ir cDm.f (cDm.n.f): une approche structuraliste”, JEA 80, 97–108. ——— 1997 “About iw and wn(n) in Late Egyptian”, ZÄS 124, 57–70. Kulikov, L. 2001 “Causatives”, in M. Haspelmath, E. König, W. Oesterreicher & W. Raible (eds.), Language Typology and Language Universals: an International Handbook vol. 2. Berlin/New York: Walter de Gruyter; Handbücher zur Sprach- und Kommunikationswissenschaft, 886–98. Lambrecht, K. 1994 Information Structure and Sentence Form. Topic, Focus and the Mental Representations of Discourse Referents. Cambridge: Cambridge University Press; Cambridge Studies in Linguistics 71.
bibliography
371
Langacker, R. 1991 Foundations of Cognitive Grammar, vol. 2, Descriptive Application. Stanford: Stanford University Press. Lavandera, B. 1983 “Shifting moods in Spanish discourse”, in Klein-Andreu, F. (ed.), Discourse Perspectives on Syntax. London: Academic Press. Layton, B. 2000 A Coptic Grammar. Wiesbaden: Harrassowitz; Porta Linguarum Orientalium. Lefebvre, G. 1955 Grammaire de L’Égyptien Classique (2nd edition). Cairo: IFAO. Levinson, S. 1983 Pragmatics. Cambridge: Cambridge University Press; Cambridge Textbooks in Linguistics. L’Huillier, M. 1999 Advanced French Grammar. Cambridge: Cambridge University Press. Li, C. 1986 “Direct and indirect speech: a functional study”, in Coulmas, F. 1986 (ed.), Direct and Indirect Speech. Berlin/New York/Amsterdam: Mouton de Gruyter; Trends in Linguistics, Studies and Monographs 31, 29–45. Lichtenberk, F. 1995 “Apprehensional epistemics”, in Bybee & Fleischman 1995, 293–327. Lichtheim, M. 1992 Maat in Egyptian Autobiographies and Related Studies. Göttingen: Vandenhoeck & Ruprecht; Orbis Biblicus et Orientalis 120. Lopez, J. 1963 “Le Papyrus Millingen”, RdE 15, 29–33. Loprieno, A. 1984 “Aspekt und Diathese im Ägyptischen”, in Junge 1984, 87–102. ——— 1986a Das Verbalsystem im Ägyptischen und im Semitischen. Zur Grundlegung einer Aspekttheorie. Wiesbaden: Harrassowitz; Göttinger Orientforschungen IV/17. ——— 1986b “Egyptian grammar and textual features”, in Englund & Frandsen 1986, 255–87. ——— 1988a “Verbal forms and verbal sentences in Old and Middle Egyptian”, GM 102, 59–72. ——— 1988b “On the typological order of constituents in Egyptian”, Journal of Afroasiatic Linguistics 1, 26–57. ——— 1991a “Focus, mood and negative forms: Middle Egyptian syntactic paradigms and diachrony”, LingAeg 1, 201–26. ——— 1991b “Topics in Egyptian negations”, in D. Mendel & U. Glaudi (eds.), Ägypten im Afroasiatischen Kontext. Aufsätze zur Archäologie, Geschichte und Sprache eines unbegrenzten Raumes. Gedenkschrift Peter Behrens. Cologne: University of Cologne; Afrikanistische Arbeitspapiere, special issue, 213–35. ——— 1995 Ancient Egyptian; a Linguistic Introduction. Cambridge: Cambridge University Press. Luft, U. 1984 “DJ.J RH.K”, in Junge, F. 1984, 103–11. Lunn, P. 1985 “The aspectual lens”, Hispanic Linguistics 2:1, 49–61. ——— 1989a “The Spanish subjunctive and ‘relevance’”, in C. Kirschner & J. DeCesaris (eds.), Studies in Romance Linguistics. Amsterdam/Philadephia: Benjamins; Current Issues in Linguistic Theory 60, 249–60. ——— 1989b “Spanish mood and the prototype of assertability”, Linguistics 27:4, 687–702. ——— 1995 “The evaluative function of the Spanish subjunctive”, in Bybee & Fleischman 1995, 430–49. Lyons, J. 1977 Semantics vol. 2. Cambridge: Cambridge University Press. ——— 1982 “Deixis and subjectivity: loquor, ergo sum?”, in R. Jarvella & W. Klein (eds.), Speech, Place and Action. London: Wiley, 101–24. Maiden, M. & Robustelli, C. 2000 A Reference Grammar of Modern Italian. London/New York: Arnold.
372
bibliography
Malaise, M. & Winand, J. 1999 Grammaire Raisonnée de l’Égyptien Classique. Liège: The University of Liège; Ægyptiaca Leondiensia 6. Martin, L. 1998 “Irrealis constructions in Mocho (Mayan)”, Anthropological Linguistics 40:2, 198–213. Meltzer, E. 1987 Review of Simpson W. 1986 Essays on Egyptian Grammar, JARCE 34, 148–51. ——— 1991 “Participles, relative forms and gemination in Middle Egyptian: a working paper”, LingAeg 1, 227–40. Mithun, M. 1995 “On the relativity of irreality”, in Bybee & Fleischman 1995, 367–88. Myhill J. & Smith, L. 1995 “Functions of obligation expressions”, in Bybee & Fleischman 1995, 239–92. Newman, J. 1996 Give: a Cognitive Linguistic Study. Berlin/New York: Mouton de Gruyter. Noonan, M. 1985 “Complementation”, in Shopen, T. (ed.), Language Typology and Syntactic Description vol. 2: Complex Constructions. Cambridge: Cambridge University Press, 42–140. Ogdon, J. 1996 “A note on and as negative morphemes”, DE 34, 61–63. Otto, E. 1969 “Bedeutungsnuancen der Verben mrj ‘lieben’ und mcDj ‘hassen’”, MDAIK 25, 98–100. Palmer, F. 1986 Mood and modality. Cambridge: Cambridge University Press; Cambridge Textbooks in Linguistics. ——— 2000 Mood and modality, Second Edition. Cambridge: Cambridge University Press; Cambridge Textbooks in Linguistics. Parkinson, R. 1997 The Tale of Sinuhe and Other Ancient Egyptian Poems 1940–1640 BC. Oxford: Oxford University Press; Oxford World’s Classics. Perdu, O. 1978 “La préposition xft et les temporelles non-concomitantes”, RdE 30, 101–14. Perschbacher, W. 1995 New Testament Greek Syntax. Chicago: Moody Press. Peust, C. 1996 Indirekte Rede im Neuägyptischen. Wiesbaden: Harrassowitz; Göttinger Orientforschungen IV/33. ——— 2004 “Das Ägyptische als afrikanische Sprache”, in Schneider, T. (ed.), Das Ägyptische und die Sprachen Vorderasiens, Nordafrikas und der Ägäis. Akten des Basler Kolloquiums zum ägyptisch-nichtsemitischen Sprachkontakt. Ugarit-Verlag: Münster; Alter Orient und Altes Testament 310, 321–407. ——— 2005: “Weiteres zur Personenverschiebung in der ägyptischen indirekten Rede”, LingAeg 13, 77–102. Pierret, P. 1874–78 Recueil d’Inscriptions Inédites du Musée Égyptien du Louvre, 2 vols. Paris: Librairie A. Franck. Polotsky, H. 1944 Études de Syntaxe Copte. Cairo: Publications de la Societe d’Archéologie Copte. ——— 1957 “The ‘emphatic’ sdm.n.f form”, RdE 11, 109–17. ——— 1960 “The Coptic conjugation system”, Or 29, 392–422. ——— 1964 “Ägyptische Verbalformen und ihre Vokalisation”, Or 33, 267–85. ——— 1965 “Egyptian tenses”, Israel Academy of Sciences and Humanities vol. II/5, 1–26. ——— 1969 “Zur altägyptische Grammatik”, Or 38, 465–81. ——— 1976 “Les transpositions du verbe en Égyptien Classique”, Israel Oriental Studies 6, 1–50.
bibliography
373
——— 1984 “Randbemerkungen”, in Junge 1984, 113–23. ——— 1987 “Egyptology, Coptic studies and the Egyptian language”, in Ray, J. (ed.), Lingua Sapientissima. A Seminar in Honour of H.J. Polotsky. Cambridge: Faculty of Oriental Studies, 5–21. ——— 1990 “‘Perfekt’ und ‘Imperfekt’: von Sethe zu Gardiner”, in Israelit-Groll, S. (ed.), Studies in Egyptology Presented to Miriam Lichtheim vol. 2. Jerusalem: Magnes Press, 768–72. Quecke, H. 1990 “Zur direkten und indirekten Rede im Koptischen”, Journal of Coptic Studies 1, 129–35. Ramsey, V. 1987 “The functional distribution of preposed and postposed ‘if’ and ‘when’ clauses in written discourse”, in Tomlin, R. (ed.), Coherence in Grounding and Discourse. Amsterdam/Philadelphia: Benjamins; Typological Studies in Language 11, 383–408. Ransom, E. 1986 Complementation: Its Meaning and Forms. Amsterdam/Philadelphia: Benjamins; Typological Studies in Language 10. Ray, J. 2000 “To be or not to be? two problems in Early Egyptian grammar”, LingAeg 7, 225–32. Reinhart, T. 1984 “Principles of gestalt perception in the temporal organization of narrative texts”, Linguistics 22:6, 779-809. Reintges, C. 1997 Passive Voice in Older Egyptian, a Morpho-Syntactic Study. The Hague: Holland Institute of Generative Linguistics; HIL Dissertations 28. ——— 2004 Review of Layton, B. 2000, JEA 90, 242–48. Ritter, T. 1992 Review of Vernus 1990, LingAeg 2, 177–87. ——— 1995 Das Verbalsystem der königlichen und privaten Inschriften XVIII. Dynastie bis einschließlich Amenophis III. Wiesbaden: Harrassowitz; Göttinger Orientforschungen IV/30. Roberts, J. 1990 “Modality in Amele and other Papuan languages”, Journal of Linguistics 26, 363–401. Roccati, A. 1979 “Abgeleitete Konstruktionen des Ägyptischen”, GM 36, 39–49. ——— 1982 La Littérature Historique sous L’Ancien Empire Égyptien. Paris: Les Éditions du Cerf. Rutherford, W. 1970 “Some observations concerning subordinate clauses in English”, Language 46:1, 97–115. Saeed, J. 1997 Semantics. Oxford: Blackwell; Blackwell Textbooks in Linguistics. Sander-Hansen, C. 1941 Über die Bildung der Modi im Altägyptischen. Copenhagen: Ejnar Munksgaard. Sasse, H. 1987 “The thetic/categorical distinction revisited”, Linguistics 25:3, 511– 80. ——— 2001 “Scales between nouniness and verbiness”, in M. Haspelmath, E. König, W. Oesterreicher & W. Raible (eds.), Language Typology and Language Universals: an International Handbook vol. 1. Berlin/New York: Walter de Gruyter; Handbücher zur Sprach- und Kommunikationswissenschaft, 495–509. Satzinger, H. 1968 Die Negativen Konstruktionen im Alt- und Mittelägyptischen. Berlin: Bruno Hessling; Münchner Ägyptologische Studien 12. ——— 1971 “sDm.t.f ‘schliesslich hörte er’”, JEA 57, 58–69. ——— 1976 Neuägyptische Studien. Die Partikel ir. Das Tempussystem. Vienna: Verlag des Verbandes der wissenschaftlichen Gesellschaften Österreichs; WZKM Beihefte 6. ——— 1986 “On tense and aspect in Middle Egyptian”, in Englund & Frandsen 1986, 297–313.
374
bibliography
——— 1987 Review of Junge 1984, BiOr 44, 617–28. ——— 1989 “Bemerkungen zum ägyptischen Verbalsystem gelegentlich zweier Neuerscheinungen (Rezensionsartikel)”, WZKM 79, 197–220. ——— 1993a “‘Emphase’ oder die Rhematisierung eines adverbialen Komplements im Ägyptischen (Rezensionsartikel)”, WZKM 83, 189–206. ——— 1993b “Die Protasis jr sdm.f im älteren Ägyptisch”, LingAeg 3, 121–35. ——— 2001 “Some remarks on a newly discovered noun clause construction of Late Egyptian”, LingAeg 9, 239–47. Scharff, A. 1924 “Briefe aus Illahun”, ZÄS 59, 20–51. Schenkel, W. 1963 “Beiträge zur mittelägyptischen Syntax”, ZÄS 88, 113–30. ——— 1965 “‘Singularisches’ und ‘pluralisches’ Partizip”, MDAIK 20, 110–14. ——— 1975 Die Altägyptische Suffixkonjugation. Theorie der innerägyptische Entstehung aus Nomina actionis. Wiesbaden: Harrassowitz; Ägyptologische Abhandlungen 32. ——— 1978 “Eine Syntax des Klassischen Ägyptisch ohne Verbalsatz”, GM 29, 105–17. ——— 1980 “Thesen zum Ägyptischen Sprachunterricht”, GM 40, 81–97. ——— 1981 “sDm=f und sDm.w=f als Prospektivformen”, in Young, D. (ed.), Studies Presented to Hans Jakob Polotsky. East Gloucester, MA: Pirtle & Polson, 506–17. ——— 1990 Einführung in die altägyptische Sprachwissenschaft. Darmstadt: Wissenschaftliche Buchgesellschaft; Orientalistische Einführungen. ——— 1992 Review of Vernus 1990, OLZ 87, 377–80. ——— 1999 “c-Kausativa, t-Kausativa und “innere” Kausativa. Die c-kausativa der Verben I.c in den Sargtexten”, SAK 27, 313–52. ——— 2000a “Die Endungen des Negativkomplements im Spiegel der Befunde der Sargtexte”, LingAeg 7, 1–26. ——— 2000b “Die Endungen des Prospektivs und Subjunktivs (cc.m=f, cc.m.w=f, cc.m.y=f) nach Befunden der Sargtexte”, LingAeg 7, 27–112. ——— 2005 Tübinger Einführung in die klassich-ägyptische Sprache und Schrift. Tübingen: Published by the Author. ——— 2006 “Prädikatives und abstract-relativisches cc.m.n=f. Beobachtungen an den Verben II.gem. und ult. n im Korpus der Sargtexte”, in G. Goldenberg & A. Shisha-Halevy (eds.), Egyptian, Semitic and General Grammar. Workshop in Memory of H. J. Polotsky. Jerusalem: The Israel Academy of Sciences and Humanities, 40–60. Searle, J. 1969 Speech Acts. An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. ——— 1975 “Indirect speech acts”, in P. Cole & J. Morgan (eds.), Speech Acts. New York/San Francisco/London: Academic Press; Syntax and Semantics vol. 3, 59–82. Sethe, K. 1899 Das Aegyptische Verbum im Altaegyptischen, Neuaegyptischen und Koptischen, vol.2. Leipzig: Hinrichs’sche Buchhandlung. ——— 1907 “Bemerkungen zur ‘Geschichte des Schiffbrüchigen’”, ZÄS 44, 80–87. ——— 1916 Der Nominalsatz im Ägyptischen und Koptischen. Leipzig: Teubner; Abhandlungen der philologisch-historischen Klasse der königl. Sächsischen Gesellschaft der Wissenschaften 3. ——— 1927 Erläuterungen zu den Aegyptischen Lesestücken. Leipzig: Hinrichs’sche Buchhandlung. ——— 1928 Dramatische Texte zu altaegyptischen Mysterienspiele II: Der dramatische Ramesseumpapyrus, ein Spiel zur Thronbesteigung des Königs. Leipzig: Hinrichs’sche
bibliography
375
Buchhandlung; Untersuchungen zur Geschichte und Altertumskunde Aegyptens 10. Shibatani, M. 1976 “The grammar of causative constructions: a conspectus”, in Shibatani, M. (ed.), The Grammar of Causative Constructions. New York/San Francisco/London: Academic Press; Syntax and Semantics vol. 6, 1–40. ——— 2001 “Some basic issues in the grammar of causation”, in Shibatani, M. (ed.), The Grammar of Causation and Interpersonal Manipulation. Amsterdam/Philadelphia: Benjamins; Typological Studies in Language 48, 1–22. Shibatani, M. & Pardeshi, P. 2001 “The causative continuum”, in Shibatani, M. (ed.), The Grammar of Causation and Interpersonal Manipulation. Amsterdam/Philadelphia: Benjamins; Typological Studies in Language 48, 87–126. Shisha-Halevy, A. 1987 “Grammatical discovery procedure and the Egyptian-Coptic nominal sentence”, Or 56, 147–75. Silverman, D. 1980 Interrogative Constructions with JN and JN-JW in Old and Middle Egyptian. Malibu: Undena; Bibliotheca Aegyptia I. ——— 1985 “Verbal nominal clauses in Middle Egyptian”, in Melanges Gamal Eddin Mokhtar. Cairo: IFAO; Bibliothèque dÉtude 97, 269–85. ——— 1986a “Determining initiality of clauses in Middle Egyptian”, in Simpson, W. (ed.) Essays on Egyptian Grammar. New Haven: Yale Egyptological Seminar; Yale Egyptological Studies 1, 34–41. ——— 1986b “Clauses in Middle Egyptian”, in Englund & Frandsen 1986, 315– 35. Simpson, W. 1972 (ed.) The Literature of Ancient Egypt. An Anthology of Stories, Instruction, and Poetry with Translations by R.O. Faulkner, E.F. Wente, and W.K. Simpson. New Haven/London: Yale University Press. Slobin, A. & Aksu, A. 1982 “Tense, aspect, and modality in the use of Turkish evidentials”, in Hopper, P. (ed.), Tense-Aspect: between Semantics and Pragmatics. Amsterdam/Philadelphia: Benjamins; Typological Studies in Language 1, 185–200. Smith H. & Smith, A. 1976 “A reconsideration of the Kamose texts”, ZÄS 103, 48–76. Smither, P. 1939 “A new use of the preposition m”, JEA 25, 166–69. Song, J. 1996 Causatives and Causation: a Universal-Typological Perspective. London/New York: Longman Sperber, D. & Wilson, D. 1995 Relevance, Second Edition. Oxford: Blackwell. Spiegelberg, W. 1923 “Der stat. constr. –qN vor folgendem Fragesatze”, ZÄS 58, 157. Steele, S. 1975 “Past and irrealis: just what does it all mean?”, International Journal of American Linguistics 41, 200–17. Strudwick, N. 2005 Texts from the Pyramid Age. Atlanta: Society of Biblical Literature; Writings from the Ancient World 16. Suzuki, S. 2000 “De dicto complementation in Japanese”, in Horie, K. (ed.), Complementation - Cognitive and Functional Perspectives. Amsterdam/Philadelphia: Benjamins, 34–57. Sweeney, D. 1986 “The nominal object clause of verbs of perception in non-literary Late Egyptian”, in Englund & Frandsen 1986, 337–73. ——— 2001 Correspondence and Dialogue, Pragmatic Factors in Late Ramesside Letter-Writing. Wiesbaden: Harrassowitz; Ägypten und Altes Testament 49. Sweetser, E. 1990 From Etymology to Pragmatics. Metaphorical and Cultural Aspects of
376
bibliography
Semantic Structure. Cambridge: Cambridge University Press; Cambridge Studies in Linguistics 54. Talmy, L. 1976 “Semantic causative types”, in Shibatani, M. (ed.), The Grammar of Causative Constructions. New York/San Francisco/London: Academic Press; Syntax and Semantics vol. 6, 43–116. ——— 1978 “Figure and ground in complex sentences”, in Greenberg, J. (ed.), Universals of Human Language vol. 4: Syntax. Stanford: Stanford University Press, 625–49. ——— 1988 “Force dynamics in language and cognition”, Cognitive Science 2, 49– 100. ——— 2000 Toward a Cognitive Semantics vol. 1: Concept Structuring Systems. Cambridge, MA/London: The MIT Press. Thacker, T. 1949 “On the structure of the expressions denoting existence and nonexistence in Middle Egyptian”, JEA 35, 31–37. ——— 1954 The Relationship of the Semitic and Egyptian Verbal Systems. Oxford: Oxford University Press. Thompson, S. 1987 “‘Subordination’ and narrative event structure”, in Tomlin, R. (ed.), Coherence and Grounding in Discourse. Amsterdam/Philadelphia: Benjamins; Typological Studies in Language 11, 435–54. Thompson, S. & Longacre, R. 1985 “Adverbial clauses” in Shopen, T. (ed.), Language Typology and Syntactic Description vol. 2: Complex Constructions. Cambridge: Cambridge University Press, 171–234. Till, W. 1933 “Der Irrealis im Neuägyptischen”, ZÄS 69, 112–17. Timberlake, A. 1982 “Invariance and the syntax of Russian aspect”, in Hopper, P. (ed.), Tense-Aspect: between Semantics and Pragmatics. Amsterdam/Philadelphia: Benjamins; Typological Studies in Language 1, 305–31. Tomlin, R. 1985 “Foreground-background information and the syntax of subordination”, Text 5:1/2, 85–122. Townsend, D. & Bever, T. 1977 Main and Subordinate Clauses: a Study in Figure and Ground. Indiana: Indiana University Linguistics Club. Uljas, S. 2000 “Of meaning and modality in Middle Egyptian object complementation”, in A. McDonald & C. Riggs (eds.), Current research in Egyptology. London: Archaeopress; British Archaeological Reports, International Series 909, 125– 34. ——— 2003 “On interclausal relations in Middle Egyptian object complementation”, in S. Bickel & A. Loprieno (eds.), Basel Egyptology Prize 1. Basel: Schwabe & Co.; Aegyptiaca Helvetica 17, 387–403. ——— 2004 “Cairo Bowl lines 7–8”, GM 201, 95–104. ——— 2006 “xpr.n and the genesis of auxiliaries”, SAK 35, 327–36. ——— In press: “Noun/personal pronoun + personal pronoun as a grammatical construction in Earlier Egyptian”, to appear in JEA 92 (2006). Van den Boorn, G. 1988 The Duties of the Vizier: Civil Administration in the Early New Kingdom. London/New York: Kegan Paul; Studies in Egyptology. Vernus, P. 1976 “Deux inscriptions de la XIIe dynastie provenant de Saqqara”, RdE 28, 119–38. ——— 1981 “Formes ‘emphatiques’ en fonction non ‘emphatiques’ dans la protase d’un systeme correlatif”, GM 43, 73–88. ——— 1984 “Sujet + sDm.f et sujet + pseudoparticipe avec les verbes de qualité: dialeqtique de l’aspect et de l’aktiosart”, in Junge 1984, 197–212.
bibliography
377
——— 1986 “Aspect and morphosyntactic patterns in Middle Egyptian”, in Englund & Frandsen 1986, 375–88. ——— 1987 “Études de philologie et de linguistique VI”, RdE 38, 163–81. ——— 1990 Future at Issue. Tense, Mood and Aspect in Middle Egyptian: Studies in Syntax and Semantics. New Haven: Yale Egyptological Seminar; Yale Egyptological Studies 4. ——— 1994 “Observations sur la predication de classe (“nominal predicate”)”, LingAeg 4, 325–48. ——— 1999 “Une formulation de l’autobiographie et les expressions avec wn et mAa”, GM 170, 101–05. Vidal, A. & Klein, H. 1998 “Irrealis is Pilagá and Toba? Syntactic versus pragmatic coding”, Anthropological Linguistics 40:2, 175–97. Wallace, S. 1982 “Figure and Ground: the Interrelationship of Linguistic Categories”, in Hopper, P. (ed.), Tense-Aspect: Between Semantics and Pragmatics. Amsterdam/ Philadelphia: Benjamins; Typological Studies in Language 1, 201–26. Waugh , L. & Monville-Burston, M. 1986 “Aspect and discourse function: the French simple past in newspaper usage”, Language 62:4, 846–77. Wente, E. 1969 “A Late Egyptian emphatic tense”, JNES 28, 1–14. ——— 1990 Letters from Ancient Egypt. Atlanta: Scholars Press; Society of Biblical Literature Writings from the Ancient World 1. Westendorf, W. 1953 Der Gebrauch des Passivs in der klassischen Literatur der Ägypter. Berlin: Akademie Verlag. ——— 1962 Grammatik der Medizinischen Texte. Berlin: Akademie Verlag; Grundriss der Medizin der Alten Ägypter VIII. ——— 1981 Beiträge zum altägyptischen Nominalsatz. Göttingen: Vandenhoeck & Ruprecht; Nachrichten der Akademie der Wissenschaften in Göttingen, I. philologischhistorische Klasse. Wierzbicka, A. 1987 English Speech Act Verbs: a Semantic Dictionary. Sydney: Academic Press. ——— 1988 The Semantics of Grammar. Amsterdam/Philadelphia: Benjamins; Studies in Language Companion Series 18. Willems, H. 1991 “The end of Seankhenptah’s household (Letter to the Dead Cairo JDE 25975)”, JNES 50, 183–91. Willett, T. 1988 “A cross-linguistic survey of the grammaticization of evidentiality”, Studies in Language 12:1, 51–97. Willis, C. 1974 An Essential Course in Modern Portuguese, Revised Edition. London: Harrap. Winand, J. 1991 “Le verbe iy/iw: unite morphologique et sémantique”, LingAeg 1, 357–87 ——— 2001 “À la croisée du temps, de l’aspect et du mode—Le conjonctif en néoégyptien”, LingAeg 9, 293–329. ——— 2006 Temps et Aspect en Égyptien. Une Approche Sémantique. Leiden/Boston: Brill; Probleme der Ägyptologie 25. Wycichl, W. 1954 “Der Aufbau der ägyptischen Pronomina nt-f, nt-s etc”, Le Muséon 67, 367–72. Zonhoven, L. 1993–94 “Polotsky, Sinuhe, negation and the sDm.n=f. On the existence of an indicative sDm.n=f in Middle Egyptian”, JEOL 33, 39–108. ——— 1996 “Studies on the sDm.t=f verb form in Classical Egyptian I: the construction Dr sDm.t=f”, BiOr 53, 613–44.
378
bibliography
——— 1997a “Studies on the sDm.t=f verb form in Classical Egyptian II: the active r sDm.t=f construction”, OLP 28, 5–31. ——— 1997b “Studies on the sDm.t=f verb form in Classical Egyptian III: the active n sDm.t=f construction”, in van Dijk, J. (ed.), Essays on Ancient Egypt in Honour of Herman Te Welde. Groningen: Styx, 383–400.
index of cited passages
379
INDEX OF CITED PASSAGES citation
line/ section
example/ published in/source of the footnote* citation
Ächt
d 1–6 h 4–5 l 4–5 n 4–6
(37) 2/n.42 2/n.42 2/n.42
Sethe, K. 1926 Die Ächtung feindlicher Fürsten, Völker und Dinge auf Altägyptischen Tongefässscher des Mittleren Reiches. Berlin: Verlag der Akademie der Wissenschaften.
Admonitions
2, 8 3, 9 4, 2 5, 3 5, 4 5, 9 6, 5 12, 6 12, 12
Gardiner, A. 1909 The Admonitions of an Egyptian Sage from a Hieratic Papyrus in Leiden (Pap. Leiden 344 recto). Leipzig: Hinrichs’sche Buchhandlung.
13, 4
6/n.50 (198) 3/n.88 2/n.79 4/n.28 (195) (337) (338) 6/n.48; 6/n.36 6/n.16
VI, pl. 21, middle 2
6/n.2
Amenemhat
IIIa–b VIIa–b VIIc VIIe VIIIa–b
(9) (108) 7/n.31 4/n.19 (98)
Amrah
pl. 33, no. 2, 4–5
6/n.30
ASAE 55
240
6/n.37
Balat-aAyn Asil 3686
1–3
(69)
Amarna
Davis, N. 1903–08 The Rock Tombs of El Amarna. London: Egypt Exploration Fund. Helck, W. 1969 Der Texte der ‘Lehre Amenemhets I. für seinen Sohn’. Wiesbaden: Harrassowitz; Kleine Ägyptische Texte. Randall-Maciver, D. & Mace, A. 1902 El Amrah and Abydos. London: Egypt Exploration Fund. Drioton, E. & Lauer, J. 1958 “Un groupe de tombes à Saqqarah: Icheti, Nefer-Khouou-Ptah, Sébek-em-khent et Ânkhi”, ASAE 55, 207–51. Pantalacci, L. 1998 “La documentation épistolaire du palais des gouverneurs à Balat-aAyn Asil”, BIFAO 98, 303–15.
∗ Numbers in brackets refer to examples cited in the main text. Numbers followed by / and ‘n.’ + number refer to passages cited in the footnotes of individual chapters. Thus e.g. ‘2/n.75’ refers to note 75 in chapter 2. The introduction is designated as ‘chapter zero’ (0).
380
index of cited passages
citation
Berlin Berlin Berlin Berlin
1157 1183 1188 1199
line/ section
example/ footnote
published in/source of the citation
10–11 2–3 A3 8 8–9 3 3–4 5 8–9 2 2 2 left 11
(300) 3/n.81 6/n.37 6/n.16 (402) 6/n.28 (159) 6/n.57 (65) 3/n.79 3/n.79 6/n.48 6/n.31
Various authors 1913–24 Aegyptische Inschriften aus den königlichen Museen zu Berlin. 2 vols. Leipzig: August Pries.
Berlin 22820
6–8
(257)
Anthes, R. 1930 “Eine Polizeistreife des Mittleren Reiches in die westliche Oase”, ZÄS 65, 108–14.
pBerlin 8869
3–4 4 6 7 9–11 12–13 13
5/n.48 6/n.21 6/n.47 6/n.10 (289) (242) 6/n.21
Smither, P. 1942 “An Old Kingdom letter concerning the crimes of count Sabni”, JEA 28, 16–19
pBerlin 9010
3 3–4 5
(260) (225) 6/n.76
Sethe, K. 1926 “Ein Prozessurteil aus dem alten Reich”, ZÄS 61, 67–79.
2/n.74 2/n.74 4/n.38; 6/n.82 4/n.38 (206) 2/n.74 2/n.74 4/n.38; 6/n.82 4/n.38 6/n.82 6/n.65; 7/n.7 (211) 2/n.74 (383) 2/n.74 2/n.67 2/n.74 6/n.85 7/n.8
Scharff, A. 1924 “Briefe aus Illahun”, ZÄS 59, 20–51.
Berlin 1203 Berlin 1204 Berlin Berlin Berlin Berlin
7311B 7732B 9571C 19286
pBerlin 10003 1 pBerlin 10016 1 3 4 4–6 pBerlin 10022 3 pBerlin 10023A 1 pBerlin 10023B 3 palimpsest pBerlin 10025 2 9 pBerlin 10026 pBerlin 10033 1 5 pBerlin 10036 1 15 pBerlin 10038B 1 pBerlin 10066 1 4
and/or: Luft, U. 1992 Das Arkhiv von Illahun. Briefe 1. Berlin: Akademie Verlag; Hieratische Papyri aus den Staatlichen Museen zu Berlin, Preussischer Kulturbesitz.
and/or: Kaplony-Heckel, U. 1971 Ägyptische Handschriften, Teil 1. Wiesbaden: Franz Steiner Verlag.
index of cited passages citation
Berlin Bowl
line/ section
published in/source of the citation
6/n.48
See Gardiner, A. & Sethe, K. 1928 in the Bibliography.
Berlin leather 1, 15 roll 2, 4 2. 13
(318) (199) 7/n.7
DeBuck, A. 1938 The Building Inscription of the Berlin Leather Roll. Rome: Pontificium Istitutum Publicum; Analecta Orientalia 17, Studia Aegyptiaca I, 48–57.
Berlin medical papyrus rto. 2, 2
(81)
Bersheh II
BH I
BM 101 BM 152 BM BM BM BM BM BM BM BM BM BM
159 225 334 239 462 471 504 567 574 579
BM 584 BM 614
2
example/ footnote
381
pl. 7, top, 7–8 pl. 21, top 6 pl. 8, A, 6 pl. 25, 74– 75 pl. 44, 7 horizontal 4 3–4 4 12–13 2 bottom 1–2 bottom 6–7 right 6–7 2 bottom 2 8 18–20 4 4–5 4 10 12
BM 805
4–6
BM 1164
3
BM 1213
vertical, 6
BM 1367
12–13
(256)
Grapow, H. 1958 Die Medizinischen Texte in hieroglyphisher Um-schreibung autographiert. Berlin: Akademie Verlag; Grundriss der Medizin Alter Ägypter V. Newberry, P. 1894 El Bersheh, vol. 2. London: Egypt Exploration Fund.
6/n.28 4/n.70
Newberry, P. 1893 Beni Hasan Part I. London: Egypt Exploration Fund.
6/n.28 3/n.52 (139) (173) 6/n.32 10/n.36 3/n.79 10/n.36 3/n.79 (174) 3/n.79 3/n.79 4/n.17 (293) 6/n.32 3/n.79 6/n.32 6/n.16 6/n.2; 6/n.28 3/n.86; 6/n.37 6/n.2; 6/n.28 3/n.86; 6/n.32 6/n.28
Various authors 1911–13 Hieroglyphic Texts from Egyptian Stelae etc., in the British Museum, vols. 1–4. London: Harrison & Sons.
382
index of cited passages example/ published in/source of the citation footnote1∗
citation
line/ section
BM 1671
6–7
6/n.114
Polotsky, H. 1930 “The stela of Heka-Yeb”, JEA 16, 194–99.
BM 1678
5
3/n.81
Fischer, H. 1964 Inscriptions from the Coptite Nome. Rome: Pontificium Istitutum Publicum; Analecta Orientalia 40.
BM 65340
7 7–8
6/n.61 (399)
Stewart, H. 1967 “Stelophorous statuettes in the British Museum”, JEA 53, 34–38.
pBM 10549
vso. 1–2
(361)
See James, T. 1962 in the Bibliography
Bologna 2
6
3/n.79
Bresciani, E. 1985 Cataloghi delle Collezioni del Museo Civico Archeologico di Bologna. Bologna: Grafis Industrie.
pBrooklyn 35.1446 Buhen I
insertion c, 4
2/n.74
pl. 62, 15
6/n.47
Buto stela of Thutmosis III 6–7
(57)
Cairo Bowl
7–8
(74)
Cairo JE 52000
col. x+3, line x+6
7/n.7
Cairo Linen
2 8–9
Carnavon tablet CGC 1419 CGC 1536 CGC 1641
9–10
6/n.57 (166); (397) (97)
14–15
(271)
5
9/n.9 9/n.9 6/n.16
Hayes, W. 1955 A Papyrus of the Late Middle Kingdom in the Brooklyn Museum. New York: Brooklyn Museum. Caminos, R. 1974 The New-Kingdom Temples of Buhen. 2 vols. London: Egypt Exploration Society. Bedier, S. 1994 “Ein Stiftungsdekret Thutmosis’ III aus Buto”, in M. Minas & J. Zeidler, (eds.) Aspekte Spätägyptische Kultur. Fest-schrift für Erich Winter zum 65. Geburtstag. Mainz am Rhein: Philipp von Zabern; Aegyptiaca Treverensia 7, 35–50. See Gardiner, A. & Sethe, K. 1928 in the Bibliography. See Griffiths, J. 1960 in the Bibliography See Gardiner, A. & Sethe, K. 1928 in the Bibliography.
See Gardiner, A. 1916 in the Bibliography Borchardt, L. 1937 Denkmäler des Alten Reiches (ausser den Statuen) im Museum von Kairo nr. 1295–1808. Berlin: Reichsdruckerei.
index of cited passages citation
line/ section
CGC 20001 CGC 20011
6 a4
CGC CGC CGC CGC CGC CGC CGC
20040 20043 20046 20057 20093 20100 20119
b1 x+10 2 1 d2 4 4 c3–4
CGC CGC CGC CGC CGC CGC CGC CGC CGC CGC
20164 20335 20341 20396 20458 20515 20516 20523 20530 20536
example/ footnote
published in/source of the citation
CGC 20712 CGC 20748
10–11 g5
6/n.114 (329); 7/n.24 7/n.24 6/n.32 3/n.79 3/n.81 6/n.48 6/n.37 3/n.79 3/n.86; 6/n.37 3/n.79 3/n.79 6/n.37 3/n.79 6/n.37 6/n.37 6/n.37 3/n.80 (178) 6/n.32 3/n.86 6/n.32 3/n.86 6/n.57 6/n.37 3/n.79 7/n.8 (189) (175) 6/n.37 6/n.37 6/n.32 3/n.86; 6/n.37 3/n.72 6/n.37
CT I
27c–28b 71b–d 78a 104/05d–e 131a–b 131b 140g 141e 142f 155g 158a–159c 159d–h
(140) (234) 6/n.16 (339) (3) 3/n.14 2/n.67 6/n.77 9/n.6 7/n.37 (44) (45)
2 2–3 9 c3 b2 6 6 1–2 1–3 4&5 4–6 CGC 20538 I d3 d3–5 CGC 20539 2. side, 9 CGC 20540 2 2–3 CGC 20541 10 CGC 20543 18–19 CGC 20567 a 1–2 CGC 20606 3 CGC 20609 2 CGC 20683 2&3 2–4
383
Lange, H. & Schäfer, H. 1908–25 Grab- und Denksteine des Mittleren Reiches im Museum von Kairo 20001–20780. Berlin: Reichsdruckerei.
DeBuck, A. 1935–61 The Egyptian Coffin Texts. 7 vols. Chicago: University of Chicago Press.
384 citation
index of cited passages line/ section
example/ footnote
published in/source of the citation
CT I
160g–161b 162d–163d 167a–b 168d–169a 169f 169d–170f 170g–i 174f 174j–175a 265g 267a–b 269i 278c–f 297a 302d–f 334/335a–c 390c–91a 391a
(46) (47) (61) (48) 2/n.88 (49) (226) 6/n.20 (50) 6/n.30 (334) 8/n.8 (344) (147) (351) (99); (331) (101) 6/n.16
CT II
2c 24c 29b 39e 40b 40h 47f 51b 102b 106b–c 125f 156b 214b 215b–216a 215b–217g 219a 221c–222a 222b–c
6/n.59 7/n.37 6/n.30 6/n.57 (156) 7/n.7 6/n.30 2/n.41 6/n.16 (207) (227) (259) 7/n.37 (350) (100) 7/n.37 (182) (129); (130) 6/n.30
254v 274/ 75c–276/77a 280/81d 334/35b–c 342b 344b 344b–345b 350a 359c–360a 375c–377c
(375) (378) (374) (385) 6/n.65 (406) (308) (67) (411)
index of cited passages citation
CT III
CT IV
line/ section
published in/source of the citation
47k 48e–49a 158a 171j–l 181b–182a 202i–j 206c–d 207d 263e 268/69a– 270/71b 316h–i 328a 332e–g 355a–b 382e–383e
6/n.82 (297) 6/n.54 (232) 7/n.37 (298) (183) 3/n.14 6/n.30
59l 75f 84h–i 84j 178h 182b–c 187a 191c 236/37b
(412) 6/n.16 (341) 2/n.67 6/n.16 7/n.37 6/n.48 2/n.153 6/n.48; 6/n.111 6/n.57; 6/n.111 6/n.111 6/n.47 8/n.8 2/n.41 6/n.50
242/43a 243b 326j 384b 385e 398/99b CT V
example/ footnote
11d 19g 21c–d 30e 39a 49b–c 49c 66e–h 80b–c 103e 104j–05b 107f 109h–j
385
(26) (323) 6/n.16 (315) (296) (409)
(273) 3/n.96 (184) 2/n.35 8/n.40 (353) 2/n.16 (33) 2/n.31; 2/n.38 (29); (32) (42) (421) (379)
386 citation
index of cited passages line/ section
example/ footnote
published in/source of the citation
CT V
110g 112f 115a–d 119/99c 123b 174c–d 241a 241c 244b–245a 247b–c 322i–j 324j 333p 349b 397m–o
8/n.13 8/n.13 2/n.45 2/n.176 6/n.30 (272) 3/n.14 6/n.16 (36) 2/n.41 (274) 2/n.41 6/n.110 6/n.30 (22)
CT VI
54l 73a
6/n.50 6/n.16; 6/n.30 (149) 2/n.39; 2/n.41 (34) 6/n.20 8/n.2 7/n.7 4/n.23 (40) (132) (77) 6/n.52 (51) (349) 6/n.30 (53) (30) 2/n.32 2/n.41 (228) 5/n.21 6/n.16 (23) 6/n.30 6/n.16 2/n.16 6/n.30 6/n.59 (340) (347)
92p 93a 93d–e 143i 158k 173r 194c 198n–p 210h–i 260b–d 269k 277q–278d 283e–g 290f 312d–f 316p–r 317b–c 318g 318j–k 318l 328b 328f–g 334o 336l 340j 344d 347f 348d–e 348f
index of cited passages citation
CT VI
CT VII
line/ section
example/ footnote
348l/348r– s/349m 348m 348p 353l–m 368l–m 371o 382e–f 393h 403i–l 408m–q 2d 3p 23o–24a 31k 32c 34a–e 54l 61m–p 62r–s 111f 157c 171g 177e 202i 210f–g 232m 228k–l 308c 353a 365g–366a 382c 429d 438c–d 463f–464b 475i–j 487c 492h
Deir elGebrawi II
pl. 8, top register, left
387
published in/source of the citation
(352) (202) 3/n.14 (128) (249) 3/n.14 2/n.67 3/n.14 (212) (28) 3/n.53 6/n.30 (213) 6/n.112 2/n.88 (41) 6/n.16 3/n.41 (148); (400) 2/n.41 (382) 6/n.20 6/n.47 2/n.177 (203) (312) (94) 6/n.72 7/n.7 (422) 6/n.57 6/n.54 (286) (146) 7/n.49 6/n.57 (342)
(407)
Davies, N. 1902 The Rock Tombs of Deir El Gebrâwi. Part II. Tomb of Zau and Tombs of the Northern Group. London: Egypt Exploration Fund.
388 citation
index of cited passages line/ section
example/ footnote
published in/source of the citation
2/n.173
1erný, J. 1937 Catalogue des Ostraca Hiératiques non Littéraires de Deir el Médineh, vol. 2 (nos 114 à 189). Cairo: IFAO.
oDeM 1406II x+3
2/n.173
Posener, G. 1977 Catalogue des Ostraca Hiératiques Littéraires de Deir el Médineh, vol. 3 (nos 1267–1409). Cairo: IFAO.
Door jamb in Luxor Museum 2
6/n.16
EAG
§ 1022
(52)
pEbers
2, 3–4 8, 15–16 8, 16 36, 14–15 37, 10–11 39, 13 40, 14 40, 18–20 41, 2 41, 16 43, 17 49, 8 51, 19 53, 10–11 56, 3 56, 20–21 66, 3 66, 4 68, 3 70, 16–17 70, 24 73, 20–21 78, 4 88, 14 88, 19 91, 7 91, 15–16 91, 16–17 92, 9 93, 12 96, 21 97, 3 99, 4 99, 5 99, 8
Grapow, H. 1958 Die Medizinischen (121) Texte in hieroglyphisher Um-schreibung (393) autographiert. Berlin: Akademie Verlag; 6/n.16 Grundriss der Medizin Alter Ägypter 8/n.8 V. (389) 7/n.26 6/n.16 (330) 6/n.16 6/n.65 3/n.98 (239) 3/n.98 6/n.114 (310) (305) 6/n.119 6/n.119 6/n.110 (311) 6/n.48 3/n.93 6/n.110 6/n.110 (8); 6/n.16 6/n.110 (248) 7/n.8 6/n.119 3/n.98 7/n.8 6/n.111 8/n.43 8/n.43 6/n.3
oDeM 133
vso. 5
Author’s handcopy See Edel, E. 1955–64 in the Bibliography.
index of cited passages citation
line/ section
example/ footnote
389
published in/source of the citation
pEbers
99, 16 100, 14 100, 16 100, 17–18 100, 21 101, 11–12 101, 13–14 101, 16 102, 1 102, 10 102, 12–13 102, 15–16 103, 6 103, 9 109, 14–15
8/n.2 8/n.3 8/n.2 (369) 6/n.76 (384) 8/n.6 8/n.6 8/n.6 8/n.6 8/n.6 (362) 8/n.8 8/n.8 (193)
Edel 1981
fig. 27, 4
3/n.54
Edel, E. 1981 Hieroglyphische Inschriften des Alten Reiches. Opladen: Westdeutscher Verlag; Ab-handlungen der RheinischWestfälischen Akademie der Wissenschaften 67.
8/n.43 2/n.177 (210) 2/n.177 6/n.36 (80) 2/n.81 2/n.177 (279) 2/n.81 2/n.177 6/n.36 2/n.81 2/n.81 8/n.6 (368) 8/n.6 (14) 2/n.81; 6/n.16 6/n.36 (118) 6/n.76 (363) (192) 2/n.81
Grapow, H. 1958 Die Medizinischen Texte in hieroglyphisher Um-schreibung autographiert. Berlin: Akademie Verlag; Grundriss der Medizin Alter Ägypter V.
pEdwin Smith 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 4, 4,
7 25 26 4–5 6 7–8 10 14 21–22 23 4–5 7 8 15 19 19–20 21–4, 1 2–3 9
4, 4, 5, 6, 7, 8,
10 21–5, 1 22 12–13 24 9
390
index of cited passages
citation
line/ section
example/ footnote
pEdwin Smith 8, 12 8, 15–16 8, 17 9, 12 10, 21–22 12, 18–19 13, 10 14, 8–9 14, 10–11 14, 14 15, 10 15, 17 16, 6–7 16, 13 16, 14–16 16, 19–20 21, 17–18 21, 19–22, 1 22, 3 22, 4 Felsinschriften
no. 400, 1–2 no. 429, no. 484, no. 504 no. 509 no. 530, no. 531
1 7 7 1–4
published in/source of the citation
6/n.16 8/n.6 8/n.6 8/n.2 (10) 2/n.81 8/n.6 8/n.6 8/n.6 8/n.3 7/n.26 8/n.6 6/n.16 8/n.6 (364) (208) (90); (261) (91) 2/n.91 6/n.47 6/n.37 (185) 3/n.79 6/n.37 6/n.53 (278) (417) 9/n.21
Hintze, F. & Reineke, W. 1989 Felsinschriften aus dem sudanesischen Nubia. 2 vols. Berlin: Akademie Verlag.
Davies, N. 1913 Five Theban Tombs. London: Egypt Exploration Fund.
Five Theban Tombs
pl. 19, 4
6/n.30
Fischer 1994
fig. 1, 3
6/n.16
Fischer, H. 1994 “A new Sixth Dynasty inscription from Naqada”, in C. Berger, G. Clerc & N. Grimal (eds.), Hommages à Jean Leclant, vol 1: Études Pharaoniques. Cairo: IFAO, 181–88.
Florence Florence Florence Florence Florence Florence Florence
vso. 7 1–7 5 3–6 4–5 6–7 x+2–3
3/n.79 (176) 3/n.79 (267) 4/n.34 6/n.55 3/n.79
Bosticco, S. 1959 Museo Archeologico di Firenze. Le Stele Egiziane dall’Antico al Nuovo Regno. Rome: Istituto Poligrafico dello Stato.
2500 2506 2561 2571 2590 6365 7599
index of cited passages citation
line/ section
example/ footnote
391
published in/source of the citation
Geneve 19583 3–4
3/n.79
Chappaz, J. 1986 Écriture Égyptienne. Geneva: Atar.
Geneve D50
4
6/n.37
Simpson, W. 1974 The Terrace of the Great God at Abydos: the Offering Chapels of Dynasties 12 and 13. New Haven/ Philadelphia: The Peabody Museum of Natural History of Yale University/The University Museum of the University of Philadelphia; Publications of the Pennsylvania-Yale Expedition to Egypt 5.
Goyon, Hammamat
no. 61, 12
6/n.28
Goyon, G. 1957 Nouvelles Inscriptions Rupestres du Wadi Hammamat. Paris: Imprimerie Nationale.
Hammamat
113, 113, 114, 114,
(131) 6/n.28 6/n.30 6/n.28
Couyat, M. & Montet, P. 1913 Les Inscriptions Hiéro-glyphiques et Hièratiques du Ouâdi Hammamat. 2 vols. Cairo: IFAO.
Hardjedef
I3
3/n.43
Helck, W. 1984 Die Lehre des Djedefhor und die Lehre eines Vaters an seinen Sohn. Wiesbaden: Harrassowitz; Kleine Ägyptische Texte.
Haskell Museum 13945
10 15 8 16
Gardiner, A. 1930 “A new letter to the dead”, JEA 16, 19–22. 1–2 3–4 6–7
Hassan 1975, vol. 2 pl. 51A pl. 4A, 3–4 Hassan, Giza II
Hassan, Giza IV
(55) (200) (348) (144) 6/n.16
fig. 208A fig. 219
6/n.16 6/n.51; 6/n.59
fig. 118, 5
(233)
Hassan, S. 1975 Excavations at Saqqara 1937–1938 vol. 2: Mastabas of Ny-aankh-Pepy and Others. Cairo: Government Printing Offices. Hassan, S. 1929–60 Excavations at Giza. 9 vols. Cairo: Government Press.
392 citation
index of cited passages line/ section
example/ footnote
published in/source of the citation
Hatnub
8, 2–4 8, 3 11, 4 16, 9–10 22, 1 22, 6 22, 12–14 22, 15 22, 19 28, 5
(394) 6/n.60 6/n.28 (327) 6/n.65 6/n.65 (319) 6/n.47 6/n.16 6/n.28
Anthes, R. 1928 Die Felsinschriften von Hatnub. Leipzig: Hinrichs’sche Buchhandlung.
pHearst
9, 8 9, 12 11, 2 11, 13 11, 18
3/n.93 6/n.16 6/n.110 7/n.6 6/n.110
Grapow, H. 1958 Die Medizinischen Texte in hieroglyphisher Um-schreibung autographiert. Berlin: Akademie Verlag; Grundriss der Medizin Alter Ägypter V.
Helck 1975
no. no. no. no. no.
7/n.6 3/n.79 (283) 7/n.55 4/n.71
See Helck, W. 1975 in the Bibliography.
Heqaib
9, 5 20, 2
6/n.111 3/n.80; 6/n.37 6/n.37 3/n.80; 6/n.37 3/n.80; 6/n.32 (266) 6/n.32 6/n.32 3/n.80
Habachi, L. 1985 Elephantine IV: The Sanctuary of Hekaib. Mainz am Rhein: Augustin.
26, 2 49, 10 57, 3–4 113, 3–4 146, 5
48, f2–3 49, d3 52, e3 61, 67, 88, 88,
f3–5 a6 2 2–3
Heqanakhte
I, rto. 3 I, vso. 13 II, rto. 31–32 II, rto. 37 III, rto. 4
(154) 3/n.46 4/n.16 3/n.53 2/n.74
See Allen, J. 2002 in the Bibliography.
Herdsman
22–23
(391)
Goedicke, H. 1970 “The story of the Herdsman”, CdE 45, 244–66.
index of cited passages citation
line/ section
example/ footnote
393
published in/source of the citation
Hermann 1940 31*, 14 47*, 10
6/n.57 6/n.57
Hermann, A. 1940 Die Stelen der thebanischen Felsgräber der 18. Dynastie. Glückstadt: Augustin.
HPB
III, pl. 6b
2/n.67
Various authors 1901–11 Hieratische Papyrus aus den königlichen Museen zu Berlin. 5 vols. Leipzig: August Pries.
Kagemni
II, 9
8/n.4
Gardiner, A. 1946 “The instruction addressed to Kagemni and his brethren”, JEA 32, 71–74
Kamose Stela x+6 x+7 x+15 x+34
6/n.44 6/n.44 6/n.55 6/n.61
Habachi, L. 1972 The Second Stela of Kamose and his Struggle against the Hyksos Ruler and his Capital. Glückstadt: Augustin; Abhandlungen des Deutschen Archäologischen Instituts Kairo 8.
Karnak stela of Mentuhetep 10
6/n.43
Vernus, P. 1989 “La stele du pharaon MnTw-Htpi à Karnak: un noveau témoignage sur la situation politique et militaire au début de la D.P.I.”, RdE 40, 145–61.
6/n.57; 9/n.7 6/n.57 (396)
Gardiner, A. & Sethe, K. 1928 Egyptian Letters to the Dead. London: Egypt Exploration Fund.
Kemit pl. 9
(190)
Posener, G. 1951 Catalogue des Ostraca Hiératiques Littérares de Deir el Médineh vol. 2. fasc. 1. Cairo: IFAO.
Khakheperraseneb rto. 6 rto. 10
(332) (371)
Parkinson, R. 1997 “The text of Khakheperraseneb: new readings of EA 5645, and an unpublished ostracon”, JEA 83, 55–68.
Kom el-Koffar A 4 9
6/n.16 (280)
Mostafa, M. 1984–85 “Erster Vorbericht über einen ersten Zwischen-Zeit Text aus Kom elKoffar, Teil 1”, ASAE 70, 419–29.
Koptos
6/n.37
Kaw Bowl
inside, 2 outside, 2 outside, 2–3
pl. 12, 2, x+6
Koptos decree I x+10
6/n.16
Petrie, W. 1896 Koptos. London: Bernard Quaritch. See Goedicke, H. 1967 in the bibliography.
394 citation
index of cited passages line/ section
example/ footnote
Krakow MNK-XI-999 10–11 12
published in/source of the citation
(255) (222)
1erný, J. 1961 “The stele of Merer in Cracow”, JEA 47, 5–9.
322, 7–9 322, 9 260, 6
(359) 2/n.173 6/n.2
145, 16–146, 1 255, 5–7 19, 8 413, 4–5
2/n.173 (429) 4/n.17 4/n.65
KRI VI
68, 1–2 671, 8 776, 5–7 816, 5–6 828, 15–16 831, 11–12 843, 10–11
(428) (360) (430) 2/n.173 2/n.173 (223) 2/n.173
Leiden AP 7 Leiden AP 72&73 Leiden F95/83 Leiden L.XI.8
5–6
(303)
middle, 1–3 2 7
(177) 3/n.79 6/n.32
LEM
47, 12 73, 11–12
2/n.173 4/n.65
Gardiner, A. 1937 Late Egyptian Miscellanies. Brussels: Fondation Égyptologique Reine Élisabeth; Bibliotheca Aegyptiaca 7.
LES
70, 12 85, 4
3/n.44 4/n.66
Gardiner, A. 1932 Late Egyptian Stories. Brussels: Fondation Égyptologique Reine Élisabeth; Bibliotheca Aegyptiaca 1.
Lesestücke
72, 80, 97, 98,
20–21 4–5 6–7 6
(95) 6/n.37 (416) 2/n.74
Sethe, K. 1928 Ägyptische Lesestücke zum Gebrauch im akademischen Unterricht. Texte aus dem Mittleren Reich. Leipzig: Hinrichs’sche Buchhandlung.
pl. pl. pl. pl. pl.
2, 2.7 8, 2.2 17, 5 18, 1 19, B2, 11
7/n.21 7/n.21 6/n.111 6/n.48 6/n.30
KRI I KRI II KRI III KRI IV
Literary Fragments
Kitchen, K. 1975–90 Ramesside Inscriptions. Historical and Biographical. 8 vols. + Translations, Commentaries and Notes. Oxford: Blackwell.
Boeser, P. 1909–10 Beschreibung der aegyptischen Sammlung der niederländischen Reichs-museums der Altertümer in Leiden. Die Denkmäler der Zeit zwischen dem Alten und Mittleren Reich und des Mittleren Reiches. 2 vols. The Hague: Martinus Nijhoff.
Caminos, R. 1956 Literary Fragments in the Hieratic Script. Oxford: Griffith Institute.
index of cited passages citation
line/ section
example/ footnote
395
published in/source of the citation
Louvre C10
x + 9–11
(27)
See Helck, W. 1975 in the Bibliography.
Louvre C11
17
4/n.16
Simpson, W. 1974 The Terrace of the Great God at Abydos: the Offering Chapels of Dynasties 12 and 13. New Haven/ Philadelphia: The Peabody Museum of Natural History of Yale University/ The University Museum of the University of Philadelphia; Publications of the Pennsylvania-Yale Expedition to Egypt 5.
Louvre C14
9
Badawi, A. 1960 “The stela of Irtysen”, CdE 35, 269–76.
9–10 13–14
6/n.16; 6/n.57 (87) (137)
Louvre C34
bottom, 1
6/n.16
Simpson, W. 1974 The Terrace of the Great God at Abydos: the Offering Chapels of Dynasties 12 and 13. New Haven/ Philadelphia: The Peabody Museum of Natural History of Yale University/The University Museum of the University of Philadelphia; Publications of the Pennsylvania-Yale Expedition to Egypt 5.
Louvre 196 Louvre C181 Louvre C202
2–3 1
6/n.37 6/n.37 6/n.59
See Pierret, P. 1874–78 in the Bibliography.
Louvre Bowl
17–18
(316)
Piankoff, A. & Clère, J. 1934 “A letter to the dead on a bowl in the Louvre”, JEA 20, 157–69.
Loyaliste
§ 3, 7/St
(434)
Posener, G. 1976 L’Enseignement Loyaliste. Sagesse égyptienne du Moyen Empire. Geneva: Librairie Droz; Hautes Études Orientales 5.
LRL
59, 1–3
(427)
1erný, J. 1939 Late Ramesside Letters. Brussels: Fondation Égyptologique Reine Élisabeth; Bibliotheca Aegyptiaca 9.
pLythgoe
rto. 7–8
6/n.114
Simpson, W. 1960 “Papyrus Lythgoe: a fragment of a literary text of the Middle Kingdom from ElLisht”, JEA 46, 65–70.
396 citation
index of cited passages line/ section
example/ footnote
published in/source of the citation
(215) 6/n.24 (419) 9/n.18 6/n.50 6/n.50 (365)
Goedicke, H. 1970 The Report about the Dispute of a Man with his Ba. Baltimore: Johns Hopkins Press.
Man and Son § 4, 3
(153)
Fischer-Elfert, H. 1999 Die Lehre eines Mannes für seinen Sohn. Eine Etappe auf dem ‘Gottesweg’ des loyalen und solidarischen Beamten des Mittleren Reiches. Wiesbaden: Harrassowitz; Ägyptologische Abhandlungen 60.
Meir I
pl. 5
Meir III
pl. 11
Meir V
pl. 31
(73); (229); Blackman, A. 1914–35 The Rock 6/n.76 Tombs of Meir. 5 vols. London: Egypt Exploration Fund. (246); (408) 9/n.9
Man and Ba
9–10 16 76–80 117–18 137 141 154–55
Meketra letter 4–5 5–6
(276) 2/n.74
See James, T. 1962 in the Bibliography.
Merikara
6/n.60 (433) 5/n.33 (113) 7/n.21 (240) (401) 7/n.21 5/n.39 (410) 8/n.4
Quack, J. 1992 Studien zur Lehre des Merikare. Wiesbaden: Harrassowitz; Göttinger Orientforschungen 23.
MFA 04.2059 vso. 3 vso. 6
7/n.7 2/n.67
Simpson, W. 1999 “The Naga-edDeir papyri”, in A. Larson & E. Teeter (eds.), Gold of Praise. Studies on Ancient Egypt in Honor of Edward E. Wente. Chicago: Chicago University Oriental Institute, 387–96.
Michaelidis Lahun letter
2/n.74
C IV, 5 C V, 7 E 17 E 25–26 E 33 E 53–54 E 66–67 E 68 E 87 E 134 E 144
1
Grdseloff, B. 1949 “A new Middle Kingdom letter from el-L§hån”, JEA 35, 59–62.
index of cited passages citation
MMA 57.95
line/ section
published in/source of the citation
6/n.48 (134)
Fischer, H. 1960 “The inscription of In-it.f, born of Tfi”, JNES 19, 258–68.
MMA 13.182.3 vertical 3–4 vertical 4–5
(25) (265)
Winlock, H. 1943 “The eleventh Egyptian dynasty”, JNES 2, 249–83.
MMA 35.7.55 10 11 14–15
6/n.111 0/n.100 6/n.28
Hayes, W. 1947 “Horemkhaauef of Nekhen and his trip to It-Towe”, JEA 33, 3–11.
MMA 65.120.2 3
3/n.80; 6/n.32
A photograph courtesy of Prof. James P. Allen
Iβ4 IIα2 IIβ1–2 IV24
(268); (304) 4/n.16 6/n.82 6/n.82 6/n.114
Vandier, J. 1940 Moaalla. La Tombe d’Ankhtifi et la Tombe de Sébekhotep. Cairo: IFAO.
no. 9, 29
(109)
Moaalla
pMoscow math.
5 8–9
example/ footnote
397
Iα2
München GL WAF 35 21–22
Naga ed-Der
pl. XV.2, vertical, 1–2
6/n.28
3/n.79
Naga ed-Deir Letter to Dead vso.
9/n.56
Neferhetep Stela
6/n.28
40
Imhausen, A. 2003 Ägyptische Algorithmen. Eine Untersuchung zu den mittelägyptischen mathe-mathischen Aufgabentexten. Wiesbaden: Harrassowitz; Ägyptologische Abhandlungen 65. Simpson, W. 1974 The Terrace of the Great God at Abydos: the Offering Chapels of Dynasties 12 and 13. New Haven/ Philadelphia: The Peabody Museum of Natural History of Yale University/The University Museum of the University of Philadelphia; Publications of the Pennsylvania-Yale Expedition to Egypt 5. Dunham, D. 1937 Naga-ed-Dêr Stelae of the First Intermediate Period. Oxford: Oxford University Press. Simpson, W. 1966 “The Letter to the Dead from the tomb of Meru (N 3737) at Naga ed-Deir”, JEA 52, 39–52. See Helck, W. 1975 in the Bibliography.
398 citation
index of cited passages line/ section
example/ footnote
published in/source of the citation
Neferty
I a–b XI d–e XIII a
(217) (78) (372)
Helck, W. 1970 Die Prophezieung des Nfr.tj. Wiesbaden: Harrassowitz; Kleine Ägyptische Texte.
Nt
40–41
2/n.4; 2/n.125; 2/n.159 5/n.3 3/n.71 3/n.71 3/n.90 3/n.90
Jéquier, G. 1933 Les Pyramides des Reines Neit et Apouit. Cairo: IFAO.
67 293 323 490 781 Ny Carlsberg Glyptotek ÆIN 964 centre, 4 Oxford Queen’s College no. 1113 Peas
3/n.79; 6/n.37
Koefoerd-Petersen, O. 1948 Les Stèles Ègyptiennes. Copenhagen: Ny Carlsberg Glyptotek; Publications de la Glyptothèque Ny Carlsberg 1. Author’s handcopy
bottom 4 bottom 4–5
6/n.37 3/n.79
B1 B1 B1 B1 B1 B1 B1 B1 B1 B1 B1 B1 B1 B1 B1 B2 B2
(237) 0/n.56 (264) (243) 3/n.46 (76) (244) (436) 6/n.30 5/n.41 (163) (151) 6/n.50 4/n.19 (301) (124) (122); (405) (381); (388) 6/n.16
62–63 42 109–10 110–11 112 114–15 134–35 146–47 157–58 211 255 268–69 273 323 351–52 117–18 119–22
R 1, 1 R 17.7
Parkinson, R. 1991 The Tale of the Eloquent Peasant. Oxford: Griffith Institute.
index of cited passages citation
line/ section
Pepi II Dakhla decree 6
example/ footnote
399
published in/source of the citation
4/n.35
Pantalacci, L. 1985 “Un décret de Pépi II en faveur des governeurs de l’oasis de Dakhla”, BIFAO 85, 245–54.
PT
16d 149a–b 204b 377a 467a 481a–b 499c 577b–d 583c 795a–b 809a–b 828a 835a 838c 944a–b 967c 990c 998 1041c 1093b 1102a 1128a–29b 1223a 1242a 1275b 1276b 1295a 1423a 1439d 1440b 1477a–b 1480b 1480c 1534a 1596b 1775b 2075a
5/n.3 3/n.16 4/n.13 4/n.78 3/n.90 2/n.47 5/n.39 5/n.3 0/n.100 (424) 6/n.78 3/n.71 3/n.71 3/n.71 2/n.47 3/n.90 0/n.100 (236) 4/n.13 0/n.100 2/n.4 3/n.16 4/n.41 3/n.71 3/n.16 3/n.16 3/n.90 4/n.13 3/n.71 3/n.71 2/n.47 4/n.78 3/n.90 3/n.71 3/n.90 2/n.88 4/n.78
Sethe, K. 1908–10 Die Altägyptischen Pyramidentexte. 2 vols. Leipzig: Hinrichs’sche Buchhandlung.
PT PII
709, +57 1055, +30
3/n.16 2/n.125
Jéquier, G. 1936–40 Le Monument Funéraire de Pepi II. 3 vols. Cairo: IFAO.
400 citation
index of cited passages line/ section
example/ footnote
published in/source of the citation
Ptahh
74–76 76 111 223 246 247 349 374 438–39 465–66 593 613–14 645
(89) 6/n.29 2/n.41 3/n.43 6/n.3 6/n.3 (398) 5/n.42 (150) (423) 6/n.61 (119) 8/n.4
pPurches
palimpsest 3–4
4/n.20
Pushkin Museum I.1.a.1137 Pushkin Museum I.1.b.32/ UCL 14326
See Allen 2002 in the Bibliography.
5–6
6/n.28
x+6
3/n.86; 6/n.37
Pushkin Museum 1695 vso. 2 pPushkin 167 fragment 2, 2 fragment 2, 3 fragment 5, 2
¥àbá, Z. 1956 Les Maximes de PtaHHotep. Prague: Académie Tchécoslovaque des Sciences.
6/n.111
8/n.8 8/n.8 6/n.50
Hojdash, S. & Berlev, O. 1982 The Egyptian Reliefs and Stelae in the Pushkin Museum of Fine Arts, Moscow. Leningrad: Aurora Art Publishers.
Posener, G. 1969 “Fragment littéraire du Moscou”, MDAIK 25, 101–06. Korostovtsev, N. 1960 “Egipetskii ieratitseskii papirus no. 167 gosudartstvennogo museja isobrasitelnih iskusstv im. A.S. Pushkina v Moskve”, in Drevnii Egipet. Moscow: Isdatelstvo Akademija Nauk USSR, Institut Vostokovedenija, 119–33.
Qurneh
pl. 3, 2, 3
7/n.8
Petrie, F. 1909 Qurneh. London: British School of Archaeology in Egypt/ Bernard Quaritch.
RAD
56, 5
4/n.68
Gardiner, A. 1948 Ramesside Administrative Documents. Oxford: Griffith Institute.
index of cited passages citation
pRam Dram
line/ section 8 15 21 46 48 51
Red Chapel
published in/source of the citation
8/n.8 4/n.51 8/n.8 8/n.8 8/n.8 4/n.51; (387) (220) 4/n.51 4/n.51 4/n.51 (218) 4/n.51 4/n51 4/n.51 4/n.51 4/n.51 4/n.51 4/n.51 (221) (219) 4/n.51
See Sethe, K. 1928 in the Bibliography
Barns, J. 1956 Five Ramesseum Papyri. Oxford: Griffith Institute.
I, Bii, 15 II, vso. 5 III, A19 III, B5 III, B10–11 III, B23 IV, C16 IV, C18 IV, Div, 2
6/n.114 2/n.88; 4/n.30 3/n.58 7/n.24 6/n.16 6/n.55 (238) 8/n.8 6/n.110 8/n.2 6/n.16
81 south 92 166, 22–23 188a, 13 301 327 349 352 355 448 460, 5 490 493 601 715
6/n.51 6/n.51 (125) 2/n.71 6/n.51 6/n.51 6/n.51 6/n.51 6/n.51 6/n.51 6/n.28 6/n.51 6/n.51 6/n.51 6/n.51
Lacau, P. & Chevrier, H. 1977 Une Chapelle d’Hatshepsout à Karnak. Cairo: IFAO.
53 64 72 76 83 87 89 91 97 101 104 107 114 117 126 pRamesseum
example/ footnote
401
I, A11 I, Bii, 10
402 citation
index of cited passages line/ section
Reden und Rufe 52 58
example/ footnote
published in/source of the citation
9/n.9 (194)
Erman, A. 1919 Reden, Rufe und Lieder auf Gräber-bildern des Alten Reiches. Berlin: Verlag der Akademie der Wissenschaften; Abhandlungen der preussischen Akademie der Wissenschaften.
Re-used block T/I.R 319 in the mortuary temple of Merenptah l. 3
(426)
Revillout 1914 26, no. 74
(431)
See Peust, C. 2005 in the Bibliography.
pRhind
no. 46, 1
2/n.153
Imhausen, A. 2003 Ägyptische Algorithmen. Eine Untersuchung zu den mittelägyptischen mathemathischen Aufgabentexten. Wiesbaden: Harrassowitz; Ägyptologische Abhandlungen 65.
Rifeh VII
31
6/n.29; 6/n.82
Griffith, L. 1889 The Inscriptions of Siût and Dêr Rîfeh. London: Trübner.
Rodin Museum 275
1–5
(167)
Various Authors 1967–68 Rodin Collectionneur. Paris: Musée Rodin.
Roumiantshev Museum 18 17/III 78 6
6/n.28
SemnaKumma
6/n.37
R.I.K. 12, 1 R.I.K. 51, 3–6
Semnah desp. 2, 8 3, 7 4, 6
(168); 6/n.37 6/n.48 6/n.48 6/n.48
Bickel, S. 1997 Untersuchungen im Toten-tempel des Merenptah in Theben unter der Leitung von Horst Jaritz III: Tore und andere wieder-verwendete Bauteile Amenophis’ III. Stuttgart: Franz Steiner Verlag; Beiträge zur ägyptischen Bauforschung und Altertumskunde 16.
Clère, J. & Vandier, J. 1948 Textes de la Première Période Intermèdiaire et de la XIème Dynastie. Brussels: Fondation Égyptologique Reine Élisabeth; Bibliotheca Aegyptiaca 10. Dunham, D. & Janssen, J. 1960 Second Cataract Forts vol. 1: Semna, Kumma. Boston: Museum of Fine Arts. Smither, P. 1945 “The Semnah despatches”, JEA 31, 3–10.
index of cited passages citation
line/ section
example/ footnote
403
published in/source of the citation
Sh.S.
47–50 56–59 60–62 89–91 130 139–46 149–52 166 179–81 186
(269) (17) (107) (376) (216) 0/n.56 (43) 4/n.49 (282) 8/n.4
Blackman, A. 1933 Middle Egyptian Stories. Brussels: Fondation Égyptologique Reine Élisabeth; Bibliotheca Aegyptiaca 3.
Sin
B7
2/n.28; 3/n.8 (120) 10/n.36 (281) (290) (415) (88) (35) 3/n.8 8/n.43 2/n.79 (325) 3/n.58 9/n.18 (155) 2/n.74 (413) 4/n.44 (437) (66) 6/n.50 6/n.30 (16) (245) 8/n.4 0/n.103 6/n.50 (6) 6/n.47
Koch, R. 1990 Die Erzählung des Sinuhe. Brussels: Fondation Égyptologique Reine Élisabeth; Bibliotheca Aegyptiaca 17.
B 17–18 B 44–45 B 68–69 B 75–76 B 77 B 106–07 B 110–11 B 111–12 B 115–16 B 126–27 B 134–36 B 144–45 B 152 B 157–58 B 181 B 182–83 B 197 B 199–200 B 214–15 B 225 B 236 B 260–61 B 280–83 B 311 R 15 R 65 R 141–42 R 142
404 citation
Sinai
index of cited passages line/ section 118, vertical 3 167, 1–3 182, 2 196, 14 200, x+1 x+9 244, 6 502, 1 510, 1 519, 4–5
example/ footnote
published in/source of the citation
3/n.79 (2) 6/n.61 6/n.28 6/n.28 6/n.28 7/n.6 3/n.81 3/n.81 3/n.81
Gardiner, A. & Peet, T. 1952 The Inscriptions of Sinai (2nd ed.). Oxford: Oxford University Press.
Siut III Siut IV Siut V
308–09 310 311 316–17 11 10 29
(12); (251) Griffith, F. 1889 The Inscriptions of Siût 3/n.4 and Dêr Rîfeh. London: Trübner. (247) 2/n.68 6/n.85 (287) 6/n.85 (277) (7) 6/n.60; 6/n.111 (288); 6/n.85 (306) (54) 6/n.85 (254) (295) 3/n.46 (322)
Siut VI
16
6/n.16
Montet, P. 1936 “Les tombeaux de Siout et de Deir Rifeh”, Kêmi 6, 131–63.
3/n.97 3/n.97
Steindorff, G. 1896 Grab-funden des Mittleren Reichs in den Königlichen Museen zu Berlin I, Das Grab des Mentuhotep. Berlin: W. Spemann; Mittheilungen aus den Orientalischen Sammlung VII.
(110) (314)
See Helck, W. 1975 in the Bibliography.
Siut I
229 267 271 280–01 282 288–89 296–97 297 297–98 298 301
Steindorff 1896 4, 3 10, 3
Stèle Juridique
15 24
index of cited passages citation
line/ section
example/ footnote
Tb
1, 23–24/ Nu pl. 12, 6 (420) 1, 24/Nu pl. 12, 6 6/n.50 1, 25 6/n.61 3/Nu pl. 38, 3 6/n.50 17, 3 6/n.48 17, 16–17 6/n.57 17, 31 6/n.48 17, 31/Nu pl. 4, 35 6/n.57 17, 34 6/n.55; 8/n.6 17, 34/Nu pl. 4, 38 6/n.57 17, 45 6/n.59 17, 56–57 8/n.6 17, 57 8/n.6 17, 105 6/n.30 17, 107–08 (370) 17, 108–09/ Nu pl. 9, 118 (367) 18, 5–6/Nu pl. 9, 5 (366) 18, 15–16 8/n.5 42, 22/Nu pl. 17, 21 (336) 57, 2 6/n.50 64, 23 6/n.48 64, 33–34 6/n.16 65, 11–12/ Nebseni pl. 66, 10–11 (157) 68/Nu pl. 18, 16 (395) 69, 6 6/n.30 72, 12–13 6/n.48 72/Nu pl. 20, 14 6/n.50 78/Nu pl. 39, 16 6/n.57
405
published in/source of the citation
Naville, E. 1886 Das Aegyptische Todtenbuch der XVIII. bis XX. Dynastie. Berlin: A. Asher & Co. also: Lapp, G. 1997 The Papyrus of Nu. London: British Museum Press; Catalogue of the Books of the Dead in the British Museum vol. 1. Lapp, G. 2004 The Papyrus of Nebseni. London: British Museum Press; Catalogue of the Books of the Dead in the British Museum vol. 3.
406
index of cited passages
citation
line/ section
Tb
78, 27/Nu pl. 40, 31–32 78, 37/Nu pl. 41, 42 86, 11–12 89, 7 90/Nu pl. 22, 4 93, 6–7 98/Nu pl. 25, 6 99, 10 99, 34/Nu pl. 64, 30 110, 15 112, 4/Nu pl. 53, 4 112, 10/Nu pl. 53, 9 126/Nu pl. 70, 14 134, 4 134, 13 136B, 10 136B, 11 144/Nu pl. 75, 33 144/Nu pl. 76, 48 144/Nu pl. 76, 49 148, 21–22 148, 22 149, 19–20 149, 56–57 149, 75
example/ footnote
3/n.98 2/n.67 6/n.110 4/n.41 2/n.41 6/n.59 7/n.7 (115) (324) 6/n.30 3/n.98; 6/n.50 6/n.110 6/n.30 6/n.30 6/n.61 7/n.21 7/n.21 3/n.14 6/n.111 3/n.98 (60) 6/n.52 (133); (205) (252) 6/n. 16
published in/source of the citation
index of cited passages citation
line/ section
Tb
149, 98/Nu pl. 86, 107 153A/Nu pl. 58, 29–30 153B, 2–11/ Nu pl. 59, 3–11 154, 4 156/Nu pl. 79, 5 167, 5 167, 6
example/ footnote
407
published in/source of the citation
6/n.16
2/n.76
2/n.76 6/n.111 3/n.98 6/n.116 6/n.116
Ti
pl. 123, middle register (38) pl. 123, bottom register 2/n.43
Steindorff, G. 1913 Das Grab des Ti. vol. 2. Leipzig: Hinrichs’sche Buchhandlung.
Tomb of Amenemhat
pl. 30C pl. 30F pl. 39, middle band
De Garis Davies, N. & Gardiner, A. 1915 The Tomb of Amenemhat. London: Egypt Exploration Fund.
Tomb of Ankhi Tomb of Merefnebef
Turin 1447 Turin Turin Turin Turin Turin Turin
1449 1534 1546 1547 1628 1903
Turin 1541
6/n.30 6/n.30 6/n.57
false door niche, 6/n.47 4 pl. 16/32, 2 3/n.81 pl. 16/32, 3 6/n.50 pl. 21/66, right 2/n.43
Goyon, G. 1959 “Le tombeau d’Ankhou a Saqqarah”, Kêmi 15, 8–22. Myśliwiec, K. et al 2004 Saqqara I: The Tomb of Merefnebef. Warsaw: Editions Neriton.
middle panel, 4 3/n.79; 6/n.32 middle panel, 9 6/n.16 7 6/n.16 10 (262) 5 3/n.79 4–5 3/n.79 bottom 2 6/n.37 4–7 (172)
Donadoni Roveri, A. 1988 (ed.), Ägyptisches Museum Turin. Das alte Ägypten: die religiösen Vorstellungen. Turin: Istituto Bancario San Paolo di Torino.
4–5
Author’s photograph
6/n.37
408 citation
index of cited passages line/ section
example/ footnote
published in/source of the citation
pTurin 54002 8–9
(24)
Roccati, A. 1968 “Una lettera inedita dell’Antico Regno”, JEA 54, 14–22.
pTurin 54003 rto. 5 rto. 11–12 vso. 10–11
(377) (31) (299)
Roccati, A. 1970 Papiro Ieratico n. 54003 Estratti Magici e Rituali del Primo Medio Regno. Turin: Fratelli Pozzo.
3/n.79; 6/n.32
Brunner-Traut, E. & Brunner, H. 1981 Die Ägyptische Sammlung der Universität Tübingen. 2 vols. Mainz am Rhein: Philipp von Zabern.
Tübingen University 458 8–9
Tübingen University 479 vertical 9 UC 14333
3/n.79
6 8 vertical x+7
6/n.28 4/n.70 6/n.24
pUC 32036
1.x+12 1.x+14
6/n.16 (79); 6/n.31
pUC 32037
rto. 3–5
(291)
Also:
pUC 32055
x+14
(284)
pUC 32057
2.7–9 2.16
(320) 7/n.7
pUC 32058
rto. 13
(414)
Collier, M. & Quirke, S. 2004 The UCL Lahun Papyri: Religious, Literary, Legal, Mathematical and Medical. Oxford: Archaeopress; BAR International Series 1209.
pUC 32124
fragment ii, 4–8 (302) fragment ii, 7–8 (313) fragment ii, 10 (164) fragment ii, 3–4 (56) h/v left, page 3, 2 3/n.14 h/v right 7 6/n.50 2.5 6/n.82 2.6 7/n.8 2.2 6/n.52 2.9 6/n.76
UC 14430
pUC 32126 pUC 32157
pUC 32158 pUC 32197 pUC 32198
13–17
(152)
Stewart, H. 1979 Egyptian Stelae, Reliefs and Paintings from the Petrie Collection. Vol. 2 Archaic Period to the 2nd Intermediate Period. Warminster: Aris & Phillips. Collier, M. & Quirke, S. 2002 The UCL Lahun Papyri: Letters. Oxford: Archaeopress; BAR International Series 1083.
index of cited passages citation
line/ section
pUC 32199
1–2 5–7 12 17 rto. 5 7 6 2 11 19 2–3 17–18 1–3 vso. 1–2 4
(292) (160) 6/n.16 (39) 2/n.171 6/n.82 6/n.76 (196) 6/n.76 (158) (418) 6/n.85 (326) 6/n.49 6/n.122 6/n.82
UCLA 97 UCLA 3769
1 3–5
6/n.32 (263)
Lutz, H. 1927 Egyptian Tomb Stelae and Offering Stones of the Museum of Anthropology and Ethnology of the University of California. Leipzig: Hinrichs’sche Buchhandlung.
Urk I
19, 21, 26, 31, 39, 39, 42, 52, 59, 60, 61, 61, 61, 62, 62, 63, 63, 79, 84, 84,
6/n.47 6/n.55 (275) 6/n.22 2/n.88 (343) (104) (165) 4/n.43 2/n.75 (62) 2/n.88 (71) (20) 2/n.85 (19) 2/n.88 (169) 6/n.47 6/n.47; 6/n.55; 6/n.68 5/n.27 3/n.14 (250)
Sethe, K. 1933 Urkunden des Alten Reiches. Leipzig: Hinrichs’sche Buchhandlung.
pUC pUC pUC pUC pUC pUC
32200 32201 32202 32203 32204 32205
pUC 32206 pUC 32210 pUC 32212 pUC 32213 pUC 32215 pUC 32291
3 14 14–15 3 1 12–14 11 2–3 16 16–17 9 14 17–18 1–3 12 2–3 11 26a–28a 1 3
85, 5–6 99, 10–11 102, 9–16
example/ footnote
409
published in/source of the citation
410
index of cited passages
citation
line/ section
Urk I
104, 12 106, 118, 119, 125, 125, 126, 126, 128, 128, 128, 129,
5 17 11 10–11 15–16 4 10 8 10–11 14–16 2–3
136, 9–11 137, 12 138 13–17 139, 10 146, 162, 180, 180, 182, 182, 182, 195, 195, 202, 204, 204, 205, 205, 212, 212, 215, 216,
6–8 11 1 7 11 14 15 1 9–10 7 9–10 10 2–6 12–14 7–8 9–11 14 5
217, 217, 217, 218, 218,
6 16 15–17 2 8–10
example/ footnote 2/n.15; (186) 6/n.21 6/n.2 6/n.54 (335) (112) 6/n.16 2/n.74 2/n.74 (18) (21) (13) (224) (68) 2/n.74 (123) 6/n.2; 6/n.28 3/n.14 (253) (86) (82) 4/n.49 4/n.49 2/n.177 4/n.23 (294) 6/n.16 (170) 6/n.47 (180) 3/n.87 (230) (231) 6/n.49 6/n.2; 6/n.28 9/n.35 3/n.87 (201) 6/n.54 (179)
published in/source of the citation
index of cited passages citation
line/ section
example/ footnote
Urk I
3/n.87; 4/n.34 (188) 221, 4 7/n.37 222, 12 (345) 223, 12–16 (438) 224, 18 6/n.54 233, 17 3/n.79 252, 3 3/n.79; 268, 13 6/n.32 6/n.55 272, 12 6/n.21 278, 9 5/n.46 282, 9–14 6/n.16 282, 12 282, 15–283, 3 (187); (235) 4/n.37 282, 17 (241) 286, 1–6 6/n.16 286, 4 4/n.21 286, 7 4/n.37 286, 13 5/n.35 286, 7–17 4/n.35 292, 12 (204) 296, 16 4/n.35 298, 7 (126); 298, 8 (141) 4/n.35 298, 15 (143) 298, 16 4/n.35 299, 17 (142) 301, 3–5 4/n.35 301, 10 4/n.35 303, 11 305, 8–306, 1 (181) 6/n.8 305, 18 6/n.16 306, 12 4/n.35 306, 13
Urk IV
3, 2 3, 4 5, 4 9, 14–16 27, 12–14 28, 14 43, 15 65,3 65,8
411
published in/source of the citation
218, 16
7/n.8 6/n.31 7/n.8 (85) (386) 6/n.28 6/n.28 6/n.30 6/n.30
Sethe, K. 1906–09 Urkunden der 18. Dynastie, vols. 1–4. Leipzig: Hinrichs’sche Buchhandlung. Helck, W. 1955–57 Urkunden der 18. Dynastie, vols. 5–6. Berlin: Akademie Verlag.
412
index of cited passages
citation
line/ section
example/ footnote
Urk IV
2/n.74 80, 8-9 (5) 81, 2-3 6/n.28 100, 3 6/n.30 116, 17 4/n.42 117, 6 3/n.52 119, 3 6/n.28 141, 7 4/n.42 148, 17 6/n.66 150, 2 (403) 150, 14–15 6/n.59 157, 7 (354) 164, 5–6 6/n.28 170, 14 (70) 181, 10–11 6/n.57 182, 3 6/n.111 220, 2 7/n.6 224, 1 6/n.51 253, 9 7/n.51 260, 6 2/n.88 267, 17 6/n.29 272, 10 273, 14–274, 3 (208) 6/n.51 278, 10 6/n.51 280, 1 6/n.51 290, 13 6/n.51 304, 4 6/n.28 312, 15 6/n.50 321, 11 6/n.28 322, 12 6/n.43 340, 1 3/n.72 341, 8 2/n.71 346, 3–4 (380) 347, 2–3 (392) 350, 8–9 6/n.50 362, 16 (92) 363, 5–6 (346) 363, 6–8 6/n.48 363, 10 (214) 363, 12 (63) 364, 1–2 (11) 364, 16–17 2/n.78 365, 10–11 6/n.16 366, 15
published in/source of the citation
index of cited passages citation
line/ section
Urk IV
367, 367, 367, 372, 377, 389, 390, 397, 429, 432, 437, 439, 480, 485, 499, 520, 520, 547, 554, 561, 567, 571, 571, 579, 590, 592, 593, 593, 606, 617, 624, 656, 687, 693, 702, 736, 736, 751, 752, 758, 766, 776, 781,
7 12 17 13–14 9 3 7 2–3 5 16 17 1 3 4–5 6 1 10 10 5 17 17 8 12 1 16 1 4–5 10–11 7 9 5 3 13 12–13 1 5 16 2 14 9 11 13–14 1–4
example/ footnote 7/n.51 7/n.21 6/n.16 (270) 6/n.51 6/n.48 6/n.59 4/n.62 2/n.67 6/n.57 6/n.50 6/n.48 3/n.44 (390) 6/n.30 6/n.30 6/n.30 6/n.61 6/n.28 6/n.51 6/n.51 6/n.51 6/n.51 6/n.51 6/n.28 2/n.67 (58) (321) 6/n.16 6/n.30 7/n.6 6/n.86 6/n.50 (435) 6/n.48 6/n.50 (356) (117) 6/n.24 6/n.30 3/n.46 (161) (317)
413
published in/source of the citation
414
index of cited passages
citation
line/ section
Urk IV
795, 807, 807, 809, 812, 812, 835,
14 8 12 10 9 13 16
836, 6 840, 12 843, 14 849, 14 852, 16 853, 13 859, 4 868, 4 868, 8 879, 4 883, 14 889, 5 890, 10–12 892, 6–7 892, 12 897, 13 943, 4–5 969, 3 978, 15 993, 9 1008, 6 1014, 16 1023, 5 1064, 6 1073, 8 1074, 12 1109, 5–8 1111, 9–13 1112, 4 1158, 17 1159, 11 1216, 6 1221, 7 1236, 1
example/ footnote 7/n.6 7/n.35 7/n.35 6/n.50 7/n.35 7/n.35 2/n.71; (105) 6/n.119 6/n.57 6/n.28 6/n.28 6/n.28 6/n.16 6/n.28 6/n.28 6/n.16 9/n.5 (106) 6/n.28 (425) (1) 6/n.31 6/n.57 (258) 6/n.48 6/n.115 2/n.67 7/n.6 6/n.30 6/n.52 6/n.30 6/n.61 2/n.71 (102) (59) 6/n.47 6/n.52 6/n.48 6/n.16; 6/n.62 6/n.30 6/n.30
published in/source of the citation
index of cited passages citation
line/ section
Urk IV
1241, 5–6 1246, 12 1255, 13 1257, 3 1260, 11 1269, 8 1270, 1 1278, 4 1279, 18 1280, 12–13 1281, 8–10 1282, 13 1286, 19 1291, 1–3 1293, 1–3 1293, 7 1294, 11 1298, 9/10 1298B, 9–14 1308, 2 1312, 7–11 1326, 13 1349, 18 1349, 17–18 1371, 16 1381, 4 1409, 3 1412, 7 1425, 13 1431, 14 1455, 18 1456, 1 1456, 2 1468, 1 1495, 1 1517, 5/6 1529, 10 1531 11/15 1531, 15 1543, 9–10 1543, 15 1560, 2
example/ footnote (103) 6/n.61 3/n.15 (127) 6/n.16 6/n.61 2/n.88 (138) 6/n.72 (111) (285) 6/n.115 3/n.32 (75) (72) 2/n.106 6/n.16 6/n.30 (145) 6/n.114 (357) (135) 7/n.6 (136) 6/n.31 2/n.74; 2/n.159 6/n.57 6/n.52 2/n.106 6/n.2 6/n.57 6/n.57 6/n.57 2/n.106 6/n.61 6/n.30 7/n.6 2/n.139 (64) (358) 7/n.7 6/n.51
415
published in/source of the citation
416
index of cited passages
citation
line/ section
Urk IV
1579, 1592, 1597, 1652, 1653, 1656, 1662, 1664, 1671, 1673, 1673, 1673, 1675, 1675, 1675, 1676, 1676, 1682, 1686, 1689, 1696, 1708, 1714, 1729, 1729, 1744, 1754, 1776, 1777, 1795, 1798, 1799, 1805, 1807, 1814, 1824, 1830, 1833, 1842, 1843, 1844, 1845, 1849,
13 7 17 8 14 4 11 13 1 6 9 14 7 10 18 12 15 13 19 9 9 10 13 6 13 5 8 14–16 17 18–19 17 4 8 21 16 10–11 10 9 5 18 15 20 13
example/ footnote 2/n.106 6/n.55 6/n.52 7/n.6 6/n.50 7/n.6 6/n.30 6/n.30 6/n.2 6/n.50 2/n.88 6/n.47 6/n.72 7/n.7 6/n.43 2/n.88 6/n.61 6/n.57 6/n.28 7/n.6 6/n.57 6/n.28 6/n.111 3/n.72 6/n.61 6/n.30 6/n.55 (93) (333) (309) 6/n.16 6/n.3 6/n.49 6/n.48 6/n.61 (355) 2/n.88 2/n.88 7/n.8 3/n.97 3/n.97 7/n.7 6/n.30
published in/source of the citation
index of cited passages citation
line/ section
Urk IV
1861, 1861, 2 1889, 1909, 1919, 1919, 1919,
Urk VII
1, 13 4, 8
example/ footnote
417
published in/source of the citation
13 6/n.51 20–1862, (328) 4 6/n.30 13 6/n.51 9 6/n.52 12 6/n.50 13 6/n.50; 6/n.61 1926, 15 6/n.61 1939, 5 3/n.97 1947, 2 6/n.52 6/n.37 6/n.51
Sethe, K. 1935 Historisch-Biographische Urkunden des Mittleren Reiches. Leipzig: Hinrichs’sche Buchhandlung.
Vernus 1976 140, example 5b
4/n.34
Wadi Hilâl
N6, 8
3/n.81
Vandekerckhove, H. & MüllerWollermann, R. 2001 Elkab VI. Die Fels-inschriften des Wadi Hilâl. Turnhout: Brepols.
Wadi el-Hudi 21, 8
3/n.14
Sadek, A. 1980 The Amethyst Mining Inscriptions of Wadi el-Hudi. vol 1. Warminster: Aris & Phillips.
pWestcar
6/n.72 (307) 6/n.114 6/n.114 (84) 7/n.24 (373) (404) 6/n.114 (162) (15) 3/n.70 (4); (191) 3/n.33 6/n.111 (116) 6/n.114
Blackman, A. 1988 The Story of King Kheops and the Magicians. Transcribed from Papyrus Westcar (Berlin Papyrus 3033). Reading: J.V. Books.
3, 2 3, 2–4 3, 10 3, 17 5, 1–5 5, 19 6, 4–6 6, 27–7, 1 7, 11 8, 3–4 8, 8–9 8, 17 9, 21–22 10, 4 11, 26 12, 3–4 12, 8–9
Vernus, P. 1976 “Le formule ‘Le soufflé de la bouche’ au Moyen Empire”, RdE 28, 139–45.
418
index of cited passages
citation
line/ section
White Chapel
59 64 67 170 182
(114) Lacau, P. & Chevrier, H. 1969 Une 6/n.28 Chapelle Sésostris Ier à Karnak. Cairo: IFAO. 6/n.16 2/n.106 6/n.51
28–30 16–17 6 back, 3–4
(171) 3/n.79 3/n.79 3/n.79
Hein, I. & Satzinger, H. 1989 Kunsthistorisches Museum Wien: Stelen des Mittleren Reiches I. Mainz am Rhein: Philipp von Zabern; Corpus Antiquitatum Aegyptiacarum 4.
Worrell 1942 203, 1–3
(432)
Worrell, W. 1942 Coptic Texts in the University of Michigan Collection. Ann Arbor: The University of Michigan Press.
Zagreb 7 Zagreb 8
3/n.79 3/n.79
Monnet Saleh, J. 1970 Les Antiquités Égyptiennes de Zagreb. Paris: Mouton & Co.
Wien Wien Wien Wien
ÄS ÄS ÄS ÄS
156 166 168 186
3 5
example/ footnote
published in/source of the citation
general index
419
GENERAL INDEX 2ae gem. roots: 270–71 A (particle): 7/n.31 Acceptance: 18, 46, 49, 62–63, 65, 64, 66, 69, 71, 72, 73, 83, 96, 207, 214, 281, 326, 335 Adjectival sentence: 2/n.166, 117, 3/ n.87, 177, 182, 6/n.54, 249, 6/n.116, 278, 307, 10/n.64 Adjective, adjective verb: 3, 177, 182, 4/n.24, 184, 187, 4/n.64, 211, 230, 231 Adjunct clause: 29, 31, 0/n.100, 97, 2/n.134, 2/n.173, 111, 2/n.193, 168, 4/n.33, 196, 215–16, 221–22, 251–53 263, 271, 280, 293, 314, 339, 345, 359 irrealis in: 120, 256–59, 339 Adverbial sentence: 13, 2/n.25, 2/n.95, 102, 2/n.146, 2/n.166, 110, 111, 117, 119, 4/n.19, 248, 287, 288, 289, 292, 294, 296, 303, 305, 306, 307, 318 Agentivity: 145, 148 Aktionsart: 2/n.179, 315, 9/n.27 Anaphoric: 3/n.29, 196, 198, 301 Anomalous roots: 29, 32, 288, 336 and passim Ante-position: 102, 103, 109, 234, 243, 262, 7/n.8, 290, 309, 303, 307, 355, 359 with ir: 102, 234, 249, 262, 298, 299 with xr: 262 Appeal to the Living: 129, 135, 157–63, 232–35 Apposition: 2/n.9, 2/n.25, 100, 4/ n.46 Arguments, argument structure: 2, 12, 40, 56, 178, 351 Aspect: 1, 8, 25, 195, 4/n.65, 200, 277, 308–34
and modality: 25, 309–21, 325–34, 340–41, 356 imperfective: 311–24, 327–28, 341, 356, 10/n.64 perfective: 311–12, 320–24, 327, 328, 331, 341 aspect locus: 323 Assertion: passim after verbs of cognition and perception: 73–84, 338 after verbs of locution: 50–56, 338 and realis/irrealis: 14–22 and xpr.n: 200–02, 304, 339 definition of: 14–15, 173–77, 335 in Egyptological literature: 23 in negated complements: 206–10, 214, 339 in predicate complements: 291, 304– 06, 340 in preposition complements: 223, 246–57, 339 in subject complements: 173–77, 179, 181–82, 339 occurrence after verbs: 41–49 refraining/withholding of: 48, 62, 90 Auxiliaries: 2, 0/n.6, 17, 23, 25, 101, 120, 180, 194, 4/n.65, 200–02, 271, 303, 306, 314, 339, 343, 345, 347, 350, 354, 358 Auxiliation: 0/n.6, 201–02 Background: 16, 84, 89, 90, 97, 175–76, 182, 186, 217–18, 230–36, 255–59, 318, 327–28, 339, 351, 355, 356, 359 Balanced sentence: 53, 10/n.38 Canonicity: 117–18 Cataphoric: 4/n.3, 198, 301, 10/n.56 Categorical: 0/n.80, 57, 281, 298 Causal clause: 168, 252–54 Causative: 0/n.89, 142–53, 3/n.46, 345
420
general index
and implicative: 147–52 and negation: 150 in Coptic: 0/n.89, 146 indirect: 144–45 with iri: 3/n.46, 3/n.52 Caus roots: 4, 29, 32, 116 ‘Circumstantial’ sDm=f: 29, 31, 0/n.95, 107, 2/n.181, 4/n.57 Cleft sentence: 0/n.100, 57, 2/n.28, 102, 2/n.175, 117, 119, 248, 278, 281, 306, 307, 352 Closure: 323–324, 327 Coding: 14, 20, 83, 114, 146, 224, 291, 321, 336 Cognitive Grammar: 27 Comment: 25 Commitment: passim and futurity: 17–18 and negation: 21 as a precondition for assertion: 14, 44–46, 49, 173–75 of the real speaker: 46–48, 55, 60, 62, 64–66, 72, 123–24, 207 Complementiser: 101–02, 108, 3/n.6, 210, 284, 286, 344, 10/n.69; see also ntt/wnt Completed, completion: 93, 149, 236, 243, 312, 9/n.7, 317, 323 Concessive: 19, 2/n.35, 176, 5/n.27, 222, 233 Conditional: 21, 1/n.27, 59, 168, 183, 5/n.27, 6/n.5, 235, 249, 273, 275, 279, 9/n.37 and irrealis: 59, 273 Conjuncts m-xt developed into: 260, 262–63, 339 prepositions used as: 4, 221 Construction Grammar: 27, 28 Contingent patterns: 262 Continuative: 168, 4/n.56, 271 Converter, converted: 9, 11, 0/n.99, 104–06, 112, 3/n.21, 6/n.71, 249, 250, 289, 296, 318, 9/n.35 Coptic: 0/n.89, 2/n.9, 115, 3/n.44, 146, 4/n.11, 4/n.45, 284, 306, 346–47 Counterfactual: 20, 2/n.35, 237–38, 255, 275 Counterparts: passim
functional: 31–34, 99, 116–18, 132, 179, 277, 336 and passim ‘syntactic’: 104 ‘Defective’ writings: 0/n.89, 32 Deixis, deictic: 41, 43, 54, 61, 169, 235, 7/n.37, 7/n.48, 336, 346 Demotic: 347 Deontic: 17, 28, 2/n.154, 126, 129, 153, 168, 229, 6/n.23, 6/n.74, 252, 253, 325, 338, 357 Derivation, derivative: 85, 333, 353 Determiners: 13 Diachrony, diachronic: 32, 35, 51, 115, 3/n.56, 159, 180, 194, 199, 202, 209, 245, 250, 259, 264, 7/n.35, 284, 332, 340, 341–47 Direct speech: 0/n.80, 38–41, 51, 7/ n.48, 346 and verba dicendi other than Dd: 2/n.8 Discourse: 19, 2/n.35, 84, 89, 92, 96, 217, 219, 224, 254, 259, 296, 327, 356 Doubling roots: 28, 32 and passim mA=f possibly specific distal: 32 Doubt: 15–16, 44, 45, 61, 122, 281, 325 Égyptien de tradition: 0/n.2 Epistemic: 17, 28, 2/n.106, 2/n.154, 121, 129, 153, 168, 174, 177, 6/n.23, 235, 252–54, 9/n.18 Evaluation, evaluative: 19, 48, 84–97 and passim Evidential: 1/n.30 Existential sentence: 102, 4/n.19, 193, 249, 8/n.6, 296 Expletive pronoun: 4/n.3, 197–99, 300 Factive: 0/n.47, 1/n.20, 1/n.28, 1/n.30, 92 Fear: 121 Figure: 6/n.95, 328 Final clause: 13, 29, 0/n.100, 106, 120, 141, 168, 5/n.3, 204, 5/n.27, 216, 225–29, 244, 256, 259, 285 Force dynamics: 27, 153, 161, 163, 191, 327
general index Foreground: 255–56 Functional Grammar: 27 Geminating sDm=f: passim absence after ntt/wnt: 7, 105–07, 249, 336 after nfr-n: 211 after non-assertive verbs: 127–70 after verbs of cognition and perception: 84–90, 94, 338 and causative: 142, 147, 152 and manipulative intent: 135–39, 164, 165, 169 and preventing verbs: 146–47, 150, 165 and realisation: 157–64, 166, 169, 336 and xpr.n: 196, 201 and success in manipulation: 132– 35, 155, 169, 338 and time-reference: 9, 88, 92, 130, 227, 238–41, 315–16, 317–20, 324–33, 340 as a ‘nominal form’: 7, 13, 29, 105– 06, 250 as ‘objective’: 23, 31 functional counterparts of: 31–34 and passim in predicate complements: 287–89, 300 in preposition complements: 223–24, 225–38, 239–41, 243–44 in subject complements: 179–80, 184–86, 190–91 modal function of: 32–34, 90, 99, 105, 116–18, 184, 282–83, 308, 319, 327–28, 330, 336 and passim Gemination as a morphological phenomenon: 32 Genitive: 0/n.8, 12, 2/n.172, 6/n.2, 7/n.35 gmi (verb): 109–15, 192, 265, 275, 285, 347 Grammaticalisation: 0/n.45, 20, 2/n.9, 99, 111, 133, 3/n.49, 146, 180, 192, 193, 200–01, 210, 5/n.25, 221, 262, 284, 306, 335, 337, 339, 345, 347 Ground: 6/n.95, 9/n.51
421
Habitual: 21, 22, 4/n.57, 238, 311, 312, 318, 320, 9/n.33, 324, 328, 359 ‘Hearsay’: 1/n.30 Hortative: 2/n.28, 60, 140 Hypotactic: 39, 50, 7/n.47 HA (particle): 275 xpr.n: 194–202, 262, 304 as an assertive marker: 200–01, 304 i-prefix: 2/n.43, 342 ib (verb): 122–24 Iconic: 39, 191, 321, 362 ii (verb): 33, 2/n.88, 2/n.106, 156, 185, 6/n.16, 314–15, 336 Illocutionary force: 14, 23, 2/n.182, 200, 207, 305 Illocutionary intention: 14, 16, 48, 49, 55, 99, 137, 177 imi (negative verb): 143, 5/n.3 Immutable roots: 28, 31, 39, 50, 92, 107, 3/n.8, 6/n.3, 8/n.2, 308, 354 Imperative: 17, 105, 140, 143, 271, 282, 8/n.22 Implicative: 0/n.47, 147–50, 152, 168, 170, 4/n.72, 203, 7/n.49, 338 Indicative: see assertion, realis, ntt/wnt Indicative sDm=f: 32, 244, 8/n.40, 9/ n.58 Indirect question: 81–82 non-verbal: 2/n.79 Indirect speech: 41, 50–72, 207, 343, 345–47 as a general signal of irrealis after Dd in Middle Egyptian: 72 in Late Egyptian: 2/n.9, 346–47 Infinitive: 0/n.3, 12, 2/n.46, 3/n.44, 3/n.47, 4/n.38, 4/n.70, 211, 212, 271, 333, 345 Information structuring: 56, 97, 259, 292, 354 ini (verb): 32, 33, 143, 312, 336 Interrogative: 21, 1/n.27, 58–61, 65, 79, 80, 105, 2/n.180, 3/n.87, 184, 4/n.43 scope of: 58–61, 65, 80, 2/n.179, 3/n.4, 210, 281, 354, 355 ir (conditional): 7/n.26, 275
422
general index
ir (topic marker): see ante-position, topic Irrealis: passim after non-assertive verbs: 125–70, 338–39 after verbs of cognition and perception: 81–101, 105–09, 338 after verbs of locution: 58–66, 67– 73, 338 and implicative: 147 and ‘unreal’: 17, 24 as a cross-linguistic variable: 20–22 as a grammatical category: 14–22 as a signal of inability and un-willingness to assert: 49 and passim definition of in Earlier Egyptian complementation: 24 distal: passim formal representatives of: 32–34, 169–70, 179, 224, 226–27, 308, 336 sDm=f with -w/-y marked as: 32, 99, 116–18, 276 and passim general division of: 24 in adjunct clauses: 120, 255–59, 339 in negated complements: 206, 210– 20, 339 in predicate complements: 292, 300, 340 in preposition complements: 223, 225–38, 255–59, 274–75, 339 in subject complements: 174–77, 179–80, 182–94 possibly in the Pyramid Texts: 35, 341–42 proximal: passim definition of: 99 gem. sDm=f marked as: 32, 99 and passim formal representatives of: 32–34, 169–70, 179, 224, 308, 336 synthetic character of: 118 is (particle): 2/n.28, 2/n.139, 2/n.146, 3/n.21, 7/n.10, 278–84, 340, 342 (i)sk/(i)sT (particles): 6/n.1 Isomorphism: 31, 80, 193, 5/n.25, 248, 7/n.23, 8/n.22, 295 iw (auxiliary): passim
and the ‘circumstantial’ sDm=f: 29 and theticity: 0/n.80 and assertion: 0/n.77, 101, 119, 202, 311, 358 and ntt/wnt: 101, 119 and second tenses: 350, 358 in Late Egyptian: 2/n.169, 111, 345 previous views on pragmatics of: 23 sentences with embedded directly: 85, 129 iwi (verb) 31, 33, 143, 185, 288, 314, 336 iwt (negative complementiser): 203–19 after prepositions: 223, 248 and modality: 23, 206, 339 and ntt/wnt: 2/n.52, 80, 206, 209, 210, 339 as a modal operator: 24 iwty (relative adjective): 348, 10/n.36 ‘Journalistic style’: 18, 1/n.26, 2/n.51 Jussive: 17 kA (initial particle): 13, 263 Knowledge: 47, 72, 83–84 Late Egyptian: 0/n.71; 0/n.77, 2/n.9, 2/n.12, 2/n.169, 111–12, 115, 3/ n.44, 195, 198–200, 6/n.86, 263, 285, 340, 343–46 Manipulation attempted: 125, 127–35, 149, 154, 157, 163, 168, 338 real: 125, 163 ‘Manner nominalisation’: 88–89 Mapping: 22, 25, 333, 336, 341 Marked, markedness: 15, 24, 33, 99–100, 115–19, 168, 180, 210, 255–57, 259, 276–77, 303, 304, 306, 309–10, 323– 24, 330, 331, 336, 341, 348, 358 Metaphor: 21, 22, 27, 210, 309, 323, 326, 329, 361 Middle Egyptian: 1, 29, 50–51, 72, 2/ n.148, 195, 245, 347 mk (initial particle): 103–04, 350 ‘m of predication’: 6/n.24, 6/n.48, 6/ n.57 Mutable roots: 4, 99, 113, 121, 122, 127,
general index 142, 152, 162, 164, 179, 193, 211, 6/n.3, 308, 336 Negatival complement: 0/n.96, 210, 5/n.32 Negative, negation: 5–6, 21, 93, 113, 203–20, 259, 280, 296, 9/n.18, 331 and irrealis: 21, 44–45, 46, 58, 120, 206, 210–20, 339 nfr-n (negation): 203–20 and passim after prepositions: 223, 238–29, 257 and ntt/wnt: 7, 210, 249, 339 as generalised irrealis: 213 as irrealis negation: 24, 206, 210– 20 negating adjunct clauses: 257 nfr pw (negative pattern): 189, 193, 5/ n.2, 249 n...is: 259, 280, 358 n-is (sDm(.n)=f): 5/n.38, 280 nn (sDm=f) (negation): 2/n.189, 179, 209, 248, 8/n.22 ‘Nominal forms’: 7–9, 12–13, 1/n.6, 88, 105, 173, 204, 213, 269–72, 295, 296, 322, 348, 351 general character of: 12–13 Nominal sentence: 2, 5, 6, 11, 2/n.25, 96, 102, 2/n.148, 105, 109, 2/n.168, 2/n.175, 113, 117, 119, 159, 177, 184, 189, 248, 278, 280, 287–307, 340, 10/n.64 predication in: 296–305 Non-geminating forms: passim after non-assertive verbs: 126–70, 187–88, 338 after verbs of cognition and perception: 81, 83, 90–94, 338 after verbs of locution: 59–60, 63, 338 and causative: 142–53 and manipulative intent: 137–38, 169 and preventing verbs: 141, 165–66 and realisation: 161–68, 171, 173 and success in manipulation: 133–34, 165, 169 and time-reference: 64, 92, 130, 134, 227, 241–42, 245, 312–15, 327 as ‘nominal forms’: 7
423
in nominal sentences: 8/n.2 in preposition complements: 223–24, 225–30, 234–38, 241–46 in subject complements: 179, 186– 94 modal function of: 32–34, 60, 63, 81, 94, 99–100, 106–108, 115–18, 276, 308 sentential status of: 103, 3/n.21 with endings -w/-y: passim absence after ntt/wnt of: 106, 249, 336 after rdi: 142–43, 152–53 status of: 29, 0/n.89, 32, 99, 106, 112, 115–18, 180, 276, 329, 330, 336 time-reference of: 245, 6/n.72, 316–17, 320–32, 340–41 Non-verbal sentences: 57, 103, 2/n.146, 119 Noun + sDm.n=f: 295 n sp: 179, 205, 209–10, 249, 9/n.20 n sDm=f: 32, 2/n.138, 2/n.175, 2/n.189, 209–10, 215, 6/n.71, 248 n sDm.n=f: 102, 2/n.189, 210, 215, 249, 271 nty (relative adjective): 51, 348 ntt/wnt: passim absence after non-assertive verbs: 121–25, 338 absence in nominal sentences: 5, 289, 296, 304 absence of as a general sign of irrealis: 24, 98–99, 125, 212, 256, 308, 335 after notionally assertive verbs: 50– 84 after prepositions: 10, 223, 246–54, 339 after verbs of cognition and perception: 73–84 after verbs of locution: 50–73 after other verbs: 94, 97–98 and ‘circumstantial’ sDm=f: 107 and evaluative use of irrealis: 84 and passim and gmi: 109–10 and xpr.n: 201 and iw: 103–04, 107, 108–09, 119 and iwt: 204, 206–10
424
general index
and is: 281, 282–83 and nty: 51, 348 and r ntt: 78, 250 and sDm=f with -w/-y: 106 and subject-anticipation: 107–08 and tense: 57–58, 59 and ‘thetic’: 56–57 and tm/nfr-n: 7, 210, 249 and wnn: 51 as complementisers: 101, 105–06 as modal operators: 24, 99, 116, 118, 125, 179, 335 and passim as ‘nominal converters’: 9–12, 106, 125, 4/n.19, 272 as signals of assertion: 24, 99–100, 105, 108, 116–18, 125, 179, 305, 335 and passim diachrony of: 50–51, 99–100 earlier views on modality: 23 etymology of: 51, 99–100 general distribution of: 10–12, 49, 125 grammaticalisation of: 99–100 in negated complements: 203–10, 219 in subject complements: 181–82, 186, 339 paradigm after: 101–09, 116–19, 248–50, 338 n-wnt (negative pattern): 192–94, 209 Obligative: 17, 229–30 Old Egyptian: 1, 50–51, 2/n.139, 2/ n.146, 127, 177, 179, 183, 4/n.24, 4/n.49, 5/n.4, 212, 6/n.1, 244, 261, 9/n.58, 343 Optative: 2/n.95 Original speaker: 40–43, 46, 54–55, 62, 76, 77, 126, 224, 342 Paratactic: 39, 51, 100, 2/n.167, 4/n.56, 201, 8/n.22 Participle: 51, 119, 159, 187, 189, 4/ n.45, 5/n.27, 348, 10/n.36 Passive: 177, 5/n.32 and .t(i)/.t(w)/.tw: see .t(i)/.t(w)/.tw Passive sDm=f: 0/n.84, 2/n.18, 102, 109, 2/n.169, 2/n.175, 117, 3/n.41, 181–82, 192, 197, 214, 248, 261, 8/ n.13
Patient: 2, 4/n.17 Polysemy, polysemic: 153, 161, 5/n.23, 222, 250 Possessive: 2/n.166 Predication: 13, 0/n.77, 25, 120, 179, 292, 296–305, 351, 352, 353 Predicative nexus: 89, 90, 4/n.38, 217, 274, 298, 307 ‘Preposition-ntt connectors’: 0/n.26, 2/ n.138, 248–50 Presupposition: 95, 179, 182, 183–86, 4/n.34, 190, 216–17, 233, 236, 251, 255, 258, 348, 352, 355, 359 and assertion: 15–16, 19, 0/n.77, 42, 48, 86, 119, 175–76, 217, 251 and ‘necessarily true’: 0/n.47, 1/ n.22 defeasibility of: 1/n.31 Prospective forms: 7, 30, 31, 106, 108, 3/n.90, 245, 8/n.2, 316 and passim as future expressions: 18, 245–46, 316, 324–25 as modal forms in previous analyses: 23, 31 as ‘nominal forms’: 13 morphology of: 29–30 Pseudo-verbal patterns: 18, 117, 249, 287, 288, 289, 291, 292, 294, 295, 296, 302, 303, 305, 306, 318, 333 pw 5, 2/n.79, 2/n.148, 2/n.167, 2/ n.168, 159, 290, 291, 293–307, 340 functioning as a modal operator: 304–05, 306, 340 semantic value of: 291, 298, 299–306, 340 Quoting: see direct speech Raising: 96–97, 2/n.167, 110–11, 145, 199 Realis: passim after verbs of cognition and perception: 73–81, 338 after verbs of locution: 50–56, 72, 338 analytic character of: 118 as a cross-linguistic variable: 20–22 as a grammatical category: 14–22 as a signal of belief, acceptance and relevance: 14, 49 and passim
general index definition of in Earlier Egyptian complementation: 24, 116, 335, 337 in negated complements: 206–10, 339 in predicate complements: 291, 305–06, 340 in preposition complements: 255– 57, 339 in subject complements: 173–77, 181–82, 339 Real speaker: 40–48, 52, 54, 58–60, 72, 77, 79, 81, 92, 124, 125, 130, 138, 157, 161, 173, 207–10, 224, 346 Reanalysis: 100, 177–78 rdi (causative verb): 32–33, 94, 124–25, 141–53, 187, 5/n.3, 312 paradigm of: 0/n.89, 142–43, 145– 53 permissive sense of: 144, 145, 150– 52 writing of: 0/n.100 r-Dd: 2/n.180, 284–86, 340, 343, 344 Relative clause: 146, 3/n.58, 168, 348–49 and modality: 0/n.62; 0/n.69, 348– 49 Relative forms: 0/n.23, 3/n.58, 7/n.17, 9/n.35, 345, 348–49 ‘abstract’: 0/n.15, 7/n.16, 348 Relevance: passim and presupposition: 16, 175–77 as a graded concept: 19–20, 327–28, 336 as a precondition for assertion: 15, 175 evaluation of: 48, 84–94, 132, 134, 175, 207, 217–20, 230–36, 243, 336 not part of modal system: 20 Reports, reported: 42, 43, 45, 1/n.26, 47, 54, 56, 61, 2/n.50, 77, 92, 130, 131, 134, 136, 137, 182, 207, 342 Result clauses: 168, 284 Rheme: 25, 352 sDm.n=f: 264–78 and passim and ntt/wnt: 268, 272–73 and tense: 277, 310 as a counterpart of both geminating
425
and non-geminating sDm=f in complementation: 274–75, 276– 77, 340 as modally unmarked: 2/n.184, 276, 331, 340 as a morphological unit: 270–72 as a ‘nominal form’: 13, 269–72 bare, after prepositions: 265–68, 274 bare, after verbs: 264–65, 273–74 in predicate complements: 288, 291 sDm.n.tw=f: 268–70, 288 sDmt=f: 7/n.35, 9/n.33, 331 the forms int=f, iit=f and iwt=f as: 33, 2/n.183, 330–31, 336 sDm.ty=fy: 159 Second tense: 13, 0/n.100, 2/n.14, 2/ n.87, 2/n.95, 103, 2/n.180, 119, 120, 3/n.31, 184, 196, 197, 201, 248, 249, 6/n.123, 268–69, 271, 273, 279, 280, 281, 285, 314, 317, 318–20, 330, 347, 349–59 direct embedding of: 128–29, 160 ‘emphatic’: 2/n.11, 2/n.22, 251–52, 318, 350, 351, 352, 353, 10/n.66, 359 setting: 5/n.27, 6/n.36, 350, 351, 352, 353, 359 Semantic roles: see Arguments Semitic: 29 ‘Sequence of tenses’: 1/n.9 Simulative: 237 Speech acts: 14, 47, 126, 135, 146, 3/ n.68, 208, 225, 252–54, 320 indirect: 137, 320 Standard Theory (ST): passim and aspect: 322 and complementation: 7–9, 103–04, 289 and modality: 23 Stative: 12, 102, 4/n.46, 248, 261, 269, 8/n.6, 8/n.8 Structuralism: 8 Subject-anticipation: 107–108, 4/n.64, 303 Subjunctive: 15, 17, 19, 44, 45, 58, 2/n.40, 71, 131, 174–76, 185, 205, 6/n.43, 225, 243, 258, 359 as ‘mood of subordination’: 0/45 ‘Subjunctive’ sDm=f: 0/n.89, 31, 106,
426
general index
108, 3/n.3, 3/n.14, 3/n.90, 6/n.68, 9/n.33 Tautology: 119, 134 Temporal clause: 168, 224, 256, 258 Tense: 1, 23, 57–58, 199, 277, 309, 310, 325, 331 and irrealis: 325–26 Theme: 25, 6/n.35, 351–54 Thetic: 0/n.80, 56–57, 59, 281, 298– 300, 301, 340 .t(i), t(w), tw (passive morphs, indefinite pronouns) 2/n.15, 2/n.63, 108, 127, 177–79, 181, 212, 264, 349, 350 tm (negative verb): 203–20 and passim absence after ntt/wnt: 7, 210, 249 after prepositions: 223, 228–29, 244, 257 and presupposition: 216–17 and resumption: 3/n.58 as an irrealis negation: 24, 206, 210– 21, 339, 358 as generalised irrealis: 213, 339 as a ‘nominal’ negation: 7, 9, 13, 204, 205, 289 earlier views on modality of: 24, 215 in predicate complements: 287, 300 negating adjunct clauses: 13, 204, 215–16, 257 negating ix/xr/kA-clauses: 13, 5/ n.11 negating relative clauses: 348
negating sDmt=f: 331 negating second tenses: 350, 358 Topic, topical, topicalised: 25, 56, 57 234, 294–95, 298–99, 301, 302–04, 305–06, 351–54, 355, 357 Ultimae infirmae roots: 4, 0/n.34, 29–34, 2/n.43, 99, 116, 127, 271, 287, 312, 322, 336, 348, 350 Ult. -w: 7/n.25 Unmarked: see marked, markedness Use versus mention: 138, 216–17 ‘Virtual clause of condition’: 120, 216, 257 ‘Virtual noun clause’: 7, 197, 4/n.78 ‘Virtual relative clause: 7/n.44, 8/n.18, 348 wh-question: 184, 5/n.41, 216, 258, 354, 355 wnn (auxiliary verb): passim and wnt: 51 as a ‘converter’: 6/n.71 as a lexical verb: 4/n.19, 4/n.46 auxiliary status of: 33, 117, 4/n.24, 192–94, 4/n.79, 318 functional role of: 33–34, 0/n.99, 318–19 problematic behaviour of: 167 time-reference of: 315–16, 318–19, 9/n.35
general index
427
INDEX OF LANGUAGES Alamblak: 21 Albanian: 18 Bargam: 21 Bemba: 43, 130–32, 134, 258, 359 Bulgarian: 1/n.13
Hebrew (Biblical): 17, 26, 259 Italian: 1/n.22, 62, 64, 174, 176, 9/n.48 Itzaj Maya: 9/n.41 Japanese: 101
Caddo: 17, 21, 6/n.43 Central Pomo: 17, 21 Coptic: see general index
Kinyarwanda: 3/n.6 Korean: 3/n.64
English: 17, 42, 46, 47, 52, 58, 61, 2/n.172, 3/n.6, 145, 153, 161, 3/ n.88, 176, 197, 200, 6/n.11, 230, 325, 342, 10/n.69 American: 0/n.67
Lango: 131–32, 134 Latin: 52, 3/n.2, 258
Finnish: 1/n.27, 2/n.172, 4/n.8 French: 0/n.58, 42–43, 1/n.22, 1/ n.26, 52, 2/n.51, 4/n.4, 175–76, 6/n.43, 6/n.91, 9/n.48 Fula: 258
Polish: 20, 1/n.22 Portuguese: 1/n.22, 1/n.23
German: 18, 45–46, 47, 52, 58, 62– 63, 2/n.51, 6/n.11 Greek Classical: 3/n.2 New Testament: 359
Mangarayi: 258 Mojave: 154–55
Romance languages: 15, 179, 361 Russian: 20, 1/n.22, 48, 3/n.3, 4/n.8, 328, 10/n.66 Scandinavian languages: 52 Spanish: 15–16, 19–20, 42, 44, 47, 48, 71, 3/n.2, 174–75, 185, 243 Swedish: 1/n.27, 4/n.8
428
general index
key to abbreviations and symbols used
429
KEY TO ABBREVIATIONS AND SYMBOLS USED∗ ١ * †
» › {} […] [consonants] ADM Adv. ASAE BiOr BSEG CdE CT DE EAG FIN GEG GM IFAO IMP INCL IND IRR JARCE JEA JEOL JNES JUSS LEg LGEC LingAeg MEg MDAIK N Nom.
Hypothetical Ungrammatical/Unattested Presupposes Implicates In transliteration, superfluous sign or signs In transliteration, lost sign or signs In transliteration, restored Admirative Adverb, adverbial Annales du Service de Antiquités de L’Égypte Bibliotheca Orientalis Bulletin Société d’Égyptologie Genève Chronique d’Égypte Coffin Texts Discussions in Egyptology Edel, Altägyptische Grammatik (see Edel, E. 1955–64 in the Bibliography) The ‘finite neutral pattern’ of realised complement situations of Bemba Gardiner, Egyptian Grammar (see Gardiner, A. 1957 in the Bibliography) Göttinger Miszellen Institut Français d’Archéologie Orientale Imperative The ‘inclusive’ of Bemba Indicative Irrealis Journal of American Research Center in Egypt Journal of Egyptian Archaeology Jaarbericht van het Vooraziatisch-Egyptische Genootschap Ex Oriente Lux Journal of Near Eastern Studies Jussive Late Egyptian Lefebvre, Grammaire de L’Égyptien Classique (see Lefebvre, G. 1955 in the Bibliography Lingua Aegyptia Middle Egyptian Mitteilungen des Deutschen Archäologischen Instituts in Kairo In transliteration, noun or personal name Nominal
∗ Omitted from this list are abbreviations for sources of Ancient Egyptian texts (i.e. citations such as ‘Urk I’ or ‘Sh.S.’). For full bibliographical details thereof, readers are requested to consult the index of cited passages.
430 OLP OLZ Or PC Ph. Pl. Pred. PT RdE REAL RecTrav SAOC SAK SS ST SUB TAM VOM Wb WGMT WZKM ZÄS
key to abbreviations and symbols used Orientalia Lovaniensia Periodica Orientalistische Literaturzeitung Orientalia Personal Communication Phrase Plate Predicate Pyramid Texts Revue d’Égyptologie Realis Recueil de Travaux Relatifs a la Philologie et a l’Archéologie Égyptiennes et Assyriennes Studies in Ancient Oriental Civilizations Studien zur Altägyptischen Kultur Same Subject Standard Theory Subjunctive Tense-Aspect-Modality Verb(s) Of Movement Wörterbuch (see Erman, A. & Grapov, H. 1926–31 in the Bibliography) Westendorf, Grammatik der Medizinishen Texte (see Westen-dorf, W. 1962 in the Bibliography) Wiener Zeitschrift für die Kunde des Morgenlandes Zeitschrift für Ägyptische Sprache und Altertumskunde
probleme der ägyptologie
The Probleme der Ägyptologie, founded in 1953 by Hermann Kees, are concerned with the religion, literature, history – political, social and economic – and language of ancient Egypt, both pharaonic, Ptolemaic and Roman, including manifestations abroad and in later times. The series accepts all serious methodological approaches. 1 Kees, H. Das Priestertum im ägyptischen Staat vom Neuen Reich bis zur Spätzeit. Indices und Nachträge. 1958. isbn 90 04 06231 9 Out of print 2 Otto, E. Die biographischen Inschriften der ägyptischen Spätzeit. Ihre geistesgeschichtliche und literarische Bedeutung. 1954. isbn 90 04 01780 1 Out of print 3 Helck, W. Zur Verwaltung des Mittleren und Neuen Reichs. Register. Zum 60. Geburtstag des Verfassers zusammengestellt von den Mitarbeitern der Ägyptologischen Abteilung an der Universität Hamburg. 1975. isbn 90 04 04362 4 Out of print 4 Kees, H. Die Hohenpriester des Amun von Karnak von Herihor bis zum Ende der Äthiopenzeit. 1964. isbn 90 04 01782 8 Out of print 5 Stadelmann, R. Syrisch-palästinensische Gottheiten in Ägypten. 1967. isbn 90 04 01783 6 Out of print 6 Velde, H. te. Seth, god of confusion. A study of his role in Egyptian mythology and religion. Reprint with corr. of the 1st (1967) ed. 1977. isbn 90 04 05402 2 Out of print 7 Guglielmi, W. Die Göttin Mr.t. Entstehung und Verehrung einer Personifikation. 1991. isbn 90 04 08814 8 8 Bonneau, D. Le régime administratif de l’eau du Nil dans l’Égypte grecque, romaine et byzantine. 1993. isbn 90 04 09687 6 9 O’Connor, D. & D.P. Silverman (eds.). Ancient Egyptian Kingship. 1995. isbn 90 04 10041 5 10 Loprieno, A. (ed.). Ancient Egyptian Literature. History and Forms. 1996. isbn 90 04 09925 5 11 Hasel, M.G. Domination and Resistance: Egyptian Military Activity in the Southern Levant, 1300-1185 bc. 1998. isbn 90 04 10984 6 12 Doxey, D.M. Egyptian Non-Royal Epithets in the Middle Kingdom. A Social and Historical Analysis. 1998. isbn 90 04 11077 1
13 Kahl, J. Siut – Theben. Zur Wertschätzung von Traditionen im alten Ägypten. 1999. isbn 90 04 11441 6 14 Grunert, S. & I.Hafemann (eds.). Textcorpus und Wörterbuch. Aspekte zur ägyptischen Lexikographie. 1999. isbn 90 04 11536 6 15 Molen, R. van der. A Hieroglyphic Dictionary of Egyptian Coffin Texts. 2000. isbn 90 04 11654 0 16 Brand, P.J. The Monuments of Seti i. Epigraphic, Historical and Art Historical Analysis. 2000. isbn 90 04 11770 9 17 Peden, A.J. The Graffiti of Pharaonic Egypt. Scope and Roles of Informal Writings (c.3100-332 bc). 2001. isbn 90 04 12112 9 18 Török, L. The Image of Ordered World in Ancient Nubian Art. The Construction of the Kushite Mind 800 bc-300 ad. 2001. isbn 90 04 12306 7 19 Moers, G. Fingierte Welten in der ägyptischen Literatur des 2. Jahrtausends v.Chr. Grenzüberschreitung, Reisemotiv und Fiktionalität. 2001. isbn 90 04 12125 0 20 Knoppers, G.N. & A. Hirsch (eds.). Egypt, Israel, and the Ancient Mediterranean World. Studies in Honor of Donald B. Redford. 2004. isbn 90 04 13844 7. 21 Müller-Wollermann, R. Vergehen und Strafen. Zur Sanktionierung abweichenden Verhaltens im alten Ägypten. 2004. isbn 90 04 13906 0 22 Morris, E.F. The Architecture of Imperialism. Military Bases and the Evolution of Foreign Policy in Egypt’s New Kingdom. 2005. isbn 90 04 14036 0 23 Török, L. Transfigurations of Hellenism. Aspects of Late Antique Art in Egypt ad 250-700. 2005. isbn 90 04 14332 7 24 Sederholm, V.H. Papyrus British Museum 10808 and its Cultural and Religious Setting. 2006. isbn 978 90 04 14349 4 25 Winand, J. Temps et Aspect en Égyptien. Une approche sémantique. 2006. isbn 978 90 04 15216 8 26 Uljas, S. The Modal System of Earlier Egyptian Complement Clauses. A Study in Pragmatics in a Dead Language. 2007. isbn 978 90 04 15831 3
issn 0169-9601
428
general index
428
general index