Linguistic Realization of Evidentiality in European Languages

IX<EXT> TOTAL

Helsinki corpus

ARCHER corpus

50 (55.6%) 10 16 24 2 (2.2%) 29 (32.2%) 17 6 6 5 (5.6%) 4 (4.4%) 90

102 (36.4%) 42 40 20 17 (6.1%) 137 (48.9%) 34 73 30 12 (4.3%) 12 (4.3%) 280

Construction Types I and III are most prevalent in both corpora, although Type I occurs more often in the Helsinki Corpus, and Type III occurs more

Evidentiality, polysemy, and the verbs ofperception

263

often in the ARCHER Corpus. Types II, VIII, and IX occur at a far lesser frequency in both corpora, but interestingly, parenthetical and external constructions occur at ahnost the same frequency in the Helsinki Corpus, and they do occur at the same frequency in the ARCHER Corpus. There is a greater disparity among the uses of Type II in both corpora, with the ARCHER Corpora showing 3.9% higher frequency of use.

5.1.2. German sehen Like its English counterpart, German sehen is the most widely attested perception verb in the examined corpora. There are 852 total attestations of sehen in the Bonner Corpus of Early New High German, 30 (3.5%) being evidential. In the Kant Corpus, 68 (3.9%) of the 1,756 cases of sehen are evidential. 3,555 attestations of sehen can be found in the Goethe Corpus, and ofthese, 274 (7.7%) express evidential meaning. Finally in the DWDS, there is a surprisingly low percentage of evidential attestations. Out of a random sample of 1,500 attestations, sehen was used evidentially in only 19 (1.3%) of these cases. There is no apparent reason why this should be the case, considering all the uses attested in earlier corpora can still be used on contemporary German. In any case, an explanation for this discrepancy lies beyond the confmes of the present study. Table 2 presents us with the breakdown of construction types in the German language corpora. Table 2: Occurrences of evidential sehen in German language corpora CONSTRUCTION TYPE I with complementizer without compl. II III infmitive past participle V VIII

IX<EXT> TOTAL

BONNER CORPUS 18 (60%) 14 4 7 (23.3%) 4 (13.3%) 3

1 (3.3%) 30

KANT CORPUS 39 (57.4%) 38 2 (2.9%) 20 (29.4%) 4 16 1 (1.5%) 3 (4.4%) 3 (4.4%) 68

GOETHE CORPUS

DWDS

73 (26.6%) 61 12 5 (1.8%) 181 (66.1%) 113 68

9 (47.4%) 8 3 (15.8%) 4 (21.1%) 1 3

13 (4.7%) 2 (0.7%) 274

2 (10.5%) 1 (5.3%) 19

264

Richard Jason Whitt

Similar to the English data, Construction Types I and III are most widely attested throughout the corpora, although- as with English- Type I enjoys greater frequency in some corpora (Bonner, Kant, DWDS), while Type III is more prevalent in others (Goethe). And similar to English, all other construction types occur at a much lesser frequency than Types I and III (although in the Bonner Corpus, Type II occurs more often than Type III). There is one anomaly in the German language corpora, particularly the Kant Corpus: the appearance of one instance of the Type V construction, which is generally restricted to object-oriented perception verbs. This will be addressed later (see Section 5.2.4).

5.2. Qualitative Analysis We now turn to an examination of the relevant construction types and the evidential meaning(s) conveyed within each of these types. In general, English see and German sehen display a remarkable similarity with on another in terms of what type(s) of evidence is/are expressed by each construction.9

5.2.1. Literal visual perception is one type of evidence that can be designated by see or sehen when it is followed by a finite complementizer clause: (30) (31)

We saw that the crocodile moved only his upper jaw. (ARCHER: < 1667laud.j 2b>) nun sah ich erst, dafi unten quervor ein ziemlich langes Biinkchen stand, worauf eine Mandoline lag. (Goethe Corpus: Dichtung und Wahrheit (1811-1833), Hamburger Ausgabe, Band 9, S. 58) 'It was then that I saw that down below a rather long bench stood diagonally in front of me, on which a mandolin was lying.'

So in (31 ), for example, visual perception allows the speaker to pinpoint and indicate the location of the mandolin. More common in this construction type, however, is the indication by the speaker (or writer) that inference has driven him/her to arrive at a certain conclusion. This inference is often based on observation, which of course involves some degree of visual

Evidentiality, polysemy, and the verbs ofperception

265

perception, so the literal meaning of see or sehen is not wholly absent when inference is being indicated: (32)

(33)

I see you are dying with curiosity to know what has excited my anger, which I consider both inquisitive and impertinent. (ARCHER: < 1889madd.d6b>) Wir sehen, daj3 die Lebenshaltung des Volkes eine bessere geworden ist, und der weitere Ausbau des Eisenbahnnetzes wird die wirtschaftliche Entwicklung auch ferner giinstig beeinjlussen. (DWDS: "Landtagsverhandlungen", in: Vossische Zeitung (AbendAusgabe) (06.03.1907), S. 4) 'We see that the people's standard ofliving has become better, and moreover, the further expansion of the railway system will influence economic development positively.'

Neither the addressee's curiosity in (32) nor the standard of living in (33) can be visually perceived directly, but some sort of action or state of affairs that is indicative of these conditions can be visually perceived, and it is this act of perception that drives the inference made in both these examples. Of a more abstract nature is the metaphorical use of see or sehen to indicate knowledge or understanding. In these cases, traces of literal perception are more difficult to pinpoint than in (32) or (33): (34)

(35)

I now see plainly that Men cannot arrive at a full Satisfaction by Riches, nor at Power by enjoying Principalities or Kingdoms, nor at Esteem and Reverence by the Accession of Dignities, nor at Nobility by Glory, nor at true Joy by carnal Pleasures. (Helsinki Corpus: CEBOETH3, Richard Preston (trans.), Boethius (1695), p. 124) Sulzers Theorie war mir wegen ihrer falschen Grundma:xime immer verhaj3t, und nun sah ich, daj3 dieses Werk noch viel mehr enthielt, als die Leute brauchen. (Goethe Corpus: Italienische Reise (1786-1788), Hamburger Ausgabe, Band 11, S. 207) 'Sulzer's theory was always abhorrent to me because of its incorrect fundamental maxims, and now I saw that this work contains much more than the people need.'

Of course, for either Boethius or Goethe to have arrived at this understanding, they must have observed or witnessed (i.e. visually perceived) events that shaped their epistemology, so even here, one cannot wholly exclude

266

Richard Jason Whitt

literal perception. After all, it is the source domain of the SEEING IS KNOWING metaphor (Sweetser 1990: 32-34; cf. Lakoff and Johnson 2003), for one knows what one has seen. Although the complementizer is present in most of the above examples, sometimes -as in (32) - it can be absent. There have been suggestions that this may be an indication of grammaticalization (see Note 6), although as far as evidential meaning is concerned, the presence or absence of this complementizer does not appear to affect the type of meaning expressed in any perceivable way. There is, however, one constructional sub-type included here that follows the structure but behaves rather differently than any other evidential construction. This is the grammaticalized use of the English present participle seeing, in which this participle behaves more like a causal conjunction (e.g. because) than an actual perception verb because the speaker/writer takes whatever propositional content falls within the scope of seeing for granted, i.e. it belongs to the domain of knowledge; this, in tum, is then described as the cause or motivation for the apodosis: (36)

(37)

Than begane he to water hys plantes, sayenge unto me, "Syr, I thowghte once never to have seene yow agayne; yow are grownowte of my knolledge; and, seynge that yt ys the wyll of God that yow shold not dye by ther crwelty, I truste that your blud shal never be requyryd at my handys. (Helsinki Cor-pus: CEAUTO 1, Thomas Mowntayne, The Autobiography of Thomas Mowntayne (1500-1570),p. 213) Seeing that diplomatic relations with Japan have been broken off, you can prevent, ifyou consider it necessary, any Japanese descent on Korea. (ARCHER Corpus: <1905pall.n7b>)

Because seeing indicates that the speaker is in possession of the knowledge, i.e. evidence, that allows the content of the protasis to be a given, this type of construction is included in the category, even though it is notably different than uses not involving subjectless present participles. But like other cases of FCCs, the complementizer need not appear in the constructions. Whatever the implications of this omission may be (a higher degree of grammaticalization?) lie beyond the confines of this study.

Evidentiality, polysemy, and the verbs ofperception

267

5.2.2. The same types of evidential meanings expressed in the construction are found in this construction as well. Only here, the exact nature of the proposition is often vaguer because the complementizer is a whword that lacks specificity. Consider (38) and (39), for example, where literal vision is indicated as evidence:

(3 8)

(39)

We were there at ye time of ye fair & so saw what there chief merchantdise consisted of namely, flax, hemp, sheep, horses, mules, & such like, & pistols also which are there made & sold in great quantities, it being a place for that sort of work ye most remarquable in France. (ARCHER Corpus: <1687ferr.j2b>) Jch habe gesehen, wie der Groj3forst Sergius von Ruj3land getotet wurde ... ("Feuilleton," in: Vossische Zeitung (Abend-Ausgabe) 03.03.1905, S. 9) 'I saw how the grand duke Sergius of Russia was killed .

So in (39), although we know the writer saw that Sergius of Russia was killed, we do not know exactly how (wie) he was killed. The case is similar in (38), where we do not learn immediately what constituted the merchandise, although this information is provided immediately after the complementizer clause. The case is similar when a general sense of observation which includes varying degrees of literal visual perception - serves as evidence: (40)

(41)

I was delighted to see how rapidly my proposition was accepted, and we made a hasty breakfast, first sending in some of our food to the other party. (ARCHER Corpus: <1892stoc.f6a>) Lasset uns bei diesem heiligen Kunstwerk, der Wohlthat, durch die unser Geschlecht ein Menschengeschlecht ward, mit dankbaren Blicken verweilen, mit Verwunderung, weil wir sehen, welche neue Organisation von Kriiften in der aufrechten Gestalt der Menschheit anfange, und wie allein durch sie der Mensch ein Mensch ward! (Kant Corpus: AA VII, Recensionen von J G. Herders Ide en zur Philosophie der Geschichte der Menschheit. Theil] & 2 (1785), S. 48) 'Let us linger by this holy work of art with thankful looks and amazement, the boon through which our species became the human race, because we see which new organization of powers begins in

268

Richard Jason Whitt

the upright form of humanity, and how only through it the person became a human!' So in (40), we lrnow the writer's proposition was accepted rapidly (but we don't lrnow how rapidly), and (41), we do not lrnow which ''new organization of powers" (neue Organisation von Kriiften) or how through these powers "the person became a human" (der Mensch ein Mensch ward), but we do lrnow that Kant claims to have observed these events occur; the whcomplementizers, which contain part of the propositional content, simply cause their respective clauses to lack specificity. The case is the same when see or sehen have metaphorical denotations of lrnowledge and understanding: (42)

(43)

When I was with him I was always puzzled and uneasy, and always wondering why on earth he had ever married Dottie or rather how Dottie had ever arranged it, but I could see why he liked the family publishing business. It was a sort of ivory tower for him a plaything with which he could do what he wanted without worrying over the financial angle and books with the Peale imprint always had artistic distinction. (ARCHER Corpus: <1951marq.f8a>) vnnd damit wir widerumb sehen/wie von der Sichtbarn Catholischen Kirchen dij3 zu verstehen seylzeugt der Hey!. (Bonner Corpus: Text 157, Johan Rosenthal, Auj3fuehrliche Widerhol- und Vermehrung der kuertzen Bedencken vom bestaendigen Baw auf! den Felsen vnd nicht auf! den Sand (1653),8.15) 'And the Lord testifies so that we see again how this is to be understood by the visible Catholic Church.'

In (42), the writer indicates that Henry Peale (the antecedent of he) liked the publishing business and that he understood this to be true. What we do not lrnow is why this was the case, even though the author indicates he also understands the reason behind Peale's affinity with publishing; he simply does not specify this fact. And in (43), exactly how (wie) the Lord's testimony clarifies what is to be understood is not indicated within the complementizer clause; that information is left elsewhere in the discourse.

Evidentiality, polysemy, and the verbs ofperception

269

5.2. 3. If there is any construction in which literal vision will be indicated as evidence, it is in the construction, which almost solely is a marker of visual evidentiality when either see or sehen is the PV Whether the NFV is an infinitive, a past participle, or a present participle (in English) makes no difference as to what type of evidential meaning is expressed; rather, the NFV is used to aspectually distinguish the action or state for which there is evidence. Consider the following: (44) (45)

(46)

(47)

(48)

I saw her pass through the room where we sat before dinner. (ARCHER Corpus: <1765bswl.x4b>) And John bare record saying, I saw the Spirit descending from heaven, like a Doue, and it abode vpon him. (Helsinki Corpus: CNTEST2, The New Testament (1611 ), I, 20) Vnd do wart ich aber verzucket, vnd sihe hi mir ston einen gar grosen man ... (Bonner Corpus: Text 231, Rulmann Merswin, Des Gottesfreundes im Oberland (1352), S. 11) 'And then I was thrilled and saw a very large man standing next to me.' Sdeath! I have heard Sense run down, and seen Idiotism, downright Idiotism triumph so often, that I cou'd almost think of Wit and Folly as Mr. Hobbes does of Moral Good and Evil, that there are no such Things. (ARCHER Corpus: <1730fiel.d3b>) Wir sehen eine neuerliche Steigerung der Provisionen (um rund 400000 M) und der Einnahmen aufZinsen- und Wechselkonto (um 1 313 000 M) hervortreten. (DWDS: ["Der Aufschwung des "], in: Vossische Zeitung (Morgen-Ausgabe) (03.03.1912), S. 10) 'We see a recent increase of provisions (about 400,000 M) and the revenues from interest- and acceptance accounts (at 1,313,000 M) emerging.'

In (44) through (46), vision provides the evidence for the propositions. But in English examples (44) and (45), a further aspectual distinction is made: the infinitive in (44) provides a non-progressive reading of the event, whereas the present participle in (45) mandates a progressive interpretation of the Spirit's descent. In (46), the infinitive is the only possible verb form because the use of a present participle in such a context in German would be ungrammatical. So whether there is to be an additional aspectual reading (or whether such a distinction in German is relevant at all) is left to

270

Richard Jason Whitt

context and reader/hearer interpretation. In (47) and (48), there is not so much a sense of direct vision as there is of observation, although in both these instances, it should be clear that vision would form at least part of the basis of observation. Hence in (47), idiotism itself cannot be visually perceived, but acts of idiotism can, and it is these acts which the writer has seen/observed that draws him to the conclusion that idiotism triumphs "so often." When a past participle appears as the NFV, two things are indicated: the resultant state of a previous action (which seems to be more in focus in (49) and (50)), or the passive framing of an event, i.e. the agent of action being omitted from the clause and patient of action appearing as the direct object instead (emphasized in (51) and (52)): (49)

(50)

(51)

(52)

Looking from where I now stood I saw on every side a thick wood pierced with avenues in a more or less perfect state. (ARCHER Corpus: 1795twin.j4b>) Als ich eines Tages die Strafte betrat, sah ich an der nahen Station der Untergrundbahn eine Menschenmenge versammelt. (DWDS: [P.B.], "Das Jahr in Paris", in: Berliner Tageblatt (AbendAusgabe) (03.03.1908), S. 1-3) 'One day when I entered the street, I saw a crowd of people assembled at the nearby subway station.' For when many things are spoken of before in scripture, whereof we see first one thing accomplished, and then another, and so a third, perceiue wee not plainely, that God doeth nothing else but lead vs along by the hand, til he haue setled vs vpon the rocke of an assured hope, that no one iote or title of his word shall passe till all be fulfilled? (Helsinki Corpus: CESERM2A, Richard Hooker, Two Sermons upon Part ofS. Judes Epistle (1614), p. 8) Wir sehen auch diese Theorie in der Praxis hinreichend bestiitigt. (Kant Corpus: AA VIII, Uber den Gemeinspruch (1793), S. 303) 'We even see this theory adequately confirmed in practice.'

The crowd of people are already assembled at the subway station in (50), and the writer indicates that he sees this current state which is the result of a previous assembling of people. In (52), on the other hand, the focus appears to be more on Kant's observation that the theory has been confirmed, but he does not indicate exactly who is responsible for the theory's confrrmation. There is indeed a resultant state indicated here as well, but there is also an additional sense of passivity that is lacking in (50). The same can

Evidentiality, polysemy, and the verbs ofperception

271

be said of (49) and (51): the focus of (49) appears to be more of a previous action resulting in avenues throughout the woods (although who is responsible for "piercing" these woods is not indicated), whereas in (51), the actual acts of accomplishments without regard to the agents of these acts are in focus. There is one other type of evidence indicated in the construction, although it occurs far less frequently than that of literal vision or general observation involving vision: the presence of an internal emotive state. Here, the speaker/writer indicates that they base their knowledge on some sort of internal compulsion rather than on external visual perception: (53)

.. .undje lebendiger irgend ein Wissen in uns wird, desto mehr sehen wir uns getrieben, es in seinem Zusammenhange aufwarts und abwarts zu verfolgen. (Goethe Corpus: Tag- und Jahreshefte (?),Hamburger Ausgabe, Band 10, S. 498) 'and the more lively any knowledge becomes in us, the more we see our selves compelled to pursue it upwards and downwards in its coherency.'

Clearly the compulsion of knowledge to engage in some sort of intellectual activity is not something to be perceived in the realm of vision, so here, Goethe uses sehen to indicate that his perception of this compulsion is internal (intuitive or emotive) rather than external. 5.2.4.

There is one instance in the Kant Corpus of sehen taking an adjectival complement, something generally reserved for object-oriented perception verbs. Here, the evidential meaning is similar to (53), in that emotions or intuition- rather than external vision- are indicated: (54)

Dadurch sehen wir uns in den geheimsten Beweggriinden abhangig von der Regel des allgemeinen Willens, und es entspringt daraus in der Welt aZZer denkenden Naturen eine moralische Einheit und systematische Verfassung nach bloj3 geistigen Gesetzen. (Kant Corpus: AA II, Traume eines Geistersehers (1766), S. 335) 'Thus we see ourselves in the most secret motivations dependent on the rule of the general will, and a moral unity and systematic

272

Richard Jason Whitt

constitution according to mere spiritual laws arises from this in the world of all thinking beings.' Dependency, something obviously not detectable by vision, is the perceived state here, and Kant indicates via a metaphorical use of sehen that he considers himself to be in this state. His use of the plural pronoun wir indicates that Kant believes his readers to share his views and find themselves in this state of dependency as well.

5.2. 5. When see and sehen occur in parenthetical constructions, they are never solely markers of direct visual perception. They can, however, indicate general observation, where vision may or may not be involved: (55)

(56)

They sometimes, I see, use the word captivate thus: "Five or six ships captivated, " "His whole army captivated " (ARCHER Corpus: <1819moor.j5b>) ,was aber fangen wir mit Felix an?" fragte Makarie, ,, welcher, wie ich sehe, mit der Betrachtung jener Bilder schon fertig ist und einige Ungeduld merken liij3t" (Goethe Corpus: Wilhelm Meisters Wanderjahre (1829), Hamburger Ausgabe, Band 8, S. 117) "'But what are we going to do with Felix?," asked Makarie, "who, as I see, is already fmished observing those pictures and is showing some impatience."'

In (56), it does seem that visual perception informs Makarie's observation of Felix's behavior. In (55), vision is involved as well, insofar as the writer is indicating what he as read, an act that requires visual perception. There can also be an added connotation of inference besides the indication of observation: (57)

This course we see hath been very effectual in a short time, with some more ripe witted children, but othres of a slower apprehension (as the most and best commonly are) have been thus learning a whole year together ... (Helsinki Corpus: CEEDUC3B, Charles Hoole, A New Discovery of the Old Art of Teaching Schoole (1660), p. 4)

Evidentiality, polysemy, and the verbs ofperception

273

Indeed, to deem whether a certain method of instruction as been effectual requires both an act of observation and subsequent inference. Finally, parenthetical constructions are capable of indicating internal states of lrnowledge and understanding metaphorically: (58)

Und hierin hat also, wie wir sehen, die Mathematik einen Vorzug vor der Philosophie, daj3 die Erkenntnisse der erstern intuitive, die der letztern hingegen nur discursive Erkenntnisse sind. (Kant Corpus: AA IX, Logik (1800), S. 23) 'And here mathematics has a priority over philosophy, as we see, for the insights of the former are intuitive, while those of the latter are, on the other hand, only discursive realizations.'

Here, Kant is simply putting forward the argument that mathematics has a priority over philosophy, and he employs the parenthetical construction involving sehen to indicate that this argument should be self-evident (i.e. understood) and to point out the inherent nature of both disciplines (i.e. aspects belonging to the realm oflrnowledge).

5.2.6. <EXT> When see or sehen occur external to the clause for which they indicate there is evidence, a variety of evidential meanings can be expressed, including direct visual perception: (59)

Oons, Madam, they're broke in to the House with Fire and Sword, I saw them, heard them, they'll be here this Minute. (Helsinki Corpus: CEPLAY3B, George Farquhar, The Beaux Stratagem (1707), p. 60)

A general sense of observation - where visual perception plays a prominent role- can also be indicated: (60)

An inquisitiue man is a pratler: so vpon the like reason, a credulous man is a deceiuer: as we see it in fame, that hee that will easily beleeue rumors, will as easily augment rumors, and adde somewhat to them of his owne, which (Tacitus) wisely noteth, when he sayth: (Fingunt simul creduntq); so great an affinitie hath fiction and beleefe. (Helsinki Corpus: CEEDUC2B, Francis Bacon,

274

(61)

Richard Jason Whitt

The Twoo Bookes of the Proficience and Advancement of Learning (1605), 21V) es ist, was ich zum voraus wuj3te, nun aber mit Augen sah: er hat eben immer gemacht, was andere zu machen wilnschten, und ich mochte jetzt nichts darilber sagen, als daj3 es von ihm ist. (Goethe Corpus: ltalie- nische Reise (1786-1788), Hamburger Ausgabe, Band 11, S. 103) 'It is what I already lmew but now saw with my own eyes: he just always did what others wished to do, and because it is from him, I wouldn't like to say anything about that now.'

Visual perception is clearly the basis for observation in (61), and it can be argued to be present in (60) as well: whether fame proves that "an inquisitiue man is a pratler" or a "credulous man is a deceiuer" can only be determined from the acts that these men engage in, acts that are perceived primarily through vision. An additional reading of inference can be included when observation is indicated: (62)

(63)

I expect every hour the definitive sentence that must separate us for ever - I am sure you feel for me - I see you do - pardon these tears! (ARCHER Corpus: <1764walp.f4b>) ,Die ist gesund und munter. - "Ja, das sah ich, als du vorbeirudertest. (DWDS: Ottomar Enking, "Die Damekower", in: Berliner Tageblatt (Morgen-Ausgabe) (08.02.1906), S. 5) "'She is healthy and happy." Yes, I saw that when you paddled by.'

In both (62) and (63), it should be clear that visual perception forms the basis of the speaker's inference: face-to-face contact in (62), and more general observation of personal demeanor in (63).

5.2. 7. Summary of see and sehen We have seen that the subject-oriented verbs of visual perception in English and German can express a remarkable variety of evidential meaning ranging from literal vision to metaphorical extensions of lmowledge, understanding, and even internal emotion. And these meanings tend to be bound to specific construction types. In both constructions involving complementizer clauses ( and ), for example, a

Evidentiality, polysemy, and the verbs ofperception

275

great variety of evidential meaning involving both direct and indirect evidence was found: literal perception, general observation, inference based on observation, knowledge, and understanding. Direct vision was almost the sole type of evidential meaning expressed in the constructions, on the other hand. And in the lone construction found in the Kant Corpus, internal states of intuition and/or emotion were indicated. General observation (where literal vision might play a role) and metaphorical extensions of knowledge and understanding were indicated in parenthetical constructions, while literal vision always seemed to play a role in external constructions, whether a single act was witnessed or a general state of affairs was observed.

6. Conclusion

I hope to have provided a general overview of how verbs of perception in English and German can express a variety of evidential meanings, a natural consequence of their inherent polysemy. It has also been shown that certain evidential meanings are restricted to specific construction types, although some types of evidential meaning do indeed appear in multiple constructions. Only the most salient similarities between English and German have been discussed here, and a number of issues are left unaddressed: how the object-oriented perception verbs also serve as evidential markers, the consequence of coupling a modal verb or cognitive verb (think, denken, etc.) with an evidential perception verb, the effects of negation, and the role of the other four sensory modalities could not be addressed here due to considerations of space. But at the least, it should be clear how fundamental perception is in shaping speaker epistemology, and it is also worth considering the role perception plays in markers of evidentiality that are not overtly indicators of sensory perception.

Notes 1.

Support for this research was made possible by the Fulbright Program and the UC Berkeley Graduate Division. In addition, I would like to thank Gabriele Diewald, Elena Smimova, Thomas Shannon, Irmengard Rauch, and Eve Sweetser for assistance during various stages of this project. Any errors are, of course, mine alone.

276 2.

3.

4.

5.

6.

7. 8. 9.

Richard Jason Whitt All of the distinctions made in perception verb typology are based on the prototypical meaning of each perception verb and not on any possible metaphorical or metonymic extensions. The deictic nature of evidentiality is best summed up by Joseph (2003: 308): "[evidential] utterances typically include indicators pointing directly to particular sources or away from potential sources, as the speaker takes a particular point of view in describing an action." I wish to emphasize that this is a tendency of evidential perception verbs, not a hard and fast rule. I did fmd a few counter-examples to this in my corpus study, but these were far and few between, so this rule holds in virtually all cases. The GerMane (German Manchester Corpus) project attempts to fill this gap with representative data from 1650 through 1800. Currently, only excerpts of newspapers are available, but plans to expand this corpus to a large variety of text types are underway. Some, e.g. Thompson and Mulac 1991 and Brinton 1996, have suggested that the omission of the complementizer in similar constructions involving cognitive verbs (guess, think, etc.) in English is part of a grammaticalization process that results in epistemic parenthetical constructions. They base their argument mainly on semantic and pragmatic criteria of spoken data. Their arguments do not seem to hold for the verbs of perception encountered in this study, although admittedly, more data would be necessary before one could argue one way or another if their conclusions are also applicable to perception verbs. I would like to thank Maurice Vliegen for bringing this construction type to my attention. For more elaborate discussions of parenthetical constructions, see essays in Debe and Kavalova 2007. I initially intended to provide quantitative data on the particular evidential meanings found within each construction type. However, such precise analysis proved impossible to apply consistently, for there are sometimes shades of multiple meanings signified in one singular instance. Therefore, I only provide an overview of the varying evidential meanings found in each construction type.

References Aikhenvald, Alexandra 2004 Evidentiality. Oxford: Oxford University Press. Brinton, Laurel J. 1996 Pragmatic Markers in English: Grammaticalization and Discourse Functions. Berlin/New York: Mouton de Gruyter.

Evidentiality, polysemy, and the verbs ofperception

277

Chafe, Wallace 1986 Evidentiality in English conversation and academic vvriting. In Evidentiality: The Linguistic Coding of Epistemology, Wallace Chafe and Johanna Nichols (eds.), 261-272. Norwood, NJ: Ablex Publ. Debe, Nicole, and Yordanka Kavalova (eds.) 2007 Parentheticals. Amsterdam/Philadelphia: John Benjamins. Diewald, Gabriele 2004 Faktizitat und Evidentialitat: Semantische Differenzierung bei den Modal- und Modalitatsverben im Deutschen In Tempus!Temporalitiit und Modus!Modalitiit im Sprachvergleich, Oddleif Leirbukt (ed.), 231-258. Tiibingen: Stauffenburg. Does, Jaap van der, and Michiel van Lambalgen 2000 A logic ofvision. Linguistics and Philosophy: 1-92. Gisborne, Nikolas 1998 The attributary structure, evidential meaning, and the semantics of English SOUND-class verbs. University College of London Working Papers in Linguistics 10: 1-25. Gisborne, Nikolas, and Jasper Holmes 2007 A history of English evidential verbs of appearance. English Language and Linguistics 11 (1): 1-29. Haan, Ferdinand de forth. Visual evidentiality and its origins. Diachronica. Harm, Volker 2000 Regularitiiten des semantischen Wandels bei Wahrnehmungsverben des Deutschen. Stuttgart: Steiner. Jakobson, Roman 1971[1957] Shifters, Verbal Categories, and the Russian Verb. Roman Jakobson: Selected Writings. Vol. II: Word and Language. The Hague: Mouton de Gruyter, 130-147. Joseph, Brian D. 2003 Evidentials: Summation, questions, prospects. In Studies in Evidentiality. Alexandra Aikhenvald and R. M. W. Dixon (eds.), 307-327. Amsterdam/Philadelphia: John Benjamins. Lakoff, George, and Mark Johnson 2003 Metaphors We Live By. With a new Afterword. Chicago/London: The University of Chicago Press. Mortelmans, Tanja 2000 On the 'evidential' nature of the 'epistemic' use ofthe German modals mussen and sol/en. Belgian Journal ofLinguistics 14: 131-148. Peirce, Charles Sanders 1998 Essential Peirce: Selected Philosophical Writings. Vol. 2 (18931913). Eds. Nathan Houser and Christian J. W. Kloesel. Bloomington: Indiana University Press.

278

Richard Jason Whitt

Schroder, Jochen 1995 Zur Verben der visuellen Wahrnehmung (Kerngruppe). In Deutsch als Fremdsprache: An den Quellen eines Faches. Festschrift fiir Gerhard Helbig zum 65. Geburstag, Heidrun Popp (ed.), 317-325. Miinchen: ludicum. Smirnova, Elena Die Entwicklung der Konstruktion wiirde + lnfinitiv im Deutschen: 2006 Eine funktional-semantische Analyse unter besonderer Berucksichtigung sprachhistorischer Aspekte. Berlin/New York: de Gruyter. Sweetser, Eve From Etymology to Pragmatics: Metaphorical and Cultural Aspects 1990 of Semantic Structure. Cambridge: Cambridge University Press. Thompson, Sandra A., and Anthony Mulac 1991 A quantiative perspective on the grammaticization of epistemic parentheticals in English. In Approaches to Grammaticalization. Vol. II: Focus on Types of Grammatical Markers, Elizabeth Closs Traugott and Bernd Heine (eds.), 313-329. Amsterdam/Philadelphia: John Benjamins. Traugott, Elizabeth Closs On the rise of epistemic meaning in English: An example of subjecti1989 fication in semantic change. Language 65: 31-55. 1997 Subjectification and the development of epistemic meaning: The case of promise and threaten. In Modality in Germanic Languages: Historical and Comparative Perspectives, Toril Swan and Olaf Jansen W estvik, 185-21 0. Berlin!New York: de Gruyter. Viberg, Ak:e 1983 The verbs of perception: A typological study. Linguistics 21: 123162. Whitt, Richard J. 2008 Evidentiality and perception verbs in English and German: A corpusbased analysis from the Early Modern period to the present. Ph.D. diss., Department of German, The University of California at Berkeley.

Evidential markers in French scientific writing: the case of the French verb voir Francis Grossmann and Agnes Tutin The present paper describes a study of the evidential functions of the French verb voir ('see') in the scientific genre, based on a corpus compiled from writings in the fields of linguistics and economics. Our objective was to answer two questions: (a) What are the main functions of this marker in this genre and how are these functions made manifest (meaning and use of voir)? (b) Do different fields have specific ways of using evidential markers? A brief presentation ofthe evidential functions of the lexicon of perception is followed by a review of the literature on the evidential meaning of voir. In order to explore this linguistic phenomenon, we compiled a large corpus of scientific writings and annotated the occurrences of voir. This approach enabled us to address the linguistic properties of voir as a statement marker and to analyze its other evidential function, that of a reference marker. We also compared the evidential functions of voir in linguistics and economics writings using a corpus drawn from these two fields.

1.

Introduction

In evidential typologies, visual evidence 1 is considered to be one of the prime sources of information. One somewhat paradoxical result of this is that certain languages with evidential systems do not have linguistic markers to indicate visual access to information - the default mode of access whereas modes such as inference or hearsay are marked linguistically. Aikhenvald 2004 found that in languages such as French and English that do not possess evidential markers in the strict sense of the word, the notions of evidentiality are sometimes used abusively; that is to say, evidential markers are grammaticalized. Excluding the specific lexicon used in evidentiality research, she did however recognize evidential strategies that are used to add evidential semantic extensions to linguistic tools with other main functions (modal markers, reported speech markers, etc.). This restrictive vision can be nuanced by arguing that, in the same way that it is useful to study the lexical methods some languages use to express aspect (even if only to observe their interactions with the values of inflectional

280

Francis Grossmann and Agnes Tutin

classes), it may also be interesting to study how these languages mark evidentiality, including through lexical means; especially as languages with evidential systems have often derived these grammatical tools from lexical markers, such as the verbs of perception. A tendency to grammaticalize lexical markers can be seen in the French language's lexicon of visual perception, particularly with respect to the use of vu 'seen' As early as the Middle Ages, vu was being used as a preposition with the meaning of eu egard a 'considering'' and, in the 17th century, vu began to be used in legal documents to signify apres avoir examine 'after having examined'. In addition, the conjunctive locution vu que 'given' appeared in the 14th century and it is still used today in some areas. It is true that this tendency to grammaticalize certain uses derived from voir 'see' 2 does not necessarily mean these uses will become part of an evidential system, as this grammaticalization is limited, for example, when voir is employed as an auxiliary to partially erase its visual aspect. Our working hypothesis is that in certain genres in French the evidential characteristic of voir is usually expressed discursively. We postulate that is particularly true in the sciences, a genre that employs very specific forms in certain uses. As a result, we felt it would be interesting to examine differences in the discursive use of evidential markers in two fields: economics and linguistics. Finally, we wanted to show how difficult it is, in certain types of written discourse, to clearly differentiate between markers used to indicate visual sources, inferences or intellectual arguments based on statements made by other parties. After briefly explaining how voir functions in French and the framework for our study (Section 2), we present the method used to analyze our corpus (Section 3) and describe the two types of evidential usage of voir (Sections 3 and 4). We conclude by contrasting the distribution of these two types of usage in the fields of economics and linguistics.

2.

Voir and the lexicon of perception in scientific genres

2.1. Perceptual paradigm Voir is used in a wide variety of forms in scientific writings. Table 1 lists some of these lexical forms. The lexical field of voir can be divided into "subjective" verbs of perception (the agent is the subject and the object "seen" is the direct object) and "objective" verbs for which the subject is the object "seen" (see Whitt 2008). However, this field also contains other syntactic categories, such as adverbs or adverbial expressions, e.g. appa-

Evidential markers in French scientific writing

281

remment 'apparently' and a premiere vue 'at first sight', prepositional forms, e.g. au vu de ces resultats 'given these results', nouns, e.g. observation 'observation', coup d'oeil 'quick look' and adjectives, e.g. visible 'visible', observable 'observable' Linguistic expressions within the lexical field of visual perception often have a hedge function, with certain verbs, e.g. apparaitre 'appear' and adverbials, e.g. visiblement 'visibly', apparemment 'apparently' being used by authors to dilute their responsibility (N0lke 1994). However, these verbs and adverbials can also have a direct evidential function (Dendale 1994). As we will see below, this hedging role does not concern the verb voir or other expressions that have a more directly evidential function, such as au vu de 'given' and a Ia lumiere de 'in the light of The role of these lexical elements is more to show that a "fact speaks for itself'; even if the access to this factual knowledge is rarely through purely visible perception. Table 1 The lexical field of visual perception in scientific writings "Subjective" verbs voir 'see' observer 'observe' apercevoir, s 'apercevoir 'perceive' regarder 'look (at)' discerner, de teeter 'discern, detect' examiner 'examine' reperer 'recognize'

"Objective" verbs apparaitre/ lafigure apparait 'appear/the figure appears ... ' sembler 'seem' reveler 'show' devoiler 'reveal' se revele Adj.) 'reveal' Adj

Adverbs visiblement 'visibly' apparemment 'apparently' a premiere vue 'at first sight' en apparenee 'apparently'

Prepositions au vu de (de ces resultats, etc.) 'given these results' ala lumiere de 'in the light of

Nouns

Adjectives

observation (les premieres observations mantrent que) 'observation the first observations show' coup d'reil 'glimpse'

visible 'visible' discernable 'discernable' apparent 'apparent' observable 'observable'

We should also mention the fact that certain lexemes in the field of voir are used to contrast appearance and reality, or, more exactly, to contrast a spontaneous but erroneous visual analysis and a deeper analysis that is contrary to appearances. A premiere vue 'at first sight' and apparemment 'apparently' are often accompanied by an adjective such as surprenant 'surprising', contre-intuitif 'counter-intuitive' or paradoxa! 'paradoxical',

282

Francis Grossmann and Agnes Tutin

as in ( 1). Here, the analysis favoured by the author is introduced by en fait 'in fact' 3 : (1)

En revanche, /'influence du revenu du conjoint est conforme a celle qui ressort de travaux analogues, mais a premiere vue plus surprenante. On s'attendrait a ce qu'un revenu du conjoint plus eleve decourage l 'activite du fait de la progressivite du systeme fiscal: les prelevements sur un meme salaire jeminin sont d' autant plus eleves que le salaire du mari est consequent, puisque la tranche marginale d'imposition est plus grande, ce qui n 'est pas le cas sur le tableau 1. En fait, le profil est perturbe par la forte homogamie sociale qui conduit les femmes diplomees a vivre avec des conjoints, eux aussi diplomes. 'On the other hand, the influence of the partner's revenue is similar to that found in other studies but, at frrst sight, more surprising. One would expect a higher partner revenue to discourage the activity because of the progressiveness of the tax system: the amount of tax paid on the wife's salary is higher when the husbands salary is high, because the marginal tax burden is larger, which is not the case in table 1. In fact, the profile is skewed by the strength of social homogamy, which means that highly qualified women tend to live with highly qualified partners' (K.IAP economics corpus)

2.2. Linguistic functioning of the verb voir In his Dictionnaire historique de la langue fran9aise, Alain Rey (2004: 4107) notes that the Latin verb videre, from which voir is derived, comes from the Indo-European root 0 weid- which indicates vision, although primarily with respect to knowledge and only secondarily with respect to perception by sight. The Latin verb, like its French successor, whether it is used intransitively or with an accusative, signifies both perceived by sight, witness and notice4 This co-existence of the perceptive and intellectual meanings of voir is the source of its uses within the scientific genre and are the subject of the present study. Numerous studies of the different uses of voir have been carried out, either as part of more general studies of verbs of perception, or separately (see e.g. Bat-Zeev Shyldkrot, 1981, 1984, 1989, 1997, Picoche, 1986,

Evidential markers in French scientific writing

283

1993, Willems, 1983, 2000a, 2000b, Franckel and Lebaud, 1990, Labelle 1996, Naukkarinen, 1997, Leeman and Sakhokia Giraud 2001, Grezka, 2006). These studies have used different theoretical frameworks to investigate the principles explaining the numerous uses of voir or the semantic variety of these uses. Franckel and Lebaud 1990, who applied Culioli's approach, showed the importance of the speaker's stance in interpreting voir, a point we feel is particularly important. Grezka (2006: 61), who took an object-class approach, highlighted the importance of the second argument for voir, as the first argument is always an animate NP Discussions have frequently looked at dichotomies, for example, between direct perception (="physiological") and indirect perception (="cognitive"), or between active perception (the perceptive act is controlled by the subject) and passive perception (perception is an experience that the subject undergoesi The roles of the co-text and of complements have been invoked to explain the passage from a passive meaning to an active meaning for prototypical ''passive" verbs of perception, such as voir and entendre. Grezka (2006: 53) noted that the co-text may govern the selection of either a passive verb of perception (voir, entendre) or an active verb of perception (regarder, ecouter). For example, in order to specify an object that demands intellectual participation, an active form of voir is selected, as in As-tu vu un film hier soir 'Did you watch a film yesterday evening?' For many researchers, (e.g. Le Goffic 1993: 250, quoted in Grezca, 2006), when voir introduces a complement clause it takes on a more cognitive meaning and signals access to knowledge. This proposition is irrefutable, although it should be realized that in many cases the cognitive dimension of voir does not entirely override its perceptual dimension. In fact, access to a piece of information is often provided by a visual clue:

(2)

Mais en regagnant sa chambre, elle a vu que la porte du balcon etait entrouverte. 'But, when she got back to her room she saw that the balcony door was ajar' (B. Friot)

This example clearly shows that when voir acquires an evidential value (i.e. when it is used to indicate an observed fact), this always implies a double predicative dimension and therefore two processes: the process expressed by the subject of voir introduced by the observer and the process that translates the observed fact, which can be carried out syntactically in different ways. Without this double structure, voir would be a simple perceptual marker and would not have any evidential value. A sentence such

284

Francis Grossmann and Agnes Tutin

asje vois un arbre 'I see a tree appears' to be a perceptual statement of fact and not a way of indicating the source or status of the information. This explains why, syntactically, the use of an evidential often involves either completive or infinitive propositions that can be regulated by voir, or other means of syntactically signalling a predication. The literature shows that an infinitive proposition favours a perceptual interpretation, whereas a completive construction leads to a more cognitive interpretation. This can be illustrated by contrasting the following two sentences: (3)

Je vois Paul partir vs Je vois que Paul part. 'I see Paul leave vs I see that Paul is leaving'

Both cases contain a double predicative structure: the structure associated with the observer who "sees", and the structure that corresponds to the observed fact, but the completive construction, which more strongly reinforces the two predications, seems to emphasize the observer's point of view. Hence, the completive construction appears to be a better candidate for use as an evidential marker. However, other structures can also be used. Thus, in a sentence such as:

(4)

On voit ici l'importance de R. 'We see here the importance of R. appliquee, 1962, val. X, 1)

(Revue de statistique

the evidential structure is expressed using an abstract predicative noun, rather than a completive. Our corpus study identified other syntactic constructions with evidential value in scientific writings.

2.3. The framework used to study voir in scientific genres Although our study is based on the above facts, our approach was much more specific, as our aim was not to explain all the uses of voir, or its polysemy; rather, it was to examine the uses of voir in scientific writings and to try and identify criteria that can be used to identify the evidential role of these uses. We postulate that voir, when it forms part of an evidential structure in a scientific text, can be subject to three independent levels of analysis: lexical-semantic, enunciative and rhetorical-pragmatic. Our lexical-semantic analysis used a similar approach to Fillmore's frame semantics. We believe that the evidential use of voir differs from

Evidential markers in French scientific writing

285

other uses in the nature of the frame elements involved in the valence of voir. These roles can be identified through the syntactic constructions employed and through the semantic characterisation of the arguments of the verb 6 The enunciative level constitutes the polyphonic space in which the scientific demonstration occurs and indicates how the semantic frames are manifested in the enunciative interplay, most notably through the personal pronouns7 There is a complex relationship between the enunciative interplay and the lexical semantic roles linked to the frame implied by the verb. As we will see with respect to voir, the figure of the scientific author8 is represented, in principle, by the personal pronouns Nous, On, and sometimes Je, which present the demonstration to a Model Reader. Although generally implicit, this Model Reader can appear in certain injunctive forms, for example, through the use of on (e.g. pour une synthese, on se report era a... ' 'fior a summary, see .... ') Lastly, the rhetorical-pragmatic level takes into account how the lexicalsemantic level is inserted into the text. We postulate that the evidential use of voir involves cognitive/rhetorical operations linked to scientific argumentation and that it frequently introduces a process of validation, thereby entering the proof system of the scientific genre. As Aikhenvald 2004 pointed out, there is no direct relationship between evidentiality and the proof or validation system. The act of specifying the source of a piece of information (evidential role) is not necessarily intended to guarantee the validity of that information. However, in certain genres of discourse, specifying the source represents a validating factor that is used argumentatively by the locutor. This is the case in scientific writing, which demands that the nature of the sources of information be specified whenever new information is added 9 Specifying both the source of a piece of information and the way that information was obtained (personal work by the author, reference to work by peers) is an important element in validating research work.

3. Methodology and corpus 3.1. The corpus Our study of voir was based on a corpus of scientific writings in French, taken from articles in respected journals, PhD theses, reports and course books, in two fields within the humanities: linguistics and economics. The journal articles were taken from the KIAP corpus, which was compiled by

286

Francis Grossmann and Agnes Tutin

Kjersti Fl0ttum and his team (see Fltattum et al. 2006) for a detailed study of auctorial presence. The items from the KIAP corpus were supplemented by items from PhD theses, reports and course books. Table 2 outlines the composition of the corpus, which contains just over 1.3 million words. Table 2. Corpus of scientific writings used in the study of the verb voir

Scientific articles (corpus KIAP) Theses, reports, course books Total

Linguistics 285 881 words

Economics 374 516 words

364 812 words 650 693 words

286 653 words 661 169 words

The corpus used was sufficiently large and varied for our purposes, which were to identifY the main semantic and pragmatic functions of the verb voir, and to determine its syntactic characteristics.

3.2. Productivity of the verb voir

The verb voir is one of the most frequently occurring verbs in our corpus of scientific writings. Table 3. The most frequently occurring verbs in the corpus of scientific writings

Verb Etre 'be' Avoir 'have' Pouvoir 'be able to' Faire 'do, make' Permettre 'allow' Devoir 'have to' Mettre 'put' Considerer 'consider' Voir 'see' Montrer 'show' Apparaftre 'appear' Prendre 'take' Donner 'give' Utiliser 'use' Correspondre 'correspond'

Economics 14709 4494 2557 783 1100 732 608 656 519 637 286 426 426 541 389

Linguistics 15518 4703 3573 1400 864 638 680 618 734 379 706 556 536 414 556

Total 30227 9197 6130 2183 1964 1370 1288 1274 1253 1016 992 982 962 955 945

Evidential markers in French scientific writing

287

As can be seen from Table 3, for economics and linguistics combined, voir is the ninth most frequently used verb and the third "full" verb (after permettre 'allow' and considerer 'consider'), if auxiliary verbs, supporting verbs and modal verbs are excluded. In linguistics, it is the second "full" verb, whereas in economics it occupies fifth place, behind permettre 'allow' considerer 'consider' montrer 'show' and utiliser 'use' Voir can therefore be considered representative of scientific writings. However, as we saw above, voir has a wide variety of uses that need to be described in more detail. The corpus was annotated for this purpose.

3.3. Annotating the different uses of voir in the corpus As was noted in Section 2, voir is a polysemic verb that is not easy to analyze and that has resisted even the best efforts to identify a common semantic core for all its different meanings (see Picoche 1986, 1993; Franckel and Lebaud 1990). For the present study we adopted a pragmatic approach, as our main objective was to isolate the evidential uses of voir. This led us to identify the main senses, which represent only a subset of the uses described in reference works (for example, in the Petit Robert dictionary or in the Tresor de Ia Langue Fram;aise ). As far as possible, we associated the main senses with precise argumental structures, in order to facilitate their analysis and to annotate them in the corpus. Five main senses were identified in the corpus and the different occurrences of these senses in the corpus were labelled using XML annotation. Voir as a statement marker, evidential, with prototypical syntactic structure: NP (hum) voit (that Sentence) I NP (abs) This use, which is very common in scientific writings, can be considered evidential. The origin of a fact, here introduced in the completive ('visage' a unfort degre d'affinite avec Ia majorite de ces cliques "face' has a high degree of affinity with the majority of these cliques', becomes obvious to the author and the reader (on) through simple observation, often of a visual support (e.g. a table).

(5)

On peut voir dans ce tableau que visage a un fort degre d'affinite avec Ia majorite de ces cliques. 'The table shows that 'face' has a high degree of affinity with the majority of these cliques'). (Linguistics corpus).

288

Francis Grossmann and Agnes Tutin

This sense is similar to the one described by Whitt 2008 as, 'an inference based on observation' We have also used this label for slightly different senses, including the one labelled 'lmowledge and understanding' by Whitt. In most cases, in a corpus, it is extremely difficult to separate these two uses 10 Voir as a reference marker, evidential, with syntactic structure: Sentence (voir NP or voir Adv). Another important function of voir- which we believe is specific to scientific writings- is its use to indicate intratextual or intertextual references. In this context, voir, used in the infinitive, is often placed in parentheses and followed by one or more intra- or intertextual references. For example: (6)

(7)

Les auteurs qui traitent ( ..) d'O. travers et au travers de abordent systematiquement ces prepositions a partir de leurs emplois spatiaux, de nature intrinsequement dynamique (voir, par exemple, Borillo 1998: 85, Spang-Hanssen 1963: 231-233) 'Authors who study "a travers" and "au travers de" systematically approach these propositions from a spatial perspective, which is intrinsically dynamic (see, for example, Borillo 1998: 85, SpangHanssen 1963: 231-233') (Linguistics corpus). On verifie, aisement, que lafonction valeur V(xt-1, Rt) est solution de /'equation de Bellman (voir annexe E) 'It is easy to check that the value function V(xt-1, Rt) is the solution to the Bellman equation ~ appendix E)' (Economics corpus)

Section 3 presents a detailed analysis of how these two evidential values of voir function. Voir in the sense of 'to examine' (non evidential) with syntactic structure: NP (hum) voir NP (-hum). This sense, which is quite common in scientific writings, is not an evidential use, as the following examples show:

(8)

Voyons les semes que le contexte du programme des Verts nous permet d'extraire a propos de ces deux courants synchroniques dans nos pays industrialises 'Let us look at the semantic features that the context of the Green Party's program allows us to extract with respect to these two synchronous currents in our industrial countries' (Linguistics corpus)

Evidential markers in French scientific writing

(9)

289

Notre objectif est maintenant de determiner les prix d'equilibre. Nous allons proceder en deux temps. Nous verrons tout d'abord l 'equilibre instantane 'Our objective now is to determine equilibrium prices. We will do this in two stages. First we will look at the instantaneous equilibrium.' (Economics corpus)

The element introduced by voir, an SN, is not a fact. It is often an object analyzed by the author during the demonstration or the study. Voir as an opinion and judgement verb (non evidential) with syntactic structures: NP(hurn) voir NP(abs) as NP(abs) or NP(hurn) voir NP(abs) PrepLoc NP(abs). In these senses (see examples below), voir can be considered a verb of opinion or judgement, and it does not have an evidential use. (10)

(11)

Les cotisations du systeme par repartition sont souvent vues comme des prelevements obligatoires sans contrepartie 'Contributions paid into a redistribution system are often seen as obligatory deductions that do not give any return'). (Economics corpus). ... l'on voit dans l'enonciation une sorte de prolongement du structuralisme ... ' ... we can see in the utterance a sort of extension of structuralism' (Linguistics corpus)

Other non-evidential uses were found in our corpus, but much less frequently. These include uses that are totally specific to the verb voir in its role as a semi-auxiliary of the passive, as in the following example: (12)

La theorie des operations enonciatives se voit intimement associee a une definition du langage "comme une activite qui modifie une situation en faisant reconnaitre a autrui une intention pragmatique" 'The theory of enunciative operations !§ intimately associated with a definition of language "as an activity that modifies a situation by acknowledging the pragmatic intention of a co-locutor"' (Linguistics corpus)

290

F rancis Grossmann and Agnes Tu tin

It is interesting to note that the corpus did not contain any examples of direct perceptual usage. Chafe (1 986) recorded a similar fmding for English.

3.4. Distribution of the uses of voir in the corpus By identifying values and annotating the corpus it was possible to determine the most common uses of voir in our corpus. A s F igure 1 shows, evidential uses account for an overwhelming majority of occurrences (75%). 600

Statement marker

Refer ence marker

Other

Figure 1. Distribution of the different uses of voir in the corpus of scientific writings (number of occurrences)

In addition, statement and reference mark ers show interesting syntactic and enunciative specificities that we will examine in the following two sections.

4.

Voir as a statement marker

In these uses, the origin of the stated fact is not directly perceptual because the author is presenting a proposition. As in the perceptual sense, the syntactic subject can be considered non-active as it receives the information, but does not really "create" it. Thus voir differs from verbs such as deduire 'deduce' or conclure 'con clude', which bring into play more complex mental operations. However, in some examples, voir is not only constative; it also includes a notion of demonstration. This is notably the case when the verb is used in the present perfect tense, as in the following example:

Evidential markers in French scientific writing

(13)

291

On a vu que couper correspond a des proces tres differents dans ses differents emplois lexicaux. 'We have seen that, in its different lexical uses, to cut refers to very different processes.' (Linguistics corpus)

It is difficult to unequivocally interpret this example out of context, but it is probable that the author has established this fact, rather than merely stating it in an earlier section of the text describing several lexical uses of the verb couper 'cut' The use of voir, rather than montrer 'show' or etablir 'establish', gives the text greater objectivity, in the true sense of the word: the facts are self evident and the responsibility of the author - which would appear greater if he/she had used terms such as on a montre 'we have shown', or on a etabli 'we have established'- is pushed to the background. Moreover, in scientific writings, voir as a statement marker occurs in specific syntactic contexts and has specific enunciative properties.

4.1. The subject of voir as a statement marker: inclusion of the "reader as witness" The subject of voir is nearly always on ('one') or nous ('we') (90.5% of occurrences in our corpus 11), as in the following example: (14)

Comme nous l'avons vu precedemment, la definition et le choix d'un concept pour l'equite posent de redoutables problemes 'As we saw earlier, the definition and the choice of a concept for equity is extremely problematical.' (Economics corpus)

The personal pronouns nous and on are not used merely out of protocol. They cannot be replaced by a first person singular pronoun, as they can in the case of other verbs, such as montrer or demontrer: (15)

(16)

*J'ai vu precedemment que couper correspond a des proces tres differents 'I saw earlier that to cut corresponds to very different processes' J'ai demontrelmontre precedemment que couper correspond a des proces tres differents... 'I showed earlier that to cut corresponds to very different processes ...

292

Francis Grossmann and Agnes Tutin

The fact that it is impossible to replace nous/on by je shows that the author is not the originator of the fact described (the je is not an active agent), but that the facts are, to a certain extent, self-evident. Furthermore, nous and on in these cases are not exclusive, as they include the community of peers: (17)

On admet couramment en epistemologie qu'il n 'existe pas de chose en soi que le scientifique pourrait apprehender sur le mode de /'observation pure. 'It is widely accepted in epistemology that there are no things in and of themselves that a scientist can understand through pure observation.' (Economics corpus)

In fact, when nous and on are used with voir to form a statement marker, they are inclusive (J + You) in that they include the reader, who is called upon to witness, alongside the author, the facts described 12 Here, voir, like other verbs of the same paradigm, is clearly used in a dialogical context (as it is when it is used as a reference marker), where the author draws the reader's attention to a group of facts that support the author's reasoning. The inclusion of the reader in verbal forms is quite rare in scientific writings, with the exception of certain imperatives 13 , as indications that the reader is being taken into account generally take the form of markers that guide the reader, such as metatextual elements d'une part 'on the one hand' ,pour conclure 'to conclude', etc.

4.2. Voir as a parenthetical Another specific characteristic of voir as a statement marker is its use as a parenthetical (on l'a vu 'we have seen', on le verra 'we will see', comme nous le voyons 'as we can see', etc), as in the following examples:

(18)

(19)

Les metaphores vives, on l'a vu, sont sensibles au contexte, et il est necessaire de connaftre les conditions d'enonciation pour pouvoir attribuer un sens precis al'enonce. 'Living metaphors, as we have seen, are context dependent, and it is necessary to know the conditions of enunciation in order to be able to attribute a meaning to the utterance.') (Linguistics corpus) Le facteur individuel incorpore en e./Jet les orientations du couple en matiere de descendance et, comme on l' a vu, celles-ci se

Evidential markers in French scientific writing

293

traduisent probablement dans les choix du menage en matiere d'habitation ou de biens durables 'The individual factor incorporates the couple's choices in terms of having children and, as we have seen, these choices are probably reflected in their choices in terms of accommodation or consumer durables ... ' (Economics corpus) These uses account for more than 20% of the occurrences in our corpus, and seem to be characteristic of scientific writings. As Rooryck (2001) showed, parentheticals are used particularly frequently with verbs expressing evidentiality. On a communicative level, the secondary predication expressed by voir reinforces its evidential function: the author provides a reminder that the facts expressed by the primary predication are justified by an observation that is made in association with the reader. In addition, the use of parentheticals helps to maintain textual coherence and the complicity with the reader by announcing information to come or by recapitulating facts that have already been demonstrated and observed. These elements help guide the reader, alongside other intertextual marks that we will look at now.

4.3. Supporting a statement: locative complements Although voir is not used as a verb of perception in scientific writings, the visual dimension is not completely absent. The author uses the text, sometimes supported by graphs and tables, to ask the reader to accept certain facts. Textual support may be implicit, in that it does not include any specific markers, but it is often given explicitly through intertextuallocative complements, such as adverbs of place ici 'here', plus haut 'above', cidessous 'below', supra 'above', infra 'below', etc. or prepositional phrases that indicate tables, graphs or other parts of the text to be examined (dans Ia section 1 'in section 1',dans le tableau 2 'in table 2', etc.). For example: (20)

(21)

On peut voir dans le tableau precedent que ce taux est du essentielle-ment aIa contribution des cliques ... 'The above table shows that this rate is mostly due to the contribution of the cliques ... ' (Linguistics corpus) Comme nous l'avons vu dans le chapitre 2, Ia proximite geographique releve de Ia localisation dans l'espace des acteurs et de leurs liens en termes de distance

294

Francis Grossmann and Agnes Tutin

'As we saw in chapter 2, geographical proxmnty decribes the localization of the parties in space and their links in terms of distance ... ' (Economics corpus) In a small number of cases the localization does not appear in a complementary clause, but in a parenthetical structure, which is a point of similarity between this use of voir and its use as a reference marker: (22)

Avant le XVle siecle, on l'a vu (chapitre I, section 2.1.), le terme d'antonomase recouvre principalement l'antonomase du nom commun. 'Before the 16th century, as we have seen (chapter I, section 2.1.), the term antonomasia mostly covered the antonomasia of the common noun ... ' (Linguistics corpus)

Like parentheticals, intratextual locative markers improve textual coherence and allow the author to guide the reader through the demonstration. These elements help support the argument by indicating the place of observation.

4.4. Summary In scientific writings, voir (and its paradigm) as a statement marker can be described using Fillmore's frames (2003). Table 4: Different uses of voir as a statement marker

Examples On voit ainsi sur Ia Figure 4 qu 'un changement de seuil peut faire basculer une configuration d'ambigui''fe 'Thus, we can see from Figure 4 that a change in threshold can clarify an ambiguous confisuration' Les metaphores vives, on l' a vu, sont sensibles au contexte, ...

Witness On

Fact un changement de seuil peut faire baseuler une configuration d'ambigurte 'a change in threshold can clarify an ambiguous configuration'

Localization sur Ia Figure 4

On

Les metaphores vives sont sensibles au contexte, ...

(not realized - the piece of scientific

Evidential markers in French scientific writing 'Living metaphors, as we have seen, are dependent on context. ..

'Living metaphors ... are dependent on context. ..

295

writing by default)

Voir brings into play the followingframe elements:

A witness: this element is generally realized using an inclusive subject pronoun (nous or on) that encompasses the author(s) and the readers. A fact: prototypically, the stated fact is expressed by a complement clause or an NP consisting of an abstract noun, but it can also appear in a principal in the case of parentheticals. A localization: Implicitly, the localization of the statement is the piece of scientific writing. Sometimes, the localization is expressed explicitly using an intratextual locative (section, chapter, figure, linguistic example). In the following section, we show that this structure can also be applied to the use of voir as a reference marker.

5.

Voir as a reference marker

The second use of voir in scientific writings is as a reference marker14 , an aspect that might seem out of place in a discussion of evidentiality. In fact, voir is commonly used in the infinitive as a synonym for cf (Latin confer = compare) 15 to direct the reader to a source. This use, which involves a double predication, differs from the typical evidential structure (described above) in appearance only. In fact, the second predication is expressed using an imperative form of voir, as the main predication corresponds to an assertion made in a locution or a section of the text to which the reference is made. This indicating function is closely linked to the use of the infinitive with an imperative value, but voir falls into a frame that is very similar to the one described in the section on voir as a statement marker. The reference introduced by voir provides the source and validates the text preceding voir. The reader is encouraged to check, to "go and see", the content of a source that supports the author's affirmation. The source can be external (intertextual use of voir) or internal (intratextual use). Peritextual content (footnotes) and parentheticals are important but slightly different means of introducing elements of proof into the reasoning. More detailed studies are

296

Francis Grossmann and Agnes Tutin

needed to compare the uses of voir as a source indicator alongside other referencing modes. The introduction of the author-date system by scientific publications allows authors to refer to sources supporting affirmations by simply adding the relevant publication date in parentheses. However, this sometimes leads to difficulties in interpretation16 , as the reference can be regarded either as a simple reference to an affirmation made by another author, or as an element that is considered to have been accepted by the scientific community. Similar ambiguities occur with the intertextual use of voir.

5.1. Voir as an intertextual reference marker When applied intertextually, voir is most commonly used to refer to other work that supports an affirmation:

L 'accroissement des inegalites face a l'emploi constitue l'un des traits majeurs de l 'evolution du marche du travail dans les annees 80. La progression de l 'ecart relatif des niveaux de chOmage des actifs qualifies et non qualifies est une caracteristique particuliere de ce phenomene inegalitaire (.voir. par exemple, GLYN [1995} 'The rise in inequalities with respect to employment is one of the main features of the evolution of the job market in recent years. The increase in the differences between the unemployment rates for qualified and unqualified workers is an important characteristic of this phenomenon of inequality (see, for example, GLYN [1995])') (Economics corpus) (23)

Here the validation is based on the knowledge accumulated by the scientific community in a certain field. In most of its uses, voir is accompanied by modifiers (in the above example, the modifier is par exemple 'for example' that reinforce the validity of the presented fact. For example, terms such as entre autres 'among others' or par exemple indicate that the fact presented by the author is so well known to the peer community that it is unnecessary to give a detailed list of references. The adverbs notamment 'notably' and en particulier 'in particular' are used to emphasise the importance of a specific contribution by highlighting the author's expertise, which allows him/her to establish a hierarchy in the sources used. Intertextual reference markers, which indicate sources of evidence, belong to a secondary enunciative level. Two localizations account for an

Evidential markers in French scientific writing

297

overwhehning majority of occurrences: parentheticals (as for the use of voir as a statement marker) and, even more frequently, footnotes (Figure 2), although, as we will see below, there is a substantial difference in their frequencies of use in economics and linguistics writings.

Figure 2. Occurrences of voir as a reference marker

A parenthetical statement marker using voir is placed immediately after the fact it supports. The sources given, which form the localization, affirm the stated fact. When voir is used in a footnote, it operates slightly differently. Like a parenthetical statement marker, a footnote number, placed at the end of the stated fact, may refer to a note providing an intertextual reference that supports the validity of the stated fact. However, some intertextual references introduced by voir also seem to have another, not entirely evidential function, in that they provide further information about a point that has not been sufficiently explained in the main text or they introduce complementary information, as in the following example. (24)

(In the body of the text) On ne saurait considerer comme relevant

tout a fait du hasard le fait qu 'a Ia fin du dix-neuvieme siecle le philosophe Victor Egger consacre un livre a Ia parole interieure, landis que dans ses le9ons sur l'aphasie Charcot met a l'ordre du jour Ia question du langage interieur et que le roman contemporain promeut le monologue interieu,-1 'We would not consider it to be pure coincidence that at the end of the 19th century the philosopher Victor Egger devoted an entire book to internal speech, while in his lectures on aphasia Charcot

298

Francis Grossmann and Agnes Tutin

raised the question of internal language and contemporary novels promoted inner dialogue 1) (As a footnote) 1. Voir notamment a ce sujet la contribution de Christian Puech dans ce recueil 'For more on this subject, see the contribution by Christian Puech in this volume' (Linguistics corpus) Parenthetical reference markers can also be used in this way, but this is much rarer.

5.2. Voir as an intratextual reference marker When used intratextually, voir refers the reader to another part of the text: (25)

Cette distinction reprend deux des fonctions (excellence et circonstances, mais pas filiation) de l'antonomase antique (voir 1.2.1.) 'This distinction takes into account two of the functions (excellence and circumstances, but not parentage) of ancient antonomasia (see 1.2.1)' (Economics corpus)

Here, the locative source is the complement of voir. This obligatory complement is sometimes realised by a locative adverb that guides movement through the text (ci-dessous, ci-dessus, infra, supra 11), as in the example above, or by a simple paragraph, section or part number. Unlike intertextual uses, the inclusion of voir (or cf) is obligatory. Locative complements are comparable to those that appear with statement markers and, in general, these two uses appear very similar in all aspects other than their syntactic structures. Hence, the following statement marker can easily be transformed into a reference marker by making a few adjustments to the syntax. Use as a statement marker (original example): Nous avons deja vu a la fin de la Section 2 que /'interdiction de (26) pratiquer la discrimination du troisieme degre entre marches a elasticites differentes a des consequences ambigues sur le bien-etre. 'We have already seen at the end of Section 2 that the ban on carrying out third degree discrimination between markets with different degrees of elasticity has ambiguous consequences on well being'). (Economics corpus)

Evidential markers in French scientific writing

299

Use as an intratextual reference marker (manufactured example): (27) L 'interdiction de pratiquer Ia discrimination du troisieme degre entre marches a elasticites differentes a des consequences ambiguiis sur le bien-etre (voir Ia fin de Ia Section 2). 'The ban on carrying out third degree discrimination between markets with different degrees of elasticity has ambiguous consequences on well being (see end of Section 2)). When used as an intratextual marker, voir is often placed in parentheses but it is relatively rare for it to be used with footnotes, at least in economics texts. References to appendices containing documentary sources are also common: (28)

... On verifie, aisement, que Ia fonction valeur V(xt-1, Rt) est solution de !'equation de Bellman (voir annexe E. 'It is easy to check that the value function V(xt-1, Rt) is a solution for the Bellman equation (see appendix E)' (Economics corpus)

Intratextual references can also be more complex, as in the case of syntactic units such as voir notre developpement precedent sur ... 'see our previous discussion of. . . where the indicator refers to ideas developed in those sections of the text as well as to a localization.

5.3. Summary To conclude, most uses of voir as a reference marker are cognitively and rhetorically similar to its uses as a statement marker, hence, their structures can be described using the same frame. The witness (the reader), who is asked to check the validity of sources, is expressed dialogically by the use of the imperative infinitive. The fact is the statement preceding the reference indicated by voir. The localization is the intra or intertextual object introduced by the marker voir. In this case, voir is obligatory, contrary to statement markers. The following examples illustrate this structure.

300 Francis Grossmann and Agnes Tutin Table 5. Different uses of voir as a reference marker Examples

Witness

Fact

Localization

On wirifie, aisement, que la fonction valeur V(xt-1, Rt) est solution de !'equation de Bellman (voir annexe E) 'It is easy to check that the value function V(xt-1, Rt) is a solution for the Bellman equation (see appendix E)' (Economics) La progression de l'ecart relatif des niveaux de ch6mage des actifs qualifies et non qualifies est une caracteristique particuliere de ce phenomene inegalitaire (voir, par exemple, GLYN [1995]) 'The increase in the differences between the unemployment rates for qualified and unqualified workers is an important characteristic of this phenomenon of inequality (see, for example, GLYN [1995] ')

=the receiver, who appears implicitly in the infinitive imperative form As above

On verifie, aisement, que la fonction valeur V(xt-1, Rt) est solution de !'equation de Bellman ' It is easy to check that the value function V(xt-1, Rt) is a solution for the Bellman equation'

AppendixE

La progression de l'ecart relatif des niveaux de ch6mage des actifs qualifies et non qualifies est une caracteristique particuliere de ce phenomene inegalitaire 'The increase in the differences between the unemployment rates for qualified and unqualified workers is an important characteristic of this phenomenon of inequality'

GLYN [1995]

6. Comparison between economics and linguistics One of our working hypotheses was that the evidential function in our corpus involves cognitive operations that are related to scientific reasoning and textual rhetoric. As methods, types of scientific object examined and modes of reasoning differ between fields, we hypothesized that the evidential use of voir would also vary according to the field. Detailed examination of our corpora confirmed this hypothesis. First, voir occurred more frequently in our linguistics corpus (513 occ.) than in our economics corpus (420 occ.), for similar-sized corpora. Two main differences in the distributions ofthe uses of voir were also noted (Figure 3). Linguistics texts contain more statement markers, whereas economics texts contain more intertextual references. However, comparable numbers of intratextual markers are found in the two fields. These results can be

Evidential markers in French scientific writing

301

interpreted in a number of ways. The greater use of statement markers by linguists may be the result of a greater need to involve the reader in the demonstration. However, this hypothesis does not seem to be supported by the occurrence of the other markers of interaction with the reader that were identified by Fl0ttum et al. 2006 in the KIAP corpus 18 For example, metatextual markers and imperatives occurred much more frequently in the economics corpus than in the linguistics corpus. Less importantly, the frequent use of statement markers by linguists probably results from the fact that linguistics authors often support their reasoning with examples in the text, whereas observations in economics are often based on linguistic objects, in particular numerical results.

59.5

Figure 3: Evidential uses of voir in economics and in linguistics (in%)

The greater use of intertextual reference markers in economics writings suggests that economists are more likely to refer to previously established knowledge in their field, whereas linguists tend to base their statements on observations contained within their own text. It would be tempting to think that economists have a more cumulative concept of science and a greater tendency to build on the work of their peers. However, the bibliographies of economics articles tend to be shorter than those of linguistics texts, at least in the case of research articles (see Fl0ttum et al. 2006: 219). Furthermore, the use of reference markers appears to be slightly different in the two fields: linguists favour parentheticals (60% of occurrences), whereas economists tend to use footnotes (66.5%), which often have other

302

Francis Grossmann and Agnes Tutin

functions that are not purely evidential. From an academic point of view, economists may have a greater need to found their work on previous research carried out within their discourse community. This hypothesis is corroborated by a study carried out on information source markers, which are much more explicit in economics than in linguistics (Garcia da Silva 2008). Nevertheless, to avoid hasty judgments and caricatural sociological interpretations, a more detailed and qualitative analysis is needed to confirm these hypotheses. Finally, although both fields appear to use comparable numbers of intratextual markers, they generally refer to different types of object. Most of the markers used in economics articles refer to graphs and tables summarizing data (Figure 1, Table 2, Graph 1), or to other sections of the text, either specifically (Section 1, Appendix D, Note 4) or generally, using "vague" adverbs of place (supra 'above', infra 'below', plus haut 'above', ci-dessous 'below', etc). In linguistics, references to tables and graphs are rare, as linguistics articles do not tend to contain such elements. On the other hand, linguistics articles do contain references to examples within the text, although cf is more commonly used in these cases.

7. Conclusion The verb voir, analyzed here within a subset of scientific writings, is only one element, albeit a very productive one, in the evidential system used in this discursive genre. Hence, it would be valuable to integrate studies of voir, and of verbs such as observer 'observe', s'apercevoir 'notice' and examiner 'examine' into more general studies of the evidential role of markers referring to visual metaphors. Analysis of our corpus has revealed a recurring semantic structure (frame) containing three main elements: a witness, a fact and a localization. However, this structure also includes a certain amount of diversity, which can be described using three main criteria: The degree of syntactic and semantic integration between voir and the stated fact. Integration is strong in completive structures indicating an observation (on a vu que 'we have seen that', etc.). In parentheticals the degree of integration is lower. There is also a third level, where voir appears in a separated structure within parentheses, in which the degree of integration is even lower. The syntactic and scriptural dis-

Evidential markers in French scientific writing

303

tance from the fact is even greater when voir is used peritextually, e.g., in note 19 · The type of localization: implicit or explicit, intratextual or intertextual. Implicit localizations (i.e., not referring to a precise location within the same text or in a precisely identified external source) and parentheticals without explicit localizations often reinforce the rhetorical or cohesive function. Localizations are always intratextual in the case of voir as a statement marker (On voit dans Duchmoll2003 que ... 'It can be seen from Duchmoll2003 that ... '). In the case of voir as a reference marker, a localization must be given, as the localization associated with voir is an essential complement and not a modifier; The status of the witness, who may have a different role in scientific reasoning. The witness is first a Being in discourse, who is associated with the co-observation of the facts presented. The dialogical dimension is central and aims to involve the reader in the demonstration. Thus, the reader is called upon to witness, either through a statement with which s/he is associated by the inclusive personal pronoun nous, or through an explicit invitation. In this second function - apparently more modest but nevertheless fundamental - the witness seems to be a mere guarantor of the information given. So what is the exact status, in evidential terms, of the verb voir? Our data support the idea of a very strong ambivalence between the verb's perceptual meaning, which although secondary remains present, and its cognitive meaning, through which it indicates the inference. In examples such as en se reportant a la figure 3b, on voit que l'espace des phases est defini par deux variables 'looking at figure 3b, we can see that the phase space is defined by two variables' (Naukkarinen 1997: 123) voir is used as an inference marker. 'When used as an intratextual marker (as well as in numerous statement uses) voir allows an author to validate a statement by inviting the reader to look at data included in other parts of the text, thereby directing the "textual circulation. 'When used as an intertextual reference marker, voir directs the reader to supporting evidence provided by the writings of third parties, and thus indicates that the 'proof' is accepted by the relevant discourse community. In both uses, the reasoning role is inextricably entwined with the evidential function. Finally, our comparison between the fields of economics and linguistics supports our hypothesis that the traditions governing the use of evidential markers in scientific writings differ between fields, even for fields within the Humanities.

304

Francis Grossmann and Agnes Tutin

Notes 1. 2. 3. 4. 5. 6.

7.

8.

9. 10.

11.

12.

13.

Note that the terms evidence and evidentiality are derived from the Latin word videre, meaning to see. See Chocheyras 1968 on this subject. For more on the use of en fait as an evidential marker, see Grossmann and Wirth2007. Leeman and Sakhokia Giraud (2001: 59) also highlight the intellectual and perceptive values of voir in French, linking them to etymology. For a summary, see in particular, Grezka 2006. We present these roles at the end of the corpus analysis. In Fillmore's framework semantics, the predicative meaning is represented by prototypical situations or scenarios (frames), which can be divided into different levels of organization. These frames are seen as places of articulation between human experience and/or the cognitive dimension and the lexical coding. The semantic roles (or frame elements) correspond to the arguments of the predicates, verbal or nominal predicates (Fillmore et al. 2003). Fillmore 1982 identified two types of frame: cognitive frames and interactional frames. Interactional frames aim to represent the interaction between the locutor and the allocutor, or between the author and the reader. From this idea we have retained the need to bring together cognitive frames (and the elements associated with them) and the enunciative and textual perspective. In order to avoid any ambiguity, it must be remembered that the notion of author comprises several components: the author as an empirical being, the institutional author, as defined in law and, finally, the figure of the author as constructed at the enunciative and textual rhetoric levels. The latter is the component that interests us here. Chafe 1986 stressed the fact that certain well-established written gemes have, over time, codified their use of very specific forms to mark evidentiality. For example, both values are possible in the phrase on voit bien dans ce schema que 'this diagram clearly shows: recognition is combined with a deduction' In most other cases, when voir is used as a statement marker, it is used in an impersonal form: II est surprenant de voir que ce type de competence est egalement produit dans l'entreprise 'It is surprising to see that this type of skill is also produced within the company' (Economics corpus) In their study of the pronoun on in the KIAP corpus, Fl.0ttum, Dahl, and Kinn 2006 also showed that the inclusive use of on is particularly linked to verbs of perception. The reader's attention can be solicited through the use of certain first person plural imperatives, as well as through pronominal and verbal forms (see Fl.0ttum et al. 2006), as in the following examples: Enfin, examinons une derniere donnee 'Finally, let us look at one last piece of data' (Linguistics cor-

Evidential markers in French scientific writing

14. 15.

16.

17. 18.

19. 20.

305

pus); Voyons brievement quelques examples 'Let us take a brief look at a few examples' (Linguistics corpus) It is interesting to note here that the verbs in the examples also belong to the voir paradigm, even if they do not have an evidential function. The Framenet project has catalogued the equivalent usage of see in English, using a frame called Source_of_information, which is defined as: "In a text, a Source_of_information is given that provides a reader of the text with further Information relevant to the text. In this frame the author and reader are completely deprofiled, with the Source_of_information made salient" Nevertheless, this deprofiling of the enunciative roles remains relative, as the role of the jussive infinitive can be interpreted as an invitation to the reader. Although the two forms are often presented as equivalents (PhD students are generally advised to choose one of the two systems), semantic motivation, which is slightly different, can explain some nuances of usage. Cf (abbreviation of the Latin confer ('compare') can also be used to introduce other elements of the same order as those already introduced; whereas voir emphasizes more clearly the desire to document the source of an affirmation. In practice, however, the two markers are often used interchangeably. On this question, see Grossmann 2002. We did not fmd passim (9a et Ia here and there') in the corpus, which allows us, according to the TLFI, "to avoid mentioning the exact localization of passages with a bearing on a given subject when a reference is given" Our corpus contains the French economics and linguistics articles from the KIAP corpus. Of course, this does not prevent the information provided in a footnote from being capital, as, according to tradition, the text provides the arguments but the footnote provides the proof by giving the sources and a commentary that makes the text clearer. For a history of footnotes, see Grafton 1997.

References Aikhenwald, Alexandra Y. 2004 Evidentiality. Oxford: Oxford University Press. Bat-Zeev Shyldkrot, Hava 1981 A propos de la forme passive se voir +Vinf. Folia linguistica XV (34): 387--407. 1984 La concurrence entre la proposition conjonctive et voir+ proposition infinitive. The FrenchRewiew, LVIII (2): 202-215. 1989 Les verbes de perception: etude semantique. In Actes du XVIJeme Congres International de Linguistique et Philologie Romanes, 282294. Tome 4. Tubingen: Max Niemeyer Verlag.

306 Francis Grossmann and Agnes Tutin La grammaticalisation des auxiliaires; le cas de voir. Scalia 10: 205224. Chafe, Wallace 1986 Evidentiality in English conversation and academic writing. In Evidentiality: the Linguistic Coding of Epistemology, Wallace Chafe and Johanna Nichols (eds.), 261-272. Norwood-New Jersey: Ablex Publishing. Chafe, Wallace, and Johanna Nichols (eds.) 1986 Evidentiality: The linguistic Coding of Epistemology. Advances in Discourse Processes. Norwood-New Jersey: Ablex. Chocheyras, Jacques 1968 Un nouvel outil grammatical en franyais moderne: le verbe voir. Le Fram;ais Moderne 36 (3): 219-225. Dendale, Patrick 1994 Devoir episremique, marqueur modal ou evidentiel? Langue Franifaise 102: 24-40. Dendale, Patrick, and Liliane Tamsmowski (eds.) 1994 Les sources du savoir et leurs marques linguistiques. Langue Franifaise 102. Fillmore, Charles J. 1982 Frame semantics. In Linguistics in the Morning Calm, Linguistic Society ofKorea (ed), 111-137. Seoul: Hanshin Publishing Co. 1985 Frames and the semantics ofunderstanding. Quaderni di Semantica 6 (2): 222-254. Fillmore, Charles J., Chris Johnson, and Miriam R.L. Petruck 2003 Background to FrameNet. In Special Issue of the International Journal of Lexicography on FrameNet and Frame Semantics, Thierry, Fontenelle ( ed.), 235-250. Oxford: Oxford University Press. Fillmore, Charles, J., and Sue Atkins 1992 Towards a frame-based organization of the lexicon: The semantics of RIK and its neighbors. In Frames, Fields, and Contrast: New Essays in Semantics and Lexical Organization, Adrienne Lehrer, Eva Feder Kittay, Richard Lehrer (eds.), 75-102. Hillsdale: Lawrence Erlbaum Associates. Flettum, Kjersti, Trine Dahl, and Torodd Kinn (eds.) 2006 Academic Voices - across languages and disciplines. Amsterdam/ Philadelphia: John Benjamins. Francke!, Jean-Jacques, and DanielLebaud 1990 Les figures du sujet, A propos des verbes de perception, sentiment, connaissance. Paris: Ophrys. Garcia da Silva, Pedro Etudes des marques de Ia filiation dans les ecrits scientifiques. 2008 Memoire de master. Grenoble: Universire Stendhal, Grenoble 3. 1997

Evidential markers in French scientific writing

307

Grafton, Anthony 1997 The Footnote: A Curious History. Cambridge: Harvard Univ. Press. Grezka, Aude 2006 Etude du lexique de la perception: bilan et perspectives. Suvremena lingvistika ('Contemporary Linguistics') 61: 45-67. Grossmann, Francis 2002 Les modes de reference a autrui: l'exemple de la revue 'Langages' Faits de langue 19: 255-262. Grossmann, Francis, and FranyoiseWirth 2007 Marking Evidentiality in Scientific Papers: The Case of Expectation Markers. In Language and Discipline Perspectives on Academic Discourse, Kjersti Fl.0ttum (ed), 202-218. Newcastle, UK: Cambridge Scholars Publishing. Labelle, Marie 1996 Remarques sur les verbes de perception et la sous categorisation, Recherches linguistiques de Vincennes 25: 83-106. Leeman, Danielle, and Madona Sakhokia Giraud 2001 Point de vue culiolien sur le verbe voir dans "Les Verbes franyais", Languefram;aise 153:58-73. Le Goffic, Pierre 1993 Grammaire de Ia phrase fram;aise. Paris: Hachette. Naukkarinen, Oili 1997 Etude de quelques verbes de perception dufram;ais. Helsinki: Yliopistopaino. N.0lke, Henning 1994 La dilution linguistique des responsabilite. Essai de description linguistique des marqueurs evidentiels il semble que et il parait que, Languefram;aise 102: 84-94. Picoche, Jacqueline 1986 Structures semantiques du lexique fram;ais. Paris: Nathan. 1993 Didactique du vocabulairefranr;ais. Paris: Nathan. Rey, Alain 2004 Reprint. Dictionnaire historique de Ia langue franr;aise, Paris, Dictionnaires Le Robert. Original edition, Paris: Dictionnaires Le Robert, 1992. Rooryck, Johan 2001 Evidentiality. Part 1. Glot InternationalS (4): 125-133. Whitt, Richard, J. 2008 Evidentiality, polysemy, and perception verbs: a corpus-based analysis of English and German. Hand-out of the conference "The linguistic realization of evidentiality in European Languages", Bamberg, 27-28 February 2008.

An interactional approach to epistemic and evidential adverbs in Spanish conversation

Bert Cornillie

This paper deals with the discourse functions of epistemic and evidential adverbs in Spanish conversation. It is shown that these adverbs not only vary in mood and relative frequency, but also have different roles in the organization of the turntaking process. By doing so, they go beyond the traditional epistemic and evidential qualifications mentioned in the existing literature. A lo mejor and igual 'perhaps' are shown to invite the interlocutor(s) to confirm or reject the view of the state of affairs (or part of it) presented, that is, the adverbs play an important role in the turn-taking process. The epistemic adverb quiza, by contrast, lacks the interactional functions typical of a lo mejor. Finally, the analysis indicates that speakers use evidential adverbs such as evidentemente to retain the turn, i.e. to keep the word.

1.

Introduction

Epistemic and evidential adverbs are generally defined as invariable expressions that qualify (parts of) a proposition in terms of the degree of likelihood that the event expressed may be real (in the case of epistemic markers), or in terms of the reference to the information that leads to a proposition (in the case of evidentials). Although this certainly holds for the bulk of the adverbs in monologic registers such as fiction or non-fiction prose, a closer look at their use in a conversational context points to several other functions in addition to qualifying the proposition ( cf. Clift 2006). This paper is concerned with disentangling the discourse functions of epistemic and evidential adverbs. Spanish has a lot of epistemic and evidential adverbs at its disposal. First, at least seven 'adverbs' express 'maybe' or 'perhaps': a lo mejor, igual, quiza, tal vez, acaso, posiblemente and probablemente. Some authors prefer to use the term 'adverbial phrase (or, in Spanish, locuci6n adverbial) for epistemic a lo mejor. Second, several adverbs (and adverbial phrases) express a specific mode of knowing: aparentemente (also al parecer, por lo que parece), evidentemente, obviamente, supuestamente, vi-

310

Bert Cornillie

siblemente (also por lo visto). 1 Note that these adverbs are to be differentiated from discourse markers such as clara, por supuesto and desde fuego meaning 'of course', which necesessarily interact with the preceding context and do not refer to a specific mode of knowing. In this paper, I will only account for the first four epistemic adverbs and the four evidential adverbs. The paper is structured as follows. I first present some preliminary remarks (Section 2) before presenting previous accounts of the adverbs and the central hypotheses of this paper (Section 3). Then we go on to discuss the data both from a quantitative and a qualitative point of view (Sections 4 and 5). In the fmal section I present a series of conclusions as well as prospects for further research.

2.

Preliminary remarks

The relation between epistemic modality and evidentiality is a heavily debated topic in the functional-linguistic literature (cf. Dendale and Tasmowski 2001; Plungian 2001; Cornillie 2007). For some authors, an epistemic value is necessarily implicated in the reference to knowledge encoded in evidential expressions, whereas epistemic expressions sometimes lack such a reference to knowledge (Plungian 200 1). Other authors argue that epistemic modality and evidentiality are two different but related categories with specific functions (Nuyts 2008; Cornillie 2009). Chafe 1986 holds an alternative view in that he establishes a broad category of evidentials which includes both epistemic and purely evidential expressions. Although the relation between epistemic modality and evidentiality is important, in this paper I will focus on what epistemic and evidential adverbs do in discourse, rather than deal with the semantic differences between them. For the purpose of this paper, it is worth recalling Chafe's 1986 comparison of spoken and written language as far as the use of evidentials is concerned. Chafe (1986: 262) mentioned that his samples of conversational English and academic writing "showed approximately the same proportion of evidential markers to the total number of words. [... ] Differences appeared, not so much in the frequency of evidentials overall, as in the frequency of specific kinds of evidentials" Hence, some evidentials will more frequently appear in written communication, whereas other ones will show up more often in spoken interaction. If we fmd different adverbs in speech and writing, the question arises whether a cognitive-functional motivation

An interactional approach to epistemic and evidential adverbs

311

for the different use of the adverbs can be found in the differences between speaking and writing. Since the two registers differ in many respects, this may be the case. First, a writer usually has much more time to produce, i.e. edit, a text than a speaker can use for producing utterances in interaction. Writers can read their texts over and make local changes so as to have the best final result. Since authors often tend to present the most nuanced view, they have recourse to specific epistemic and evidential expressions involving specific modes of knowing. Register-related preferences may also give rise to interesting alternations with the same verb. In previous research I have accounted for (i) the tendency of using evidential auxiliaries in written Spanish (e.g. parecer 'to seem'+ infinitive) to the detriment of other forms with a finite que-clause and for (ii) the fact that the auxiliary + infinitive does not show up in spoken language. The answer lies in the observation that the qualification expressed by the auxiliary is based on inferences from reasoning only, whereas the parece que-construction can also involve inferences from direct evidence (Comillie 2007a,b). Secondly, speakers are in direct, face-to-face interaction with the interlocutors or co-participants. It can be expected that the involvement in a social activity causes speakers to pay more attention to speaker-hearer related experiences. As a matter of fact, this leads me to hypothesize that evidentials used in conversation refer more frequently to evidence shared with co-participants. The preceding observations entail to the following questions. (i) with regard to written language, does more time to deliberate and to edit mean more inferences on reasoning and less circumstantial inferences? (ii) with regard to spoken language, does direct experience in interaction lead to more circumstantial inferences? Due to space restrictions, in this paper I will only address the second question.

3. Previous accounts and hypotheses Since the publication of Chafe's seminal paper on evidentials, back in 1986, the existing literature has not satisfactorily dealt with register-related topics in epistemic modality and evidentiality. Moreover, in-depth studies of modality in conversation are rare. In this respect, the field of Spanish linguistics is no exception. The New Reference Grammar of Modern Spanish mentions that a lo mejor is "confmed to spoken language or informal styles" (Butt and Benjamin 1994: 244) but does not specifY whether quiza or tal vez are more frequently used in one or the other register. Moreover,

312

Bert Cornillie

no specific interactional functions are attributed to the individual adverbs. The Gramatica Descriptiva de Ia Lengua Espanola (1999) only dwells on the modal adverbs briefly, comparing them with other adverbs. No attention is paid to their use in day-to-day Spanish. Whereas for English conversation there is Kiirkkiiinen's 2003 book on the expression of epistemic stance, the field of Spanish conversation studies does not have any publication that comes close to it. In line with Kiirkkiiinen 2003, Fernandez Sanmartin 2006 presents a comprehensive inventory of the epistemic and the evidential expressions that appear in the informal conversations compiled by the Val.es.Co group. It is a first step towards a better understanding of the use of different expression types in Spanish conversation, although does not deal with the dialogic context in which the expressions are used. Hence, a clearly interactional approach to Spanish epistemic and evidential expressions is still needed. Now, if the interaction has not been a central topic, what are then the approaches adopted for the study of Spanish modal adverbs? Many authors discuss the mood alternation with some of the adverbs, although some authors observe that with epistemic adverbs a different mood selection does not alter the likelihood expressed (Branza and Delbecque 2008). Most grammars point out that a lo mejor and igual combine exclusively with the indicative mood whereas, when they precede the verb, tal vez and quiza usually combine with the subjunctive mood. Yet, no satisfactory account is given for this mood distribution. In his Gramatica Comunicativa del espanol, Matte Bon (2000: 257-258) distinguishes between the subjunctive and the indicative in terms of the "hip6tesis rematicas" (new information) and "hip6tesis tematicas" (given information) the adverbs introduce. The adverbs quiza(s) and tal vez are said to qualify given information, whereas a lo mejor would only introduce rhematic hypothesis or new information. In (1), Speaker B takes the tum from speaker A and attempts to explain the situation given by speaker A. (1) A. Nada ... lmposible: esta comunicando sin parar. (Matte Bon 1999: 257) 'nothing ... Impossible: they do not stop talking' B. A lo mejor tienen el te!efono estropeado. 'Perhaps their telephone is broken' When speakers use quiza and tal vez with a fmite verb in the subjunctive mood, as in (2), they formulate a hypothesis based on thematic or given

An interactional approach to epistemic and evidential adverbs

313

information. Hence, the information was previously mentioned to the speaker but was not communicated to the co-participant before. (2)

A. ;,Sabes? Quiza cambie de trabajo. (Matte Bon 1999:258) 'You know (already)? Maybe (s)he changes job'

Quiza(s) and tal vez can combine with a rhematic hypothesis, but, according to Matte Bon, do never have the same rhematic force as a lo mejor. By using this adverb, the speaker "gives the impression of presenting hypotheses that come up at the moment of formulating them" (Matte Bon 1999: 258). 2 As it will be shown in Section 5, a lo mejor does more: the speaker not only presents a hypothesis but also attempts to know what the coparticipants think about it. Moreover, a lo mejor functions in (1) as an epistemic marker with an inferential dimension, somewhat similar to inferential deberlmust. By contrast, quiza in (2) is not used with an interactional purpose, nor does it have an inferential dimension. Unlike the epsitemic markers, evidential adverbs do not witness mood variation (cf HaBler 2004). The indicative is the standard mood with these expressions. Since this paper combines the study of epistemic and evidential adverbs, it is worthwhile replacing a lo mejor and quiza by the evidential adverbs under examination here. In (3) a lo mejor can be readily substituted by aparentemente 'apparently', obviamente 'obviously' and evidentemente 'evidently', but conflicts with supuestamente 'supposedly'

(3) A. Nada ... lmposible: esta comunicando sin parar. 'nothing ... Impossible: they do not stop talking' estropeado. their telephone is broken' - Aparentemente (inferential - circumstantial) - ??Supuestamente (reported- knowledge) - Obviamente (inferential - knowledge) - Evidentemente (inferential - knowledge)

B. . ................. tienen el telefono

The fact that aparentemente fits in the open slot in (3) can be explained by the new situation the speakers are confronted with. Speaker A and Speaker B did not expect to be unable to call somebody. The inference is made on the basis of their observation that they cannot call. When obviamente and evidentemente are used, Speaker B refers to a situation which (s)he has already experienced on another occasion or which (s)he is familiar with through other means. Supuestamente, by contrast, excludes personal expe-

314

Bert Cornillie

rience and reports the observation made by someone else. It is clear that such a reading clashes with the pragmatics of the context. In sum, the test with the four evidential adverbs points to variation in the mode of knowing, which can be inferential or reported, and the type of evidence, which can be circumstantial, i.e. direct evidence, or knowledge, i.e. indirect evidence. In (4), none of the four evidential adverbs fits in the slot occupied by the epistemic adverb quiza. The first person singular of the verb cambiar may also clash with the evidential qualification following the question ;,sabes? 'You know?'

(4) A. ;,Sabes? ••••••••••••• cambia de trabajo. 'You know? .............. (s)he changes job' - ??Aparentemente (reported- circumstantial) - ??Supuestamente (reported- knowledge) *Obviamente (inferential - knowledge) *Evidentemente (inferential - knowledge) The test in (4) shows two things: (i) the discourse context of epistemic marker without an inferential dimension does not allow for evidential expressions. It tells us that evidentiality and epistemic modality are two different functional categories; (ii) there is a conflict between introducing new information to co-participants and using adverbs that hint at knowledge shared with them. With regard to (ii), there is a slight difference between aparentemente-supuestamente and obviamente-evidentemente. The former can be used if the speaker frames the information as previously unknown to him/her, whereas the latter do not allow for such a reading. Whereas the mood alternation in terms of information structure is quite convincing, an interactional approach will have to disentangle the discourse function of adverbs in a conversational context. In a dialogic setting, information structure can be broadenend with the criterion of shared or personal information. Hence, the difference between the subjunctive and the indicative mood should not only be seen in terms of theme-rheme but also be examined from the point of view of the speaker's on-line planning to ensure a reaction or a response from the co-participants (Kiirkkiiinen 2003: 83). The two following hypotheses can be derived from the preceding observations: Hypothesis 1: the epistemic adverbs a lo mejor and igual present a hypothesis to be confirmed or refuted by the interlocutor, whereas quiza and tal vez stick to the speaker's epistemic qualification of the proposition ra-

An interactional approach to epistemic and evidential adverbs

315

ther than engaging in interactional strategies. This opposition could be seen as a functional elaboration of the theme-rheme opposition. Hypothesis 2: the evidential adverbs aparentemente and supuestamente emphasize the evidence, but differ as to the type of conclusion (new or reported) they can encode. The former adverb is expected to perform both readings, whereas the latter only can be reported. Obviamente and evidentemente, by contrast, focus the attention on the interlocutor or coparticipant. Not the new or the reported conclusion but the on-line planning (reasoning) behind is emphasized. Before I go on to test and subsequently verify these hypotheses in the qualitative analysis in Section 5, I will first present a quantitative overview of the relative frequency of the adverbs in different corpora.

4. A quantitative view of the epistemic and evidential adverbs In this section I present figures that illustrate the frequency of the adverbs in different corpora. In order to roughly distinguish between registers, we will start with the Corpus del espaiiol, which is a large on-line corpus containing more Latin American than peninsular Spanish. Its section of contemporary Spanish consists of the three subsections: spoken language, literature and texts. Table 1. Frequency distribution of the epistemic adverbs %

Spoken % Literature % Text %

a lo mejor

quiza(s)

tal vez

1030 31.94 443 13.98 33

1004 31.13 967 30.52 622

1191 36.93 1758

3.10

58.35

55.49

411 38.56

3225 100 3168 100 1066 100

Table 1 starts from the following question: which of the three adverbs is preferred in the spoken register, in literature and in non-fiction prose, respectively? Note that igual is not taken into account due to technical problems. 3 It can be observed that both in fiction (literature) and in non fiction prose (text) there is a clear preference for one of the three adverbs. Fiction authors seem to have more often recourse to tal vez (55.49%), whereas, in non-fiction prose, quiza(s) is most frequently used (58.35%). In the latter

316

Bert Cornillie

case, a lo mejor can hardly be considered a real option (3.10%). Surprisingly, in the spoken subcorpus the three epistemic adverbs are equally frequent, with tal vez being slightly more frequent. Table 2. Frequency distribution of the epistemic adverbs

a lo mejor % Quiza(s) % tal vez %

Text

Literature

sEoken

33 2.19 622 23.99 411 12.23

443 29.42 967

68.39

37.29

38.72

1758

1191 35.45

1030 1004

52.32

1506 100 2593 100 3360 100

Since the register corpora all comprise approximatively 5.150.000 words, we can also look at how an adverb is distributed over the various registers. The figures in Table 2 allow us to check the degree of entrenchment of an adverb in a certain register. For instance, whereas Table 1 shows more or less the same frequency in the spoken subcorpus, Table 2 indicates that only a lo mejor has a prototypical use in spoken language. Tal vez, by contrast, confirms its high frequency in literature observed in Table 1. Finally, despite the fact that quiza(s) is the speaker's first choice in texts, the adverb is more often found in literature and in spoken language. Let us now examine along the same lines the frequency distribution of the four evidential adverbs under examination. First of all, it should be observed that these adverbs are far less frequent than the epistemic ones. Table 3 shows that the opposition between aparentemente - supuestamente and evidentemente - obviamente is reflected in their respective frequency in non-fiction prose. In fiction prose, aparentemente and evidentemente are the preferred evidential adverbs. Finally, evidentemente is by far the most frequent evidential adverb in spoken Spanish. Table 3. Frequency distribution of the evidential adverbs

Press

Ae.arentemente

Sue.uestamente

174

162 30.86 35 10.70 60 6.38

33.14 literature

157 48.01

Spoken

111 11.81

Evidentemente Obviamente 97 92

18.48 99 30.28

596 63.40

17.52 36 11.01 173 18.40

An interactional approach to epistemic and evidential adverbs

317

The frequency distribution in Table 4 confinns the tendencies observed in Table 3: aparentemente and supuestamente are more frequent in the written register, whereas evidentemente and obviamente are much more frequent in the spoken corpus. Table 4. Frequency distribution of the evidential adverbs

Aparentemente % Supuestamente % Evidentemente % Obviamente %

Press

Literature

SEoken

Total

174

157 35.52 35 13.62 99 12.50 36 11.96

111 25.11 60 23.35 596

442 100 257 100 792 100 301 100

39.37

162 63.04

97 12.25 92 30.56

75.25

173 57.48

Given that the written genres differ among them, it may be expected that considerable variation can be witnessed within the spoken genre. Indeed, the broad term "Spoken language" covers a whole range of spoken subregisters such as political debates, telephone calls, interviews and conversations. If we really want to study interactional contexts, we have to base our analysis upon purely conversational data. As for spoken Spanish, there are at least two high quality conversational corpora from Spain, respectively from the University of Valencia and the Autonomous University of Madrid. The so-called Corpus del Habla Culta is a third option, albeit less suitable for interactional analysis because of the interview format. Interviews cannot be considered an equivalent of spontaneous speech. Corpus de conversaciones coloquiales. Val.es.Co (Valencia). +/- 100.000 words Corpus oral del castellano. Conversation section. UAM (Castilia). 252.349 words Corpus del habla culta (Castilia: 138.539 words/America: 841.644 words) Let us look at the frequency distribution of the epistemic verbs in these different corpora. Several observations can be made: (i) the informal Val.es.co. corpus has the highest percentage of a lo mejor; (ii) igual is the second most frequent epistemic adverb in informal peninsular Spanish; (iii)

318

Bert Cornillie

quiza(s) is far more frequent in the peninsular part of the Habla Culta Corpus; (iv) the American part of the Habla Culta Corpus has the highest percentage of tal vez, which is almost absent in the corpora of peninsular Spanish. Table 5. Epistemic adverbs in different corpora

Val.es.Co (Valencia) +/-100.000 words % DAM (Castilia) 252.349 words % Habla Culta (Castilia) 138.539 words % Habla Culta (America) 841.644 words %

a lo mejor

Quiza(s)

tal vez

lgual

Total

93 79.49

4 3.42

0

20 17.09

117 100

131 69.31

14 7.41

43 22.75

189 100

57 50

56 49.12

0 0.88

114 100

116 25.84

166 36.97

167 37.19

0

449 100

0.53

Table 6. Evidential adverbs in different corpora

Val.es.co (Valencia) +/-100.000 words % DAM (Castilia) 252.349 words % Habla Culta (Castilia) 138.539 words % Habla Culta(America) 841.644 words %

aparentemente

supuestamente

evidentemente

obviamente

0

0

3 100

0

3 100

2 8.70

0

19 82.61

2 8.70

23 100

0

0

2 100

0

2 100

20 25.32

10 12.66

41 51.90

8 10.13

79 100

Table 6 shows that also in conversation evidential adverbs are less frequent than the epistemic ones. In all of the four corpora, evidentemente is the

An interactional approach to epistemic and evidential adverbs

319

most frequent evidential adverb. Somewhat surprisingly, several adverbs do not show up in the peninsular conversation corpora, whereas they do in the Latin American interview corpus. In sum, a lo mejor is by far the most common epistemic adverb in informal and interactional contexts, igual being the second option. As for the evidential adverbs, evidentemente is clearly the preferred expression. It is striking that several adverbs are really marginal in conversation. The contrast with their frequency in interviews corroborates the assumption that interviews differ from spontaneous talk-in-interaction. There is the guidance by the interviewer, and the interviewee usually attempts to elaborate larger sequences of discourse. In other words, the interviewee is keen to retain the turn.

5. A qualitative view of the adverbs

In this section we look at the discourse functions of the epistemic and evidential adverbs attested in the Val.es.Co corpus. I will focus on the contrast between a lo mejor and quiza, and also discuss the case of evidentemente. The aim is to present a functional account of the differences in frequency distribution. The main claim is that people use most those adverbs that serve interactional strategies best (a free interpretation of DuBois 1985). The discourse contexts that surround a lo mejor allow us to differentiate between its various discourse functions. The analysis below will indicate that the adverbs under examination do much more than conveying epistemic readings. That is, they go beyond the traditional focus on the evaluation of the likelihood. Table 7 Different readings of a lo mejor Discourse functions Alignment (fmward-type) Aligmnent (backward-type) Attenuation Situational Purely epistemic Intentional Approximative

Examples 54 9 8 7 6 5 4 93

% 58.06 9.68 8.60 7.53 6.45 5.38 4.30 100

320 Bert Cornillie From Table 7 we clearly observe that with a lo mejor speakers combine epistemic qualifications and alignment strategies with the co-participant. Example (5) illustrates such a context. (5) 1 2

3 4 5 6 7

8

9 10 11

Fragment 1 (Val.es.Co 2002: 117 -lines 1456-1457): G: Ia primerat quitar el contacto 'the first [question] take offthe contact' E: Ia primeral quitar el contacto en teorial![o sea pues} 'the first take off the contact in theory so well' L: [pues/no lo entiendo} 'well I don't understand' G: [no no no no no o s(e)a} 'No no no no no actually' en teoria y en practical!o sea§ 'in theory and in practice actually' E: [pero a lo mejor noes lo primero que haces} 'but perhaps this is not the first thing you do' L: [((a lo mejor explota))] 'perhaps it explodes' G: inormal! [que puede explotar el cochel puede arder1=} 'normal the car may explode may burn' E: [si/llevas raz6n tit} 'yes you are right' G: =sino quitas al contacto Ia bateria estaa- estafuncionando§ 'if you don't take off the contact the battery is is working' E: ((y va y hace pumba)) And goes and does 'baing'

In (5) speaker E suggests a hypothesis to speaker G which (s)he wants to be taken into account. Interestingly, the epistemic value does not correspond to doubt about the proposition that follows, but should rather be seen as a reply to speaker G. Furthermore, Speaker L elaborates the hypothesis of speaker E and directs it to speaker G, who fmally takes the turn back and answers affirmatively. Hence, the two suggestions introduced by a lo mejor provoke the co-participant (i.e. speaker G) to take the turn. As a matter of fact, using a lo mejor can be considered a discourse strategy of the speaker to achieve alignment with the co-participant. This phenomenon can be termed forward-type of alignment (DuBois 2007). Let us now discuss the other functions of a lo mejor. In addition to a forward-type of alignment, the adverb also serves to confirm alignment that

An interactional approach to epistemic and evidential adverbs

321

is already achieved in the previous turn. This use is called the backwardtype of alignment. (6) 1 2

3 4

5

6 7

8 9

10

11 12

13 14

15 16

17 18

Fragment 2 (Val.es.Co 2002: 100-101) E: § [pero yo por ejemplo} tengo problemas de cultural 1mt I for example I have cultural problems' que si que entiendo que- no debes ser asi pero lo soy! 'which oh yes I understand you shouldn't be so but I am' no se por que/yo por ejemplo- a mime repugna laa 'I don't know why I for example I disgust' homosexualidadlyo estoy totalmente en contrail 'homosexuality I am completely against' [pero no (( ))=} 'but not' L: [;,pero por que?} 'But why' E: =no lo selno se por quelnunca lo he descubierto/1 'I don't know I don't know why I never discovered it' de verdadl hay ahi una diferencia entre la mente de ((un 'really there is a difference between the mind of a' sex6logo)) y de un homosexualj/1/hombre/ 'sexologist and a homosexual you know' casos de homosexualidad concretos no conozco ninguno 'concrete cases of homosexuality I don't know any' va(le)- yo que sellos aparentes eso que- uno que' okay- dunno those who manifest the thing that' entonces[de tio a tio eso a mi-J 'well from guy to guy for me' G: [eso- e- e- es por eso} 'This is it is therefore' porque no has conocio ningUn caso§ 'because you haven't known any case' E: es por eso a lo mejorl 'it is because of this perhaps' porque imaginate que si yo tuviera-§ 'because imagine that ifl had' L: si estuvieras con un tio de esosj§ 'if you were with such a guy' E: yo creo que ahora- a lo mejor si co- si I think that now perhaps if I if

322 Bert Cornillie

19 20 21

22 23 24

25 26 27

28

29 30 31 32 33

[a lo mejor conozco=] 'perhaps I know' L: [jgMg]_no me aaa] 'perhaps it doesn't--- me' E: = (( )) y el- a lo mejor digo i}oder!llpero de verdad, and he perhaps I say waw/jeezes but really' pero yo de- de entrada lo rechazo§ 'but I from the beginning I reject it' G: §yo cuando- cuando ibaa§ 'I when when I went' E: §yo que sel me da mucho§ 'dunno it causes me a lot' G: § cuando iba al institutoJ yo 'when I went to the secondary school, I' [tenia unn- un amigo=] 'I had a friend' E: [o sea/nose] 'well dunno' G: =que era- que me llevaba cantidad de bien con el ;,no? 'who was, whith whom I was on very good terms, you know' ibamos siempre tresjuntos ;,no? y uno de ellos eraa!luno 'the three of us were always together and one of them was one' de los otros dos era marica/lmaric6n§ 'one of the two others was a fag ... gay' E: § ;,pero lo reconocia el como tal?§ 'but did he recognize it as such' G: § lo reconocia ellperoo el no se comportaba- no se 'he recognized it but he did not behave' comportaba con- con los amigos que tenia dee- ni se pasaba 'he did not behave with the friends he had nor did he" ni se [comportaba mal=] 'nor did he misbehave ' E: [yayaya] 'I see I see'

In line 15 of example (6), speaker E confirms speaker G's hypothesis cautiously by means of a lo mejor. The primary focus is on the alignment between speakers, though an attenuative dimension can also be observed. Interestingly, there are contexts in which we find attenuation without clear alignment strategy. In line 18, for example, a lo mejor precedes the condi-

An interactional approach to epistemic and evidential adverbs

323

tiona} clause introduced by the si conjunction, and thus weakens or attenuates the condition. Speaker E wants to keep her freedom and, hence, does not want to be held responsible for the consequences of her conditional reasoning. This use is in line with other face-saving strategies. In line 18, the repeated a lo mejor is part of the conditional protasis. In this context the adverb has a situational use, in that it has scope over a potential situation sketched in this first part of the conditional. In the main clause (apodosis) we also fmd a lo mejor. Since in this context the adverb combines with a verb in frrst person, it points to the speaker's intentions, hence the term intentional use. A lo mejor is followed by is a longer pause which indicates that the co-participant can take the tum. Since (s)he does not take it, the speaker goes on. Both the situational and the intentional dimension focus on the co-participant in that the speaker expressed that (s)he takes into account the co-participant's opinion and feelings. An approximative use of a lo mejor before numbers is also observed in the corpus, as exemplified in (7). (7) 1 2

3 4 5 6

7

8 9

10

Fragment 3 (Val.es.Co 2002:128 -line 231) A: [pero e-) pero es que//es que es el gasto que lleva 'but but you know the expenses which brings about' el ascensor/1/es la cantidad de luz quee!gasta§ 'the elevator is the amount of electricity that it consumes' M: §sf yes A: luego el engrase/luego las ave[rias] 'then the grease and then the damages' M: [no/eso es] una cosa que se tiene 'right, this is a thing that one has' que- que- que!alimentarj/se tiene que- que/que cuidarj 'to to to nourish one has to to to care' /claro eso es [(( ))] 'of course that's right' A: [pero si} ahora pagamos mil pesetas de luz de 'but if we now pay thousand pesetas for the electricity of escalera luego tendrian que sera lo mejor DIEZ MILI/I 'the stairs later on it must be perhaps ten thousand' [e- es un gasto enorme] 'this is an enormous cost'

324

Bert Cornillie

Finally, note that the examples of pure epistemic modality lack any interactional dimension. The function of a lo mejor in these contexts is the judgment of the likelihood. On the other hand, uses as in (8) could also be seen as part of a personal dialogue. Evidence for this can be found in the verb decir 'say' (8) 1 2

3 4 5 6

7

8 9

Fragment 4 (Val.es.Co 2002: 83-line 48) G: hombrel {. ..}he llegao y he llamao all- ahi al- al veinticinco 'Listen I arrived and lrnocked a ... there .... at the 25' no- no abria nadie ;,no? al veinticinco§ 'No nobody opened .. no. at 25' L: § ;,pero sabias que era el veinticinco? 'But did you lrnow that it was the 25' G: silyo sabia que eraajleste pisoj yy bueno pues ... por lo que 'Yes, I lrnew that it was .. this apartment and well from what I' me acuerdo yo de orientaci6n y tallsabia que mas o menos 1 remember about the direction and so I lrnew more or less that' era ... aqui ;,no? y he llamao y como no abria nadie yo digo 'It was here I lrnocked but as nobody opened I say' a lo mejor noes aqui.. y he llamao ahi alla(d)o y tampoco estaban 'perhaps it is not here and I rang at the house next to, and there weren't at home either' E: ;,si?l 'Oh yes?' G: pues vaya 'A pity'

Let us now tum to quiza(s). For reasons of space, the examples of quiza(s) will not receive the same attention as those of a lo mejor. One clear difference we can mention from the outset is that quiza does not show up at the end of the turn, that is, it does not invite the co-participant to take the tum. This adverb seems to serve to modalize the speaker's own discourse without aiming at interventions of the co-participant. Hence, quiza is not primarily directed at alignment, as shown in (9). (9)

B: A:

Fragment 5 (Val.es.Co 2002: 337-line 63) § ;,pero podeis ampliar la memoria real 1 en corto plazo? 'But can you amplify the real memory in short term' yolo llevo pidiendo hace un aiio [(RJSAS) (())} I am asking it since a year [laughter]'

An interactional approach to epistemic and evidential adverbs

B:

A:

B:

325

[(( ))} buenolpues a ver si podemoos/1/hacer algo aquilpor lo menos transitoriamentel;,no? 'Well, well let's see if we can do something here at least as a transition /okay?' paraa!poder resolver un poco el programa porque quizaa/si estos ee- trabajos de Tese6 corto no se les hiciera 'to be able to solve a bit the program because maybe if to these eh' 'works ofTese6 corto weren't applied' suapautllpues noo- no habria tanto problema ino? 'swap-out well no-- there wouldn't be such a problem, right?' por ejemplo 'For example' o sea que Ia degradaci6n es demasiado fuerte debido a esto ;,no? aa-a que§ 'Let's say that the degradation is too strong due to that, isn't ah that'

As example (1 0) confrrms, quiza appears in longer turns than a lo mejor. By means of this adverb the speaker suggests an idea which (s)he can refute or attenuate subsequently. Not quiza(s) but other elements serve to organize turn completion: intonation breaks, exclamative utterances and tag questions give the co-participant the chance to intervene. (10) 0:

A:

0: A:

Fragment 6 (Val.es.Co 2002: 363-line 534) = tu familia/dado que tu eres experta en esos temas les ira todo muy bien ;,no? 'your family /given that you are an expert in these topics they will do well, won't they' bueno/mira [veras- veras Ia cuesti6n familiar=] 'Well /look you see- you see the family question' [;,o a Ia familia nose puede enseiiar?J 'Or can't the family be taught?' = nolo sea!por ejemplo yo 1 lo que esta claro y hoy lo comentaba con otra amiga 1 yo no tengo que convencer a nadie 1lo seale! quee- yo lo que puedo explicar es algo 1 y Ia persona pues recapacita y si le va 1 oo parece- le parece posible yy aceptable 1 pues valelentonces tengo en mi casa doss chicoss de ingenieria!con lo cual estos son mas escepticos ;,no? 0 sea decirltal!PERO por ejemplo pueslquiza porque me ven y soy su madre

326 Bert Cornillie

B: C:

'more sceptical aren't they?that say like BUT for example well maybe because they see me and I am their mother' pues entran j nose como decirte ino? mi hija no ... mi hija si (( )) 'they come in I don't know how to tell you, you know my girl don't you .. my girl yes' tranquila es- es mujer y entonces esa cosa pues le apetecely mi marido me respeta un monton 1lse le ponen asi unos ojitos viendo Ia cantidad de personas que vienen y luego le dicen i hay que ver tu mujer! y aquel diciendo ivaya tela! o sea [(RJSAS)} [(RJSAS)} [(RJSAS)} y yo sin saberlo

Finally, we will examine the evidential adverbs in their interactional contexts. Despite the fact that they are very infrequent in the Val.es.Co-corpus, the three examples of evidentemente give us a good idea of how they ftmction in conversation. By means of evidentemente the speaker is clearly engaged in keeping alignment with the co-participant: the speaker indicates that (s)he is aware of the co-participants potential objections. Thus, the adverb preempts the co-participant's next turn. The two examples of evidentemente in example (11) illustrate this pattern. Interestingly, the intonation break that follows the adverb is a starting point for elaborating the main argument with which the speaker tries to convince the co-participant. Hence, the break not only does not lead to turntaking, the reverse is true: it gives the speaker the occasion to keep the turn. (11) A:

B: A

C: A: B:

Fragment 7 (Val.es.Co 2002:128 -line 486) § es una decision de administracionlpura y simple 'it is a administrative decision pure and simple' pero no esta basada en ningun estudio previoo§ 'but is not based in any previous study' § no j evidentemente!si que esta basado en una situacion anterior donde solamente habia uno [y entonces=} 'No obviously (yes) it is based in a situation in which there was only one and then' [y entonces} 'and then' = habia un colapso total 'There was a collapse' mm

An interactional approach to epistemic and evidential adverbs

A:

327

entonces el problema quee nosotros hemos detectado/es un problema que hemos detectao evidentemente/ 'Hence the problem that we have detected ouservelves is a problem that we have detected obviously' entonces/llo que hicimos fue/ee/de uno pasar a treslpero es un problema de capacidad y de prestaciones... no se si me explico/ normalmente/si lo hiCJEses por prestaciones no tendrias tres/ tendrias cincuenta ... porque el acceso es totalmente aleatorio ... entonces claroljpf!llme explico ;,no?

Evidentemente contrasts with both a lo mejor and quiza. This evidential adverb differs from a lo mejor in that it is part of the speaker's strategy to keep the tum, and it differs from quiza because of establishing mental contact with the co-participants. In other words, by means of evidentemente the speaker lets the co-participant know that (s)he is still developing his/her idea. After this qualitative analysis we now come back to our initial question of Section 2 concerning the type of inferences. It turns out that the evidential adverb mainly involves inferences from reasoning. Yet, the reasoning is not so much triggered by sensorial, visual or auditive stimuli but is part of the discourse planning. It can be argued that there is a semantic shift from a purely evidential reading (i.e. something that can be clearly perceived) to a discourse related reading (i.e. something that is clear according to discourse patterns). Interestingly, the other evidential adverbs do not readily allow for such a reading. It is now clear that in conversation adverbs do not readily express evidential qualifications that are based on circumstantial inferences. This confirms that conversation is very much focused on the attitudes of speech participants rather than on integrating new knowledge from the immediate physical context. Now, given the previous observations, what can we tell about the epistemic adverbs? As shown in Section (2), a lo mejor- igual are suitable for expressing an inferential dimension, whereas quiza and tal vez do not usually invite for such a reading. The fact thatthe former are more frequent in conversation than the latter leads us to conclude that the addition of a (circumstantial) inference to an epistemic evaluation facilitates the interaction. The reason why epistemic adverbs experience a shift to an evidential dimension, whereas evidential adverbs shift away from it is a topic for further research.

328

Bert Cornillie

6. Conclusions

This paper has shown that epistemic and evidential adverbs deserve to be examined in terms of their discourse functions in conversation. The relative frequency in different corpora has turned out to be indicative for these functions. The quantitative corpus analysis has shown that a lo mejor is the preferred epistemic adverb in spoken language, in general, and in conversation, in particular. In (Latin-American) interviews (Corpus del Habla culta), we observed the relatively frequency of quiza(s) and tal vez. The evidential adverbs aparentemente and supuestamente contrast with evidentemente and obviamente in that the former are far less frequent in spoken language than the latter. In conversational Spanish a lo mejor reveals to have several functions other than being an epistemic qualification of the proposition. The organization of alignment appears to be the most prominent one in the Val.es.Co corpus. A lo mejor is not only used to introduce hypotheses with new information, but also invites the interlocutor(s) to confirm or reject the view of the state of affairs (or part of it) presented, that is, the adverb plays a role in the tumtaking process. The relative frequency of igual goes along the same line. Thus, the above-presented analysis gives a functional explanation for the claim that with a lo mejor (and igual) the indicative mood correlates with new hypotheses. Interestingly, the corpus data of quiza contrast with those of a lo mejor, not so much because of the thematic hypotheses, but rather because of the longer turns in which quiza appears. The marginal frequency of quiza in conversational Spanish illustrates that it is less interactional than a lo mejor. The relative frequency of quiza(s) in interviews is due to the longer turns, in which the speaker does not invite for turntaking. In sum, we believe that the analysis presented in this paper demostrates that an interactional approach can account for the different mood and frequency distribution of the epistemic and evidential adverbs.

Notes 1.

In recent publications more attention has been paid to adverbial phrases such as por lo visto 'apparently' and al parecer 'seemingly' (Gonzalez 2006; Fant 2005, 2006).

An interactional approach to epistemic and evidential adverbs 2. 3.

329

In Spanish: "a lo mejor da la sensaci6n [ ... ] de presentar hip6tesis que se le ocurren al hablante en el momento mismo de formularlas." Since this form also refers to the adjective with the meaning 'same', it is extremely frequent (more than 3500 examples). Yet, only a tiny minority of these many examples have a modal reading. Further research on igual is needed.

References Branza, Mircea. D., and Nicole Delbecque 2008 Variaci6n modal con los adverbios de duda en espafiol. In Studii de Lingvisticii ~i Filologie Romanicii, A. Cunita, C. Lupu and L. Tasmowski ( eds.), 58-71. Bucharest: Editura Universitafii Bucure~ti. Briz, Antonio and Grupo Val.Es.Co. 2003 Corpus de conversaciones coloquiales. Madrid: Arco-libros. Butt, John, and Carmen Benjamin 1994 A New Reference Grammar of Modern Spanish. 2nd ed. Great Britain: Edward Arnold. Chafe, Wallace 1986 Evidentiality in English Conversation and academic writing. In Evidentiality: the Linguistic Coding of Epistemology, Wallace Chafe, and Johanna Nichols (eds.), 261-72. Norwood, NJ: Ablex. Clift, Rebecca 2006 Indexing stance: Reported speech as an interactional evidential. Journal of Sociolinguistics 10: 569-595. Cornillie, Bert 2007a Epistemic Modality and Evidentiality in Spanish (semi-)Auxiliaries. A Cognitive-functional Approach. (Applications of Cognitive Linguistics 5.) Berlin/New York: Mouton de Gruyter. 2007b On the continuum between lexical and grammatical evidentiality. Evidence from Spanish. Italian Journal of Linguistics 19 (1 ): 108129. 2009 Evidentiality and epistemic modality: on the close relationship of two different categories. Functions ofLanguage 16 (1): 44-32. Dendale, Patrick, and Liliane Tasmowski 2001 Introduction: Evidentiality and related Notions. Journal of Pragmatics 33 (3): 339-348. DuBois, John W. 1985 Competing motivations. In !conicity in Syntax, John Haiman (ed.), 343-366. Amsterdam/Philadelphia: John Benjamins. 2007 The Stance Triangle. In Stancetaking in Discourse, Englebretson, Robert ( ed.), 139-182. Amsterdam/Philadelphia: John Benjamins.

330 Bert Cornillie Fernandez Sanmartin, Alba La expresi6n de Ia modalidad epistemica en el espaiiol 2006 conversacional. Ma Thesis. Department of Spanish Philology. Santiago de Compostela: Universidade de Santiago de Compostela. Gonzlilez Ramos, Elisa 2005 Como eludir responsabilidades sobre lo dicho: los signos ''por lo visto" y "al parecer" (analogias y diferencias en su empleo actual). Espaiiol actual: Revista de espaiiol vivo. 84: 153-158. HaBler, Gerda 2004 El uso evidencial de adverbios modales. In Algunos problemas especificos de Ia descripci6n sintactico-sem(mtica, Juan Cuartero, and Gerd Wotjak (eds), 229-244. Berlin: Frank & Timme Verlag. Karkkainen, Elise 2003 Epistemic stance in English conversation. A description of its interactional functions, with a focus on I think. Amsterdam/Philadelphia: John Benjamins. Matte Bon, Francisco 2000 Gramatica Comunicativa del espana!. Madrid: Edelsa. Nuyts Jan 2008 The "one-commitment-per-clause" principle and the cognitive status of qualificational categories. Linguistics 47 (1): 141-171 Plungian, Vladimir A. 2001 The Place of Evidentiality within the Universal Grammatical Space. Journal ofPragmatics 33 (3): 349-357. Tomer Castells, Sergi 2005 Aspectos de Ia semantica de los adverbios de modo en espaiiol. Ph. D. diss. Department of Linguistics. Barcelona: Universitat Pompeu Fabra.

Revelative evidentiality in European languages: linguistic marking and its anthropological background Alexandra Kratschmer and Adrienne Heijnen

In this article we aim to present the results to date of an ongoing interdisciplinary research project on dreams and dreaming, focusing primarily on the linguistic aspects, i.e. the marking of the source and manner of acquisition of dream material in dream accounts, pointing at relevant relations to the cultural background of the speakers. 1 The project is executed by a linguist (AK) and an anthropologist (AH). The objective of this article is threefold: a) empirical (as we draw a first inventory of linguistic markers used in European languages in order to present information as "dreamed", diachronically and synchronically, helping us to seize the complex semantic structure of this domain), b) theoretical (as we wish to show that the semantic domain of evidentiality, revelative but also in general, will draw profit of the integration of a ,rather undetected aspect, namely the interaction between the human receiver and a human/divine/personified source of information) and c) methodological (as we seek to use our results on the linguistic and anthropological aspects of dream treatment to illustrate the heuristic potential of cross checking "claims" made by language structure about non linguistic matters (culture, cognition, etc.) in the relevant discipline, in our case anthropology.

1.

Revelative Evidentiality

We start out with a definition of our concept of revelative evidentiality (Section 1), continue with considerations regarding the heuristic value of our interdisciplinary approach (Section 2), introduce briefly our written and oral corpora (Section 3), after which we present the empirical linguistic and anthropological results of our studies to date and address possible relations between these (Section 4). In our conclusions (Section 5) we return to our previously established threefold objective, namely rendering account for the empirical, theoretical and methodological results of our studies.

332

Alexandra Kratschmer and Adrienne Heijnen

1.1. A first working defmition Following Aikhenvald's seminal monography on evidentiality in the languages of the world, "Evidentiality is a linguistic category whose primary meaning is source of information" (Aikhenvald 2004: 3). The author clearly states that this term should be used uniquely to refer to grammatical means of coding the source of information (Aikhenvald 2004: 11). In the mean time, an increasing number of scholars, especially those working on European Languages, insist on "considering evidentiality as a more general functional category whose scope includes not only grammatical but also lexical phenomena" (Squartini 2007: 1). For example, Squartini (2008: 918) aims at "exploring how an integrated account of both grammatical and lexical evidentiality can contribute to a better understanding of the whole domain" [our emphasis]. In what follows, we will use the term of evidentiality noting the semantic domain of information source as marked by linguistic means. These linguistic markers can be grammatical, lexical or even discoursive (contextual). 2 We will hence speak of grammatical vs. lexical vs. discoursive evidentiality marking. Thus, the inclusion of the option of discoursive evidentiality marking even goes further in challenging the original3 understanding of the category of evidentiality than e.g. Squartini cited above. 4 Within this domain, another distinction, advocated for by e.g. Squartini (2008: 918), is useful for our purpose, namely "mode of knowing" vs. "source of evidence": the "mode of knowing" is always "dreaming" in our case, but the "source of evidence" can be linguistically represented as the dreaming subject him/herself or as a more or less unspecified source outside the subject. Note here that the dreaming subject might be distinguished from the waking subject in certain societies, especially in epistemological terms (Heijnen forth.). In languages with grammatical evidentiality marking, one can typically observe the use of different markers for different types of information sources, with different degrees of specification depending on the language. The recurrent semantic parameters are visual, non-visual sensory, inference, assumption, hearsay and quotative (Aikhenvald 2004: 63-64). Some languages do also have markers for endophoric evidentiality that "refer to the speaker's own physiological or psychological states" (Plungian 2001: 354, as well as personal communication). As early though as in 1910, Boas inventarizes a suffix with a special evidential meaning 'see in a dream' in Kwakiutl. In 1986, Jacobsen suggests that this suffix could be related to a formative suffix meaning 'have a dream' in Makah. It was actually Roman

Revelative evidentiality in European languages

333

Jakobson, who coined the term "revelative evidence" for such phenomena in 1957 (Aikhenvald 2004: 345i Aikhenvald discusses the grammatical marking of that kind of information source in her monography. She remarks that in different languages different types of knowledge and experience are associated with conventionalized sources. In Jarawara, Quechua, Tuyuca and Tatuyo dreams are treated on a par with directly observed reality, in Shipibo-Konibo, Yukaghir, Cree and Modem Eastern Armenian dreams are represented as being outside conscious reality and are coded as reported or non-firsthand evidential, while in Tucano and Tariana ordinary humans' dreams are coded as non-visual evidential, but shamans' as visual evidential (Aikhenvald 2004: 380-381). Thus, dreams are very often marked by means normally used to refer to other kinds of evidential categones. Since we interpret "evidentiality" as a semantic category, we treat information revealed in dreams as a special evidential category meriting its own attention. What interests us primarily is which linguistic means (grammatical, lexical, or discoursive) can be used in different languages to mark this semantic category, whether dream marking draws on other evidential categories (e.g. the visual one) and/or non-evidential semantic categories (e.g. epistemic modality). Our working definition of revelative evidentiality reads broadly "information linguistically marked as created inside the mind of a subject without direct input from the outside world (which is not necessarily the view of the experiencing subject, see below), e.g. dreams, visions/revelations, hallucinations, inspiration/ideas, etc.", information revealed in a dream being a very important subcategory, in fact the category this article will focus on.

2. The interdisciplinary point 2.1. Basic methodological assumption Conventionalized and grammaticalized linguistic expressions "make claims" about cultural experience and knowledge and cognitive processes. While inferring cultural and cognitive circumstances directly from linguistic structures can lead to erroneous conclusions, we see testing "claims made by language" against anthropological and (neuro-, psycho- etc.) cognitive empirical data to be a powerful heuristic tool. The importance of this becomes obvious when considering Aikhenvald's conclusions about dream

334

Alexandra Kratschmer and Adrienne Heijnen

theories among the Yukaghir, mentioned in Section 1.1. While Aikhenvald claims that for the Yukaghir, dreams are not equal to conscious reality, basing herself on the older work of Jochelson 1905, recent anthropological study undermines her point. Willerslev, who has conducted ethnographic fieldwork among the Yukaghir in the 1990s and 2000s, writes that, for the Yukaghir, dreaming involves a conscious experience of the ayibii6 the soul, often unfolding as sexual engagement with spiritual beings (Willerslev 2007: 175). In the ongoing project we intend to test "claims" made by language on the cultural-cognitive basis of dream experiences against anthropological data on culture related perspectives to dreams and dreaming in order to see whether there can be detected any convincing relations between language use and cultural ideas and practices, and, in the positive case, whether there is any plausible way in determinating a causal directionality from culture to language or vice versa. At the present point of the investigation there seem to emerge certain patterns, even though they lack statistical significance.

2.2. Expected benefits At a general level, an interdisciplinary investigation will further our understanding of the relation between thought, practice and language and will be able to put NeoWhorfian approaches to the thought/language problem to the test. At the linguistic level, it will further our understanding of the semantic category of evidentiality (also regarding such factors as e.g. interactivity, see below) and of its cultural and cognitive basis, besides helping to inventorize its linguistic realization options. At the anthropological level as well as that of cognitive sciences, it will further our understanding of the use of linguistic material as an important heuristic source. Methodologically, in anthropology, "what people do" often predominatess "what people say", while anthropologists depend increasingly on semi-structured interviews due to a change of research conditions. These new settings urge for a reconsideration of the relation between language use, cognition, culture and social practice.

Revelative evidentiality in European languages

335

3. The empirical data 3.1. The linguistic corpus

3.1.1. Bible versions ofJoseph's story Our linguistic corpus is meant to cover both diachronic and synchronic data. As a diachronic corpus, we chose different versions of Joseph's story (Old Testament, Genesis 37-44) 7 in different European languages (as well as the Biblical Hebrew original) through history. This method allowed us to compare variations on a constant theme, notably a story focusing on dreams and dream accounts, both interlinguistically and language internally through time. We have to date analyzed the following versions of Joseph's story: Biblical Hebrew (original= 0, B.C., exact period heavily discussed), Classical Greek (Septuagint = SE, B.C-A.D, exact period equally discussed), Latin (Vulgata = V, 4th cent. A.D.), German (1545/modemized = G1, 1871 = G2, 1905 = G3, 1912 = G4, 1951 = G5, 1984/99 = G6, 2000 = G7), Dutch (1618/modemized = DU1, 1734 = DU2), Danish (1550 = DA1, 1647 = DA2, 1931 = DA3, 1992 = DA4, 2002 = DA5), Icelandic (1815 = IC1, 1841 = IC2), Faroese (1949/74 = FA), French (1707/44 = FR1, 1744/1881/86/1996 = FR2, 1859/1991 = FR3, 1894/1910 = FR4,), Italian (1649 = IT1, 1821 = IT2, 1925 = IT3, 1974 = IT4, 2007 = IT5) as well as Spanish (1569 = S1, 1909 = S2, "modem"= S3, 1986/97 = S4, 2005 = S5). We have used the English spoken American Standard Version (ASV) for glossing purposes if convenient.

3.1.2. Interviews with informants We have to date registered oral dream accounts and qualitative data regarding language use in dream telling, combined with linguistics tests on grammaticality, discourse compatibility and paraphrasing dream related expressions, among Icelanders, Faroese, Dutchmen, Germans as well as Italians. The main criterium for selecting informants was that the group should allow for comparison between the various languages and societies. The informants are men and women between 20 and 35 years of age, who study various disciplines at institutions of higher education in Denmark. Whereas our corpus of Icelandic dream accounts possesses a clear quantitative significance (cf. Heijnen 2005a, forth.), our Faroese, Dutch, Ger-

336

Alexandra Kratschmer and Adrienne Heijnen

man and Italian oral corpus to date consists of qualitative data collected through focus group interviews and semi-structured interviews. Even though our data is extensive enough to observe certain patterns, it still lacks quantitative statistical weight.

3.2. Anthropological methods and data collection With the help of focus group interviews and semi-structured interviews, we have collected data on the cultural perspectives and practices related to dreaming, dream telling and interpretation among Faroese, Italians, Germans, Dutchmen and Icelanders. For the Icelandic case, our data is extensive. Heijnen collected the data during fieldwork in Reykjavik and South Iceland during the period 1996 to 2008, using the methods of participant observation (including semi-structured interviews) and literature studies. Participant observation allows for collecting data on the context in which dreams are remembered and narrated. To obtain historical depth in the analysis and to reduce personal bias, Heijnen collected numerous dream accounts written from the 13th century until the present day8 Thus, the anthropological data consists of nearly 1000 Icelandic dream accounts, covering a period of almost 1000 years (Heijnen 2005a, forth.).

4. The results to date

4.1. Linguistic results 4.1.1. Issues of revelative evidentiality in Joseph's story across time and across languages During our work with the biblical material, the following cardinal points of revelative evidentiality emerged as subject to variation as well as pattern forming: reference to the dream experience as such marking the beginning of the dream account marking of the dream account while unfolding it combined forms where at least two of the aspects just mentioned are fused marking the ending of the dream account

Revelative evidentiality in European languages

337

the revelative element the etymology of the lexemes '(a) dream',' to dream' Not all of these cardinal points have to date been subject of systematic data scrutiny, therefore, we will focus here on the reference to the dream experience (4.1.1.1.), the marking of the beginning of the dream account and of its unfolding (4.1.1.2.) and fmally the revelative element (4.1.1.3.).

4.1.1.1. The reference to the dream experience as such In this area, the questions which arise are those of verbal valence and of the semantic roles selected by the verb 'to dream' as well as those of the choice of verb in combination with the noun 'a/the dream' Clear diachronic and synchronic typological patterns can be observed which also can be blurred by obvious traductological issues. In our corpus, the following constructions and diachronic patterns are found: 'to dream a dream': this is the only documented form used in the Biblical Hebrew original9 and is, probably by strong traductorial influence, present in all our corpus-languages, except Latin (translated from Classical Greek, not from the Hebrew original) and Icelandic (on the special status of Icelandic, see below) 'to see a dream': this is the by far preferred construction in Classical Greek10 and the only documented construction with the noun dream in Latin; the Latin use of this construction is not pursued by the Romance Languages, nor does it exist in the Germanic Languages 'to see something' with reference to dreaming is documented in the Latin Vulgata 'to dream' (absolute use) or 'to dream + direct object' (typically occurring in constructions like 'I/he dreamed something' or with completive clauses as 'tell me what you have dreamed', 'listen what I have dreamed', 'I have dreamed that'): this construction is documented in German, Danish, Dutch, Faroese, French, Italian and Spanish remains active throughout history 'to have a dream': Dutch and Faroese, and Spanish show a stable (and exclusive) use in our corpus 11 of 'to have' in combination with 'a dream' and especially for Spanish this construction is by far the preferred one in our corpus, compared to 'to dream (a dream)'

338

Alexandra Kratschmer and Adrienne Heijnen

German and Danish on the one hand and French and Italian on the other undergo a characteristic diachronic development towards constructions suggesting increased subject participation (cf. 4.1.1.1.1) Icelandic shows an extremely stable use throughout history of a double accusative construction (cf. 4.1.1.1.2)

4.1.1.1.1. The subject participation axis In our corpus, there can be observed that earlier Bible versions of German and Danish make extensive (though not exclusive) use of constructions with the dreaming person in a dative/oblique 12 case (it dreamed [to] me, a dream [nom.] dreamed [to] me): Es [nom.] hat uns [dat.] getriiumet 'it has dreamed [to] us' (40:8, G1), Was ist das fur ein Traum, der [nom.] dir [dat.] getriiumet hat? 'what kind of dream is that, that has dreamed [to] you?' (37:10, G1), Mig [obl.] dmmde ocsaa '[to] me [it] dreamed also' (40:16, DA1), Da dmmde oss [obl.] baade en nat huer sin dmm [nom.] 'it dreamed [to] both of us one night a dream' (41:11, DA1), which later on recede in favor of constructions with the dreaming person as a nominative subject: [... ] der Pharao [nom.] triiumte 'the Pharao dreamed' (40:8, G2), Hor dog, hvadjeg [nom.] har dmmt! 'listen what I have dreamed' (37:6, DA3). Parallel to this, a representation of the dreaming subject as a nominative 'possessor' ('I had a dream') is documented from the first sources onward and remains strong throughout history: [... ] hatte Joseph [nom.] einmal einen Traum 'had J. once a dream' (37:5, G1), [... ] haffde Joseph [nom.] en gong en dmm 'had J. once a dream' (37:5, DA1). On the other hand, our corpus shows a development, concerning French and especially Italian, from a productive 'to have a dream' (Joseph eut un songe13 'J. had a dream', 37:5, F4; Giuseppe ebbe un sogno 'J. had a dream', 37:5, IT3 14) to a more and more frequent use of 'to make a dream' (Nous avons fait un songe 'we have made a dream', 40:8, F2; Giuseppe fece un sogno 'J. made a dream', 37:5, IT5). In both languages, 'to dream' with a nominative agentive dreaming subject is documented throughout history (J'ai aussi songe 'I have also dreamed', 40:16, F1; sogno di nuovo 'he dreamed again', 41:5, IT1). Looking at these data, language use seems to make a claim that the dreaming subject, in a way consistent for several European language communities, represents itself on an "activity" or "subject participation" axis, as moving towards more subject participation/responsibility: starting as an

Revelative evidentiality in European languages

339

experiencer, ( s)he becomes an owner and finally a creator of his/her dreams (see Table 1). Table 1. Diachronic development along the subject participation axis (simplified) >> owner >> languages experiencer 'I had a dream' German, Danish 'It dreamed to me' 'A dream dreamed to me'

French, Italian

'I had a dream'

creator 'I dreamed' 'I made a dream' 'I dreamed'

The validation by historical and present day anthropological empirical material of this "claim made by language" about a development of the cultural attitude to dreams and dreaming in these linguistic communities represents an important heuristic opportunity, which we will follow up in 4.2.2.2 and 4.2.2.3 resp.

4.1.1.1.2. The double accusative construction in Icelandic

Both in our corpus, but also in the remainder of the thousand years of Icelandic language history, documented in the prolific written source material (cf. Heijnen 2005a, forth.), a strong persistence in the use of the construction '[It] dreamed me [accusative] a dream [accusative] [evt. for somebody/something [dative]]': Mig [ace.] dreymdi draum [ace.] [!Yrir einhverju [dat.]] is found. 15 Based on different theoretical backgrounds, different syntactic interpretations can be given of this construction (Hrafnbjargarson/Tromso, personal communication). While traditional grammar considers it as a double (accusative) object construction, Generative Grammar (e.g. Hrafnbjargarson 2004: 133) analyzes it as a construction with a human accusative subject and nonhuman accusative object. These two different interpretations make, as we see it, different, but equally intriguing claims about how the dreaming subjects experience their role in a scenario which is equally complex in both cases. According to the traditional (double object) analysis, there is an unidentified agentive entity (the implicit nominative subject of the verb), a "power", exercising a direct influence (marked by the direct objects) on the dreaming person as well as on the dream (its content), sometimes regarding a third party (the prepositional dative object). The verb 'to dream' shows a complex valence pattern

340

Alexandra Kratschmer and Adrienne Heijnen

with a series of actants (1 nominative, 2 accusatives) and the frequent presence of a standardized circumstant (a prepositional dative), mirroring a complex constellation of involved entities. According to the Generative Grammar analysis (human accusative subject), the dreaming person combines "agentive" (subjecthood) and patient/theme features (accusativity), "acting" and "undergoing the action" at the same time. The dream itself is (equally) under direct influence of the action/the agent. 16 Also in this interpretation 17, there is a complex constellation of entities involved. We will show below that this complex constellation of involved entities in dream experiences can indeed be confirmed as an important underlying metaphysical stance of Icelandic culture (cf. 4.2.1 and 4.2.2.1 ).

4.1.1. 2.

The marking of beginning of the dream account and of its unfolding

Due to lack of space we have to limit ourselves to a very synoptic presentation of the linguistic phenomena regarding these two cardinal points, reserving the extensive discussion of the extremely rich data material for another occasion. As some of the following phenomena can mark both the beginning and the unfolding of the account, a collective treatment is by no means unnatural.

4.1.1.2.1. Visual evidentiality This is by far the most dominating evidential domain in our corpus. Here, a pervasive role is played by the 'behold'-marker which goes back to the Biblical Hebrew original hinneh 18 and which is subsequently found in all corpus languages except Faroese, e.g. German (siehe, as in 37:7, G2). It has to be noted though that the visuality conveyed by the 'behold' -marker is a phenomenon that obviously arose due to a traductorial problem: most target languages lacked a comparable focus marker, and the marker with a similar textual deictic function at hand was one conveying the idea of summoning the gaze of the listener ('behold'). The traductorial pressure to conserve this marker (but in a new disguise) seems to have been rather strong. Three of our Germanic corpus languages (namely German, Dutch and Danish) share the feature that they allow for perception verbs to be fol-

Revelative evidentiality in European languages

341

lowed by a subordinate clause introduced- instead of by 'that' (dass, at, dat)- by the manner conjunction 'how' (wie, hvordan/hvorledes, hoe dat), expressing not a manner, but the intensive perception of an action in progress or a situation under development (e.g. Duden [Drosdowsky 1984: 686], for German). While there are no German occurrences of this phenomenon in our corpus, both one Dutch and one Danish version (DAl) use the 'how'-connector, not after a perception verb, but after 'had a dream', probably capable of implying a context of (visual) perception, in the same verse 19 (41:1): hadde Pharao eenen droom hoe dat hy stand aan het waater, DU2. Another aspect of visuality in dream accounts is what we call tableau painting, grammatically marked by imperfective verb forms (see 5.1.1.2.4.).

4.1.1.2.2. Inferential evidentiality and epistemic modality intertwined In our corpus, it is often verbs or complex expressions of epistemic modality that are used to mark the beginning or the continuation of a dream account. In the European Languages represented in our corpus, these markers appear to carry systematically an evidential nuance inherent to the lexerne/construction (cf. Nuyts 2001: 64-72, 122-129, 205-208 on modal adverbs and adjectives, mental state predicates as well as modal auxiliaries in German and Dutch; Cappelli 2007, on English verbs of cognitive attitude) or depending on the grammatical and/or semantic context (cf. Kratschmer 2006 on the broad, context sensitive variety of evidential nuances of the Italian epistemic verbs sembrare and parere, both 'to seem', as well as Comillie 2007 on Spanish parecer 'to seem', equally expressing different nuances depending on the syntactic context). 20 Typically, the epistemic verbs and expressions found in our corpus are of that latter kind, i.e. expressing a clearly subjective judgment, sometimes combined with an element of epistemic reserve in the sense of marking that the content of the narration is not unambiguously assessable. It has to be noted that these epistemic verbs are not attested in the Hebrew original, but find their way into the text already in the Greek Septuagint (r.fJflf!V oimen 'to believe', 'to seem'). They are consequently consistently used in the Latin Vulgata (puto 'to mean', 'to judge', 'to consider as'), in the German (diinken, archaic 'to seem'), Dutch (dunken 'to seem', 'to think), Danish (tykke (archaic), syntes, both 'to mean'), Icelandic (ao pykja 'to find', 'to feel'' 'to seem', 'to like'; ao pykjast 'to claim'' 'to pre-

342

Alexandra Kratschmer and Adrienne Heijnen

sume', 'to profess'), Faroese (tao var (fyri) mtEr, sum 'it was (to) me, as (if)') French (sembler, roughly 'to seem'), Italian (parere 'to seem') as well as in the Spanish Bible versions (parecer 'to seem'). The evidential nuance carried by these lexemes is mostly inferential, in rare cases possibly visua1. 21 In a series of versions, a comparative connector is the inherent part in an expression of epistemic reserve, as in the subordinate connector following the verb 'to seem' This is true for Danish som 'as' following syntes 'to seem' (an archaic combination, e.g. 40:10: Oc (mig siuntis) som det blef gmnt22 , DA2), as well as for Faroese sum 'as' following 'it was [to] me': Tao var mtEr, sum vit bundu bundi uti aakrinum23 , 37:7, FA.

4.1. 1.2. 3. Alternative reality

Dreams can be marked as alternative reality by various means, as e.g. a comparative connector 'as if' 24 This is documented in the following Icelandic version of 40:9 based on pykja 'to seem' followed by sem 'as (if)' and the subordinated verb in the preterite subjunctive, a non-realis mood: Mer patti i svefninum sem vinviour sttEoi [pret. subj.] forir framan mii5 , IC1.26 The alternative reality aspect of the narrated dream can be marked in Icelandic, not only by the preterite subjunctive mood following 'to seem' just mentioned, but also by this mood following mig dreymdi '[it] dreamed [to] me', as in mig dreymdi ao vinviour vreri gegnt me?7 (40:9, IC2). 28 The same is true for German, where the 2nd subjunctive is used: Mir hat auch getriiumet, ich triige drei weij3e Korbe auf meinem Haupt? 9 ( 40: 16, G 1). 30

4.1. 1.2.4. Narratological aspects

A "typical" feature of Romance dream telling is the use of a characteristic aspectual pattern in the preterite verbal conjugation, at least following grammaticography. 31 This use is opposed to conventional narrative aspect use which combines perfective (marking actions as single and concluded) and imperfective preterite (marking actions or states as ongoing at reference time). The oniric preterite consists of casting the whole account in the imperfective aspect, which is often explained by the imperfect's potential of expressing "alternative realities" 32 It can, on the other hand, be used as a mere (non fictional) narrative marker, supposed to "immerge" the listener

Revelative evidentiality in European languages

343

into the action, typically in history writing or modem journalist style (Dardana, Trifone 1995: 354; Togeby 1982: 319). 33 Interestingly enough, the use of the oniric imperfective aspect in our Romance Bible versions is not systematic at all, compare e.g. the two competing version of Spanish 37:7: [ ... ] y he aqui que mi gavilla se levant6 [perfective preterite] [... ] (S4 + S5) vs. [... ] y he aqui que mi manojo se levantaba [imperfective preterite] [ ... ] (S 1-3). 34 One can conclude that the oniric imperfect, claimed by Romance grammaticography to be a definitorial feature of dream telling, is, as far as our corpus is concerned, regularly overridden by the "conventional" narrative aspect use. 35 This question of the stability of use of the oniric imperfect has also been part of our native speaker investigation with Italian informants (see 4.1.2.3.). (Visual) evidentiality and concerns of story telling go hand in hand, when the teller makes use of the narrative strategy of painting a "tableau" for his/her public's (inner) eye. This is done in several of our corpus Bible versions by verbal markers of imperfective aspect which represent the action "as it unfolds", either in the form of present participles as in the original Hebrew version, in the Greek Septuagint, the Latin Vulgata as well as our Icelandic versions (bl6mstrandi 'blossoming', 40:10), as a gerund as in Spanish (estabamos atando 'we were binding'; 37:7, S4 + S5), or as finite verbal forms as the Classical Greek imperfective aspect: (which stands in opposition to the aorist, coding perfective aspect) 36 , as well as the Latin imperfective preterite. 3 7 The imperfective markers in our corpus are the only grammatical markers of (revelative) evidentiality, but they are not autonomous. They require a context marked as a dream beforehand and can therefore only be classified as combinatorial markers. The context is created by lexical elements, which obligatorily have to be members of the lexical family 'a dream'/'to dream' The same is true for the above mentioned lexical markers, namely the epistemic verbs.

4.1.1.3.lnteractive evidentiality Finally, concerning the aspect of dreams being experienced as divine revelations, it turns out that evidentiality can and should be regarded from an interactive perspective where information can be passed and received in a deliberate and conscious manner.

344

Alexandra Kratschmer and Adrienne Heijnen

The idea of revelation is verbalized three times in our written corpus in Gen 41. Joseph explains the divine origin and intention of the Pharao's dreams to the Pharao: [... ]God has told to Pharaoh what He is about to do (Gen 41 :25); [... ]God has sho\\11 to Pharaoh what He is about to do (Gen 41:28). The Pharao in his turn speaks about Joseph's dream interpreting gifts equally as being divine revelations: Since God has informed you of all this [... ] (Gen 41:39; all ASV, all emphases ours). 38 In both constellations there is a relevator (God) who reveals something (the target) to the other (the Pharao or Joseph), and in all cases the action is seen from the perspective of the agent (God). Table 2. Revelation: the interactive patterns

level of interaction aspect focussed on visual perception

expressions: tokens

'to point/indicate': 11 3 rosli;ev SE 5 zeigen G 1, 4-7 1 anzeigen G5 2 indicare IT4,5 'to move the target into the field of 'to hold up/forward': 3 perception of the other and hold it 1 hiGGid 0 there' 2 ostendere V 'to cause the other's perception of the 'to make see': 10 target' 1 her' a0 2 sehen lassen G2,3 1 lade skue DA3 2 lade se DA2,4 1 ltita sja IC2 1 lata siggja FA 2 {air voir F 1,2 'to remove an obstacle in order to 'to unveil/to reveal': 3 enable the other to perceive the target' 1 enthullen G7 2 a[slere DA5 'to arrange the naked(= non-hidden) 'to render bare in the target in the open(= in the visual field open': 3 of the other)' 2 abenbare DA3,5 1 opinbera FA 'to direct the other's gaze to the target by one's finger'

Revelative evidentiality in European languages

345

'to shed light on the target in order to 'to declare/to make enhance the other's perception of it' clear, to illuminate': 5 1 birta ICl 1 augrysa IC2 3 declarer Fl-3 'to arrange the target in a way that 'to exhibit/display': 15 enables the other to perceive it' 1 vertoonen DU1,2 1 syna ICl 1 montrerF3 4 mostrare ITl-3 8 mostrar Sl-5 'to manifest': 5 'to make the target touchable to the hand of the other' 5 manifestare IT1,4,5 symbolic behavior symbols used 'to give signs': 2 2 significare ITl ,3 'to communicate verutterance act bally (to tell)': 1 1 fortrelle DA4 'to deliver a message': content 20 3 verkundigen G 1,4 6, 8 kundtun G 1-6 1 ankUndigen G7 2 verkondigen DU1,2 1 kond doen DU2 1 kundgere DA3 1 booa ICl 1 kunngjora IC2 2 anunciar S4,5 cognition 'to make know': 21 enrichment of the other's cognitive universe 1 hOch"''.' 0 1 te kennen geven DUl 2 give til kende DA1,2 3 lade vide DA1,2,4, 1 lata vita FA 6faire connaftre Fl-4 2 far conoscere IT2,3 5 hacersaber Sl-5

346

Alexandra Kratschmer and Adrienne Heijnen

In Table 2 the verbs and expressions used in the single versions of our corpus text are listed in conceptual groups following their etymology39 The etymologies of the used verbs and expressions very clearly contain the concept of interaction. As one can se, in our corpus, the level of visual interaction is by far the dominant one (55 vs. 23/symbolic behavior vs. 21/cognition). Inside the visual group, we find three dominating subgroups: 'to indicate' (11), 'to make see' (10) and 'to display' (15). Inside the symbolic behaviour group, 'to deliver a message' is the dominating choice (20 of 23 ). The cognition level, despite of disposing of only one realization pattern: 'to make know', is nevertheless important (21). There are no absolute patterns regarding the distribution of the above conceptual options of revelativity over the different languages in different periods (e.g. typological constraints regarding interaction levels), but one can detect a marked preference for the concept 'to deliver a message' in the Germanic (18 of 20) and a marked preference for 'to display' (13 of 15) in the Romance and a certain preference for 'to make know' (13 of 21, thereof one in the Hebrew original) again in the Romance languages. On the basis of these data, the interactive aspect seems crucial in evidentiality matters and we wish to make a plea for an integration of interactive aspects in evidentiality studies in general, that is, in all studies on information source marking. At that point, we even want to raise the question, not at all rhetorically meant, if one can find, in the languages of the world with grammatical evidential marking in the Aikhenvaldian sense, different evidential markers for 'overhearing something' vs. '(inter-)actively being told' or ' (visually) perceiving' vs. 'being shown'

4.1.2. Issues ofrevelative evidentiality in modern European languages form the native speakers 'perspective We repeat that we possess statistically relevant material only for the Icelandic situation at this point of the project (viz. Heijnen 2005a; 2007, forth.). The data material for the other communities has to be regarded as a first insight that has to be corroborated by a far larger amount of data.

Revelative evidentiality in European languages

347

4.1.2.1. Linguistic results ofparticipant observation among and interviews with Icelandic native speakers Our Icelandic corpus consists of approximately 80 semi structured interviews, conducted between 1996-2009 (Heijnen 2005a, forth.); a collection of interviews primarily carried out by the Icelandic ethnologist Hallfreour Om Eiriksson40 and the earlier listed literary sources (cf. n. 4). In line with the investigations for this particular project, we also conducted two interviews with Icelandic students in Denmark in the age group 25-35. Our data show a strong continuity in the use of the construction: mig dreymdi draum 'it dreamed me a dream', since the 13th century41 at least. However, the in note 11 mentioned inversion of nominative, accusative and dative42 is to some degree present in our material. Our modem data demonstrates that the double accusative construction mig dreymdi draum is especially consequently used among Iceland's older population. In addition, the mixing up of accusative and nominative occurs primarily in the perfect tense, mig hefur dreymt 'it has been dreamed me a dream', people using eg hef dreymt 'I have dreamed' instead, and in the general mann dreymir 'it dreams one' Thus, the application of nominative eg dreymdi 'I dreamed' occurs, but is rare. We have not encountered Icelandic native speakers who consequently apply nominative in the marking of dream accounts. Having analyzed numerous dream accounts, we found the following epistemic verbs: the most important being ao pykja 'to find' 'to feel', 'to seem'' 'to like'' ao pykjast 'to claim'' 'to presume'' 'to profess'' 'to make believe', 'to pretend' and ao .finnast 'to seem', ' to think, 'to like'. Ao pykja and ao pykjast are especially found in (older) written material and in dream accounts of the older population. The verb ao .finnast is less formal and more commonly used nowadays. These epistemic verbs are often followed by a designation of place43 Our data shows that dream accounts are told in the past tense and in the "historical" present. In the latter case, the dreamer enhances the recalling of the dream by "reexperiencing" the dream, as well actualises the content for the audience. The use of the "historical" present is often related to intense dream experiences, such as nightmares. In addition, the subjunctive mood occurs frequently in the modem dream accounts (see the example in 4.2.2.1), in a similar way as we noted for the biblical references (see 4.1.1.2.3).

348

Alexandra Kratschmer and Adrienne Heijnen

4.1.2.2. Results of interviews with German native speakers We interviewed one female and two male university students from the Northern and Eastern part of Germany aged between 21 and 23 years. We started with linguistically oriented single interviews and proceeded with a focus group interview to obtain anthropological data. Two of the students had never heard the construction mir [dat.] hat getraumt ('it has dreamed to me') and experienced it as outright ungrammatical, the last one remembered reading the form in older texts44 For all of them ich habe getraumt ('I have dreamed') was the only form they claimed to use. 45 They spontaneously observed that the dative construction would vehiculate the idea that the dream was coming from outside themselves, while on the contrary they felt it was their own mental product. This is clearly a case of what we would like to call "norm-violation induced transparency", i.e. the "literal'' semantic interpretation of constructions that depart form the norm. This phenomenon has its counterpart in "norm-compliance induced opacity", i.e. that fact that conventionalized ("salient") non-"literal'' meanings are automatically accessed as whole units without their constitutional parts being analyzed each for themselves. 46 That means that the fact, that the dative is experienced as expressing extra-subject dream origination does not imply that the form with the nominative experiencer is automatically and consciously used and interpreted as expressing intra-subject dream creation by the naive language user. Inspired by the biblically documented use of the archaic epistemic verb diinken (in G 1 and G4) for marking the beginning of the dream account, the subjects' readiness to interpret the modern equivalent scheinen 'to seem' not as an epistemic reserve marker, but as a mere marker of revelative evidentiality, was tested, but the subjects could actually not decide on the question in a conclusive manner. Inspired by the biblical use of the visual perception verb (ich/er sah '1/he saw') in dream accounts, we tested the natural interpretation of this verb in dream accounts either as a marker of revelative evidentiality or as a marker of straightforward visual perception as an ingredient to the dream action (on a par with other events occurring in the dream). Here, the subjects agreed on only accepting the latter. A phrase lch sah, wie ich Fahrrad fuhr 'I saw how I rode the bike' was not interpretable as 'I dreamed that I rode the bike' to them, but only as 'I saw myself riding the bike (from outside, as one can experience that in a dream)' Asked to produce an actual dream account, we registered both "historical" present (actualizing the content for the teller and the audience) and

Revelative evidentiality in European languages

349

past tense forms. One subject spontaneously produced a subordinate clause in the second subjunctive, as documented in G1 and G4 (cf. 4.1.1.2.3): ich habe getriiumt, ich ware [2nd subj.] in einem Turm 'I have dreamed I were in a tower', thus marking the alternative reality status of the dream. Directly questioned, all subjects accepted that use. To conclude, it can be said that the results of our interviews did not contradict the diachronic evolution documented by our Bible studies, i.e. that the verb triiumen 'to dream' is becoming more and more agentive, meaning that it is more easily combined with a nominative agent than with a dative experiencer.

4.1.2.3. Results of interviews with Italian native speakers We interviewed three female university students aged between 21 and 31, originating from Northern, Center and Insular Italy respectively. All linguistically oriented interviews were single interviews; two of the informants gave a group interview, while one informant, for schedule reasons, also responded to the anthropological questions during her individual interview. They all came to the conclusion that the noun un sogno 'a dream' was absolutely incompatible with avere 'to have' and always requiredfare 'to make', thus confirming the diachronic tendency announced by our Bible corpus. According to two of them (the third one not being present), avere 'to have' was compatible though with un incubo 'a nightmare', or with un sogno premunitorio 'a premonitory dream' 47 Again they spontaneously observed, that 'to have' would indicate that the dream was not the product of their own mind, as they felt it was and as 'to make' appropriately verbalized according to their view, while the premonitory dream could be seen as a message from elsewhere. Again, we have to suspect a possible "norm violation transparency" -effect here. 'To dream about somebody' can only be construed as sognare qualcuno 'to dream somebody', and one subject spontaneously produced the sentence una mia arnica[ ... ] sognava se stesso correre 'a [girl] friend of mine dreamed herself [to] run' 48 The epistemic verb sembrare 'to seem' is still possible as a dream account marker (it was not produced spontaneously though), but the linguistic tests made it very clear that this function is only accessible after the introduction of the dream concept by the noun sogno or the verb sognare.

350

Alexandra Kratschmer and Adrienne Heijnen

The informants were asked to give an actual dream account as well as to create a fictive dream account on the basis of a prefabricated mini-plot in key words about a dog chasing them (i.e. containing actions succeeding each other and thus potentially reflecting clearly any oniric aspect use). Both in the genuine dream accounts and the fictive ones, the subjects used mainly either present tense or perfective tense forms, with only isolated occurrences of the oniric imperfect (and only by one informant). Asked about it post hoc, they knew the concept and judged dream accounts based on the oniric imperfect as "more beautiful'', "more poetic", "more correct", "more story", while the normal aspect was "more direct", "more effective", "quicker", obviously meaning that it did not require a genre related elaboration effort. The informants confrrmed further that the Italian conjunction come 'how' following vedere 'to see' cannot be used to emphasize the perceptual process regarding actions in progress as the Dutch, Danish and German equivalents can, but only can refer to the manner in which an action is performed (hence the lack of it in our biblical corpus). We can conclude that, also what our Italian informants are concerned, the tendencies found in our diachronic Bible corpus were at least not contradicted. This was true for the weakening of the expression avere un sogno 'to have a dream' in favour of fare un sogno 'to make a dream', as well as for the not at all unchallenged position of the oniric imperfect as an obligatory marker of dream accounts.

4.2. Anthropological results

4.2.1. Written data- The Icelandic case Supported by the written sources, listed in note 4, we argue that the idea of dreaming being a means to experience temporal and spatial realms that are hidden from the waking mind can be traced continuously from the first Icelandic sources written down in the 13th century until the present day. Christianity, foreign influence and modernisation have, of course, generated change in Icelandic society, and it should be noted that the narrative form, dream content and the actors involved in dreaming, dreamtelling and interpretation vary from period to period. However, the basic theory of dreaming being a revealing activity is persistent. In the Old Norse literature, dreaming is primarily related to revealing the fate of the dreamer or the persons in his or her direct environment. In

Revelative evidentiality in European languages

351

the 16th to 18th century, many dreams are revelations, as we lrnow from other European Christian divine literature and provide guidance in Christian ethics, while few dream references found in the rimur (metrical romances) and folktales witness of dreaming as a gateway to hidden realms. In the literature, the realness of other realms is often proved by the dreamer's ability to bring something from these realms to the waking world. This can be lrnowledge, objects (e.g. Skioarima in Homan 1975: 197-216) or a change of physical state, for example by becoming pregnant (e.g. Kotludraumur infP.JE 1954-1961,1: 59). Two surveys, conducted respectively in 1974 (Haraldsson 1978) and in 2006 and 2007 (Amalds et al. 2007), show the persistence and even slight increase of an inclination towards considering dreams as revelations in present day Iceland. People were asked whether they considered themselves being berdreyminn49 , the results being 36% of positive answers for 1974, 39% for 2006 and 42% for 2007 (Amalds et al. 2007).

4.2.2. Oral data 4.2.2.1. Results of interviews, dialogues and participant observation among Icelandic native speakers

Also our ethnographic data shows the strong continuing presence in modem Iceland of the theory that dreaming does not involve a withdrawal within the self, but social engagement whereby hidden lrnowledge is revealed (Heijnen 2005a, 2005b, in press). The following example from one of Heijnen's keyinformants, who worked at the time of the dream as night watch at a home, illustrates the revelative role of dreams in contemporary Iceland: Mig dreymdi ao eg vreri50 i vinnunni og eg var ao klifra upp einhvern vegg, ofan a bakinu amer sat su sykursjuka, og pao sem eg var ao gera, var ao eg var ao lyfta henni upp i sykri, fyrir ofan vegginn var einhverskonar op, eoa gluggi og eg purfti ao kama henni pangao inn, fyrir innan gluggann var starfsf6lk, eg mann bara einn starfsmanninn og pao var hun Rosa. [It dreamed me a dream that I was at work and I was climbing up a certain wall. On my back was the person with diabetes and what I was doing was, that I was raising her level of insuline. High at the wall was a kind of radiator, or window, and I needed to get her in there. On the other side of the window were staffmembers, I only remember one staff member and that was Rosa.] (Icelandic man, b. 1969, Reykjavik, 21 January 2000)

352

Alexandra Kratschmer and Adrienne Heijnen

The dream refers to a Diabetes patient. The informant said that the night he dreamed his dream, this patient got an insulin chock due to a wrong level of medication. Other dreams about this patient followed this first dream, encouraging the dreamer to ask for further medical examination. Our data contain many examples, where people follow the wellbeing of others through their dreams and gain knowledge about their own life or that of others. However, despite for persistence of the theory that dreams are revelative, Heijnen's ethnographic studies in the period 1996-2009 demonstrate an increase in the inclusion of personal feelings and emotions in Icelanders' dream accounts and interpretations, as well as a slight change in epistemology. Rather than experiencing the dream solely as originating from an external source, a few dreamers mentioned that they picked up unconsciously hidden hints in daily life, which were responsible for the generation of revealing dreams during the night. Against the background of Aikhenvald's claims about Yukaghir dream theory, we argued earlier (see 2.1.) that to understand the interrelation between local attitudes and practices concerning dreaming, on the one side, and language use, on the other, as well as to support our thesis that evidentiality should be approached from an interactive perspective, it is crucial to consider notions of personhood (including ideas about the "soul") and agency. As such, the Icelandic dream theory, which recognizes that dreams are not taking place within persons' head but involve engagement with other realms, has radical implications for understanding the use of linguistic markers of evidentiality. While according to a mainstream Euroamerican psychological point of view, dreamers create their dreams and are the only agents in the dream, in Iceland the agents in dreams might be the dreamer, but also the dead and other beings who appear in dreams, with or without counterparts in the waking world. Telling a dream is therefore not necessarily an act of uttering personal sentiments, anxieties or perceptions, but might solely involve the communication of information (visually experienced in the dream) about the life of others. Also, as for the Yukaghir, while the dreamer might be observed passively lying asleep, his or her "soul" might be actively engaged in conscious "frrst-hand" experience in the dreaming realm. Thus, to conclude, our data shows a strong parallel between the persistence of the double accusative construction mig dreymdi draum 'it dreamed me a dream' and the continuing importance of the theory that dreams have the potential to reveal knowledge, otherwise hidden for the waking mind51

Revelative evidentiality in European languages

353

4.2.2.2. Results of interviews with German native speakers For our subjects, dreams are products of their own psyche, processing phenomena of real life impressions and (pre)occupations. They are not revelations coming from whatever source outside themselves. Even as products of their own psyche, dreams do hardly directly influence their choices in the woken life, at most could give occasion for reflection. Nevertheless, they all knew other persons from their nearest community, most female, to whom dreams serve as guidelines or objects of speculation or preoccupation, either because of their supposed psychoanalytic value or their status of being messages from supernatural realms. At least what our few subjects is concerned, we can conclude that the linguistic and the anthropological data provided by them are highly consistent: the attitude towards dreams and dreaming as phenomena produced by the dreaming subject's psyche seems to be perfectly mirrored by the diachronically successful verbal valency structure where the verb triiumen 'to dream' selects for a nominative (agentive) subject representing the dreammgperson.

4.2.2.3. Results of interviews with Italian native speakers As it was true for our German subjects, our Italian subjects equally saw dreams as products of their own minds. One of our subjects reported having had several premonitory dreams having come true later on. The source of these dreams was seen by her as a supernatural sphere consistent with her catholic faith, a conceptualization not infrequent in Italian society where Catholic faith is a social parameter of a certain weight. Nevertheless, she claimed to conceive herself as the creator of the actual dream, of the form it takes, while it is uniquely its content, the message that comes from a source outside her. Italy has further a strong tradition, originating from the town of Naples, of dream interpretation regarding guessing the right numbers for different lottery institutions, the so called Smorfia. 52 Our subjects that had not experienced premonitory dreams did not attribute any importance to dreams, but again knew people that did, be it for psychoanalytic or religious, revelative reasons (catholic or less orthodox, as in the case of believers in the Smor.fia). Also concerning our Italian subjects we could conclude that the cultural attitude towards dreams and dreaming and the linguistic realization of the reference to this phenomenon is consistent to a certain degree, taking our subject's metaphysical subtle-

354

Alexandra Kratschmer and Adrienne Heijnen

ties into account: the dreaming person creates the dream (entirely or at least what its actual form is concerned, in case it is revealed from a supernatural source outside the dreamer), which can be said is mirrored by the loss of the construction form 'to have a dream' in favour of 'to make a dream'

5.

Conclusions

5.1. Empirical conclusions As far as the linguistic data analyzed is concerned, it is no surprise that no independent grammatical markers for revelative evidentiality could be found in the European corpus languages, Classical Greek, Latin, German, Danish, Dutch, Icelandic, Faroese, French, Italian and Spanish, nor in the Biblical Hebrew part of our written corpus. The only grammatical markers found were dependent on a previous marking of the oniric sphere by a lexical element of the etymological field of dream or dreaming. These markers were ones of imperfective verbal aspect, either as finite inflection or as participial/gerund forms, representing the dream content either as belonging to a sphere of alternative reality or as an action unfolding before the eye of the teller as well as the listener ('visual evidentiality'). Visual evidentiality in general proved to play a dominant role in European revelative evidentiality. This aspect can be marked by the use of perception verbs or conjunctions emphasizing the perceptive process (as in Dutch and Danish). As a traductorial heritage from the Hebrew original, the 'behold'-marker equally belongs to this sphere. As typical for European languages, inferential evidential aspects are often intertwined with aspects of epistemic judgment. Epistemic reserve is a part of dream accounts in the form of verbs like 'to seem' or expressions as 'it was (to me)' What the creative source of the dream is concerned, we can distinguish between linguistic structures that seem to mark an attribution of creative power and thereby responsibility for the dream to the dreaming subject and those seeming to mark its attribution to a metaphysical realm outside the dreaming subject. While a language as Icelandic shows a diachronically rather stable pattern of the dreaming subject in oblique case position, suggesting exposure to the action and not acting, we can pin down a diachronic process towards a marking that suggests more subject participation and

Revelative evidentiality in European languages

355

thereby responsibility in languages like German, Danish, French and Italian by the use of nominative subjects and/or the creation/production verb 'to make' With the exception of the lexical family 'to dream', 'a dream', revelative evidentiality in European Languages is not a semantic domain that has a specialized formal apparatus of markers at its disposition, but which presents itself more like a dynamic system that has to be constructed to the purpose, building on elements of visual and inferential evidentiality, of epistemic reserve and of comparison, and in some cultures of a seeming interaction between the dreaming subject and another realm. It is furthermore absolutely depending on the explicit lexical marking of the concepts 'a dream' or 'to dream' What the anthropological background of these data is concerned, we could witness a certain consistency between dream conceptualization "suggested" by construction types in a certain language and the cultural attitude of (some of) its speakers, a consistency that by no means should be considered as self evident though. This was true e.g. for Icelandic, whose double accusative construction matches perfectly well with the Icelandic attitude to dreaming as messages received, by the dreaming person, from another realm. German and Italian, for which there could be observed a diachronic development from linguistic constructions suggesting less subject participation and hence responsibility for the dream towards constructions seemingly mirroring a higher degree of both, speakers agreed on experiencing dreams as creations of their own mind, at least what their outer form was concerned. This field of investigation deserves by all means a larger scale attention, whereas the outcome hardly can be predicted at this point.

5.2. Theoretical conclusions In so far, as revelative evidentiality is perceived, in many historical and some contemporary cultures, as a phenomenon implying a metaphysical source outside the dreaming subject to whom a message is revealed, the interactive aspect of revelative evidentiality emerged as crucial. A large inventory of metaphors or metonymies for this passing on of information between the revelator and the dreaming subject could be inventarized in our corpus. These metaphors and metonymies stem from the sphere of again -visual perception, symbolic behavior and cognition.

356

Alexandra Kratschmer and Adrienne Heijnen

These data suggest that the interactive aspect might be important in evidentiality matters in general, and we would like to encourage typologists to follow this lead. It might be rewarding to see whether one can find, in the languages of the world with grammatical evidential marking, different evidential markers for e.g. 'overhearing something' vs. '(inter-)actively being told' or '(visually) perceiving' vs. 'being shown'

5.3. Methodological conclusions Based on the assumption that testing "claims" made by conventionalized and grammaticalized linguistic expressions against anthropological and, by the way, (neuro-, psycho- etc.) cognitive, empirical data represents a powerful heuristic tool, we have collected some first data regarding "claims" made by language on the cultural-cognitive basis of dream experiences and anthropological data on culture related attitudes to dreams and dreaming. We have focussed, in the first place, on Icelandic, German, and Italian. Independently of any concrete outcome, which in our case actually happened to show a certain consistency between linguistic representation and cultural attitude, we consider this method to be a fruitful guideline to both linguistic and anthropological studies. The merits of this method are twofold: furthering and preventive. On one hand, this method helps linguistics and anthropology (or cognitive sciences) to fertilize each other, taking advantage of the other discipline's results and emulating them by their own. On the other hand, this method helps linguists not to jump to cultural or cognitive conclusions based on mere linguistic data, while it helps anthropologists not to ignore relevant linguistic clues, circumventing the occasional "blind spots" of both disciplines. On the long run, an interdisciplinary investigation will further our understanding of the relation between thought, practice and language and will be able to put NeoWhorfian approaches to the thought/language problem to the test. At the linguistic level, it has proved able to further our understanding of the semantic category of evidentiality, especially regarding the factor of interactivity.

Revelative evidentiality in European languages

357

Notes 1.

2. 3. 4. 5.

6. 7.

8.

We are deeply indebted to the following scholars for sharing their expertise on the following languages: Dan Enok Serensen!Aarhus: Biblical Hebrew; Steffen Krogh/Aarhus: Biblical Hebrew & Early New High German; Jette PersianilAarhus: Classical Greek; Per Overgaard/Aarhus: Latin; Hemik Jergensen!Aarhus: Early Modem Danish; Sten Vikner/Aarhus: Early Modem Danish & Icelandic; Gunnar Hrafu Hrafnbjargarson!Tromse: Early Modern Danish & Icelandic; Susana Femandez/Aarhus: Spanish; Karsten Hvidtfeld/Aarhus: Modern German & Modem Danish, Peter Bakker/Aarhus: Old and Modem Dutch. All responsibility for errors is of course ours. We are equally deeply indebted to our Icelandic, Faroese, Dutch, German and Italian informants, as well as to an anonymous reviewer whose observations helped us to enhance the focus of this paper and to elaborate on some crucial issues. As we will show in the case of dream accounts. In the Aikhenvaldian sense. We thank the anonymous reviewer for pointing out to us, how far we actually stretch the understanding of this category. It was also Jakobson who, on that very same occasion, first used the term "evidential" in the way we understand it now, as a generic term for information source. Until then, it was understood as used by Boas, namely 'something for which there is evidence' (Aikhenvald 2004: 13). The ayibii can be literally translated as shadow and be physically manifested in dreams (Willerslev 2007: 57). The quoted examples stem from the following different dream accounts (which is sometimes also true for our exemplification of a particular phenomenon, when different languages or even different Bible editions in the same language choose different marking strategies for a particular verse): Joseph's direct speech account to his family about his own two dreams (37:6-7 and 37:9), the chief butler's direct speech account to Joseph (40: 9-11), the chief baker's direct speech account to Joseph (40: 16-17), the account of the Pharao's two dreams as told by the narrator's voice (41:1-7) as well as the Pharao 's own direct speech account of his two dreams to Joseph (41 :17-24). Literature studies included the following Old Norse sources, containing ca. 530 references to dreams (Kelchner 1935: 3). Eddukvreoi (the Poetic Edda); Landnamab6k (Book of settlements); islendingab6k (Book of Icelanders); Konungasogur (Kings' sagas); Fornaldarsogur (Sagas of antiquity); islendingasogur (the Icelandic family sagas) and Sturlungasaga. For the period 14th19th century, mainly rimur (metrical romances); vitranir (epiphanies) and folk tales were used and, for the modem society, personal collections of dreams; diaries; letters and dreamfora at radio, television and in newspapers, as well as surveys on beliefs and practices around dreaming and related phenomena

358

9. 10. 11. 12.

13.

14.

15.

16. 17.

18.

19. 20.

Alexandra Kratschmer and Adrienne Heijnen

(Haraldsson 1978; Arnalds et al. 2007 and the survey Simi 61. Draumar, fyrirburoir, spadomar oflceland's National Museum, 1985). Steffen Krogh!Aarhus, personal connnunication. Jette Larsen Persiani/Aarhus, personal connnunication. For different data regarding Faroese cf. note 46. The oblique case in Danish, a phonetically merged form of both prior accusative and dative, has probably to be analyzed as deriving from a historical accusative in this case, especially given the Icelandic situation, usually representing a conservative condition af Scandinavian; Steen Vikner/Aarhus, personal connnunication. In our Bible corpus Fl-F4, we exclusively found the "old or literary" (cf. Larousse [Guilbert 1977], s.v. songe) songe '(a) dream' and sanger 'to dream', but not reve and rever, the unmarked forms in contemporary French (cf. Larousse 1977, s.v. reve). The same version shows 'to make a dream' systematically in direct speech, reflecting the innovativity of the construction, felt probably as more informal at that time: il sogno che ho fatto 'the dream that I have made', 37:6, but also in 40: 8 and twice in 41 : 11, IT3. But there have been exceptions already in Old Icelandic, with a growing tendency from the middle of the 19th century (Jonsson, Eythorsson 2005: 224f.): it is true for many verbs with human semantic roles in accusative and dative case that speakers use nominative or dative for accusative and nominative or accusative for dative, or even dative or accusative for nominative; among a longer series of concerned verbs, 'to dream' is one of the most stable ones though, still standing strong with its use of the dreaming subject in the accusative, and dative preferred over nominative as an alternative to accusative (Jonsson, Eyth6rsson 2005: 232). Note also the form mig dreymdi sjalfa mig 'me [ace.] dreamed myself [ace.]', 'I dreamed about myself Which is completely ours; for e.g. Hrafnbjargarson!personal connnunication, there is no clear semantic role assignable to the human accusative subject in this or any of the other occurrences of double accusative constructions in Icelandic. Hinneh is a focus marker used in direct speech and narratives to draw the listener's or reader's attention to the piece of information following it. Its etymology is heavily discussed, but does as far as is known not comprise the idea ofvisuality (S.0rensen 2003: 418-419) ASV: Pharaoh dreamed: and, behold, he stood by the river. [lit. 'P. had a dream how he stood by the water'] Cf. also Plungian (2001: 354): "[ ... ]an evidential supplement can always be seen in an epistemic marker"; for this author, the opposite does not automatically hold in all languages, but following him, there exist "modalized evidential systems", viz. systems where evidential markers (grammatical, lexical, or

Revelative evidentiality in European languages

21.

22. 23. 24.

25. 26.

27. 28. 29. 30.

359

else) must code aspects of epistemic modality at the same time. This view of Plungian's is critized by Aikhenvald (2007, note 3), who argues that "in the same language one evidential may have en epistemic extension, and another one may not"). The inferential domain can comprise subdomains as circumstantial inferences (based on circumstantial clues), generic inferences (based on world knowledge) and conjectures (lacking any sensory evidential basis; cf. Squartini (2008: 924) on Romance languages). Concerning the exact evidential nuances of sembler/paraftre in French, cf. N.0lke (2001 :30), of sembrare/parere in Italian, cf. Kratschmer 2006 (synchronically) as well as Squartini 2006 (diachronically), and ofparecer in Spanish, cf. Cornillie (2007:118). ASV: and it was as though it budded [lit. 'and it seemed to me as though it became green']. ASV:for, behold, we were binding sheaves in the field [lit. 'it was [to] me as [ ... ]']. In a similar way as regarding the 'behold'-marker, this could equally be a traductorial issue: the original Hebrew form in the relevant verses is the conjunction k, with the possible readings 'as if, but also purely temporal 'when' Clines (1998: 348) gives as an example specifically Gen 40:10 and translates it into and when it budded. Note that the ASV writes and it was as though it budded. The Septuagint shows hosper, which can have the same two readings (comparative and temporal) ask and therefore did not cause problems of translation. The choice of the comparative reading 'as if in certain translated versions could therefore be due to inaccurate translation. ASV: In my dream, behold, a vine was before me [lit. 'It seemed [to] me in the sleep that[ ... ]']. This example combines the alternative reality reading with an epistemic verb. This observation together with the fact that parallel contexts also can be presented as contexts of mere epistemic reserve (e.g. 40:10: Oc (mig siuntis) som det blef grent, DA2), confirms, as we see it, the parallel analysis, proposed in Kratschmer (in press), of contexts of epistemic reserve and of alternative reality comparison (illustrated by two possible readings of the Italian verbs sembrare/parere 'to seem'), attributing their difference to a mere degree of epistemic force ('epistemic factor'): compare 'the computer seems broken' (epistemic reserve) vs. 'the computer seems sentient' (alternative reality comparison). ASV: In my dream, behold, a vine was before me [lit. 'it dreamed me that a vine were before me'] .. This use is also documented in our modern oral corpus, see 4.2.2.1. below. ASV: [... ]I also was in my dream, and, behold, three baskets of white bread were on my head. This use is equally documented in our modern oral copurs, see 4.1.1.2. below.

360

Alexandra Kratschmer and Adrienne Heijnen

31. E.g. Dardano and Trifone (1995: 354) for Italian, and Fernandez Ramirez (1986: 282) for Spanish. 32. Cf. Dardano and Trifone (1995: 354). The imperfective preterite is also used as a fiction marker in children's play (Fernandez Ramirez 1986: 275-276., Togeby 1982: 315), as well as in the French prostasis of the present irrealis hypothetic period: Si j'avais [imperfect preterite] de /'argent, j'acheterais [present conditional] une voiture 'if I had money, I would buy a car.' (Italian and Spanish use the imperfective subjunctive in this context). 33. Marking thus punctual, concluded actions: fr./sp./it. Colombe!Colombo naissaitlnacialnasceva [imp.] en/ne/1451, 'Columbus was born in 1451' 34. ASV: and lo, my sheaf rose up. 35. The interference of narrative conventions with evidentiality is also noted by Aikhenvald (2004: 380): "The narrative genre as a macro-convention typically overrides all our preferences. Narrative conventions thus serve to narrow down the polysemy of evidentials depending on person and other factors." 36. Jette Larsen Persiani!Aarhus, personal communication. 37. Per Overgaard/Aarhus, personal communication. 38. As correctly pointed out by the anonymous reviewer, these contexts do equally challenge the boundaries of the category of evidentiality, namely its (presupposed?) defmition as marking the source of a piece of information presented as a full proposition expressed in clausal form. Our contexts enumerated here do admittedly show pronominal combinations ('all this') and pro-phrases ('what he is about to do'), hence not semantcially full propositions but rather propositional variables (nominal and clausal). Nontheless, we still consider these contexts relevant to evidentiality matters, by two reasons. The first reason is that comparable contexts exist where the same revelative verbs are followed by semantically and syntactically full clauses, as in e.g. II Kings 8:13 (Jehovah hath showed me that thou shalt be king over Syria.; ASV). The second reason is that even in languages with grarnmaticalized evidentiality marking, these marker do not only occur in the standard contexts of full clauses, but can be attached to pro-phrasal elements, e.g. an interrogative marker standing in for a proposition (which by definition is uncomplete due to its interrogative status) as in the following example from Kham (mentioned in Aikhenvald 2004: 248: ''the question is directed by the speaker to the addressee (second person), but it 'has its origin outside the speech situation': the author of the question is 'he or she' "): karao di why REP [= reportative; note by AK + AH] '(He or she wants to know) why' 39. Information source marking verbs like speech verbs are often taken into consideration in evidentiality studies only when standing e.g. in parenthetical form, not when taking a clausal complement. It appears that this is due mostly to practical reasons, the body of data simply threatening to grow to unsur-

Revelative evidentiality in European languages

40.

41. 42.

43.

44.

45.

46.

47.

48. 49. 50.

361

mountable dimensions. This, of course, is not a valid reason for not investigating these structures at all, hence our decision to include revelative verbs here. We always chose the original etymology, even though not all lexemes are etymologically transparent any more to non linguist nativespeakers, as e.g. dan. abenbare 'to render bare in the open', being more or less interpreted as an abstract 'tell a secret', the same way as engl. reveal is no longer felt as to express 'to remove a veil from' by non linguist native speakers. This collection is part of the Oral Archives of the Ami Magnusson Institute of Icelandic Studies. The construction appears for example in LandnamabOk (Book of Settlement) of which the oldest manuscript survived dates from the 13th century. However, it is generally thought that LandnamabOk was created in the 12th century. This mixing up of nominative and accusative is recognized in Iceland as a widespread grammatical deficiency and is designated in Icelandic with the general term polfallssyki, literally, 'the illness of accusative' Also inversion between nominative and dative is a recognized problem and called pagufallssyki 'illness of dative' The following examples illustrate this: [ ... ] ao pau hj6n voru nyhirttuo; patti konunni sem maour ka:mi ao rnmstokknum. Pessi maour biour hana kama jlj6tlega a fa:tur (II>JE, I, 1954-1961: 17). '[ ... ] when the couple just had gone to bed, it seemed to the woman that a man had come to the bedside. This man asks her to get up quickly';[ ... ] svo dreymir mig nokkuo eftir, bara skamm eftir jaroforina, pa finnst mer ao eg fari gegnum dimm gong. '[ ... ] then it dreamed me, somewhat later, just after the funeral, it seems to me that I am going through a long dark corridor' (quote from an interview with an Icelandic woman of70 years old, 2001). He was unsure though whether it should not rather be an accusative: mich [ace.] hat getriiumt, which gives associations to the Icelandic case variations mentioned above. As a more anecdotical note, the linguist in this collaboration (AK), remembers her own mother using mir hat getriiumt still about 20 years ago, experiencing it as slightly affectated at that time, while her mother now claims never to use anything else but ich habe getriiumt. Regarding the concept of saliency, cf. e.g. Giora 1997; regarding a discussion of the concept of literal/"minimal" meaning (and among others saliency), cf. e.g. Ariel 2002. A statement not all Italians want to subscribe to though (Valentina Bambini, Pisa, personal communication). Cf. the parallel Icelandic construction mentioned in note 12. Ber ('naked' or 'revealed') dreyminn (masc. nom. adjective of the verb ao dreyma 'to dream') means to be able to dream dreams that reveal knowledge on past, present and future events; cf. the interactive evidential pattern 'to render bare in the open' mentioned in 5.1.1.3.

362 Alexandra Kratschmer and Adrienne Heijnen 51. Note, here, the use ofthe subjunctive mood (cf. 4.1.1.2.3). 52. The Faroese situation provides an interesting parallel: In Faroese, a development from the construction meg droymde 'it dreamed me [ace.]' to eg droymde 'I dreamed' can be detected (cf. Poulsen 1997 as well as Jonsson and Eythorsson 2005: 227). This linguistic change seems to run parallel with a change in attitude towards dreaming, whereby dreaming increasingly is thought to take place within people's head. Even though, for some, dreams have the potential to reveal knowledge, the successful integration of Christian doctrines in Faroese society has defined revealing dreams as superstition (see e.g. Joensen [1975: 139-147] for information on the role of dreams in Faroese society). 53. Following this tradition dating back to the 16th century (Cosentino [2003: 1]), different dream contents are associated with different numbers, conceived as secure bets for the next lottery extraction (e.g. mulino ad acqua 'water mill' = 7, mulino a vento 'wind mill'= 77; Cosentino [2003: 333]).

References Aikhenvald, Alexandra Y. 2004 Evidentiality. Oxford: Oxford University Press. 2007 Information source and evidentiality: vvhat can we conclude? In Evidentiality between lexicon and grammar; Mario Squartini (ed.), 209-282. Ariel, Mira 2002 The demise of a unique concept of literal meaning. Journal ofPragmatics 34: 361-402. Arnalds, Asdis A., Ragna Benedikta Gar6arsdottir and Fri6rik Jonsson 2007 Konnun a islenskri pjootru og truarviohorfum. Reykjavik: Felagsvisindastofuun, Haskola islands. Capelli, Gloria 2007 "J reckon I know how Leonardo da Vinci must have felt Epistemicity, evidentiality and English verbs of cognitive attitude. Pari: Pari Publishing. Clines, David J.A.I. (ed.) 1998 The dictionary of Classical Hebrew. Sheffield Acedemic Press. Comillie, Bert 2007 The continuum between lexical and grammatical evidentiality: a functional analysis of Spanish parecer. In Evidentiality between lexicon and grammar; Mario Squartini (ed.), 109-128. Cosentino, M. ( ed.) La vera smorfia napoletana. Sogni e numeri per vincere al lotto. 2003 Florence: Giunti.

Revelative evidentiality in European languages

363

Dardano, Maurizio, and Pietro Trifone 1995 Grammatica italiana con nozioni di linguistica. Terza edizione. Bologna: Zanichelli. Drosdowsky, G-unther ( ed.) et al. 1984 Duden- Grammatik der deutschen Sprache. 4. Auflage. Mannheim: Dudenver lag. Fernandez Ramirez, Salvador 1986 Gramatica Espanola. 4. El verbo y la oracion. Madrid: Arco Libros. Giora, Rachel 1997 Understanding figurative and literal language: The graded salience hypothesis. Cognitive Linguistics 7: 183-206. Guilbert, Louis ( ed.) Grand Larousse de Ia langue franr;aise. Tome sixieme. Paris: Libraire Larousse. Haraldsson, Erlendur 1978 Pessa heims og annars. Konnun a dulramni reynslu fslendinga, trUarviohorfum og pjootru. Reykjavik: Bokaforlagio Saga. Heijnen, Adrienne 2005a Dream Sharing in Iceland. Ph.D. diss. Arhus: University of Aarhus. 2005b Dreams, Darkness and Hidden Spheres. Paideuma 51: 193-207 Stregkoder, sparegrise og barnedab. Det islandske selv i 2007 entrepretWrskab og navngivning, Tidsskriftet Antropologi 55: 21-35. in press Name-giving by the dead. An argument against the geneticisation of relatedness, Social Anthropology. forth. The Life of Dreams. Sharing Sensory Experience in Iceland. Princeton: Princeton University Press? Hrafnbjargarson, Gunnar Hrafn 2004 Oblique Subjects and Stylistic fronting in the History of Scandinavian and English: The Role ofIP-Spec. Ph.D. diss., Arhus: University of Aarhus, Aarhus. Jochelson, Vladimir Illich 1905 Essay on the grammar of the Yukaghir Language. American Anthropological Supplement 7: 369-424. Joensen, Joan Pauli. 1975 F rereske sluppfiskere.Etnologisk undersegelse af en erhvervsgruppes liv. Torshavn: Feroya FrMskaparfelag. Jonsson, Johannes Gisli, and Th6rhallur Eythorsson 2005 Variation in subject case marking in Insular Scandinavian. Nordic Journal ofLinguistics 28 (2): 223-245. Kelchner, Georgia Dunham 1935. Dreams in Old Norse literature and their affinities in folklore. Cambridge: Cambrodge University Press.

364

Alexandra Kratschmer and Adrienne Heijnen

Kratschmer, Alexandra Che te ne sembra? Semantica e pragmatica delle costruzioni italiane 2006 con sembrare/parere. In Atti del XVI Congresso dei Romanisti Scandinavi, Copenaghen e Roskilde, 24-27 agosto 2005, Michel Olsen, and Erik Zwiatek (eds.). http://www.ruc.dk/cuid/publikationer/ publikationer/XVI-SRK-Pub/MMH/MMH07-Kratschmer in press Categorisation vs comparaison une question de quantification epistemique. Modele interpretatif semantico-pragmatique modulaire des constructions italiennes avec sembrare /parere. Cahiers Chronos. Nelke, Henning 2001 La dilution linguistique des responsabilires. Essai de description des marqueurs evidentiels il semble que et il paraft que. In Le regard du locuteur 2. Pour une linguistique des traces enonciatives, Henning Nelke (ed.), 17-34. Paris: Kime. Nuyts, Jan 2001 Epistemic Modality, Language, and Conceptualization. Amsterdam/Philadelphia: John Benjamins. Poulsen, J6han Hendrik W. (ed.) 1997 Feroyskoroab6k. http://obg.fo/ Plungian, Vladimir A. 2001 The place of evidentiality within the universal grammatical space. Journal ofpragmatics 33: 349-357. Squartini, Mario 2006 Hearsay and quotatives in the diachronic evolution of the Italian pare/sembra evidential pair. Abstract at the 39th annual meeting of the Societas Linguistica Europaea (SLE), 30th august - 2nd september 2006, Bremen. Investigating a grammatical category and its lexical correlates. In 2007a Evidentiality between lexicon and grammar; Mario Squartini ( ed.), 1-6. (ed.) Evidentiality between lexicon and grammar. Italian Journal of 2007b Linguistics 19 (1). 2008 Lexical vs. Grammatical evidentiality in French and Italian. Linguistics 46 (5): 917-947. Serensen, Dan Enok Bibelsk-hebraisk grammatik. Kopenhagen: Hovedland. Togeby, Knud 1982 Grammaire Franr;aise. Vol. II. Copenhague: Akademisk Forlag. Willerslev, Rane Soul Hunters. Hunting, Animism, and Personhood among the Siberia 2007 Yukaghirs. Berkeley: University of California Press.

Revelative evidentiality in European languages

365

Source material: Bible versions -books Danish 1550

1647 Dutch 1734

Icelandic 1841

Faroese 1949

Kong Christian the Third's Bible translation, primariylbased on Luther's German version. Facsimile-Edition. Det Dansk Sprog- og Litteraturselskab. 1917. Copenhagen. Bishop Svane's revised version of Hans Povl Resen's 1607 translation, based on the original languages. Copenhagen. Nederduytse bijbel. Dat is bet oude en nieuwe Testament na de Oversetting van D.M. Lutherus met Aile desselfs gelykluydende Texten, en Inhoud der Kapittele tot Dienst van de Christelyke Gemeentes, toegedaan d'Onveranderde Augsburgsche Confessie van verscheyde Drukfouten vebetert. t' Amsterdam gedrukt voor bet Lutherse Weeshuys en zijn te Bekoomen by de Eygenaar van 't Prvilegie van bet Psalm boek. Amsterdam 1734. Biblia. l>oo er heileg ritning. I 5ta sinni utgefin, a ny yfirskoooo og leiOrett, a0 tilhlutun ens islen6ska Bibliu felags. Vi6eyar Klaustri. 1841. BiBLIAN HIN HEILAGA SKRIFTIN i cy6ing Victor Danielsens fra 1949©. Endursko6a0 1974©. Hin Feroyski Bibliugrunnurin eigur upphavsrrettindini til hesa cy6ingina av Bibliuni.

Source material: Bible versions -online http://scripturetext.com [as accessed in december 2007; bibliographical indications as listed on the website, with varying degree of informativity, authors' comments in square brackets]: Hebrew Orig. Westminster code with vowels and consonants: The BHS Hebrew Bible, Leningrad Codex has been made available through the hard work of several organizations, including the Westminster Theological Seminary, University of Pensylvania CCAT, German Bible Society, and Unbound Bible. Septuagint The Septuagint /LXX /Greek Old Testament has been made available through the hard work of several organizations, including the Univer-

366

Alexandra Kratschmer and Adrienne Heijnen

Vulgata German 1545

1871 1905 1912 1951 Danish 1931

Dutch 1618

French 1707

1744 1859 1894 Spanish 1569 1909 Mod.

sity ofPensylvania CCAT, German Bible Society, and Unbound Bible. Biblia Sacra Vulgata [translated from Greek by Hieronymus, 4th cent. AD] Luther: Made available in electronic format by Michael Bolsinger at http://www.luther-bibel-1545.de (see here for the most recent versions in text and HTML format). It was converted to SWORD format by Matthias and Joachim Ansorg. This text has been modernized, and the Johannine Comma added to 1 John 5. 1871 Elberfelder Bible 1905 Elberfelder Bible 1912 Lutherbibel 1951 Schlachter Bible Det Gamle Testamente af 1931: copyrighted Etext (c)1931 Det Danske Bibelselskab. The Old Testament of 1931: copyrighted Etext (C) by Det Danske Bibelselskab 1931. (The Danish Bible Society). The New Testament is Public Domain. Staten Vertaling: BUBEL DAT IS DE GANSE HEILIGE SCHRIFT DOOR LAST VAN DE HOOG-MOGENDE HEREN STATEN-GENERAAL DER VERENIGDE NEDERLANDEN EN VOLGENS HET BESLUIT VAN DE SYNODE NATIONAAL GEHOUDEN TE DORDRECHT IN DE JAREN 1618 EN 1619 UIT DE OORSPRONKELIJKE TALEN IN ONZE NEDERLANDSE TAAL GETROUWELIJK OVERGEZET- BEVATTENDE AL DE KANONIEKE BOEKEN VAN HET OUDE EN NIEUWE TESTAMENT Based on electronic edition from http://www.coas.nl/ bijbel. La Bible David Martin, 1744. Downloaded from http://desmond. oshea.free.fr/BibleMartin!Bible%20Martin%20 1744/Bible%20Marti n%201744.html French: Osterveld 1744. Text dovvnloaded from www.theophilos.sk French: Darby. Text dovvnloaded from www.theophilos.sk. Version Louis Segond 1910 (LSG) -- Texte libre de droits. Spanish: Sagradas Escrituras (1569) Spanish: Reina Valera (1909) Spanish: Modern

Revelative evidentiality in European languages

1986

2005

367

Las citas biblicas son tomadas de La Biblia de las Americas © 1986, 1995, 1997 by The Lockman Foundation, La Habra, Calif, http://www.lockman.org. Usadas con permiso. Las citas biblicas son tomadas Nueva Biblia de los Hispanos © 2005 by The Lockman Foundation, La Habra, Calif, http://www.lockman. org. Usadas con permiso.

Italian 1649 1649 Italian Giovanni Diodati Bibbia 1927 Italian: Riveduta Bible (1927) English [for glossing purposes only] 1960 New American Standard Bible Copyright© 1960, 1962, 1963, 1968, 1971, 1972, 1973, 1975, 1977, 1995 by The Lockman Foundation, La Habra, Calif. All rights reserved. For Permission to Quote Information visit http://www.lockman.org http://www.dbg.de/[as accessed in nov. 2007 and jan. 2008] German Bibeltext in der revidierten Fassung von 1984. Herausgegeben von 1984 der Evangelischen Kirche in Deutschland. Durchgesehene Ausgabe in neuer Rechtschreibung. © 1999 Deutsche Bibelgesellschaft Stuttgart. Die Lutherbibel wurde in den Jahren 1952-1984 uberarbeitet. Gute Nachricht Bibel. Revidierte Fassung der ,,Bibel in heutigem 2000 Deutsch" Durchgesehene Ausgabe in neuer Rechtschreibung. © 2000 Deutsche Bibelgesellschaft Stuttgart. http://www.bibelselskabet.dk/[as accessed in feb. 2008] Danish 1992 Bibelteksten er hentet fra den autoriserede oversrettelse, © Det Danske Bibelselskab 1992. http://www.hverdagsbibel.dk/ [as accessed in feb. 2008] Danish Hverdagsbibel [Bibel in "every day language"]: den nye, gennemre2002 viderede udgave af hele Bibelen pa Hverdagsdansk udkom den 1. november, 2007. Den kan lreses online sammen med en rrekke andre oversrettelser fra International Bible Society pa www.bibleserver. com. http://www.snerpa.is/netlbiblia/biblia.htm [as accessed in dec. 2007] Icelandic Hio islenzka bibliufelag. Biblian Heilog ritning. 1815. [identical to 1815 the following 1966-edition: Biblia. Pao er heilog ritning. Ny pyoing ill frumma1unum. Reykjavik: A kostnao hins islenzka bibliufelags. Prentsmiojan Oddi H.F. 1966] http://www.laparola.net/[as accessed injan. 2008]

368

Alexandra Kratschmer and Adrienne Heijnen

Italian 1821 1974

Italian 2008

Nuova Diodati, Revisione (1991), edizione La Buona Novella Brindisi. [modernized Diodati-version, under continuous revision] La versione C.E.I., seconda edizione (1974), a cura della Conferenza Episcopale ltaliana. La Bibbia di Gerusalemme ha lo stesso testo, con l'aggiunta di alcune note prese dalla Bibbia di Gerusalernme francese. Euna versione approvata dalla Chiesa cattolica romana http://www.biblegateway.com [as accessed in jan. 2008] La versione C.E.I., a cura della Conferenza Episcopale ltaliana.

Subject index Adposition 65-67, 99, 109 Adverb 65f., 69, 78, 90, 92f., 98, 109, 112,115,117,134,143,149-152, 202,225,234f.,282f.,292,298,300, 304,311,314-318,320f.,325,328330,343 Agglutinative 7, 69, 72, 111, 118, 192 Ambiguity 8, 133, 138, 152-154, 177180, 191, 199 Anthropology 333, 336, 358 Auxiliary 67, 71, 78, 81f., 84, 138, 282, 289,291,313 Basque 133-155 Complementation 100, 190, 204, 217, 222,251 Complementizer 65-67, 77, 89, 92, 99f., 107f., 112, 117, 190, 205, 210, 213215,217,259,264,268-270,278 Conjunction 67f., 76, 98, 108, 114, 131, 155, 189,202,261f.,268,325,342, 352,356 Construction4, 18, 29, 40, 61, 66, 70, 75f., 82, 107, 117, 166, 176,210, 241,251,257,262,286,313,339341,357 Croatian 95, 98, 119 Danish 337, 339, 343, 356 Deixis 9, 229, 232, 239, 244 Embedding 165, 168,171, 174f., 182, 189, 191,207,213,216{. English 2, 5, 9, 24, 48, 85-87, 94, 101, 114, 133, 140, 163, 169, 191, 194,

200,202,251,254f.,259f.,264f., 278, 312, 337 Epistemic modality 46, 61f., 161f., 221223,230,235f.,308,312,321,339 Evidentiality auditory 33, 35, 91, 236 direct 26, 28f., 32, 34-37,41, 90, 169,222,228,272,279,309 hearsay 7, 10, 25, 29, 35f., 61-116, 134f., 138, 141, 144-146, 148150, 187, 192, 211,233, 236, 252,277,330 firsthand 28, 35, 37,331 indirect 26, 28f., 34-37,41,46, 69f., 74, 78, 82, 86, 91, 106, 115, 169, 191f., 272 inferential8, 10, 23, 30, 34, 36f., 40, 48,63,69f., 72, 74,77-80,84, 86, 91, 106, 112, 117, 154f., 182, 192-195, 205, 207, 209, 222, 311{.,325,339,352 interactive 227, 341, 344, 349, 353 personal 26, 29-32, 36f., 45, 223, 311 quotative 29, 35, 78, 98, 100, 102, 104,113,192,211,215,237,330 reportive/ reportative 29, 35-37, 39, 48, 61f., 66, 69, 74, 76, 78, 80, 82-85, 91, 93, 99, 103f., 108, 112, 115, 117, 160, 166-168, 179f., 192,194,198,201,209 revelative 25, 329, 334, 341, 346, 351 sensory 33f., 37, 330 third-hand 35-36 visual35-37, 40,223,231,236,277,

370

Subject index

331, 338f., 351 Evidentiality marking discoursive 330f. grammatical12, 16, 19f., 30, 36, 38, 63-65, 67, 74, 78, 82, 135, 140, 146, 165, 171, 191, 221,247, 330f.,341,351 lexical 3-8, 18, 62, 65f., 108, 115, 133, 135, 146, 149f., 165, 277, 330,341 Faroese 333, 335,338, 351 French 48, 80, 154, 204,234, 277, 283, 335f.,339,352 German 23, 48, 68, 77, 79, 81f., 85, 88, 103, 117,159f., 166f., 171,182,239, 247f.,250,254-256,260f.,267,272, 333,335,338f.,345f.,350f. Grammaticalization 3f., 22, 63, 6, 110, 125,264,278 hnperfect69, 73, 77,237f.,240,340f., 347 Italian 78, 80, 86, 112,230, 333, 335f., 339f., 346f., 350f. Lexicalization 2f., 63 Metaphor 251,253, 263, 266, 271, 353 Modal verb 3, 166, 182,273,282 Parenthetical66, 92f., 105f., 136, 152, 165,169, 174f., 178,231,258f.,270, 290f.,293,295,300 Particle 5, 7, 40, 65f., 76, 80, 92f., 96, 98f., 102, 105, 107, 111-113, 116, 135,140, 147,225f.

Perception verb 247-251, 256, 259, 269, 273,338,346,351 Person 42f., 72, 88, 104, 149, 254, 283, 289 Polish 84, 89f., 94, 97, 103, 106, 108, 112, 117 Polysemy 84,248, 251f., 252,259,273, 282 Portuguese 77, 80 Predicative 65f., 69, 88f., 112, 114, 206, 212, 214, 282 Presupposition 165,178-181,184 Reasoning 30, 35, 37, 192, 208, 290, 293,298,301,309,313,320,325 Romance 48, 67, 77, 79, 82, 87, 112, 115,117,221,230,235,335,340, 344,348 Russian 25, 31, 90, 97, 99, 102, 106, 108, 110, 112, 117 Scope 63, 65, 67, 89, 92,96-99, 105, 108, 135f., 150,152-154,198,264, 321,330 Spanish 80, 86, 94, 133-149, 153, 237f., 242,307,309f.,313,315,335,339, 349,351 Strategy, evidential10, 62, 135 Subordination 188, 190f., 208,214,219 System, evidential: modalized 7, 48 non-modalized 7, 48 Temporality 1, 4, 239 Tense 4, 7, 19f., 22f., 42, 49,70-74,78, 80f., 87, 118, 139, 191, 193,200, 203,228,242,288,344

Subject index

Turkish 8, 19, 27, 31, 63, 69f., 74, 111, 187,189-192,200,202 Volition 7, 44f., 81, 87f., 166, 249

371

Witness 28, 31, 37, 72, 102f., 136, 159, 192f., 195, 226, 263, 273, 280, 290, 292,297,301,311,352