Solomon A. Birnbaum: a Lifetime of Achievement: A Collection of His Papers (Palaographie)

Salomo/Solomon A. Birnbaum Ein Leben für die Wissenschaft/A Lifetime of Achievement II: Paläographie/Palaeography

Ein Leben für die Wissenschaft/ A Lifetime of Achievement Wissenschaftliche Aufsätze aus sechs Jahrzehnten von/ Six Decades of Scholarly Articles by

Salomo/Solomon A. Birnbaum herausgegeben von/edited by Erika Timm in Zusammenarbeit mit/with the Collaboration of Eleazar Birnbaum und/and David Birnbaum

II: Paläographie/Palaeography

De Gruyter

ISBN 978-3-11-025227-9 e-ISBN 978-3-11-025228-6 Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http://dnb.d-nb.de abrufbar.

© 2011 Walter de Gruyter GmbH & Co. KG, Berlin/Boston Gesamtherstellung: Hubert & Co. GmbH & Co. KG, Göttingen Satz: Michael Trauth, Trier

∞ Gedruckt auf säurefreiem Papier Printed in Germany www.degruyter.com

Contents

Preliminary Note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

VII

Solomon Birnbaum’s Life and Work . . . . . . . . . . . . . . . . . . . . . .

IX

Scholarly Articles by Solomon Birnbaum A. The Palaeo-Hebrew Script . . . . . . . . . . . . . . . . . . . . . . . . . . . 28. 29. 30. 31.

1

Table of Semitic Alphabets. 10th – 5th Centuries B. C. E. . . . On the Possibility of Dating Hebrew Inscriptions . . . . . . . . . The Dates of the Gezer Tablet and of the Samaria Ostraca . . . The Palaeo-Hebrew Ostraca Sherds and a Fragment of a Stele, all from Samaria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32. The Lachish Ostraca . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33. The Leviticus Fragments from the Cave . . . . . . . . . . . . . . . .

3 7 13 17 53 79

B. The Hebrew Script . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

89

34. The Development of the Hebrew Scripts . . . . . . . . . . . . . . . . 35. Methodology in Hebrew Palaeography . . . . . . . . . . . . . . . . . 36. Notes on the Internal and Archaeological Evidence Concerning the Cave Scrolls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37. The Dates of the Cave Scrolls . . . . . . . . . . . . . . . . . . . . . . . . 38. How Old Are the Cave Manuscripts? – A Palaeographical Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39. The Qumran Scrolls and Palaeography . . . . . . . . . . . . . . . . . 40. The Date of the Isaiah Scroll . . . . . . . . . . . . . . . . . . . . . . . . . 41. The Date of the Habakkuk Cave Scroll . . . . . . . . . . . . . . . . . 42. The Date of the Covenant Scroll . . . . . . . . . . . . . . . . . . . . . . 43. The Age of the Cave Scrolls: Two Letters to the London Times 44. The Date of the Hymns Scroll . . . . . . . . . . . . . . . . . . . . . . . . 45. The Date of the Incomplete Isaiah Scroll from Qumraˆn . . . . 46. Bar Kosba (Kokhba) and Akiba . . . . . . . . . . . . . . . . . . . . . . . 47. Akiba and Bar Kosba (Kokhba) . . . . . . . . . . . . . . . . . . . . . . .

91 137 147 153 157 179 191 197 207 217 219 231 241 253

VI

48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61.

Contents

The Beth Mashku Document . . . . . . . . . . . . . . . . . . . . . . . . . A Fragment in an Unknown Script . . . . . . . . . . . . . . . . . . . . An Unknown Aramaic Cursive . . . . . . . . . . . . . . . . . . . . . . . . The Negeb Script . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Kephar Bebhayu Conveyance . . . . . . . . . . . . . . . . . . . . . The Kephar Bebhayu Marriage Deed . . . . . . . . . . . . . . . . . . . A Sheet of an Eighth Century Synagogue Scroll . . . . . . . . . Die Vielfalt der hebräischen Schrift im Mittelalter . . . . . . . . Die Schrift des Dukus Horant . . . . . . . . . . . . . . . . . . . . . . . . Manuscripts in Old Yiddish . . . . . . . . . . . . . . . . . . . . . . . . . . Design for an Ornamental Ashkenazic Printing Type . . . . . . Das Datum des Codex Zimt Sand (Maase Bukh) . . . . . . . . . Palaeographical Examination of the ‘Klein Manuscript’ . . . . A Scrap from a Hebrew Manuscript of a Kita¯b al-Mala¯him / {Kita¯b fı¯ ilm al-jafr} . . . . . . . . . . . . . . . . . . . . . . . . . . .˙. . . . 62. The Etymology of the Term Mashait . . . . . . . . . . . . . . . . . . .

257 273 277 299 333 363 377 385 387 395 401 403 409 417 425

Index of Persons (in Chapters 28– 62) . . . . . . . . . . . . . . . . . . . . . 455

Preliminary Note

When many of the articles in this volume were first published, Hebrew palaeography was not yet considered a reliable scholarly discipline, in particular by certain well-known Biblical scholars and archeologists, whose views were all the more strongly expressed because they disagreed on the dating of the Dead Sea Scrolls etc. Shortly after the Dead Sea Scrolls were discovered in 1947, but long before their publication, photographs were sent to Dr. Birnbaum to estimate their dates and authenticity. Relying on the palaeographic method that he had developed, he established dates which, though initially disputed by some scholars, were later confirmed by Carbon 14 physical testing and external archeological analysis. His dating of the Scrolls is now universally accepted by scholars. His most significant major publication on palaeography is The Hebrew Scripts (2 vols., London 1954–1957 and Leiden 1971), containing hundreds of illustrations of dated Hebrew writing, a detailed palaeographical examination of each letter of the alphabet in each illustration, and an explanation of its place in the overall pattern of the development of Hebrew scripts. More recent advances in the field of palaeography, based largely on Dr. Birnbaum’s pioneering principles (whether acknowledged or tacitly accepted) have resulted in further progress. His methods, though revolutionary when he first enunciated them, are now widely regarded as norms, which have in some areas been further refined by new generations of researchers who have to a great extent, so-to-speak, stood upon his shoulders. The present Editors believe that Dr. Birnbaum’s articles should be read in the context of the time when they were written. Consequently the Editors have not attempted to update them in the light of later developments in the field. For a more detailed description of Dr. Birnbaum’s work on Hebrew palaeography see below, p. XX –XXIV . We would like to take this opportunity to express our gratitude to the libraries and archives which provided photographs for the illustrations appearing in this volume. Eleazar Birnbaum, David Birnbaum

Solomon Birnbaum’s Life and Work1

Overview: Solomon Birnbaum is the undisputed pioneer in two major, closely related fields of research: in historic Yiddish linguistics and in the palaeography of Hebrew and other Jewish languages. He was born in Vienna on December 24, 1891, where he attended high school, with the final two years in Czernowitz where he obtained his baccalaureate (‘Abitur’) in 1910. He then studied architecture in Vienna, but also began to devote himself to Semitic studies. From 1915 he served in the Austro-Hungarian army for three and a half years (achieving the rank of lieutenant with distinction). After the war he resumed his oriental studies, from which he graduated with a doctorate in Würzburg in 1921. From 1922 to 1933 he taught Yiddish at the University of Hamburg and became interested in Hebrew palaeography. In 1933 pressures from the Nazi regime forced him to emigrate to London with his family, where he taught Yiddish and Hebrew palaeography at two Schools of the University of London until his retirement in 1957. In 1970 he and his wife Irene moved to their sons in Toronto, where he died on December 28, 1989. – In detail:2 1

2

The works of Solomon A. Birnbaum (occasionally: S. A. B.) briefly quoted below are listed in detailed form in the bibliography (volume I, p. XXIX – XLVIII ) under the year of publication; the Yiddish language articles written in Hebrew script (in the second part of the list), are referred to with an “h” before the year. Of the numerous articles about S. A. B., which provide information about various aspects, we wish to quote the following: “Birnbaum, Solomon Asher”, in: Encyclopaedia Judaica. (Second Edition, Detroit 2007), with literature. – Hugh Denman, “Terumato shel Shlomo Birnbaum labalshanut hayidish”, in: Ha-Sifrut 10 (1986), p. 252–262 [Hebr.]. – Walter Röll and Erika Timm, “Laudatio”, in: Verleihung der Würde eines Ehrendoktors der Universität Trier an Professor Dr. Salomo Birnbaum, 4. Juni 1986. [ Trier 1986], p. 10–14, reprinted as “In Memoriam Salomo Birnbaum”, in: Jiddistik Mitteilungen 3 (1990), p. 16–22. This Laudatio has been partially incorporated into the above article. – Shmuel Hiley, “Solomon A. Birnbaum”, in: Dov-Ber Kerler (ed.), History of Yiddish Studies. Winter Studies in Yiddish 3 (1991), p. 3–13. – Dovid Katz, “Shloyme Birnboym. 1891–1989”, in: Oksforder Yiddish 2 (1991), p. 271–276 [In Memoriam]. – Peter Freimark, “Juden in der Hamburger Universität”, in: Eckart Krause [et al.] (eds.), Hochschulalltag im Dritten Reich. Berlin / Hamburg 1991, p. 125–147 (especially 129–133 and 144–145). – Utz Maas, Verfolgung und Auswanderung deutschsprachiger Sprachforscher 1933–1945. Vol. I, Osnabrück 1996, p. 201–204, Art. “Salomon

X

I.

Solomon Birnbaum’s Life and Work

Yiddish Philology

Solomon Asher Birnbaum3 was the oldest son of the writer and publisher Dr. iur. Nathan Birnbaum.4 Many of his ancestors had been rabbis in Germany, in Galicia and Hungary and had thus been native speakers of Yiddish. However, like his father, Solomon grew up in Vienna with German as his mother tongue. Nathan Birnbaum (1864–1937) was one of the leading intellectual spokesmen of Judaism of his time. As a young man, he had turned his back on assimilation and henceforth resolutely regarded the Jewish people as a nation, not as a merely religious community. With this in mind, he had coined the term ‘Zionism’ long before Theodor Herzl.5 Unlike Herzl, Nathan Birnbaum wished to achieve complete, ethno-cultural equality for his people in the diaspora, at first within the multicultural Austro-Hungarian empire. In this context he

3

4

5

A. Birnbaum”. – Art. “Solomon A. Birnbaum” (contributed by David J. Birnbaum), in: John M. Spalek / Sandra H. Hawrylchak, Guide to the Archival Materials of the German-speaking Emigration to the United States after 1933. Vol. 3.1, Bern [et al.] 1997, p. 57– 67 [about The Nathan & Solomon Birnbaum Archives (now located in Toronto), which were established by Solomon B.]. – Christopher M. Hutton, Linguistics and the Third Reich. London / New York 1999, especially “Solomon Birnbaum”, p. 197–205. – Art. “Tsu Sh. Birnboyms elftn yortsayt”, in: Afn Shvel 320 (2000), p. 1–8. – Susanne Blumesberger, Art. “Birnbaum, Salomo”, in: Internationales Germanistenlexikon 1800–1950. Vol. 1, Berlin / New York 2003, p. 190–192. – Kalman Weiser, “The ‘Orthodox’ Orthography of Solomon Birnbaum”, in: Jonathan Frankel (ed.), Dark Times, Dire Decisions. Oxford 2004, p. 275–295. – David Birnbaum, “Der Nosn un Shloyme Birnboym-Arkhiv in Toronte”, in: Afn Shvel 344–345 (2009), p. 41– 44, David Birnbaum, “The Nathan and Solomon Birnbaum Archives, Toronto”, in: Kalman Weiser / Joshua A. Fogel (eds.): Czernowitz at 100. Lanham MD 2010, p. 181–188. When writing in German, B. signed Salomo (occasionally: Salomo A.) Birnbaum, and in English he used the signature Solomon A. Birnbaum (rarely without the “A.”). In Yiddish it was Shloyme Birnboym (or, using his own transcription, S´loimy Biirnboim). Basic information about Nathan Birnbaum is contained in Art. “Birnbaum, Nathan”, in: Encyclopedia Judaica. Jerusalem 1972 (Second edition, Detroit 2007), complete with literature. According to David Birnbaum, this article, which is marked with “Ed.”, was written by S. A. B. – Furthermore: S. A. B.: “Nathan Birnbaum”, in: Leo Jung (ed.), Men of the Spirit. New York 1964, p. 519–549, which also contains quotations from obituaries (1937) by a dozen illustrious contemporaries. – Joshua A. Fishman, Ideology, Society and Language: The Odyssey of Nathan Birnbaum. Ann Arbor 1987. [With 15 selected articles by N. B. from the period of 1890 to 1931 (translated into English).] 1890 in the journal published by him entitled Selbst-Emancipation, vol. 3, No. 4 dated May 16, 1890 (and others); cf. S. A. B. 1964 (above, n. 4), p. 523.

Solomon Birnbaum’s Life and Work

XI

discovered the Yiddish language as the unifying force of Eastern European Jewry. In 1908 he convened the now legendary conference on the Yiddish language in Czernowitz, in which Solomon also participated at the age of 17.6 Eventually Nathan Birnbaum distanced himself from Zionism and turned to Orthodox Judaism. Many aspects of Solomon’s ideas were rooted in those of his father, though not in the form of a Publizist but rather that of a scholar. Solomon started learning Yiddish while still in high school. From 1906 onwards he began to correspond with his father in Yiddish.7 In 1907 he published a story by David Pinski, which he had translated into German, in the Neue Zeitung (Vienna) under the pseudonym Ben Acher 8 and in 1913/14 he published an extensive series of Yiddish songs and excerpts from Mendele in the journal published by Fritz Mordechai Kaufmann entitled Die Freistatt, in a scholarly transcription designed by him, complete with translation. At an early age he displayed the aptitude and skills to produce accurate and stylistically valid translations of Yiddish literature into German: he practised this skill mainly in the nineteen-twenties, when he created his excellent Mendele translations – among others.9 Birnbaum’s first major scholarly work was a Yiddish grammar, which he completed in 1915 and which was published in Vienna three years later.10 To date, this first work has not received the theoretical 6 7 8

9

10

Cf. David Birnbaum in: Afn Shvel (2009), p. 42 f., also in: “Czernowitz at 100”, p. 183 f. (both above, n. 2). Information kindly provided by David Birnbaum. I. e. ‘the son of Mathias Acher’, his father’s pseudonym. Until the 1930s (at least) Solomon used a variety of different pseudonyms (easily recognizable by anyone), including for instance Shaban / ñ`baw (= Shlomo Asher ben Nathan), and also rwa (Asher); sometimes only the letters Sch. / .w or B. / .b; in the case of very short contributions the words wnÅ rbÅ (‘human being’). In the list of publications (in volume I, p. XXIX –XLVIII ) under 1907, 1913/14, 1920, 1924 (and 1961 reprint). Conversely he translated Max Brod’s novel Tycho Brahes Weg zu Gott into Yiddish (h1921). The Yiddish translation of Bilder aus Palästina by Davis Trietsch (1911 and later edd.), which was published in Yiddish in 1921 by the same publisher without identifying the translator, was, according to David Birnmaum, in fact done by S. A. B. He also translated parts of Glückel Hamel’s Yiddish memoirs (1691–1719) into modern Yiddish and excerpts of these were published in the magazine Der Jid (1922). The manuscript was completed in the summer of 1915. In the foreword to the fourth edition B. describes the difficulties in getting it published during the wartime period of 1916 to 1918. In many library catalogues the date of publication is erroneously stated to be 1915.

XII

Solomon Birnbaum’s Life and Work

recognition it deserves as a milestone in the history of scholarship. It was the first coherent scholarly grammar of the Yiddish language ever produced. Due to its informational content this is a book which, even after many years of use, one continues to admire for the soundness and accuracy of each of its many observations. In particular, from a historical point of view, it is an early masterpiece of a strictly synchronicimmanent description of language. With admirable assurance Birnbaum managed to withstand the temptation of permitting his descriptive categories to be influenced by the German written language, by school grammar of Latin or by diachronic considerations. And last but not least: his classification of East Yiddish dialects outlined here continues to be of fundamental relevance to this day.11 In sum, the book has retained all its freshness. Solomon Birnbaum obviously possessed the ideal prerequisites to produce such a successful creation: an opus distinguished by such amazing assurance in its overall concept as well as by such soundness of detail could only be written by an individual who displayed towards the Yiddish language the unconditional love associated with homecoming, but coupled with the objective detachment of the non-native speaker. The work was re-printed by the publisher Helmut Buske in Hamburg in 1966, and up to 1988 there were three further reprints. It is currently out of print. A few years years after producing the Grammar, while still recovering from his war wounds in hospital, Birnbaum wrote his second monograph in 1918:12 another synchronic, but far more detailed description of the Hebrew-Aramaic component of the Yiddish language. This aspect of the Yiddish language was described here for the first time ever, in no more than fifty pages. Nevertheless this is neither a superficial outline nor merely a commendable basis for future work. The material penned by Birnbaum almost ninety years ago, without the benefit of relevant scientific tools and under unusual circumstances, can still claim complete relevance to this day. When he completed regular courses of oriental studies at the universities of Vienna, Zurich, Berlin and Würzburg after the war, this thesis, with some minor revisions, was accepted as a doctoral dissertation in Würzburg in 1921 and was printed in 1922. In volume I of the present publication two previously unpublished 11

12

Already correctly evaluated by Max Weinreich, Shtaplen, Berlin 1923, p. 14– 15. Cf. also Marvin Herzog, “Channels of Systematic Extinction in Yiddish Dialects”, in: For Max Weinreich on his Seventieth Birthday. The Hague 1964, p. 93–107, especially p. 93. It is this book that Birnbaum wrote while in hospital (not the Grammar, as is frequently reported).

Solomon Birnbaum’s Life and Work

XIII

articles (Chapters 3 and 4, from the period prior to and around 1930) are included, “Der Vokalismus des Talmuds in aschkenasischer Überlieferung” and “Hebräisch und Jiddisch”, which revisit the topics of the dissertation and reinforce the earlier findings with comprehensive material. After obtaining his doctorate Birnbaum moved to Hamburg, where on the initiative of the Germanist Conrad Borchling he was offered a lectureship for Yiddish at the university as part of the “General Lectures” for the winter semester 1922/23. He remained in this position for more than a decade and, looking back, he said that here for the first time “the Yiddish language had been taught as a subject at a modern German university”.13 In addition to lecturing14 Birnbaum intensified his Yiddish research activities, focused primarily on history, which he continued up to an advanced age. These are so extensive that we are unable to acknowledge them in detail here. In subsequent decades he wrote a significant number of encyclopaedia articles; it should be emphasized that by this means basic information about the Yiddish language and literature was disseminated far more effectively than previously. He has no equal in his German language contributions towards improving the general knowledge and reference options in this field.15 We wish to highlight some topics of major relevance, which are also discussed in volume I. First: Etymology (Chapters 15, 16, 17; cf. also Chapter 62 in volume II); its main thrust is to caution against careless derivation from the Hebrew. Second: Origin and Age of the Yiddish language (Chapter 8), in which Birnbaum, for the first time, traces the history of Yiddish as far back as the 9th / 10th century.16 Third: Yiddish and the German dia13

14 15

16

“[. . .] die jiddische Sprache im Rahmen einer modernen deutschen Universität Lehrgegenstand wurde”. Cf. Freimark (above, n. 2), p. 130: from a letter by Birnbaum to Freimark dated 12.7.1983. The topics of the lectures (see p. 299 in volume I) indicate the broad spectrum of his current research projects. Cf. Chapter 5 in volume I, also in the list of publications under 1929, 1931, 1932; English 1987. According to information supplied by David Birnbaum, S. A. B. wrote about 160 mostly short articles on Jewish personalities and terminology for the encyclopaedia Großer Herder (4th edition, Freiburg i. Br. 1931–1935), which were partially altered by the publishers, often contrary to the author’s intention. In his personal copy S. A. B. usually listed the articles contributed by him and in some of them the editorial revisions were identified. In the Großer Herder, articles generally remain unsigned. A short summary already appears in Chapter 5 (paragraph entitled “History”) of 1929; cf. also the relevant paragraphs 1974, p. 1–16, and 1979 (Survey),

XIV

Solomon Birnbaum’s Life and Work

lects (Chapters 9, 10, 11). Fourth: Linguistic analysis of a variety of Old Yiddish texts (Chapters 7, 11, 12, 13, 14).17 In this context Birnbaum’s discoveries about the historical development of phonetics have gained increasing importance. The article on “Geschichte der u-Laute im Jiddischen” (Chapter 11) deserves special mention.18 Here Birnbaum, with the aid of extensive documentary evidence spanning eight centuries, convincingly substantiates his theory that one of the most striking characteristics of Southern East Yiddish, namely the fact that / i / and / i: / corresponds to mhd. u and uo, originated in Germany, but subsequently disappeared from there after a pronounced initial phase (in the 14th –16th century).19 One particular issue was of special concern to him since the beginning of his Yiddish studies: A radical reform of orthography, which was ‘wild’ at the time, as well as a phonetically appropriate transcription of the spoken language into Latin script. In this context, right from the start his approach was established »on the basis of the u-dialect [i. e. the pronunciation in the Middle and Southern regions of East Yiddish] which comprises the majority of the speakers,20 is associated with the classical authors of the language and predominates in the theatre«21 – as opposed to Yivo’s efforts towards a standardisation of the language which started shortly afterwards, where preference was given to the northern pronunciation under the influence of the intellectuals from the North. For Birnbaum the ‘orthography / transcription’ complex was associated with strong ideological emotions. In his view the Yiddish language was

17

18 19

20 21

p. 44–57, as well as h1931, h1966 and, incidentally, several articles on Old Yiddish. Regarding Old Yiddish texts see also the list of publications under 1932, h1922 (Glikl Hamel), h1931, h1932 and h1938 (songs), h1939 and h1941 (letter), h1964 (psalms) as well as the relevant paragraphs 1974, p. 63–104, and 1979, p. 145–189. 1934 in Yiddish, 1981 in expanded form, in German. Since then the relevant trends in Late Medieval German have been explored in more detail in Germanistic research (s. Ernest Beyer, La palatalisation vocalique spontane´e de l’Alsacien et du Badois. Sa position dans l’e´volution dialectale du germanique continental. Strasbourg 1964), as a result of which a genetic relationship between the development of Yiddish and German can no longer be dismissed. In the decades prior to the Second World War “approximately three quarters” (as early as 1918 in the Grammar, p. 16, and elsewhere later). 1923 (“Übersicht”), p. 122: “[. . .] auf der Grundlage des u-Dialekts aufgebaut, der die Mehrheit der Sprecher umfaßt, dem die Klassiker der Sprache angehören und der auf dem Theater herrschend ist”; correspondingly mentioned in the Grammar, p. 16, and elsewhere later.

Solomon Birnbaum’s Life and Work

XV

fundamentally tied to the orthodox religious and cultural traditions of Ashkenazic Judaism, its yidishkayt, and this was most likely to be found in the u-dialect regions.22 Without traditional ‘Jewishness’ the language would inevitably be exposed to non-Jewish outside influences, resulting in stunted growth.23 This attitude is the reason why Birnbaum was never prepared to accept an artificial standard – on any level of the language – which was not based on the tradition of the majority.24 Firstly, orthography. Efforts to bring about an orthographic reform were being made since the end of the 19th century, more or less independently from one another, in many different places,25 effectively since the 1920s. Three systems remained historically relevant: (1) The Yivo-system, generally accepted and in use nowadays, which was developed by a standing commission after 1926 and was adopted and published in 1937 in the Takones fun yidishn oysleyg,26 (2) the Soviet system27 and (3) Birnbaum’s system.28 Birnbaum saw the necessity for a reform29 ever since he began working on the Yiddish language. It was his aim to revive the natural historical development of orthography, which had been interrupted by the actions of the Maskilim, and, starting from the u-dialects, to create, 22

23

24 25

26 27 28

29

Cf. Chapter 10 in volume I (with bibliographical comments by the author in note 3); otherwise predominantly in Yiddish, for instance h1925 “Yidish un yidishkayt” (English translation 1969), h1929 “Loshn, oysleyg un asimilatsye”, h1930 “Yidishkayt un loshn”, h1931 “Ge ule fun loshn”. Cf. the sensitive description of an Orthodox Jew, Shmuel Hiley (above, n. 2), p. 9 f., also Shikl Fishman in: Afn Shvel (2000), p. 5; detailed in Kalman Weiser (above, n. 2), p. 279 passim. Birnbaum considers the influence of the German language to be the most insidious, since it impedes the natural development of the language; cf. for instance his article in the list of publications under h1939, entitled “A verter-reshime, Daytshmerish-Yidish”, reprinted in Afn Shvel (2000), p. 6–8. Cf. Chapter 18 in volume I and, for instance, 1979 (Survey), p. 100. Illustrated in detail by Mordkhe Schaechter, Der eynheytlekher yidisher oysleyg. New York 1999; source material in: Joshua A. Fishman, Never Say Die! The Hague / Paris / New York 1981. As summarized by Schaechter (above, n. 25) with the text of the Takones. Cf. Birnbaum, Chapter 19 in volume I; cf. also Schaechter (above, n. 25), p. 18–26. Regarding Birnbaum’s views on orthography cf. Chapter 5 (p. 55–58) and Chapter 6 in volume I; furthermore, list of publications under 1913/14, 1918 (Grammar) p. 11–33, 1979 (Survey) p. 197–215; in Yiddish: h1924, h1929, h1930 (several times), h1931, h1932 (verter-bikhl), h1953, h1977. In 1933 (Chapter 6 in volume I, p. 62 f.) Birnbaum gave a concise and clear description of the differences between the five orthographic systems in current use.

XVI

Solomon Birnbaum’s Life and Work

according to scholarly principles, a new ‘interdialectal’ spelling system which could be read by any speaker of Yiddish in his own dialectal pronunciation. He considered a precise identification of the vowel qualities (and quantities) with the aid of diacritical signs to be the most important requirement. He designed such a system in 1915 for his Grammar and continued to refine it over several decades up to the Survey in 1979, in which he renounced several suprascripts,30 which could be typed or printed only with difficulty. The system as such remained substantially the same from 1915 to 1979, especially with regard to the identification of the stressed vowels.31 In 1930, at a teachers’ conference in Rabka,32 Birnbaum’s orthography, which he had just presented in the brochure Yidishkayt un loshn (Warsaw 1930), was accepted for use in the orthodox Bajs-Jakow 33 [Beth Jacob] Schools and shortly afterwards also officially for the publications of the Bajs-Jakow publishing house; it was also used in orthodox periodicals such as Jeshurun (Warsaw) and Kinder-gortn (Lodz), as well as various orthodox daily newspapers.34 Following the destruction of Jewish life in Eastern Europe, Birnbaum’s system became virtually extinct, but he himself continued to use it until the end of his life and attempted to persuade editors to use it for his articles, with limited success.35 30 31

32

33

34

Cf. for instance in volume I, p. 56 (Chapter 5 of 1929) with p. 65 ff. (Chapter 6 of 1933). Cf. the transcription table in his Grammar (1918) p. 11–18 with Survey (1979) p. 197–215 including annotations. He vacillated about the designation of secondary syllable vowels by an Ayin (1918 ff.) or Yud (from 1925), finally permitting both in some instances (examples in Survey p. 201–210, and elsewhere). – For additional details see Denman 1986, p. 253–255; Hiley 1991, p. 4 and passim; Weiser 2004, p. 279 passim (all above, n. 2). Cf. also Schaechter (above, n. 25), p. 33. Birnbaum reported about the outcome of the conference in Rabka (approx. 50 km south of Cracow) in 1930 in German in Der Aufstieg, and several times in Yiddish, for instance in the Oylem-bleter (see h1930). Birnbaum’s transcription. – The orthodox educational network of the BethJacob-Schools for girls was founded in 1917 by Sarah Schenirer (1883–1935) and in 1935 comprised 227 schools with 27,000 pupils in Poland alone. Cf., for instance, the Encyclopaedia Judaica. Second Edition, Detroit 2007, vol. 3, art. “Beth Jacob Schools” and vol. 18 “Schenirer, Sarah”. Cf. Schaechter (above, n. 25), p. 33. – In practice, the application of the system was not so easy for the printers; a close inspection of publications, even those produced by the Beth-Jacob publishers, frequently discloses that they did not comply with the directions; there is also the fact that not all diacritical symbols were available for certain type sizes.

Solomon Birnbaum’s Life and Work

XVII

Equally, Birnbaum’s efforts to create an optimal scholarly transcription system into Latin script continued for over seven decades.36 His system was established, once more, on the basis of the language spoken by the majority in the u-dialect regions, but was intended to be applicable to all dialects. On the other hand, the very fact that it is designed to transmit the most subtle differences between the individual dialects means that it is complicated and cannot easily compete against the Yivo-system, which simply reflects the phonemic system of ‘Standard Yiddish’. In this case, far more than in the orthography system, Birnbaum made several fundamental changes since the first version he introduced and applied in Die Freistatt (1913/14)37 – compare this first version of the system, for instance, with the one in his Grammar (1915/18), in the article entitled “Die Umschrift des Jiddischen” (1933),38 reprinted in volume I as Chapter 6, and in the Survey (1979, especially p. 200–223).39 For this reason it is not possible – and would be a historical falsification – to impose a standard form of transcription on the articles contained in volume I, in view of the fact that they originate from different eras.40 This is even more applicable to the presentation of Old Yiddish texts in Latin script. 35

36

37

38 39 40

Successfully though, for instance, in 1964 in the Festschrift for Max Weinreich, see the charming presentation by Shikl Fishman in Afn Shvel (2000), p. 4. In the Yivo publications (after 1926) Birnbaum had to endure the use of their unwelcome orthography even in his own articles. According to the author (see volume I, p. 62, n. 1) “the first words printed in his transcription system go back to the year 1907” [in German]. He was probably referring to the individual words in the translation of Pinski’s narrative “Durchgesetzt” in the Neue Zeitung in 1907. Regarding the final stage of this development see the Editors’ foreword to Chapter 6 in volume I. Cf. the annotated pronunciation table p. 56–59 as well as the articles on p. 412– 414 and 588–591. Already at this stage Birnbaum emphasized that his transcription was specifically designed for the purpose of providing crossdialect legibility and should not be regarded as meaning to suggest that the Hebrew script should be abolished, since he “rejected this idea which has recently arisen” (p. 59, “[da er] diese letzthin aufgetauchte Idee ablehne”); cf. also Birnbaum’s retrospective 1987 (volume I, p. 297). This article assumes a central position in the overall development and has been frequently cited in the research literature. Additional articles on the subject ‘transcription’ can be found under h1929, h1944, h1947 in the list of publications. Unfortunately Birnbaum did not elaborate in sufficient detail on his ideas regarding transcription and especially on his views concerning priority as applied to technical matters. For this reason readers of volume I wishing to discover details of the transcription will have to research the various stages of development on their own.

XVIII

Solomon Birnbaum’s Life and Work

After Birnbaum concentrated more fully on the study of Old Yiddish texts (from 1932) he also felt the need to enrich Germanistic research by new insights into the Yiddish language, especially relating to phonetic developments,41 while continuing to highlight the distinctive individuality of the language. In this context he considered it unavoidable (at least for philological purposes) to present the texts in Latin script, depending on the language level involved. He rightly criticized the transcriptions published previously as being unsystematically oriented to German spelling.42 In fact he thought that “familiarity with older German linguistic usage, knowledge of Judaism [and] of the problems associated with old Yiddish spelling ought to be combined”43 when preparing transcriptions of Old Yiddish texts. Obviously one cannot expect to have a uniform transcription for earlier states of the language spanning several centuries, due to the fact that considerable changes occurred not only in pronunciation (with regional differences) but also in spelling. On the occasion of his edition of psalm 6 in thirteen versions from several centuries (Chapter 12 in volume I) Birnbaum strongly emphasized that the transcriptions should be viewed with caution and should not be used as a basis for the phonological research of Old Yiddish. Nor was it possible – solely due to space considerations – to justify the transcription for each individual text. This would require devoting an entire article to each of the texts. He stated that the transcription should be “a compromise between the intrinsic phonetics, a pure transliteration of the characters, and the desire to avoid technical problems in printing.”44 He continued constantly to search for an optimal compromise.45 41

42 43

44 45

An opportunity rarely taken advantage of in German philology and comparative literature to this day, even though quite a few annotated editions of Old Yiddish texts have meanwhile become available. This mainly referred to the book by Staerk and Leitzmann; cf. Birnbaum’s (otherwise positive) review of 1924. In 1932 on the occasion of the edition of the treatise on bloodletting (Teuthonista, p. 198, “[vielmehr müssten sich] Bekanntschaft mit älterem deutschen Sprachleben, judaistisches Wissen [und] Erkenntnis der Probleme der alten jiddischen Schreibung vereinen”). Volume I, bottom of p. 178 f., similarly Chapter 14, p. 207, on the occasion of the edition of 37 text samples from six centuries, and elsewhere. Evidently with the intention of avoiding Old Yiddish texts looking like German texts. – Already in 1932 (in other words, still in a German-language context) two highly different systems existed (compare the edition of the treatise on bloodletting [in Teuthonista 8] with the rendition of psalm 6 in Chapter 12 of volume I). Later, in an English-language context, all relics of

Solomon Birnbaum’s Life and Work

XIX

Towards the end of the 1920s, Birnbaum was inspired by a major plan about which he had been approached by Heinz Kloss, a researcher of minority languages and by Franz Thierfelder, a linguist concerned with cultural policies. It involved the establishment of an institute of closely related languages, which was to focus on Dutch, German, Yiddish, Afrikaans, Frisian and Pennsylvania-Dutch and in which Birnbaum was to run the Yiddish department.46 In 1930 Birnbaum prepared a draft program for this, which was divided into three parts: ‘Information’ (e. g. about the cultural position of Yiddish in various countries, the Yiddish school and educational systems, the publication of Yiddish books in transcription), ‘Research’ and ‘Teaching’, whereby the preparation of teaching materials was to take precedence initially, but a synchronic and diachronic philology in its entire spectrum was envisaged – objectives which have remained wishful thinking in Yiddish research to this day – including for instance a comprehensive historical dictionary, an etymological dictionary, dictionaries of dialects, as well as an extensive historical grammar of the Yiddish language.47 The institute of closely related languages was never established. However, Birnbaum continued to work on his plan for an Institutum GermanoJudaicum. The appeal which he sent at the beginning of 1933 to nonJewish German scholars, general linguists, and Hebrew and Old Testament experts in Germany, Austria and Switzerland, received a response in the form of signatures and specific words of approval from 60 eminent professors.48 However, in the atmosphere of the beginning Nazi period this project, too, was doomed to failure. The extraordinarily fertile Yiddish research of the Hamburg years was later included in the two books mentioned previously: In 1974 the

46 47 48

German orthographic traditions were removed (ch for x and kÆ / Ó was replaced by x; for some time v and w for bÆ / u and ù were represented by v or w turned upside down [cf. p. 91 f. and 208 f. in volume I], then by f or v [p. 210 ff.]). What is unusual and difficult to understand is the choice of s´ for /ʃ/ to represent a Ï w, where previously sˇ had been used (in the case of Modern Yiddish already as early as 1954, cf. for instance p. 279 in volume I; see also p. 221 ff. [1965]). The development of the designations of Old Yiddish vowel qualities and quantities can only be understood by studying them closely, compare the stages illustrated in Chapters 12 (1932) and 14 (1965) with the level in 1979 in the Survey. Cf. Chapter 20 (from 1987) in volume I; see also in the list of publications under 1972. Cf. in the list of publications under 1930 (“Die Stellung der jiddischen Sprache”), especially p. 361–364. See also Freimark (above, n. 2), p. 130. Cf. the list of names and comments p. 301–304 in volume I.

XX

Solomon Birnbaum’s Life and Work

book entitled Die jiddische Sprache. Ein kurzer Überblick und Texte aus acht Jahrhunderten was published in German in Hamburg,49 and in 1979 a greatly expanded English edition entitled Yiddish, A Survey and a Grammar 50 was published in Toronto. This type of phenomenological overview of the Yiddish language, which is based almost exclusively on the author’s own research, represents the lifetime achievement of the Yiddish scholar Birnbaum.

II. Hebrew Palaeography (and ‘Jewish Languages’) While still in Hamburg Solomon Birnbaum – guided by his desire to date medieval Yiddish manuscripts and find their place of origin51 – took up palaeographic studies. In the spring of 1929 he submitted an 800-page paper with the title Die nordjüdischen Kursivschriften. Eine Studie zur hebräischen Buchstabengeschichte to the Faculty of Philosophy as his professorial (‘habilitation’) dissertation. In this study he covered the Hebrew scripts from the Middle Ages and the Early Modern Era, used in Western and Eastern Europe. Despite several positive assessments the habilitation did not materialize, obviously for reasons of university politics. As P. Freimark pointed out, this would have been the first habilitation in Jewish studies at a German university.52 In the spring of 1933 Birnbaum left Hamburg and escaped to London with his family. After several difficult years he experienced a turning point after meeting Norman Jopson, a professor of Slavonic and comparative studies. The latter arranged for Birnbaum to receive an invitation to give a series of lectures, which provided an entree to the University of London. From 1936 until 1957 he taught Hebrew palaeography and epigraphy at the School of Oriental and African Studies and from 1938 he also taught Yiddish at the School of Slavonic and East European Studies.53 Although he continued to contribute to Yiddish 49 50 51 52 53

Once again from the publisher Helmut Buske; two additional reworked and expanded editions were published in 1986 and 1997. A selection of 37 of the 75 text samples in the Survey can be found in Chapter 14 in volume I. Cf. The Hebrew Scripts I (1971), p. 22, also David Birnbaum in Afn Shvel (2009), p. 42, and “Czernowitz at 100”, p. 183 (both above, n. 2). For details of the procedure see Freimark (above, n. 2), p. 125–147. According to information provided by Eleazar and David Birnbaum, S. A. B. was seconded from the university during the Second World War to the British Postal Censorship, specializing in Yiddish and other Jewish languages. In this way he discovered at an early stage that the Holocaust was actually taking

Solomon Birnbaum’s Life and Work

XXI

periodicals in Eastern Europe from England,54 the main focus of his research activities over the next two decades was in the area of Hebrew palaeography. By contrast with his studies in Hamburg, he continually expanded the geographical areas of the script provenances in all directions, going back to the earliest traditions of Hebrew scripts. Since most Jewish cultural groups used the Hebrew alphabet when writing their everyday languages, Birnbaum’s interest was also directed towards Jewish languages other than Yiddish through his interest in palaeography. An impression of this is provided in Chapters 1 and 2 as well as Chapter 21, where the Jewish languages from the Eastern edges of Europe are specifically identified and illustrated by examples of texts.55 Chapters 21–23 and 25–27 furthermore demonstrate an intensive scholarly preoccupation with Judezmo (the Jewish-Spanish language,56 which Birnbaum later called Jidic), the Bukharic language (the Jewish-Persian of Central Asia)57 – both areas for which very little groundwork existed at the time. Furthermore, Chapter 24 provides insight into an uncompleted handwritten study of about hundred pages, which examines the difference between the language of the Jews of the Maghreb (which Birnbaum called Maaravic), particularly in the City of Fez, with the language used by the Muslims in the same city. During the 1930s his opus magnum The Hebrew Scripts became ready for publication, but could only be published after the end of the Second World War.58 In the foreword to the first volume Birnbaum relates the dramatic printing history, which started in 1939 before the outbreak of the war and ended in 1971. The illustrations were lost in the upheaval associated with the beginning of the war. On the other hand, the delay provided Birnbaum with the opportunity to integrate his analyses of the Dead Sea scrolls and other documents, which had been

54 55 56 57 58

place and he did what he could to persuade the British authorities in power to take some action. In 1945 he compiled an English-Yiddish Phrase Book at the request of the Jewish Committee for Relief Abroad, which was used by staff members of the aid organizations in the post-war years. See the numerous articles in the list of publications starting from h1933, in orthodox as well as Yivo-journals. Additional articles on Jewish languages in the list of publications under 1944, 1947, 1951. Birnbaum did not like the linguistic terms ‘Jewish-plus’ or ‘Judaeo-plus’ (see, for instance, Chapter 2, p. 7). An explanation of the linguistic term can be found in Chapter 1, p. 3 (and Chapter 2, p. 8). In the list of publications under 1954 (Part II, The Plates) and 1971 (Part I, The Text ).

XXII

Solomon Birnbaum’s Life and Work

discovered from 1947 on. On account of these new finds the work had to pass some unexpected tests which, from today’s vantage point, it passed with flying colours. Although some of his date determinations initially met with doubts, since at that time palaeography was not yet regarded as a reliable scholarly discipline by theologians and archaeologists, they have since then been verified and generally accepted by experts with the aid of the radio carbon method and more recent archaeological analyses. The most important individual research of Hebrew (and Aramaic) documents from the Dead Sea and of Old Yiddish manuscripts which had only recently come into the field of research, as well as of several other documents which were remarkable for various specific reasons, were assembled by the author himself for the present volume II.59 They must be regarded as groundbreaking pioneering work within the context of the period in which they were composed. For this reason, the present Editors have refrained from making any attempt at modernizing them in the light of subsequent research. The Hebrew Scripts contain around 400 representative examples (in the form of facsimiles of handwriting and alphabetic and comparative tables) showing the history of Hebrew script going back over 3000 years, each accompanied by detailed palaeographic descriptions of the letters and their significant characteristics classified within the general development pattern of the Hebrew scripts. With this extensive study Birnbaum transformed Hebrew palaeography from mere intuition into a systematic and comparative discipline. On the one hand, due to the abundance of material at his disposal, he was able to apply quantitative methods to the field of palaeography for the first time. On the other hand, he recognized that as a rule the palaeographic relationship between two Jewish cultural groups is based on their regional history. He was able to demonstrate that up to approximately the 13th century there was virtually no difference between the script of Northern French and German, i. e. Ashkenazic Jews, and he utilized this realization to support a strong argument that the original homeland of most German Jews was Romanic and not Byzantine-Slavic.60

59

60

On the documents from the Dead Sea: Scrolls, Chapter 33, 36– 45 (on methodical principles, Chapter 38, and especially Chapter 43); everyday documents, Chapters 46–53; on Old Yiddish manuscripts: Chapters 56, 57, 59. Fundamental points already made in 1931, “Aschkenasische Handschriften. Woher stammen die deutschen Juden?”, see also The Hebrew Scripts I, paragraphs “Zarphatic Type” and especially “Ashkenazic Type”.

Solomon Birnbaum’s Life and Work

XXIII

Even today, 40 years after completion of the publication of The Hebrew Scripts, it is possible to establish from the state of research that Birnbaum’s monumental work continues to be the only historically accurate overall representation of Hebrew palaeography and for some purposes the only useable representation at all. In the intersubjective comprehensive documentation of the conclusions his work has set new standards for the entire field of palaeography, nor just the Jewish one. In view of the sheer volume of Hebrew manuscript material61 now available it has become obvious that this gigantic subject could not possibly be handled by a single individual. Nowadays scholarly palaeography is in the hands of an Israeli-French joint project, which was established in 1965 and is amply staffed and financed.62 In addition to an abundance of manuscript publications arising from this project,63 a detailed codicological database (SfarData) was set up under the leadership of Malachi Beit-Arie´, which has made it possible, by means of a sophisticated retrieval system, to classify undated Hebrew manuscripts with more chronological and geographical precision.64 In recognition of 61

62

63

64

At the Institute of Microfilmed Hebrew Manuscripts in Jerusalem microfilm copies of most of the world’s major manuscript collections are available, a total of about 100,000 items. (Website: http://www.jnul.huji.ac.il/imhm/). Sponsored by the Israel Academy of Sciences and Humanities (in conjunction with the Jewish National and University Library) in Jerusalem and by the Institut de Recherche et d’Histoire des Textes (Centre National de la Recherche Scientifique) in Paris. I mention here only the two major serial publications Manuscrits me´die´vaux en caracte`res he´braı¨ques portant des indications de date jusqu’a` 1540, ed. by Colette Sirat and Malachi Beit-Arie´, 3 vols., Paris / Jerusalem 1972–1986, and Codices hebraicis litteris exarati quo tempore script fuerint exhibentes, ed. by M. Beit-Arie´ [et al.], 4 vols., Turnhout 1997–2006. The computer database “SfarData” was presented by Malachi Beit-Arie´ several times, for instance in a graphic summary (with additional related literature), in his book Unveiled Faces of Medieval Hebrew Books. The Evolution of Manuscript Production – Progression or Regression? Jerusalem 2003, p. 16 f. – In future, this database will become an important tool for the modernization or recompilation of manuscript catalogues, among others. The first occasion when Beit-Arie´ was able to use the database to assist with a palaeographical and codicological description, was for the Hebrew Manuscripts in the Biblioteca Palatina in Parma. Catalogue. Ed. by Benjamin Richler. Palaeographical and Codicological Descriptions by Malachi Beit-Arie´. Jerusalem 2001. Since then, also Hebrew Manuscripts in the Vatican Library. Catalogue. Compiled by the Staff of the Institute of Microfilmed Hebrew Manuscripts, Jewish National and University Library, Jerusalem. Ed. by Benjamin Richler. Palaeographical and Codicological Descriptions by Malachi Beit-Arie´ in Collaboration with Nurit Pasternak. Citta` del Vaticano 2008.

XXIV

Solomon Birnbaum’s Life and Work

Birnbaum’s standard work, Beit-Arie´ writes: “[. . .] But not before the pioneering publication of S. A. Birnbaum’s The Hebrew Scripts [. . .] did Hebrew palaeography become an independent discipline. One cardinal aspect of it, the typology and evolution of the script, was for the first time studied systematically”.65 ***** In the course of his long career as a researcher Solomon Birnbaum established a terminological framework as well as many other basics for two major scholarly disciplines, for the benefit of future generations. However, he never established a ‘School’ in the narrow sense of the word. He largely distanced himself from the ‘business of scholarship’, even from the Yivo Institute.66 Yet he invariably acted as an indefatigable and unbiased adviser. Anyone who approached him with a technical question was astounded by his virtually inexhaustible knowledge and an impressive willingness to impart information and to engage in dialogue.67 He will go down in the history of scholarship as an individualist of exceptional personal nobility. Even after the Holocaust he continued to keep up a correspondence with German scholars. In the 1960s, when a new young generation at the University of Hamburg under the leadership of the German scholar Walter Röll began to concentrate with renewed interest on the work started by Birnbaum, Birnbaum participated from London as a generous and friendly adviser. These friendly ties were maintained when Birnbaum moved to live with his sons in Toronto in 1970 and part of the Hamburg team relocated to the newly founded University of Trier, where they immediately started to establish Yiddish studies within the framework of Germanic studies. In 1986 Solomon Birnbaum was honoured by the University of Trier with an honorary doctorate and for the conferral he returned to Germany for the first time – at the age of 95. The university paid tribute to Birnbaum as an important scholar, a steadfast 65 66 67

Beit-Arie´, The Makings of the Medieval Hebrew Book. Studies in Palaeography and Codicology. Jerusalem 1993, p. 7. See also the article devoted to Shloyme Birnboym in Afn Shvel 2000 (above, n. 2), especially p. 1 f. The archive established by S. A. B. called The Nathan & Solomon Birnbaum Archives (Spalek / Hawrylchak 1997 [above, n. 2], p. 57 ff.) bears eloquent witness to this: here, approx. 20,000 letters from and to S. A. B. dealing with linguistic, palaeographic and general Judaistic subjects are archived.

Solomon Birnbaum’s Life and Work

XXV

representative of supranational ideals and a person of impressive kindness. In his acceptance speech he emphasized “the feeling of satisfaction at seeing how much times have changed and that nowadays so much – relatively – is being done to continue what I and a few others started in those days”.68 Since 1986 the situation has definitely improved all over the world in both disciplines.69 No doubt Solomon Birnbaum would be very pleased about this. Erika Timm (Translated by Ruth Segal, Toronto)

68

69

“[. . .] das Gefühl der Genugtuung, wie sehr sich die Zeiten geändert haben und daß heute so – verhältnismäßig – viel getan wird für das, was ich und wenige andere in jenen Zeiten angefangen haben”, in: Verleihung der Würde eines Ehrendoktors [. . .] (above, n. 2), p. 26. – In Hamburg the name of Solomon Birnbaum has also been honoured: The “Salomo-Birnbaum-Bibliothek” was established at the initiative of the late Dr. Günther Marwedel and nowadays contains about 3,000 volumes of Yiddish texts and studies. The “SalomoBirnbaum-Gesellschaft für Jiddisch in Hamburg e. V.” was established in 1995 for the purpose of promoting Yiddish language, literature and culture (website: www.birnbaum-gesellschaft.org). Regarding palaeography see above; as for Yiddish studies, nowadays the range of activities in research and teaching – in Europe (even Eastern Europe) as well as America, Israel and even in China and Japan – can hardly be grasped. In Germany alone two chairs have been established: In 1990 at the University of Trier and in 1996 at the Heinrich-Heine-University in Düsseldorf. Yiddish studies are also represented in a more modest format at other universities in Germany (permanently in Berlin, Bonn, Frankfurt am Main, Hamburg, Halle-Wittenberg, Heidelberg, Munich, Potsdam, Stuttgart and Tübingen), as part of the departments of German Studies, Slavic Studies, History, Jewish Studies or Theology.

Scholarly Articles by Solomon Birnbaum

A. The Palaeo-Hebrew Script

000002

Birnbaum 01.04.11 18:17

000003

28.

Birnbaum 01.04.11 18:17

Table of Semitic Alphabets 10th – 5th Centuries B. C. E.

The Table, drawn by Solomon A. Birnbaum, was first published in: The Ancient Near East in Pictures Relating to the Old Testament, by James B. Pritchard. Princeton 1954, p. 88.1 (Carved letters are shown in outline, letters written in ink are shown solid.) {In the present volume the table appears reduced in size. The Key to the table has been added by Eleazar Birnbaum (first column: no. in table; second column: title; third column: date B. C. E.; fourth column: illustration number in The Hebrew Scripts (HS) by S. A. Birnbaum, 2 vols., Part I: The Text, Part II : The Plates. London 1954–1957 / Leiden 1971. For illustrations of documents containing the drawn alphabets, nos. 4, 5, 7, 13: see Chapter 34 in the present volume, figs. 1, 3, 4, 7 (pp. 101, 103, 104, 106); no. 16: Chapter 33 (p. 80). For further discussion of documents containing the alphabets: cf. Chapters 29–33, see also 34 / 35.

1

{For special bibliographical information see pp. 281–282.}

4

[ 28 ]

Key to Table 1

Ahiram Inscription

ca. 975

HS 01

2

Yehimilk Inscription

ca. 950–930

HS 04

3

Samaria Ivories

ca. 900–875

HS 09

4

Gezer Calendar Tablet

ca. 875–850

HS 2

5

Moabite Stone Mesha Stele

ca. 830

HS 013

6

Kilamuwa Inscription

ca. 825

HS 014

7

Samaria Ostraca

ca. 775–750

HS 4–11

8

Shema Seal

787–744

HS 3

9

Bar Rakab Inscription

733–727

HS 018

10

Siloam Inscription

716– 686

HS 14

11

Nerab Stelae

ca. 1st half of 6th cent.

HS I, col. 123

12

Pharaoh Letter

605

HS I, col. 122

13

Lachish Ostraca

597–588

HS 23–26

14

Jewish Seals

6th century

HS 19, 26*, 37

15

Meissner Papyrus

515

HS 122

16

Leviticus Fragments

ca. 450

HS 28–31

17

Elephantine Papyri

ca. 400

HS 147

18

Eshmunazar Sarcophagus

5th century

–}

000005

Birnbaum 01.04.11 18:17

5

Table of Semitic Alphabets 10th – 5th Centuries B. C. E.

000006

Birnbaum 01.04.11 18:17

000007

29.

Birnbaum 01.04.11 18:17

On the Possibility of Dating Hebrew Inscriptions*

The determination of undated inscriptions in the original Hebrew script is certainly not an easy task. Should we, therefore, not attempt to tackle it at all, and rather wait1 until new finds provide us with new material? A real change in the situation could only be brought about by much better and more extensive ones than those hitherto brought to light. Are we justified in reckoning with the likelihood of such finds? Is it not at least as reasonable to expect that the present number of early inscriptions will not be very considerably increased? And if this be so, should inadequacy of available material be permitted to act as a permanent bar to research? Have others allowed it to do so, in palaeography or in any other fields? The answer is, no. North Semitic. In the present case the investigator has, too, a possibility of somewhat enriching his resources by drawing on non-Hebrew North Semitic writing. It cannot seriously be contended that – as far as the early centuries are concerned – the scripts of Phoenicia, Syria and Moab may not under any circumstances whatsoever be ‘lumped together’ with the script of Israel and Judah. Actually, palaeographers do not separate them. The reason is apparent. If, for example, we put the alphabets of the Ahiram, Abibaal and Elibaal inscriptions, of the Arslan Tash Ivories, and of the Mesha Stele / Moabite Stone side by side with that of the Gezer Tablet, we realize that there is no difference in the basic forms or in the style. Those differences which do appear are connected with the chronological factor. Whatever regional differences there may have been between * First published in: Palestine Exploration Quarterly 76 ( July – Oct. 1944), pp. 213–217. – {Illustrations of inscriptions mentioned in this chapter appear as follows: (1) In the present volume. Gezer Tablet / Calendar in: Illustrations to Chapters 34 / 35. Mesha Stele / Moabite Stone in: Illustrations to Chapters 34 / 35. Samaria Ostraca in: Illustrations to Chapters 34 / 35; Chapter 31. (2) In The Hebrew Scripts by Solomon A. Birnbaum, vol. 2, London 1954–57. Ahiram Tomb Inscription: plate 01. Abibaal Inscription: plate 05. Elibaal Inscription: plate 06. Arslan Tash Ivories: plate 012. Siloam Inscription: plates 14, 35.} 1 Dr. Diringer thinks we should: “the extant material of early Hebrew epigraphy is not yet sufficient” (PEQ 75 [1943], pp. 50–54, the quotation p. 54). The present notes have been occasioned by his article.

000008

8

Birnbaum 01.04.11 18:17

[ 29 ]

the various North Semitic groups must have been very small indeed, since we cannot ascertain what they were, from our material. Hence, we need not, on the score of incongruity, shrink from utilizing documents other than Judaean and Israelite, for we are not violating the principle that comparisons should only be made within a group. Scantiness of Material does not altogether prevent us from observing the main development, even though it restricts our observation of transitional stages. But from one point of view this might even be said to be of advantage. Where all the stages are known, they are so near to each other that the forms have naturally many features in common. This complicates the work of comparison, since, in the development of script, regular movements of the whole alphabet do not take place. It is the individual letters – sometimes certain features, of a group – which undergo changes. And they do so at different rates of speed. In one and the same period and in one and the same documents, letters which have changed very little during a given time, appear by the side of others which have developed far more quickly. The investigator must not consequently be misled into ascribing too great an age to the document. The absence of connecting links has the effect of throwing the differences into clearer relief, and of thereby, in a way, simplifying the task of comparison. It is thus quite possible that certain conclusions drawn from a few documents may be as dependable as if they had been based on a far greater number. For example, the six inscriptions used for dating the Gezer Tablet / Calendar2 suffice to show the growing measure of similarity between them and the Tablet, as we pass from the eleventh century of Ahiram’s tomb3 inscription to the ninth century of the Mesha Stele. Sometimes, still fewer even will suffice. For the dating of the Samaria Ostraca,4 use was made of two inscriptions only. From the regnal years on the Ostraca we know that these must have been written during one of the reigns of Ahab, Jehu, Jehoahaz or Jeroboam II . This knowledge would enable us to establish precise dates, provided we were in a position to relate the script of the sherds to that of dated inscriptions belonging to those times. The Mesha Stele (M) and the Siloam Inscription (Si) are available for that purpose. Both are well dated documents, 2 3 4

PEQ 74 (1942), pp. 104–108 {= Chapter 30 in the present volume}. {Later S. A. B. dated the document ‘ca. 975’ (see The Hebrew Scripts, vol. 1, no. 01; see also above, Chapter 28, no. 1). – E. T.} Above, n. 2, pp. 107–108.

On the Possibility of Dating Hebrew Inscriptions

9

the former from about 850, the latter from the end of the eighth century B. C. E. If, on comparison, we found that the script of the sherds was distinctly nearer to M than to Si (or vice versa), then we should be justified in assuming a closer temporal relation between the Ostraca and the reign during which M was written, than between them and the reign during which Si was written (or vice versa). In other words, if the script of the sherds resembles M, then the period of Ahab or Jehu – or in a lesser degree, Jehoahaz – would be probable; if there were greater similarity with Si, then the period of Jeroboam II would be the likelier. It would appear, therefore, that wealth of material is not always essential to reaching reliable results. Certainty of Date. Even the principle that comparison may be based only on documents whose date is beyond doubt, needs qualification. In favourable circumstances, even material of controversial date can be very useful, e. g., the Ahiram inscription (ibid.). Some scholars have ascribed it to the middle of the thirteenth century, on account of the Ramses II fragments. But, as Spiegelberg points out, these cannot be taken as indicating the age of Ahiram, since the tomb was plundered in the eighth-seventh century.5 It seems preferable to accept Albright’s date (“probably not far from 1100 B. C. E.”).6 The idea in using the sarcophagus alphabet is not, of course, to take it as a starting point from which to reckon out the time distance to the script of the Tablet. The Ahiram alphabet is taken merely to represent the upper time limit, since comparison shows that the Gezer forms are clearly later and cannot belong to the Ahiram period – wherever that might be placed. Inscription and Manuscript. With so little material available, it is natural to make use of everything there is, and to turn to handwriting, too, for comparison. In theory this should of course be avoided, since all material imposes on the writer a particular technique, which may result in a difference of style or even in changes of the basic form of letters. Actually however, with the exercise of due care, the inherent danger of comparing incongruous scripts need not work out to be so great in practice as it would appear to be in theory.

5 6

{Wilhelm Spiegelberg, “Zur Datierung der Ahiram-Inschrift von Byblos”, in: ˙ 735–737. – E. T.} Orientalistische Literaturzeitung 29 (1926), pp. {W. F. Albright, “A Neglected Hebrew Inscription of the Thirteenth Century B. C.”, in: Archiv für Orientforschung 5 (1939), pp. 150–152, quotation p. 152: “we can hardly date [the Ahˆıraˆm inscription] earlier than the end of the period ˙ assigned it above, that is, about 1100 B. C.” – E. T.}

000010

10

Birnbaum 01.04.11 18:17

[ 29 ]

If, for example, we put the Mesha and Siloam alphabets side by side with that of the Samaria Ostraca, we see that there is a general difference. It is one of style and is occasioned by the use of the pen for the Ostraca and of the chisel for the others. In handwriting, a strong contrast develops between up- and downstrokes. In the one, the writing edge of the pen is parallel to the movement of the hand, thin lines being the result, in the other, there is an angle of 90 degrees, resulting in thick lines. But that is all. The writing of the Ostraca was not an ornamentally developed manuscript hand. The essential shape remained unchanged. Hence, comparison as to their respective stages of development is possible. * * * In the following, attention is given to the details brought up in Dr. Diringer’s aforementioned article. North Semitic. “The forms, with all their details, of the Ahiram characters [are not] the original forms of the north-Semitic letters” {p. 52}. I have nowhere in my article said or implied that they are. Curving. Curves do not in themselves constitute a proof of either earliness or lateness of development. Both in cursive and ornamental development, straight lines become curved and vice versa. Gezer Letters. Daleth. The right side is not “practically vertical”. It slants to the right, it is perfectly parallel with the downstrokes of the Heth in the line above. (In the examples of Resh, the downstrokes are by no means all vertical. Forms like these are just indefinite, intermediate between a slant to the left or to the right. This may be said of Samekh, too.) Waw. The top part of this letter in the fifth line is not, even “roughly speaking”, a semi-circle. It is similar to all the other Waws of the Tablet. Their top parts are all narrow and high, sometimes these are straightlined, or even completely angular. This is in clear contrast to the wide form of Ahiram, with its curved lines. Kaph. In certain respects, the forms in the Tablet and in the Mesha Stele differ, but in both, the right hand stroke is projected downwards. Mem. There is no reason whatsoever for assuming that the lowest stroke was initially of the same length as the others and that it was only lengthened later. The letter clearly has the younger form: a top part and a shaft. Variants. It is not a feature peculiar to the Gezer Tablet that practically no letter appears twice in the identical shape. The variants, e. g., in the Ahiram inscription, are even more numerous.

000011

Birnbaum 01.04.11 18:17

On the Possibility of Dating Hebrew Inscriptions

11

The Gezer variants cannot be taken as “a proof that the forms of the letters were still unstable in details” {p. 53}. The man who wrote the Tablet does not appear to have been a professional engraver. His writing was thus a kind of cursive. We need only look at anything we write ourselves – and paper and pen are much easier materials to handle – in order to realize that we rarely shape a letter twice in precisely the identical way in one and the same line. Nobody would, however, therefore jump to the conclusion that the Roman characters of to-day are still unstable in detail.

000012

Birnbaum 01.04.11 18:17

000013

30.

Birnbaum 01.04.11 18:17

The Dates of the Gezer Tablet and of the Samaria Ostraca*

I. The Gezer Tablet The Gezer Tablet {see p. 101 in the present volume} cannot be dated from archaeological evidence. Let us see what aid towards determining its age can be derived from palaeography. There has been considerable deviation in the results arrived at by epigraphists on this score – the dates ranging from the eleventh century B. C. E. to about 900 B. C. E. This is not surprising since the material available for comparison is only very scanty. The following lines are an attempt, through a new, detailed comparison in the course of which a few recent finds have been utilized, once more to tackle the problem of their date. The documents we shall utilize are: the Byblos inscriptions of Ahiram (A; early eleventh century)1 and of Abibaal and Elibaal (AE ; middle and second half of the tenth century), the Samaria Ivories (SI ; first half of the ninth century), the Arslan Tash Ivories (AT ; middle of the ninth century) and the Mesha Stele (M; about 850). Aleph: In A, the oblique strokes, though converging, stop at the vertical and do not meet; they are curved. In the later inscriptions, they cross the vertical and meet; they are straight; the main stroke is longer than in A, and these characteristics are to be found also in G [= our Gezer Tablet], where, too, the angle of the cross strokes is much more obtuse than in M. Daleth: In AE , the right side is vertical; in the later inscriptions and also in G, the right side is slanting. In AE and M, neither of the sides is produced downwards, which is also the case in G but not in SI and AT . Waw: In A, the top part consists of a semi-circle; in SI , one example of it is still wide, both halves having nearly turned into straight lines; in M, the top is similar but narrower, which is also sometimes the case in G, where, too, there are forms with straight lines. Only in A does the downstroke turn left. * First published in: Palestine Exploration Quarterly 74 ( July – Oct. 1942), pp. 104–108. – {For illustrations of inscriptions mentioned in this chapter see the introductory footnote to Chapter 29 in the present volume.} 1 {Later S. A. B. dated the document ‘ca. 975’ (see The Hebrew Scripts, vol. 1, no. 01; see also above, Chapter 28, no. 1). – E. T.}

000014

14

Birnbaum 01.04.11 18:17

[ 30 ]

Zayin: In A, the two horizontals are short, in SI and AT they are long. The connecting stroke is usually vertical, but in SI a slanting one also occurs and it is this form which we find in G. Heth: G agrees with A, SI and AT as against M, where the right hand vertical is prolonged downward. Yodh: In A, the letter is a wide curve, halved by a short horizontal, and has a horizontal base; in AE , the top is nearly angular. In SI , AT and M it is completely so. The middle stroke is usually nearer the top stroke; both features are characteristic for G, too. The base is horizontal, as in A, in AE and AT ; it slants upwards in SI and M; G has both upward and downward slants. Kaph: In A and AE , all three strokes are of equal length, in AT the middle stroke, in M the right hand stroke is lengthened downwards. G agrees with M. Lamedh: In A, the long stroke is slightly curved, in M, it is strongly curved; in AT , it is straight; the latter is also the case in G. The angle at the bottom is rather obtuse in A, but is acute in AT , as also in G. Mem: In A, the lowest stroke is of the same length as the others, being just one line of the zig-zag. In the other inscriptions it is prolonged downward and this too, is the case in G. The downstroke is straight in G but in AT there is, besides this form, one whose end bends to the left, as in M. Pe in A is practically crescent shaped, the upper part being just slightly less curved that the lower; in SI , the curve is broken; in M, the top is angular; G has the same feature but the left stroke of the angle is curved, whereas it is straight in M. Sadhe and Qoph in G are as in AE and M. Looking through these notes, we find that the difference between G and A is very marked: as many as about half of the letters of the alphabet have had to be discussed. Hence, the early part of the eleventh century seems to be out of the question for G. Aleph, Mem, Sadhe and Qoph would suggest that G cannot be older than the middle of the tenth century; Daleth, Zayin, Yodh and Mem would suggest that it is not older than the first half of the ninth century; Waw, Kaph, Lamedh and Pe would seem to bring it down to the middle of the ninth century. Looking now for characteristics pointing to a terminus ante quem, we come upon features of Aleph, Daleth, Heth and Mem which we no longer encounter after the middle of the ninth century. Thus our two lines of delimitation meet at this same date. For the sake of convenience we might tabulate the results of our comparison as follows. When some characteristic of a letter points to a

000015

Birnbaum 01.04.11 18:17

The Dates of the Gezer Tablet and of the Samaria Ostraca

15

definite chronological relationship between G and the other inscriptions, we shall indicate this by putting the approximate date of the inscription in question with a dash either before or after, e. g., “Yodh 875–” means: “The particular detail or characteristic of Yodh in the Gezer Tablet corresponds to that of the Yodh in the Samaria Ivories – it is younger than the forms of the Ahiram Sarcophagus and the Abibaal and Elibaal inscriptions; it therefore suggests a date for the Gezer Tablet not before or not much before the first half of the ninth century”. (937 – AE , 875 – SI , 850 – AT , 850* – M.) Aleph Daleth Waw Zayin Heth Yodh Kaph Lamedh Mem Pe Sadhe Qoph

937– –850* 875– –850* 850*– 875– 850* 875– 850*– 850– 850– 937– –850 850*– 937– 937–

The latest date on the left side and the earliest date on the right side indicate the period in which the Gezer Tablet was written. These are identical and give the middle of the ninth century.

II. The Samaria Ostraca We did not include the Samaria Ostraca in our comparison for two reasons. Firstly, because their date is too uncertain. For the archaeological evidence bearing on their date offers us four possible periods: that of Ahab (867–854), of Jehu (842–815), Jehoahaz (814–798) or Jeroboam II (783– 743). By utilising the Ostraca we should be trying to find the x of Gezer by means of the y of Samaria. Secondly, if possible, it is of course, preferable to avoid comparison between inscription style and handwriting. The position is, however, different when attempting to date the Ostraca. No handwritten material near their period is available for comparison,

000016

16

Birnbaum 01.04.11 18:17

[ 30 ]

and so we have no alternative but to make use of inscriptions. There are, in all, not more than two inscriptions we can utilise. But they have the advantage of standing at the beginning and at the end of the period in question: the Mesha Stele and the Siloam inscription (Si; of about 700). Are the forms of the Samaria Ostraca (S) nearer M or nearer Si? Aleph: The cross strokes of S are almost parallel, as in Si, but, on the other hand, they meet as in M. Heˆ: S resembles Si more than M. Waw: S and Si are identical as against M. Zayin: There is no Zayin in M, but, for the sake of completeness, we might perhaps compare G; S goes together with Si, as against G. Heth: S resembles Si more than M. Yodh: The angles in S and Si are on the whole more acute than in M. Kaph: S and Si have the downstroke bending to the left, in M it is straight. Lamedh: The slant is more like that in Si than that in M. Mem: The top parts of S and Si resemble each other with their parallel; M is different. Nuˆn: In S, the hook as a rule does not go to the top of the downstroke, in M it does. Samekh: There is no Samekh in Si; S does not agree with M. Ayin: S and Si have an oval in a slanting position, M has a circle. Sadhe: S has the duplicated form of Si, not the archaic form of M. Qoph: One of the forms of S agrees with the form of Si, as against M. This comparison would seem to leave us no doubt that S is nearer to Si than to M and that the Ostraca belong to one of the later possible periods. The similarity is, in fact, so great that they can safely be ascribed to the latest period, that of Jeroboam II , i. e., to the years 774, 773, 768 and 766 B. C. E. This dating would also prevent there being too wide a gap between their time and the time of the Lachish Ostraca. If the Samaria Ostraca were put at the early 9th century, the small degree of development from this date to the Lachish Ostraca of the early 6th century would be somewhat surprising.

000017

Birnbaum 01.04.11 18:17

31. The Palaeo-Hebrew Ostraca Sherds and a Fragment of a Stele, all from Samaria*

A. Ostraca The sherds dealt with in this chapter were first published by E. L. Sukenik. His untimely death has caused the present edition to be placed into other hands. He was – as Mr. Crowfoot told me – no longer satisfied with his original readings and would have suggested a number of changes had he lived to carry through a revision. Many letters on these sherds are indistinct or mutilated and require detailed discussion before a reading can be arrived at. In the transliteration a line over a letter signifies uncertainty of reading, a double line indicates a higher degree of uncertainty.1 Bibliography A: Eleazar L. Sukenik, “Inscribed Hebrew and Aramaic Potsherds from Samaria”, in: PEFQS 65 (1933), pp. 152–156. B: Eleazar L. Sukenik, “Inscribed Potsherds with Biblical Names from Samaria”, in: PEFQS 65 (1933), pp. 200–204. C: David Diringer, Le Iscrizioni anticho-ebraiche palestinesi. Florence 1934. D: Eleazar L. Sukenik, “Potsherds from Samaria, Inscribed with the Divine Name”, in: PEFQS 68 (1936), pp. 34–37. E: W. E. Staples, “A Note on an Inscribed Potsherd”, in: PEFQS 68 (1936), p. 155. F: Eleazar L. Sukenik, “Note on a Fragment of an Israelite Stele Found at Samaria”, in: PEFQS 68 (1936), p. 156. G: William F. Albright, “Ostracon C 1101 of Samaria”, in: PEFQS 68 (1936), pp. 211–215. H: Eleazar L. Sukenik, “A Further Note on an Inscribed Potsherd”, in: PEQ 69 (1937), pp. 140–141. J: Shemuel Yeivin: Toldot ha-ketav ha- Ivri, vol. 1: Ad hurban Bayit Rishon. ˙ Jerusalem 1939. K: Sabatino Moscati: L’Epigrafia ebraica antica 1935–1950. Rome 1951. * First published in: Samaria-Sebaste. Reports of the Work of the Joint Expedition in 1931–1933, and of the British Expedition in 1935. Vol. 3: The Objects from Samaria, by John W. Crowfoot [et al.]. London, Palestine Exploration Fund, 1957, pp. 9–34, plates 1– 4. 1 This device was first used in my article in Palestine Exploration Quarterly (PEQ) 71 (1939), pp. 20–28 and pp. 91–110 {= Chapter 32 in the present volume}.

000018

18

Birnbaum 01.04.11 18:17

[ 31 ]

Some of the Documents used in the Study of the Script B. C. E. AshD EgD ElP ElS Esan Ged GezT HarO Jaaz LachO lmlk Sen Shal ShemS Sil 2

3 4 5 6

7 8 9 10 11 12 13 14

15 16 17 18 19 20 21

Asˇsˇur Documents2 Egyptian Documents5 Elephantine Papyri6 Eliakim Stamp7 Esangil Endorsement8 Gedaliah Seal9 Gezer Tablet10 Harvard Ostraca11 Jaazaniah Seal13 Lachish Ostraca14 lmlk jar handle stamps (periods I–III ) Documents, Senacherib period16 Documents, Shalmaneser period17 Shema Seal18 Siloam Inscription20

659;3 c. 6604 third century 495– 400 c. 590 330 586 c. 875 77012 c. 590 c. 59015 726–589 705– 681 727–722 c. 76419 c. 70021

Mark Lidzbarski, Altaramäische Urkunden {aus Assur. Leipzig 1921} (Ausgrabungen der deutschen Orientgesellschaft {E, 5. Wissenschaftliche Veröffentlichungen der Deutschen Orient-Gesellschaft} 38). Berlin, Museum: V. A. 7498. Berlin, Museum: V. A. 8384. S. A. Birnbaum, The Hebrew Scripts, nos. 148–150, 154–161. A. H. Sayce and A. E. Cowley, Aramaic Papyri {discovered at Assuan. London 1906}; Eduard Sachau, Aramäische Papyrus {und Ostraka aus einer jüdischen Militär-Kolonie zu Elephantine. Leipzig 1911}; Birnbaum, op. cit., nos. 122– 147. Birnbaum, op. cit., nos. 21 and 22. Mark Lidzbarski, Ephemeris {für semitische Epigraphik. 3 vols., Giessen 1902–1915}, vol. 2, p. 202. {David Diringer, “On Ancient Hebrew Inscriptions Discoverd at Tell edDuweir (Lachish)” – Part II, in:} PEQ 73 (1941), pp. 89–109, pl. VIII. 5. Birnbaum, op. cit., no. 2. George A. Reisner [et al.], Harvard Excavations {at Samaria, 2 vols., Cambridge 1924}; Birnbaum, op. cit., nos. 4–9. Round figure for the years 773/772, 772/771, 767/766. Birnbaum, op. cit., nos. 19 and 20. Harry Torczyner, {Lachish. I:} The Lachish Letters. {Oxford 1938}; Birnbaum, op. cit., nos. 23–26; idem, “The Lachish Ostraca”, in: PEQ 71 (1939), pp. 20–28, 91–110 {= Chapter 32 in the present volume}. Round figure for the period 597–588. Corpus Inscriptionum Semiticarum (= CIS). Vol. 2, {Paris 1889 ff.,} nos. 10, 15, 16, 38. Ibid., nos. 2–7, 11. Birnbaum, op. cit., no. 3. Round figure for the period 782–746. Birnbaum, op. cit., no. 14. Round figure for the period 716– 684.

000019

Birnbaum 01.04.11 18:17

I. The Material in Palaeo-Hebrew Script Sherd 1 (plate I. 1) m l w k r b u b w q h m y u h // k r b || › m r y w h n m i |

1 2 3

brks˙ lm ˚ brk // hw mhqs˙ bw ymnhs˙ rm› |||

The contents are by no means clear. Perhaps the fact that each of the three lines is written by a different person22 may help us to discover the meaning of the ostracon. Line 1: We could take brk as the name of the addressee and s˙lm as the greeting addressed to him: baruk, sˇalom ‘Baruch, greetings!’ On comparison with the Lachish Ostraca this interpretation does not commend itself. Most of these begin with the greeting (IV, V, VIII, IX ), one with the address of the sender (III ), and two start with the designation and name of the recipient (II, IV : l dny y ws˙ ). Of these, the latter is the nearest in meaning to the interpretation of brk as ‘Baruch’. It differs, however, in an important respect. In LachO, the recipient is not directly addressed, but l is placed before his name. It may therefore be preferable to take brk as the greeting: baruk ‘be blessed!’ The next word would then presumably not be another greeting – baruk, sˇalom ‘be blessed! peace’ – but the name of the addressee: sˇallum (or sˇillem). Line 1 would accordingly read: ‘Greetings, Shallum!’ This result – arrived at before examination of Sherd 4 – would be borne out by that ostracon. Finally, the possibility must be mentioned that the line contains no greetings but consists of two names: ‘Baruch, Shallum’. Line 2: It would seem natural to read brk as a name: baruk ‘Baruch’. The numeral sign after it would connect its bearer in some way with two units of something – we do not know what. If we were to read brk as barka ‘thy grain’ (Yeivin) we would be exchanging our knowledge of the addressee’s name for that of the material. As the seventh sign is not a resh we cannot read ro¯ ¯ım ‘shepherds’. This sign not being pe either, we cannot read p m. It would be very difficult to interpret too. Happa am signifies ‘this time in particular’ as distinct from other times, it is not the colourless expletive ‘now’. As a matter of fact, the phrase that has been suggested: ‘now pay attention’, would be a sort of expletive, and we would not expect to meet with that kind of thing in the short and necessarily elementary text usual on an ostracon. The letter in question 22

This is obvious from the script, and can be verified by a comparison of the individual letters (see pp. 21–24).

000020

20

Birnbaum 01.04.11 18:17

[ 31 ]

being most probably a waw, we seem to have here a phonetic spelling of the personal pronoun, third person masculine. The subsequent letters m, if read im ‘with’, would be compatible with such a reading, and the next letters hqs˙ too, would fit in, if read haqqasˇ ‘the straw’. Such a word would not be surprising in a context containing the word ‘barley’ (line 3). The last two letters of the line could then be read bo ‘in it’, ‘in him’, and this might conceivably refer to the field. The second half of the line is very crowded, especially at the end. I rather think no writing is missing on the left and the surface of the edge has only splintered off.23 Anyhow, if haqqasˇ is correct, and b means ‘in’, then w cannot be the first letter of a word. Could we have here a haplography for hqs˙ s˙bw, the second s˙ being the relative pronoun? This word is considered to be late, but on the other hand, we have it in Phoenician (usually with a prosthetic aleph). Line 3: If my suggestion that nothing is missing on the left of line 2 is correct, then the first letter of line 3 could not be the final one of the last word in line 2, and in that case the reading min ‘from’, ‘of’ would be impossible. Ymn hs˙ rm ‘he shall count the barley’ or ymnh s˙ rm ‘he shall count barley’ does not seem too likely in a document of this kind where we expect the registration of facts rather than references to the future. In addition, it would be striking in such a short text to employ quite unnecessarily the word ‘counting’ – the idea being already expressed in the figure connected with the word ‘barley’. It appears preferable to regard ymnh and the word brk above it as parallel to each other, i. e., ymnh would be the name Imnah (Gen. 46: 17, etc.). In connection with the name Baruch we have an indication of a quantity (‘2’); followed by a word denoting a material (‘straw’). Likewise – although in different order – the name Imnah is connected with an object (‘barley’) followed by an indication of quantity (‘13’, or ‘3, 8, 23’ etc.). Tentative Translation and Interpretation: 1 Greeting, Shallum! 2 Baruch: 2; the one (?) with the straw. 3 Imnah: barley: 13 (or 3, 8, 23, etc.). Perhaps our document is a receipt made out by two people – that would fit in with the fact that lines 2 and 3 are written by different hands. It would have been handed to a man in charge who forwarded it to Shallum and therefore wrote line 1, the formula of address. Or it might 23

In line 1, however, a rather bigger bit has come off, although even that must have been a very narrow splinter.

000021

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

21

conceivably be an account of work done by two workmen, written down by themselves, and sent on by their supervisor or foreman. Bibliography: A, 152–154; C, 71–72; G; J, 131–133; K, 37–39. The Date of the Script Line 1: B: The downstroke bends just slightly to the left without forming a proper base. This characterises it as a cursive form, the first of its kind being known to us in the pen-written cursive of HarO. But even there the form with a proper base preponderates. In LachO there is only one example of the non-cursive type. K: The downstroke sweeps considerably to the left. As we meet with this feature for the first time in HarO, our form is not likely to be earlier than the eighth century. The left stroke, slanting down to the right meets the shaft at right angles. After this form had been reached there was no further change in the engraved type, while in the pen-written cursive the direction of this stroke changed completely between the times of HarO and LachO. In view of the cursive influence apparent in both, the absence of such influence in our kaph might possibly indicate a considerably earlier date than LachO – not far from HarO. Although most specimens in HarO have the three strokes joining at one point, it does occasionally happen that the inner one runs to the middle of the left one, as in our form. Thus HarO would provide our terminus a quo. L: The top stroke is straight, which is on the whole a late characteristic and to be met with from HarO onwards. The bottom stroke slants down to the right – a rare feature. Perhaps this form is a final development of the type where the curve had become angular. A datable document with a specimen of this kind is ElS. Har 4925 has the same slant but the main stroke is curved. The angular transition form mentioned above is very frequent in HarO; we get it, too, in Sil, but it is not present at all in LachO. M: The downstroke bends considerably to the left, pointing to the same date as in K. S˙: The two halves are not symmetrical, the right inner stroke does not issue from the top of the left inner one but from a point slightly lower down. Hence our ostracon would not precede the eighth century; before that we have only the symmetrical form. This W-type is the prevalent one in HarO, but there are already cases where the left inner stroke runs to roughly the middle of the right one; in Sil we have exactly the same form as in our ostracon; in LachO, apart from occasional W-forms, the right inner stroke runs to about the middle of the left one ( W); this form is a cursive counterpart of the Sil type. Hence our ostracon would not be distant in time from Sil.

000022

22

Birnbaum 01.04.11 18:17

[ 31 ]

Line 2: B: The first specimen has a base and does not give us any chronological clues. The second specimen tallies with that of line 1 (see discussion there). H has a cursive feature which we first meet with in HarO: the left top stroke is produced to the right. Another cursive trait is that the bottom stroke is curved – or curls up at the end – and is longer than the other two. We find this neither in HarO nor in LachO but it occurs in Sil, lmlk III, some undated seals, and Jaaz. W: Although the top of the first specimen might possibly indicate a connection with one of the specimens in GezT, it is unlike the other waws there. Our form corresponds to the later type which first makes its appearance in HarO. From the specimen at the end of the line, it is not possible to decide whether the top corresponds to the form of HarO or to that of the next stage, represented in Sil and LachO. To judge by the first specimen, it looks as if it was the later stage that we had here. K tallies with that of line 1 (see discussion there). M: As to the downstroke see the discussion of kaph in line 1. The left part consists of two ticks; the top strokes are parallel, the bottom ones look like a continuation of each other, but on closer inspection we see that they are not made by a single stroke. Our form tallies with that on many seals, jar handle stamps (especially lmlk II and III), LachO I, and is very like those on some of the other LachO. It seems to be a transition stage between the two-tick type of HarO and the Sil type, where the base strokes have been fused into one. At the same time the relative position between this horizontal stroke and the downstrokes has changed: the former is no longer a base but crosses from centre to centre. What appears to have happened is that the formal tick type was influenced by the new cursive one-stroke type: the base started slanting, but then became horizontal after the manner of the cursive form; finally, the slanting bit was omitted, the writer moving his pen back, in an upward direction, on the downstroke, or else just omitting the slanting stretch, and starting the horizontal higher up straight from the downstroke. Ayin stands in between the slightly wider and not very sloping form of HarO, and the narrower and more slanting one of LachO. Q: Our earliest specimen of this type occurs in HarO. It is also the form of Sil, ElS and of half the specimens in LachO. But in none of them is it open like here. The rest of the LachO specimens are open and appear to represent a more advanced stage of development than the type in our ostracon. Hence this might be earlier than LachO. S˙: The right inner stroke seems to issue from the middle of the left one as in Sil.

000023

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

23

Line 3: H: See line 2. Y: The base is exceedingly long. It ends as a tick, a not infrequent feature in HarO which we do not meet with in later documents. The length of the base connects our letter with Sil. M: The downstroke is extraordinarily long, which places our letter closer to Sil than to HarO. At the end it sweeps vigorously towards the left. This again indicates a date after HarO, as Sil has an analogous feature. Our form is an emphasised and slightly angularised version of that in Sil. The top part consists of two ticks. Hence a date not too far from HarO is suggested, as even at that time the later form is already beginning to appear. However, in some of LachO, the earlier type is still recognisable. N: As to the downstroke, see the preceding letter. The top part is a rectangular tick, while elsewhere it forms an acute angle. The middle stroke slopes down to the right – a specimen of this kind occurs in LachO. As far as I can make out on the photo, this middle stroke runs to the top of the downstroke. Although that is a feature of the original form, it is absent in HarO and Sil (where the downstroke starts slightly higher up). It appears again, however, in Jaaz, ElS and LachO. In the latter, most specimens have no top stroke. This might perhaps indicate that our ostracon was earlier. Ayin tallies with that of HarO in having a curve on the right and an angle on the left. The form is consequently wide. Sil has a similar but more triangular one. As in HarO, our specimen has a fairly vertical stance. R: Comparison with the specimens in lines 1 and 2 shows that the length of the downstroke – although already very considerable there – is doubled by the writer of the third line, thus placing our ostracon much nearer to Sil than to HarO. The angle on the left has been rounded off, the triangular outline yielding to a semi-circular one. The top stroke tends to curve in HarO. From Sil onwards the curve becomes a regular feature, in LachO it is generally more pronounced than in Sil. Our ostracon is therefore closest to LachO in this respect. S˙: The right inner stroke has disappeared. This form corresponds to those in lines 1 and 2 – it is more cursive than they – and gives us Sil as the terminus a quo. We shall now tabulate our results, substituting figures for the general indications given above.

000024

Birnbaum 01.04.11 18:17

24

[ 31 ]

Lines

1

2

3

B

770–

770–

–

H

–

770– 700–590

770– 700–590

W

–

770– (700–)

–

Y

–

–

770 700

K

770– 770 770–

770– 770 770–

–

L

770– 590 –700

–

–

M

770–

770– after 770 before 700

N

–

–

700 590 before 590

Ayin

–

after 770 before 700

770 770

Q

–

770– before 590

–

R

–

–

700 770– 590

S˙

800– 700

700

700–

700 700 –770

The latest terminus a quo is 700, the earliest ad quem is 770. This gap of about two to three generations is, of course, not a contradiction, as these termini are only very approximate, and the generations overlap. What we can say is that our ostracon is likely to have been written at some date within that period. This is confirmed by certain features of mem and ayin. The features pointing to the earlier limit, and those favouring the later one, are of fairly even weight, so that we may perhaps not be far out if we think of a date equidistant from the extreme dates, i. e., round about 735 B. C. E.

000025

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

25

Sherd 2 (plate I. 2) To judge by the drawing every letter is perfectly clear: ywiuil

There seems to be no doubt that we have here a name preceded by the possessive l. That gives us an additional name for our ancient Hebrew onomasticon. It is the inverse of Ys˙ yhw. Translation: ‘Jojesha’s’. Bibliography: B, 203; C, 311.

Sherd 3 (plate I. 3)

] ar“‘ [ ] h ar“‘ z y l

r¯ r¯ l z h [] [ ’ ’ We seem to have here one of the names zr (Ezer, I Ch. 7: 21; 12: 9; Azzur, Jer. 28: 1) or zrh (Ezrah, I Ch. 4: 17). The latter appears less likely as the little hole after the resh might be a word separator. If we were to read z , it would make Uzza, or a hypocoristic name (Eza, Azza). The reading zr(y)h(w) is hardly likely. There is certainly not enough room for a Y, and we would have to assume a slip on the part of the writer; furthermore, the Harvard ostraca have the form -yw, not -yhw. Translation: ‘Ez...’s’. Bibliography: B, 201; C, 308–309. The Date of the Script H provides no chronological clue. Z: There are no tails. Our latest specimen of this early form occurs in HarO, according to the hand copy of HarO 37. Elsewhere in HarO the lower stroke has a tail. The time of HarO is therefore our terminus ad quem. L is an angularised variant of HarO and thus tallies with Sil. Ayin is narrow, as in HarO and Jaaz. Tabulation: Z L Ayin

–770 770–700 770–590

This suggests a date not much after 770 – say about 750 at latest.

000026

Birnbaum 01.04.11 18:17

26

[ 31 ]

Sherd 4 (plate I. 4)

]xw ] u” i l q ] (k)‘ m s ] – ra ] n |||

1 2

s˙ h [ ˙ qlyw ¯¯ [

3 4 5

s m (k)‘ [ r –[ || n [ |

To make out the meaning from this narrow strip of letters seems too much to expect. If there really was a K in line 3, it might perhaps give us the name Semachiah, recorded in the Bible and the Lachish Ostraca. Bibliography: A, 154–155; C, 73. The Date of the Script Aleph: It is hardly possible to say whether the two transverse strokes join left of the shaft, or whether what we see belongs to the upper stroke only; further, whether the two strokes are intended to be parallel or not. What we have here seems to be a transition form between the earliest type with converging strokes, as in HarO, and the later, with parallel ones, as in Sil, Jaaz, LachO. Y: There is no base, and the transverse stroke seems actually to cross the downstroke. Thus our form tallies with that of LachO. L: Sharply pointed forms occur at all periods, and angularisation of the round form is the rule in HarO, and remains so until Jaaz and Ged. An almost identical form with that in our ostracon – wavy bottom stroke bending up on the right to form a straight line – occurs in the bt lmlk graffito of Lachish,24 which, because of its mem, suggests a date hardly much later than HarO. Hence our ostracon would belong to the same period. M: The top part is a practically complete W. This is usual in lmlk I but HarO, Har 2854 and ShemS already have the abbreviated type. Sil shows that late in the century even that form had undergone a great deal more development. This suggests that by then the W-type had gone right out of use, and that our form would hardly be very much later than HarO. The downstroke is not particularly long, which points to it being earlier than Sil, ElS, Jaaz. N: As to the two top strokes, see Sherd 1. Regarding the downstroke, see here in mem.

24

PEQ 73 (above, n. 9), pl. IX.

000027

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

27

S: In HarO the final tail is still very small. The earliest dated specimen showing it to have grown to reach the shaft is in LachO, but the fact that we meet with it in the hmlk smk seal25 from Lachish – and ˙ cursive development and adoption usually there is a time lag between into formal writing – might mean that it is appreciably older than LachO. Q: The construction of the top part, as far as can be made out from the photo, is that of Square Hebrew, except that at the top on the left there is a slanting stroke of considerable length, instead of a small kind of serif. Thus it differs from the corresponding type in HarO in having a sharply pointed angle instead of the left top curve. According to the photo it seems possible that the slanting stroke is the end of a curve which would, more or less, restore the original outline. The angle is characteristic of LachO. As LachO employs a still further developed type, side by side with this one, there might be a chance of our ostracon belonging to an earlier period, before that later form had developed. Tabulation: Aleph

after 770 before 700

Y

590

L

770

M

770 before 700

N

590 before 700

S

before 590

Q

after 770 before 590

This table indicates a date between 770 and 700. In order to find a likely point within this period we must take into account the pull exerted by various features towards the earlier limit, on the one hand, and towards the later, on the other. This would give the M a somewhat greater significance, and place our ostracon slightly nearer to HarO than to Sil – say round about 745.

25

S. H. Hooke, {“A Scarab and Sealing from Tell Duweir”,} in: PEFQS 67 (1935), pp. 195–197.

000028

Birnbaum 01.04.11 18:17

28

[ 31 ]

Sherd 5 (plate II. 5) lph ” [ ˙ The spacing makes it likely that none of the last three letters begin a new word, and that none of the first three letters end a word. Thus these four letters form a ‘word’ by themselves, or at least the main part of a ‘word’, since Hebrew words are generally short. If – as is not unlikely in this kind of document – it were a name, the l could have been preceded by an aleph, or a theophorous element could have followed ph , with l expressing ownership. But it is equally possible that ph is a ˙ hypocoristic, as Sukenik suggests, from a root ph , pwh etc. ˙As he ˙ 26 and three pointed out, this name occurs in the Elephantine˙ papyri 27 times in Palmyrene.

] a“ x p l

Translation: ‘Peha’s’. ˙ Bibliography: A, 152; C, 70. The Date of the Script Aleph: The traces seem to indicate that the cross strokes were parallel. Not knowing what were the conditions on the left of the shaft we can only say that the form on our sherd appears to tally with that usual in Sil and after that. The length of the shaft also connects it with Sil. H and L: See Sherd 9. P: The downstroke is very long indeed, indicat˙ing a date not too far from Sil. Tabulation: Aleph

700– 700

L

770–700

P

700

This table seems to suggest a date round about 725 B. C. E.

26 27

A. [E.] Cowley, Aramaic Papyri {of the Fifth Century B. C., Oxford 1923}, no. 40. Jean Cantineau, Inventaire {des Inscriptions de Palmyre. Beyrouth 1930}, vol. 4, p. 25.

000029

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

29

Sherd 6 (plate II. 6)

] dl

ld [

Sukenik suggested the beginning of a proper name, evidently taking the l to signify ownership, and the second letter (which he read r ) as the first of a name.28 Bibliography: A, 155; C, 73–74. The Date of the Script D: The left stroke does not project on top. This is practically the only type in HarO. Our ostracon cannot be much later, since from Sil onwards the left stroke always juts out. As in Sherd 7, the right stroke is of considerable length, and accordingly indicates a date in the neighbourhood of Sil. L provides no chronological clues. A date halfway between HarO and Sil suggests itself – say around 735 B. C.E.

Sherd 7 (plate II. 7) brk hz ˙ The suggestion that brk be taken as baruk ‘Baruch’ raises the question as to what the name was intended to convey. If it referred to the owner of the vessel, or to the addressee of a letter we would have expected it to be introduced by a l, as on Sherd 6, and regularly elsewhere. Hence the interpretation as a participle is preferable. To see in it the predicate of the subsequent subject – which is obviously the name Ahaz – and to explain this sentence as a votive text: ‘Blessed be Ahaz (by God)’, is grammatically possible, of course,29 but it does not seem right. We know of no Hebrew votive inscriptions. More natural would be the interpretation as a greeting. Again, it is grammatically possible to take brk as a construct and translate: ‘Blessed by Ahaz’. But that could hardly be a greeting. It would be more natural to take the words as a nominal sentence: ‘May Ahaz be blessed’ or ‘Mayst thou be blessed, Ahaz’. This would make a suitable beginning for a letter. To the objection that the writing would have been smaller and the available space altogether more economically used if a longer message had been intended, zxakrb

28

29

Names starting with r are infrequent. In HarO there is only rp (no. 24); the first letter of line 4 in HarO 1 can hardly be a resh. In the Bible we have during the period of the Divided Monarchy the name Rehoboam. There is a corresponding expression in Gen. 14: 19 (though not in the mouth of an Israelite).

000030

Birnbaum 01.04.11 18:17

30

[ 31 ]

the answer may be that the writer very soon realised he had begun too big, and that he had stopped and made a new start on another sherd. That would then explain why the rest of the surface is empty. Tentative Translation: ‘Greetings, Ahaz!’ Bibliography: B, 202–203; C, 309–310. The Date of the Script Aleph: The horizontal strokes are parallel, the short one on the left stands in between the two on the right. We meet with a form like this on some seals and it appears to occur sometimes in Sil. The parallelism of the strokes is already firmly established by the time of Sil, while in HarO there is rarely a trace of any trend towards this development. That would point to our ostracon being closer in time to Sil than to HarO. B: The downstroke slopes well to the left; we first meet with this in HarO. The left part is not an angle but a curve: the top triangle has become almost a semi-circle. We find the first traces of this development in ShemS and HarO. Thus our ostracon would be appreciably later. On the whole, the triangular structure continues in existence permanently, although, after Sil, the left top stroke is always more or less curved. Z: The top stroke has no tail, a feature which does not occur in our material after HarO. The lower stroke has a considerable tail. This development had already started by HarO; there are some very long specimens in LachO and Jaaz – side by side with shorter ones – so that our form might place the ostracon not far from that period. K: The downstroke is very long which brings us within the neighbourhood of Sil. At the end it bends to the left – an angular correspondence to the forms from HarO onwards. What on the photo looks like a continuation upwards, parallel to the downstroke, is probably not part of the letter. The middle top stroke runs to the centre of the left one; the earliest dated example of this feature is HarO 22. R: The top stroke is very slightly curved. Traces of this development are to be met with in HarO, and after Sil it is the rule. Tabulation: Aleph

700–

B

770– after 770

Z

–770 770– 590

000031

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

K

700 770– 770–

R

770–

31

The date of our ostracon would thus be between 770 and 700. The features favouring the earlier limit and those pointing to the later one, are of fairly even weight, so that a date equidistant from the termini would not be unreasonable – say around 735.

Sherd 8 (plate II. 8) almdl

ldml

The initial l indicates ownership and dml is a hypocoristic of dmlyhw. This name is known from two seals where it was read, in spite of Clermont-Ganneau’s and Torrey’s correct readings, as Remaliah/o, because the latter name occurs in the Bible. Clear and unambiguous palaeographical evidence30 was set aside for arguments based on extraneous evidence. The root dml occurs in Arabic. Besides an agricultural meaning there are these: ‘to put a thing into a right, or proper, state’; ‘prepare, improve it’; ‘make peace, effect a reconciliation, or adjust a difference between the people’; ‘heal a wound’. That one of these should form part of a theophorous name seems to present no difficulties. But how is the presence of this root here to be explained? Could its absence from the Bible and from post-Biblical Hebrew literature be just accidental? Translation: ‘Demala’s’. Bibliography: B, 200; C, 307–308. The Date of the Script Aleph: The shaft is very long, suggesting a date for our ostracon in the neighbourhood of Sil. The cross strokes form an angle on the left of the shaft. Our latest dated specimens having this feature are in HarO. Right of the shaft, however, the cross strokes are parallel, as from Sil onwards. As in Sherd 9 we have here a transition form between HarO and Sil, or rather, between HarO and Sherd 9, where the hand first travelled back on the top stroke. According to this our ostracon would be closer to the time of HarO than of Sil. 30

In the seal of dmlyhw bn nryhw there was actually an r for comparison.

000032

Birnbaum 01.04.11 18:17

32

[ 31 ]

D: The left stroke juts out on top. In HarO, among the many specimens of the letter, there are – if the hand copies are correct – only one or two possible indications of a development in this direction. By the time of Sil however, the fully developed type is the only one. The right stroke runs rather far down. The beginnings of this feature are to be seen in HarO, but not until the time of Sil was it well developed. L: See Sherd 9. M: The downstroke is very long, as in Sil. It is almost straight, and from its end a straight line bends sharply upwards at a small angle. This is an angularisation of the Sil type. Both the curved and angular types persist. Thus Sil is the terminus a quo for our type. The top consists of two ticks, written separately: the bases are at very slightly different angles; the downstrokes are parallel. This form is almost identical with the one we saw in Sherd 1, line 2. Tabulation: Aleph

700 –770 770

D

between 770 and 700 ” ” ” ”

L

770–700

M

700 700– between 770 and 700

The termini lead to a date round about 735, and there are no features important enough to modify it towards a slightly earlier or later time.

Sherd 9 (plate II. 9)

] lxl ba

lhl [ ˙ b

Line 1: As lhl is no root, and as there would be enough room before the first lamedh˙ for the end of a preceding letter, the absence of anything makes it possible to assume that this is the beginning of the text, and that the l indicates possession. hl would then be the first two letters of a ˙ name, and sadhe has been suggested by Sukenik as the third radical. If the trace following the second lamedh were a stroke, it could not belong to a sadhe. It has no tail, and a tailless sadhe would be of so early a date as to be incompatible with the late forms of the other letters. As that

000033

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

33

trace could only be part of an aleph, hl would be a hypocoristic name. If however it is not a stroke at all, our˙two letters could be the beginning of Heles or any of the names that start with hl. ˙ ˙2: It has been suggested that b is the ˙end of another name (i. e., Line presumably the father’s), the beginning of which is in line 1. This is possible but doubtful. Among the few graffiti of this kind we have three where the sherd is not broken off after the owner’s name, and in all these cases (the next ostracon, lmlkrm of the Harvard excavation and lsmk of Tell el Hesy) no writing at all follows. Hence it would perhaps be safer not to expect a patronymic here. I have no suggestion for line 2. Translation: ‘Hl...’s’. Bibliography: B, 201–202; C, 309. The Date of the Script Aleph: The left half of the lower cross stroke runs back into that of the upper one, and, just before reaching the shaft, begins sloping down to the right and continues in that direction for half its right side; finally it turns upwards running parallel with the upper cross stroke. We have here a confirmation of our suggestion in Sherd 4. This seems clearly a transition from the earlier type to that with parallel cross strokes, and would suggest a date for our ostracon after HarO and before Sil. B: The left stroke of the top part is only slightly curved, but the base is rounded, and the triangularity of the top part has, strictly speaking, gone. The first traces of this development are to be seen in ShemS and HarO, so that our ostracon might be appreciably later. The left stroke does not issue from the top of the downstroke but starts higher up. Specimens of this kind occur in HarO 22 (and in the Shallum ben Adoniah seal),31 but, being so exceptional, cannot safely be used as a chronological clue. Nor may any inferences be drawn from the fact that such a one occurs on the coins of the First Revolt. The base is very long, connecting our form with that of Sil rather than HarO and ShemS. H: The right shaft projects downwards only, the left one upwards ˙ i. e., cursive development has (I) shortened the right shaft on top only, so that the upper transverse stroke can be written together with it, with one lift of the pen, and has (II) shortened the left shaft at the bottom, so that the lowest transverse stroke can be written together with it, likewise with one lift of the pen. The same top is to be found in HarO, but for the 31

{E. J. Pilcher, “Signet with Old-Hebrew Inscription”, in:} PEFQS 51 (1919), {pp. 177–181, illustration} p. 178.

000034

Birnbaum 01.04.11 18:17

34

[ 31 ]

bottom we have no earlier specimen than in LachO. But perhaps we ought not to draw any chronological conclusions from this feature seeing that elsewhere (on the ivories from Samaria32 and Arslan Tash33) it is already present in the ninth century. L: The bottom part tallies with the forms in HarO and Sil; it is rounded, in contrast with ShemS and HarO 2854 on the one hand, and Jaaz and Ged on the other, where we have either an angularisation of the round form, or the old, straight bottom line. Accordingly our form would belong to the period between HarO and Sil. Tabulation: Aleph

after 770 before 700

B

after 770 700

L

770–700

This table points to an intermediate date between 770 and 700. There is nothing to tip the balance in favour of either an earlier or later part of this period so that we might place it approximately half way, i. e., around 735.

Sherd 10 (plate II. 10) n h ( b)‘ [ m

] (b)‘ h n m

If there was a b before the h it would not have formed a root together with the two subsequent letters: bohen ‘thumb’ is surely most improbable. Nor does bahem/n ‘in them’ appear likely; the same applies to taking hm/n as a suffix of another word. The only words to end in hm/n are: zhm, yhm/hmm, nhm, sˇhm, thm, khn – all unlikely candidates. Could hm/n be the end of a name? A name has a good chance of occurring on these ostraca. The only name to end in hm is ‘Abraham’, which is hardly possible, and there are no names ending in hn. A possibility we have not yet considered would be to take hm/hn as a word by itself. The reading as a (separate) pronoun can be discarded as improbable. Thus we are left with hin ‘Hin’. Experts would presumably be able to calculate whether the vessel of which our fragment formed part could really 32 33

Samaria-Sebaste (above, introductory note) vol. 2, pl. XXV , 8. Franc¸ois Thureau-Dangin, Arslan Tash. Paris 1931, p. 91, {fig. 33,} no. 2.

000035

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

35

have contained that measure (about 6 litre). But, of course, hin might have been preceded by ‘half’ or ‘quarter’. The Date of the Script H provides no chronological clues. N tallies with the specimen in Sherd 1, except that the bottom here is completely angular. Accordingly the ostracon might perhaps be assigned to about 700.

Sherd 11 (not illustrated) The outer surface is covered with scratchings, some of which resemble Hebrew characters. They seem to be the work of a child. Bibliography: A, 155, fig. 4. The Date of the Palaeo-Hebrew Sherds We have examined the ostraca independently of each other, so that our results were based on the comparative material only. In no case did we take in account the dates arrived at for the other sherds. If, in the end, we find that all these dates belong to one and the same period – around 750, 745, 735, 725 – it is not unreasonable to see in this a mutual confirmation, and to assign them with some measure of confidence to about the third quarter of the eighth century B. C. E.

Concluding Note No general conclusions can be drawn from our material because it is too varied in nature. Half of the sherds (nos. 2, 3, 5, 8, 9, and, probably, 6) are not ostraca (i. e., sherds used for writing) – the letters were incised when the vessels were still whole. The words indicate a connection – expressed by l – between the vessel and a certain person: the owner. It seems impossible to say whether the vessels might have contained offerings. Three sherds are ostraca. Two of them are communications of some kind (nos. 1 and 7). No. 4 might be a list. No. 10 cannot be classified. One document (see section B, p. 47) is of an official character: the stele fragment comprises part of an inscription. Most of the names are known to us from the Bible, some occur also in the Harvard ostraca: Ahaz (p. 29) [2 K. 16: 1; HE 2]34 or Ahzai [Neh. 34

{HE + number = Harvard Excavations (above, n. 11), vol. 1, Part IV.1 Inscriptions, pp. 232 ff.}

000036

Birnbaum 01.04.11 18:17

36

[ 31 ]

11: 13; HE 25] or hzyw [corresponding to Ahaziah, 1 K. 22: 40]; Helez ˙ HE 30 etc.] or Helkai [Neh. 12: 15] or hlqyw (p. 33) [1 Ch. 2: 39; [corresponding to Hilkiah, 1 Ch. 26: 11]; Ezer (p. 25) [1 Ch. 4: 4 or˙ Neh. 12: 42] or Azzur [Neh. 10: 18] or Ezra [1 Ch. 4: 17] or Uzza [Ezr. 2: 49; HE 1]. Three names are non-Biblical: Demala (p. 31) – hypocoristic of Demaliah, known from seals; Jojesha (p. 25) – a completely new name; Peha (p. 28) – known from Elephantine. ˙

II. Sherds with Letters in Aramaic Script (a) Incised and Stamped Material Sherd 12 (plate III. 12)

] d“ b y i” z”

z¯ y¯ b d¯ [

What was the purpose of this stamp? It is a very unusual one. Jar stamps are placed on the handles, this one is on the body of the vessel. Does it tell us the name of the maker? Having decided against Y as the first letter and H or W as the second, we cannot read yh bd or yw bd which are not only unknown to us but would hardly seem to make possible word forms: if the Divine element comes first, the second part is either a verb (e. g. Jojakim) or a predicative noun (e. g. Joab). Now a name ‘Yw/yh is servant’ is obviously out of the question. And it would certainly be going too far, if in order to have a verb as the second element, we interpreted bd as an Aramaic word and assumed an Aramaic name, ‘Yw/Yh has made’, corresponding to Asael in meaning. If our reading zy bd is correct, its obvious interpretation would be zy bd, i. e., zi abad ‘(that) which he made’. This would presumably refer to the jar, and the name of the maker would follow on bd. Would not that be too long a text for a plain stamp? However, zi could also be taken as a nota genitivi and it would correspond to the l of the Hebrew stamps (lmlk, ly znyhw, etc.). What follows – bd – would then be a name, and we would naturally think of supplementing it by adding two letters, giving us: bdyh (or bdyw). That would leave the stamp with a short text, and in a traditional form: zy bdyh/w/hw ‘Obadiah’s’, i. e., either ‘belonging to Obadiah’ or ‘made by Obadiah’. However, this reading and interpretation are just a suggestion. The Date of the Script B: The top part of lapidary B remains closed – although it is no longer triangular – until the first half of the sixth century (604–561; CIS

000037

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

37

II, no. 54). In the cursive, however, we find a slightly open form already in Shal, and a few decades later it is wide open on top, in Sen. This period, then, constitutes the terminus a quo for our ostracon. Here the left stroke is almost vertical, joining the right one below the middle of the line height, as in our earliest specimens. Later – especially after 500 – this form is very rare, the left growing increasingly oblique until it becomes almost horizontal. It finally joins the top of the right stroke, which is by then somewhat lower than the line ceiling. Thus we might perhaps think of a date not after 500. D (if this be the correct reading) leads to the same result. Z: Our earliest specimen of the Z-form occurs on the Zakir stele35 (first quarter of the eighth century), our latest in Esan. Y: Specimens with the bottom stroke sloping down to the right are rare, e. g., 674 (CIS II, no. 39c), 482 (Ibid., no. 122).36 Whether they are only accidental or whether it was from this form that the simple yodh of later times developed we do not know. In none of the specimens is the angle between the main stroke and tail as obtuse as on our ostracon, and this might perhaps be considered as a point in favour of assuming a development which ended in a straight stroke. If that were the case we could take 674 as the terminus a quo. The terminus ad quem would be furnished by the time of Esan. Ayin is, by the time of Shal, imperfectly closed, and in Sen it is already open. This then is our terminus a quo. Tabulation: B

687– – c. 500

D

687– – c. 500

Z

800 –330

Y

674– –330

Ayin

682–

That would bring us to a date in the sixth century B. C. E. 35 36

Henri Pognon, Inscriptions se´mitiques {de la Syrie, Me´sopotamie et de la re´gion de Mossoul. Paris 1907}, plates 9–10. In Palaeo-Hebrew, too, we find a similar form at various times: GezT, ShemS, now and then on the coins (second century B. C. E. – second century C. E.).

000038

Birnbaum 01.04.11 18:17

38

[ 31 ]

Sherd 13 (plate III. 13)

] hkhgl[

]lghkh [

The sequence hkh suggests the word hkhn, i. e., hakkohen. The word preceding it might then have been a name. The only name to end in lg is Peleg. But we would hardly expect it here. Alternatively, if lg were not the end of a name but of another noun, or simply a word by itself, there would be log ‘log = 1/2 a hin’. This would make sense on a jar. For we must remember that our sherd is not an ostracon but that the text was on the jar before it was broken. That is clear from the way the writing is arranged on the sherd, and is confirmed by the way the appearance of the letters is described by Sukenik: “freely incised before burning”. But what is lg hkhn ‘the log of the kohen’? Lgh might be the end of the name Bilgah. If the reference were to one of the twenty-four priestly divisions, the question would arise as to how the word could be connected with the next one – koh[en] – which indicates an individual. And kh must be so read because, apart from khy (which seems impossible here) there is no other root starting with the sequence kh. To take that as the end of a word and read hikkaˆ ‘he smote’, or similarly, is obviously out of the question. Apart from this grammatical difficulty, is it likely that there would be a Jerusalemite institution in Samaria? It would therefore, be preferable to take Bilgah as the name of an individual. That would, however, still be inferior to the reading lg hkhn. Sherd 15, and possibly 12, being in Aramaic, our document might be in that language, too. The first H would then have to be linked with the preceding letters, resulting in palga ‘the half’. If the next letter constitutes the beginning of a root, we have only khy, khl or khn to choose from. An inscription on a vessel (not on an ostracon) could hardly contain a verb (khy, khl). We are thus once more left with only khn, and the word would be kahanaˆ. However, palgaˆ kahanaˆ is out of the question. To read pe˘lag as Hebrew – as would be natural at a later time – would, presumably, be too bold here. And what would be pe˘lag hakkohen ‘the part / share / due / half of the kohen’? Bibliography: D, 35–36; E, 155; H, 140–141. The Date of the Script G provides a terminus ad quem: this form precedes the third century B. C. E. because it was then that the circumflex shape gave place to a new construction – the left stroke now issued from the right one, at a point in the middle.

000039

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

39

H: The earliest specimen approximating to this type is known to us from Shal, although the earlier type still occurs in Nerab (first half of the sixth century). But the horizontal position of the left top stroke is not found before about the fourth century. Not until the third has it become general. As to the gap between the two left strokes, we have an example of it as early as 727–722 (CIS II, no. 6), but only by the first century B. C. E. did it become a common feature. The lower stroke, if prolonged to the right, would meet the upper one at the point where it joins the shaft, or near it. After the third century this form is very rare, for already in ElP the top of the lower stroke had started moving to the left, until, finally, in the second century B. C. E. it became parallel with the right stroke. Thus we might take the third century as our terminus ad quem. The top stroke is not on the line ceiling but considerably below it. Analogous constructions reach into the third century C. E. (Dura) but start going out of use by the first century. K: As to the top part, the earliest specimen of approximately this type occurs at the end of the seventh century (Pharaoh Letter,37 of 605) and in the following half century (Nerab I). The earliest example of the waviness of the downstroke is on the Meissner papyrus (515).38 The downstroke provides us with a terminus ad quem: by the time of EgD the end of the stroke has turned, with a wide curve, to the left; simultaneously it has greatly shrunken in size, and even shortish forms occur.39 37

38

39

A. Dupont-Sommer, {“Un papyrus arame´en d’e´poque saı¨te de´couvert a` Saqqara”, in:} Semitica 1 (1948), pp. 43– 68; A. Bea, {“Epistula aramaica saeculo VII exeunte ad pharaonem scripta”, in:} Biblica 30 (1949), pp. 514–516; A. Malamat, {“The New Aramaic Saqqaˆrah Papyrus from the Time of Jeremiah” [hebr.], in:} Ydy t hhbrh h bryt lhqyrt rs ys˙r l w tyqwtyh {= Bulletin of ˙ ˙ the Israel Exploration Society} 15 (1949), pp.˙34–39. {Hans Bauer / Bruno Meissner, “Ein aramäischer Pachtvertrag aus dem 7. Jahre Darius’ I”, in:} Sitzungsberichte der preussischen Akademie der Wissenschaften {phil. hist. Klasse}, Berlin 1936, pp. 414– 424. The fourth letter of the sherd cannot be a W. Such a top part as here does not occur in W after the eighth century, and is, therefore, incompatible with the horizontal stroke of the H because this feature starts in the fourth century. That top part would also be incompatible with the form of the second letter if – as has been suggested – that were a Y, this circumflex type having only started in the third century. Apart from that, the downstroke of a W is not as long as the one in the letter here. Its length is obviously not just accidental – all the characters on the sherd are well-shaped and carefully incised. The letter is double as long as all the others: the lower half was clearly meant to be a descender. Hence our sign can only be a K. – As to the second letter, if it were a Y it would not fit in with the K – there would be about a century of development between the two letters (see above). Furthermore, the stance of

000040

Birnbaum 01.04.11 18:17

40

[ 31 ]

Tabulation: G

– 4th cent.

H

4th cent. – 727– – 3rd cent. – 245 C. E.

K

605– 515– – 4th cent.

According to this, we get the fourth century B. C. E. as the approximate date of our ostracon.

Sherd 14 (plate III. 14)

] hil[

]lyh [

Sukenik’s suggestion that what we have here is the Divine Name Yh – parallel to the Megiddo Jar and the Harvard find lyw – is hardly acceptable: we would expect lyw. There might have been other letters on the right or left. Bibliography: D, 35. The Date of the Script H: The earliest specimens of this type appear in AshD and CIS II, no. 43; the latest specimens seem to be those of CIS II, no. 142 (third century). Y: The complete old form survived until Esan, which is therefore our terminus ad quem. L: The latest specimen without a tail to my knowledge occurs in EgD, although at that time the new form was already prevalent. The wide curve at the bottom is more characteristic for late ElP than for earlier papyri. Thus c. 400 might form a terminus a quo.

the letter would be uncharacteristic of Y (but characteristic of G). – In addition, it might be mentioned that the archaeological context as well as the script of the other sherds preclude an eighth-century date and indicate one in the fourth. – In short, it is quite impossible to read our word as the Tetragrammaton and to base conclusions on that. – Nor is the reading Judah ({F. M. Cross, “The Oldest Manuscripts From Qumran”, in:} Journal of Biblical Literature 74 [1955], p. 156, n. 17) correct. – The length of the fourth letter is as impossible for the supposed D as it was for the supposed W. The date suggested here – “ca. 200” – is at the very least a century too late.

000041

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

41

Tabulation: 660–3rd cent.

H Y

– 330

L

– 3rd cent. c. 400–

The latest a quo and the earliest ad quem result in the second quarter of the fourth century as the approximate date of our sherd.

(b) Pen-Written Material Sherd 15 (plate III. 15)

] x a r d n i z a b c n  1 n s b z y n d r h [ ˙ ˙ ] ||| ||| ||| (?) – – – 2 – – – (?) ||| ||| ||| [ The meaning ‘plantation, vineyard’ for nsb is deduced by Sukenik from that of ‘plant, shoot’ – he compares the ˙word nsbt in the Bodl. papyrus ˙ MS. Aram. a. l. (P), translated ‘farm’ by Sayce. The nsb of the in˙ or ‘he has scriptions means stele. The third word is beyond doubt ‘vow’ vowed’. The suggestion that the text is in the nature of a label indicating the producer and the year when the contents of the jar – most likely wine – were produced, is based on the erroneous reading of the third and fourth words as the Accadian name: Nabu-ahhe-(...), from which the final – ˘ ˘the elements are not separated, verbal – element is missing. In a name while in the present case there is clearly a vacant space before the aleph, to separate it from the preceding letter. Since we have no name here, the second word is presumably the relative pronoun and the third a verb. So far the translation would therefore be: ‘The nsb which h [] has vowed – ˙ be the correct 9[’. From this we may infer that ‘vineyard’ ˙could hardly translation of nsb . For a label the text ‘The vineyard which h... has ˙ ˙ of the) 9 (th year)’ seems to me too complicated. vowed – (produce In addition the word ‘year’ is only assumed on the basis of a most uncertain trace. Finally, the arrangement of the writing suggests that it was made on the sherd – i. e., not to serve as a label on the jar while it was still whole. It appears that the label idea ought to be discarded altogether. The strokes are probably numerals expressing a quantity. The quantity of the liquid in the jar could hardly be intended because that is given by the size of the jar. In connection with the word ‘vow’ it would not be

000042

Birnbaum 01.04.11 18:17

42

[ 31 ]

unreasonable to think of a sum of money. Our result would then be that h vowed money for a nsb , to the value of 9 (?) . . ., but we would, ˙ ˙ what a nsb was. It could, of course, be a unfortunately, still not know stele. But of what kind? However that ˙may be, the sherd seems likely to be a registration note. Since the text was most probably written on the sherd (not on the jar), it would be reasonable to assume that only little is missing at the end of line 1. Hence we may perhaps think of a name like Ahiya, Ahiyahu, etc.40 ˙ ˙ Tentative Translation: 1 The stele (?) which Ah(ijah?) vowed 2 9 (?) Bibliography: A, 155–156. The Date of the Script Aleph: While in ElP the left stroke is only a little stump, it is longish in our first specimen and long in the second; in that respect they tally with the forms in use from EgD to the first century C. E. B: The stroke begins with a tick, a development we first meet with in ElP. The downstroke and ‘base’ are curved. They do not form a right angle as in EgD and later. Our specimen would therefore be of an earlier date. D: The top part is a rather narrow tick like that in ElP. It is an earlier type than the EgD form, where the tick has become very obtuse-angled to the point of being, at times, curved or straight. Z is a straight stroke of line height, a form known to us from the time of AshD to that of EgD. H: Similar types first appear in Shal. The straight, high connecting bar˙ is used – although not frequently – in ElP and EgD. Y: The simplified type appears for the first time in AshD, sometimes41 side by side with the older form. The left stroke still issues distinctly from the middle of the main stroke; not much later it moves somewhat higher up, and in ElP a position close to, or at, the top preponderates. The letter is rather small; in late ElP a taller type appears, and by 40

41

In his article “Die Rolle Samarias bei der Entstehung des Judentums” (Kleine Schriften zur Geschichte des Volkes Israel, vol. 2, München 1953, p. 323), Albrecht Alt uses Sherd 15 as evidence (and basis?) for “die fremde Oberschicht” in Samaria. The erroneous reading Nabu-ahhe- of the ed. pr. and its ˘ ˘ painting an ethnic and interpretation of nsb as ‘plantation’ misled Alt into ˙ for which there is no support in our sherd. sociological picture Berlin, Museum: V. A. 7498.

000043

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

43

the time of EgD this has become prevalent. In the second century the left part begins to shrink. Our ostracon has the tall type with a long left stroke. It thus belongs to the period between late ElP and EgD. N: In ElP the wavy line is the only type, but by the time of EgD waviness is as a rule reserved for the final position. In our ostracon we can still detect on top the merest trace of what was formerly a curve. The bottom curve has been much reduced in length and is now in a horizontal position. Specimens of this type date from the time of EgD. S: In ElP the main stroke is a curve; in EgD we have a wavy ˙ specimen. In ElP the right part is generally shaped like a circumflex, i. e., issuing from the main stroke; in EgD we have a curve which runs in the opposite direction, i. e., towards the main stroke. In ElP there is as yet no indication of a base; in EgD there is a short but distinct base. Our letter has all the features of EgD. R: The conditions are similar to those described for the development of daleth. Tabulation: Aleph

3rd cent. –

B

5th cent. – – 4th cent.

D

5th cent. – – 4th cent.

Z

7th cent. – – 3rd cent.

H ˙

5th cent. – – 3rd cent.

Y

late 5th cent. – – 3rd cent.

N

3rd cent. – 3rd cent. –

S ˙

3rd cent. – 3rd cent. – 3rd cent. –

R

5th cent. – – 4th cent.

The latest terminus a quo and the earliest ad quem indicate the two half-centuries round the year 300 as the date of the writing. The a quo is of slightly lesser weight because the absence of fourth century material

000044

Birnbaum 01.04.11 18:17

44

[ 31 ]

makes it possible to assume that the form in question was already in existence before the third century. Hence it might be preferable to assign our ostracon to the second half of the fourth century B. C. E.

Sherd 16 (plate III. 16) t“ h“

||| • —“

t‘ n” w

˜‘ All we can say is that this ostracon possibly contains a date. The two signs after s˙nt – if that is the correct reading of the word – are presumably numerals, which is corroborated by the final strokes whose role as ‘3’ can hardly be doubted. It does not seem unreasonable to take the first of the signs as ‘10’ and the second as ‘5’. This would not tally with the system of HarO, where, according to the generally accepted interpretation, ‘10’ is • and ‘5’ is —. Our signs are hardly likely to have developed from these because, even if the writers of our Aramaic ostraca were the descendants of those who wrote HarO, the old system of numerals had gone out of use together with the old script. Hence the two signs in question ought to be Aramaic. The first, if it were a horizontal stroke only, would tally with the Aramaic numeral ‘10’. If our sign were —, its Aramaic nature would be doubtful, although it could be derived from the Aramaic variant with a little tail, as in the papyri. If, however, our sign were ˜, the connection with Aramaic would be still more unlikely. If the next symbol, the circumflex, were ‘5’, this would again cut out Aramaic, where five unit strokes are used. (Later Aramaic, Palmyrene and Nabatean have special signs for ‘5’, which cannot be connected with our circumflex.) Similar difficulties would arise if we wanted to ascribe to the two signs the values ‘100’ and ‘10’. On grounds of general likelihood it seems preferable for the time being to assume the lower value, ‘18’, because – if it gives us a date – ‘113’ cannot be a regnal year. Tentative Translation: 1 ? 2 18th (?) year (?) Bibliography: A, 155–156. The Date of the Script H: See the discussion of Sherd 13. N: The top part has disappeared (as far as can be made out from the reproduction). This does not occur

000045

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

45

in ElP but it is usual in EgD. In the latter the long downstroke is very unusual. Hence an earlier date for our ostracon is suggested. S˙: Our form does not tally with that of ElP which is pointed, narrow and straight-stroked. It resembles some of the EgD specimens which have greater width, the bottom angle rounded, and curved strokes. There is as yet no tendency to produce the left stroke downwards, a feature appearing in and after the second century. T: The right stroke is long, not very much shorter than the left one, i. e., it is of line height. This feature develops in ElP and becomes fixed. The left stroke is very long, a characteristic which has disappeared by the second century. Tabulation: H

4th cent. – – 3rd cent.

N

3rd cent. – – 4th cent.

S˙

3rd cent. – – 3rd cent.

T

5th cent. – – 3rd cent.

This points to a date very roughly about 300 B. C. E. – say 350–250.

Sherd 17 (plate III. 17) q“ p d’ a‘ s w‘ i” [ ] s r’

1 2

p q¯

d’ s [ ] y¯ s˙’ s ’ r’

The Date of the Script D: Would tally with ElP and EgD. Y: Likewise. S: In ElP the left stroke is still on top, joined to the end of the horizontal. In our specimens it has moved down, the horizontal issuing from its middle. In EgD specimens it has, generally, moved down even lower. According to this, our ostracon would belong rather to the fourth than to the third century. P tallies with ElP and EgD. Q and R would do so too. Tabulation: D Y S P

5th – 3rd cent. 5th – 3rd cent. 4th cent. 5th – 3rd cent.

000046

Birnbaum 01.04.11 18:17

46

[ 31 ]

Q R

5th – 3rd cent. 5th – 3rd cent.

Accordingly, our ostracon may have been written in the fourth to third century B. C. E. The Date of the Aramaic Sherds Our general remarks in regard to the date of the Palaeo-Hebrew Sherds (see p. 35) apply also to that of the Aramaic ones, except that their smaller number, their inferior state of preservation and the nature of the third century material are factors combining to make for less accurate results. The fourth century B. C. E. seems to me a likely time for these sherds (apart from no. 12).

Concluding Note Half the sherds are ostraca (15–17), the others were inscribed (13, 14) or stamped (12) before the baking. No. 12 appears to be the manufacturer’s stamp, no. 13 might refer to the quantity contained in the vessel, no. 14 is too short to tell. The same may be said of ostraca 16 and 17. No. 15 is a docket for an offering. A difficult question arises out of these sherds. They come from Samaria of the fourth century, when the Samaritan community was in its formative period. The Samaritans use a direct continuation of the Palaeo-Hebrew script up to this day. The sherds are, however, written in the Aramaic alphabet. Are we to infer from this fact that those inhabitants who wrote them were not Israelites-Samaritans? On the other hand, we know extremely little about conditions at this period. We must reckon with the possibility that the general expansion of the Aramaic script had embraced Samaria, too, superseding Palaeo-Hebrew there, as it had done in Judaea, with the difference that the Samaritans continued to employ the old writing for the Pentateuch. In short, the script of the sherds does not tell us who wrote them. Does the word koh[en](?) of Sherd 13 provide a possible clue? It is not, of course, specifically Hebrew or Israelite-Jewish but if the sherd were pagan we might think there would be a greater likelihood of our finding the root kmr here – as we do in Aramaic inscriptions.

000047

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

47

B. The Fragment of a Stele (plate IV)

]/ rwa [

]

s˙ r / [

There is no reason why we should not read this as a˘sˇe˛r. The word separator shows that the original contained some more text. The script being of a monumental nature points to our fragment having formed part of an inscription of some importance. It is our first piece of real inscriptional material since the Siloam Tablet (and the badly preserved remnants of two other inscriptions from Siloam).42 Translation: ‘which’/‘who’. Bibliography: F; J, 131–134. The Date of the Script Aleph: As we have on the right of the shaft only the upper cross stroke, we do not know whether the cross strokes are parallel or not – an important criterion for the age of an aleph. On the left, according to the photo, there was a lower cross stroke, but it seems first of all to run back on the upper stroke and later only to slant down, forming a very small angle with the upper stroke. See Sherd 9. R: The downstroke is very long, connecting our ostracon with Sil. The top part almost coincides with that of Sherd 1, line 3, but our specimen being written in a lapidary script, is less cursive, i. e., it is not a semi-circle. However, its date would be approximately the same as that of the R in Sherd 1, line 3. S˙: See Sherd 1, line 1. Tabulation: Aleph

after 770 before 770

R

700 770–

S˙

700

Aleph indicates that our fragment is somewhat nearer to the upper limit of the period 770–700, perhaps round about 745. Although this is a lapidary script the letters are of very moderate size: the line height is only 12–14 mm., the descender about 32 mm. That is 42

Cf. N. Avigad, {“The Epitaph of a Royal Steward from Siloam Village”, in:} Israel Exploration Journal 3 (1953), pp. 137–152.

000048

48

Birnbaum 01.04.11 18:17

[ 31 ]

not much in comparison with the line height in the preceding ostraca which is from 4 to 8 mm., or HarO with from 7 to 10 mm. The line height in Sil is about 10 mm., (descender: c. 17 mm.) in the Mesha inscription c. 18–20 mm.

000049

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

Plate I. Palaeo-Hebrew Ostraca from Samaria

49

000050

50

Birnbaum 01.04.11 18:17

[ 31 ]

Plate II. Palaeo-Hebrew Ostraca from Samaria

000051

Birnbaum 01.04.11 18:17

The Palaeo-Hebrew Ostraca Sherds

Plate III. Ostraca from Samaria in Aramaic Script

51

000052

52

Birnbaum 01.04.11 18:17

[ 31 ]

Plate IV. Fragment of Israelite Stele from Samaria

000053

32.

Birnbaum 01.04.11 18:17

The Lachish Ostraca*

The edition1 of the Lachish Ostraca is a work of great merit, showing Professor Torczyner’s wide learning and ingenuity. He himself, however, expresses the hope that the future may bring further researches and discussion towards clarifying what is obscure. The first part of our essay deals with the problem as to the connection between the documents and is partly based on the results of the second part, which consists of palaeographic and other notes.

I. The Connection between the Documents Two main lines of investigation are here followed: An examination of the external characteristics of the ostraca (the writing and the material of the sherds) and a comparison of their contents. The former allows us to make some use of the small fragments, too, while our comparison of the contents is virtually confined to a few tolerably well preserved sherds, containing a fair amount of text.

The Scribes Ostracon I Three letters in this ostracon differ from their representatives in nearly all the other sherds: Nu¯n, Yodh and to some extent Beth. Nu¯n consists of three strokes, the two top strokes forming a sharp angle. The Yodh ends in a sharp angle towards the right. Beth in most cases has a small base. These features are characteristic of an older stage of development, as can be seen from the corresponding forms of the Samaria Ostraca.2 This does not, however, imply that Ostracon I cannot be contemporaneous with our other sherds. The older forms can be explained as elements of a more elaborate style embedded in the general cursive style. These * First published in: Palestine Exploration Quarterly 71 ( Jan. and Apr. 1939), pp. 20–28 and pp. 91–110. {Some slight revisions were made by the author in 1988.} 1 {Harry Torczyner, Lankester Harding, Alkin Lewis, J. L. Starkey, Lachish. I: The Lachish Letters. London 1938.} 2 {See Chapter 30 in this volume.}

000054

Birnbaum 01.04.11 18:17

54

[ 32 ]

more elaborate forms agree, of course, with those of epigraphic style. Apart from these features the writing of Ostracon I does not differ in type from that of the other ostraca and they do not therefore suggest the existence of an elaborate style of writing, side by side with the plain cursive. Only such an assumption would make it theoretically possible to connect Ostracon I with the other sherds, because then it could have been written by one of their scribes. Ostracon II 3 It would follow from the foregoing observations that the writing of this document differs from that of Ostracon I. As to its relation to the other sherds, see nos. III , ff. Ostracon III 4 It is written by another hand than I and II . This is clear from the difference in the way the lines deviate from parallelism, in slant and in the form of every letter. Contrast, e. g., Zayin (II ,4 and III ,18), Mem (II ,3, and III ,20), Resh (II ,4, and III ,21), Ayin (II ,3, and III ,20) etc. Ostracon IV 5 The writing differs from that of I to III in the deviation of the lines from parallelism, in slant and in the form of the letters. Contrast, e. g., Aleph: in III it has long cross strokes, near each other, in IV they are short and wide apart; for Nu¯n contrast II ,1 and IV ,4; for Zayin II ,4 and IV ,12, for Waw II ,2 and IV ,2 etc. Ostracon V The hand of this sherd, too, differs from the preceding ones. Contrast, besides, the general characteristics, e. g., He¯ (II ,2, and V,9), Samekh (III ,19, IV ,6,9, and V,6) etc. Ostracon VI In comparing this sherd with the others we shall pay special attention to Ostracon II , as it comes from the same pot, and to III , as Torczyner thinks the hands of III and VI closely resemble each other.

3 4 5

{See S. A. Birnbaum: The Hebrew Scripts (= HS). 2 vols., London 1954– 1957 / Leiden 1971, plate 24.} {See HS 23. See also Chapter 34, p. 106, fig. 7.} {See HS 25, 26.}

000055

Birnbaum 01.04.11 18:17

The Lachish Ostraca

55

As to the deviation of lines from parallelism: In II there is a tendency for them to droop downwards just at first but to become horizontally parallel afterwards; on the right where they droop, the lines are narrower than on the left. In Ostracon VI , on the other hand, the lines run practically straight. The difference between thin and thick strokes seems to be more marked in II than in VI . The triangle Resh of II has a sharp angle at the left and is wide, while that of VI is much narrower across and less pointed. Mem of II ,3, has two distinct narrow top angles, the whole upper region of the letter is pointed, while in VI ,2, the top consists of an indistinct bow formation. The tops of the Waws are very different: in II the top consists of a small thick slanting stroke going from right to left, which is met half-way by a small stroke coming from the left. The first mentioned of these lines goes to the top of the shaft. In VI this line is thin and lies right across the top of the shaft, sticking out on both sides. Thus, in spite of its bad state of preservation, and the fact that there are, on the other hand, similarities between II and VI – the group lw and the form of Zayin – it appears that these two ostraca were not written by the same hand. The slant is different in III and VI , which is very clear, e. g., in the case of Lamedh. As to forms, it is enough to point out the great difference between the Zayin in III ,13, and VI ,2,10, the Mem in III ,18– 21, and VI ,2, the Qoph in III ,18, and VI ,7. In comparing IV with VI , see the shape of the Lamedh, which in IV is angular or curved, the top being much thicker than the bottem, while in VI it is rather straight and of more even thickness. Contrast also Mem (IV , II , and VI ,2), Samekh (IV ,6, and VI ,4), Qoph (IV ,9, and VI ,7) etc. Finally for V and VI it will suffice to contrast the specific form of Samekh in V,6, with that in VI ,4, so also Zayin of V,10, with VI ,2,10. As a result of these comparisons we can state that Ostracon VI was not written by one of the hands of Ostraca I to V. Ostracon VII Of this nearly effaced sherd, only a few letters are legible. The writing seems to differ from that of the rest of the ostraca, contrast, e. g., Beth (II ,4, and VI ,6, and VII ,6), Shin (II ,1, and VII ,6). Ostracon VIII The obverse of this sherd offers even less material for comparison. If those traces which are still visible were originally signs in the same writing as the word on the reverse, then there is certainly no similarity between the writing of our sherd and that of the others from the same

000056

Birnbaum 01.04.11 18:17

56

[ 32 ]

pot, because the Yodh and Nu¯n of the reverse correspond to the type (but not the hand) of Ostracon I. Ostracon IX The writing differs from that of the other ostraca, contrast, e. g., Mem (I,1,4,5, and IX ,1,2), He¯ (III ,17–21, and IV ,1–13, and IX ,1), Resh (II ,5, and V,10, and VI ,13, and IX ,8), Shin (VII ,6, and IX ,8), Yodh (VIII , obv. 2, and IX ,1). Ostracon XI To judge from the Samekh of line 5, it is written by a hand other than III , IV , V, VI . Ostracon XVI Written by a different hand from the preceding sherds; contrast, e. g., the Mem of line 1 with that sign in I,1,5, in II ,3, in III ,18–21, in IV ,2,11, in V,3, in VI ,2, in VII ,6, and in IX ,2. Ostracon XVII The writing of this sherd reminds one of that of Ostracon I. The Beth in line 1 has a distinct base, the Yodh in line 2 has a bottom hook, the Nu¯ns in lines 2 and 3 have an angular top part. It is written, however, by a different hand. Ostracon XVIII The writing differs from that of the other sherds, including those from the same pot. Contrast, e. g., Resh with that letter in I,1, in II ,6, in III ,19, in IV ,11, in V,10, in VI ,13, in VIII ,5, in IX ,8, in XIII ,2, in XVI ,4; Ayin with that letter in XIII ,1, and in XVII ,1. Result The result of the foregoing comparisons is unfortunately negative: we have not been able to establish any connection between the ostraca in regard to their senders, as they all seem to be written by different hands. If we accepted Torczyner’s theory that with the exception of XIII , the letters originate from one man, Hoshaiah, then so much at least would follow from our results: that most of them would have to date from rather different times since we can hardly imagine that the commander of a little place would have a great choice of secretaries at one and the same time.

000057

Birnbaum 01.04.11 18:17

The Lachish Ostraca

57

The Material of the Sherds The Description of the Potsherds, given by Mr. Harding,6 contains the important feature that the sherds II , VI , VII , VIII and XVIII come from the same pot. It must, however, be remarked that VII and VIII , which were near neighbours in the pot, according to the picture on page 220 {of Torczyner’s book}, differ considerably in the description, while II and VI , also near neighbours, are described in almost identical terms. As VI , XVIII and VIII are immediate neighbours, it might be deduced that the letters written on these four sherds emanated from the same place and quite probably, although not necessarily, from the same person. Of all the other sherds only IV and V are described in nearly identical terms. But as I have not seen no. IV and no indication is made of their having come from one pot, we have no reason to assume the probability of their having a common sender, and must leave it at the possibility.

Comparison of the Contents a) Formal Elements The introductory formulae of the letters supply us with some material. There are three formulae: · · · inda la xlw · · · Ó[dby] · · · huhi ymwi

in II and VI , in III , in IV , V (?), VIII and IX .

Although it is conceivable that one person should use various introductory formulae, it is not very probable. The power of tradition in such things was certainly even stronger in those times than in this highly individualistic age, with its ideal of change, and even we cling rigidly to certain formulae at the beginning and end of official letters. Thus the multiplicity of formulae in the Lachish Letters probably corresponds to a multiplicity of writers (or senders), not less than three, but there might have been more as identity of formula does not, of course, denote identity of authorship. There is also the possibility that a difference in the formula denotes difference in position or rank in the case of the addressee. This would either mean that the letters have different addressees or that the formula is to be taken to be relative, i. e., that persons of different status wrote to the same man. 6

{See above, n. 1, p. 184.}

000058

Birnbaum 01.04.11 18:17

58

[ 32 ]

b) Material Elements We shall now turn to the matter of the letters. Ostracon I This document consists of a list of names, none of which occur in any of the other sherds. It, therefore, gives no clue as to a possible connection with them. Ostracon II 7 This letter resembles III and IV in that it is addressed to one who is obviously the writer’s superior in social standing. As it does not deal with a defence of its author against some accusation, it cannot be paralleled in this respect with III or others. Ostracon III 8 It is like II , IV and VI in regard to the social relationship between recipient and writer, and like VI in a passage containing an oath and the words arq and rps. The illegibility of the traces and the doubtfulness of Torczyner’s conjecture, because it does not fill in the space satisfactorily, do not allow us to accept the name wuai as a connecting link with II and VI . Ostracon IV 9 Like II , III , V , and VI in regard to social relationship of recipient and writer. In this letter only it seems to be a subordinate writing to one in higher office. If it is of military contents, which is quite possible, then it would afford a parallel with the preceding one. There is nothing exculpatory in this letter, to act as a link with III . Ostracon V Like II , III , IV and VI in regard to the relationship of writer to recipient. There is no curse or exculpation to link it up with III . The improbability that the conjectured uhikms is right, robs V of any connection with IV . Ostracon VI Like II , III and IV in regard to social relationship of writer and recipient, and like III in a passage containing an oath and the words arq and rps. 7 8 9

{See HS 24.} {See HS 23.} {See HS 25, 26.}

000059

Birnbaum 01.04.11 18:17

The Lachish Ostraca

59

The editor’s interpretation of our document being based on improbable and uncertain premisses, there is no connection between this letter and III . Ostracon VII The conjectured Tobshillem affords no trustworthy link with I. Ostraca VIII and IX Nothing in the few signs on these fragments indicates any connection with other sherds. Ostracon XI This fragment may possibly be connected with IV , through the name Semachiah. Ostracon XII It is impossible to connect this fragment with any of the letters. Ostracon XIII It is very improbable that there is a connection between this letter and Ostracon IV through the name Semachiah. Ostracon XVI The existence of the word abn on this tiny fragment appears to be extremely doubtful, so it cannot be said to supply us with a link with III . Ostracon XVII As the fragment does not contain the word uhimri there is no connection with I. Ostracon XVIII The possibility that Õlw means Tobshillem hardly suffices to connect this fragment with I, and still less with VII .

Inferences It is not the purpose of this chapter to evolve a new theory as to connection between the documents but to provide material for further research. For the number of uncertainties and cases of doubt arising from Torczyner’s readings points not only to the necessity of new explanations and interpretations of detail, but some of them would seem

000060

60

Birnbaum 01.04.11 18:17

[ 32 ]

to shake the very basis of his theory about a one-man correspondence, exculpatory in nature and centering around the prophet Uriah. There is in our ostraca no ñtnla ñb uhirbki to suggest the Biblical rubky ñb ñtnla and the story of the prophet Uriah whose name does not occur in the letters. The reading abn is improbable in Ostracon XVI and the traces supposed to be uh are illegible. The theory about the court-martial on Hoshaiah, in connection with Uriah’s fate, is also open to objection. What are the charges brought against Hoshaiah? One accusation seems to be connected with his military activities, another with his having cursed the king. Both would be material for court-martialling Hoshaiah. The main charge, however, is that he has done something whereby the prophet Uriah has been endangered. But the danger to Uriah came from the king. Now even if we assume that Ya ush, as a faithful Jhwh-worshipper, would secretly act against the king’s command in the affair of Uriah, how could he court-martial a subordinate officer or official for doing something he knew to be in conformity with the king’s policy? Even if all the other judges had been on his side, would not Hoshaiah have simply appealed to the king for help?10 Let us now sum up. The ostraca were found in one place (except XVI and XVII ). It is therefore not improbable that they were received by one person. In the letters II to VI the formulae of address are more or less preserved and we are entitled to recognize in the recipient a man of some sort of importance. According to II and VI his name was wuai. Letters III and IV would enable us to assume that he was an official, as the text seems to indicate administrative, political or even military contents. The fact that they were probably addressed to one person gives us no hint, of course, as to the number of senders. From their contents it would be reasonable to connect III and VI as coming from the same person, but it is conceivable that two persons should swear an oath in connection with the same matter, or that one passage is only a quotation from the other. Between IV and XI the name Semachiah is possibly a link. From the formal characteristics of the contents it is possible that II and VI originated from one source, and IV , V, VIII , IX from another source. The letters II and VI have not only the same addressee and the same introductory formula, but they are part of the same pot. These therefore are two ostraca which we can with much probability assume were sent by the same man. In contents they show no connection but this is, of course, not unusual in any way. 10

See our short review of the edition, in: Bulletin of the School of Oriental Studies 9 (1937–39), pp. 807–810, especially p. 809.

000061

Birnbaum 01.04.11 18:17

The Lachish Ostraca

61

I should like in conclusion to appropriate Mr. Harding’s words: “No doubt there are many with better claims to the title xqph (the open-eyed) than I have, and we look to them to provide us with fresh food for thought”.

II . Palaeographic and Other Notes Ostracon I Line 1: Mem and Resh as well as He¯ and Waw run into each other. This cannot properly be called a ligature because the characters remain independent of each other, and their shape is not affected. This is one of the characteristics of cursive writing. One often comes across letters touching each other, e. g., Aleph and Lamedh (II ,1), Aleph, Daleth, Nu¯n (II ,1), Pe and Resh (III ,19), Lamedh, Heth, He¯ (III ,21) etc. ˙ according to Torczyner, “not Line 4: The thin stroke of the Taw is, on the level”, i. e., it is not one single stroke but written with two lifts of the pen. This seems very improbable. It appears to be simply a wavy line. Lines 4 and 5: The Resh differs from that in l. 1, the downstroke being much smaller and turning towards the left in the end. Nevertheless, the signs can hardly be read as Daleths, which have pointed tops, the left stroke being prolonged upwards after it is met by the right stroke. This reading (which is also Torczyner’s) is textually corroborated by the sign following the letter in question in line 4 which seems to be rather Mem than Nu¯n, so that uhimri appears preferable to uhindi, a name known from the Elephantine papyri, but there possibly identical with hinzai. A textual corroboration is to be found in line 5 where the reading Daleth would result in a name – uhidn – which is not known from other sources and which would be very difficult to explain etymologically. The following transliteration, like the subsequent ones, gives only such signs as can actually be seen and letters which can be palaeographically inferred from existing traces, but nothing which has its source in the contents, even where it is very probable. One line over a letter denotes probability, and two lines signify possibility, both in varying degrees. Effaced or illegible letters are indicated by dashes, corresponding to the probable number of such letters.

000062

Birnbaum 01.04.11 18:17

62

[ 32 ]

Transliteration of Ostracon I .

uhirmg

1

.

’ i uhin z a

2

” ” i” . nb . ” b g“ x uhin z a

3

‘ i“ nbuh ix ” jbm uhimr

4

uhirn . nb . uhintm

5

uhilchnb

. mlwbjnb

Ostracon II Line 5: The reading rkzi, as Torczyner points out, is impossible. The second sign cannot be Zayin as is clear on comparison with the Zayin in line 4. The sign in question is much too slanting; instead of a tail it has a big shaft; instead of the long, more or less straight top bar it has a thick angle; the top of the letter does not belong to this angle, i. e., to the left part, but is the top of the shaft, apparently severed from it by a scratch on the surface, stretching upwards, outside the sign, if what looks like a downstroke, is not ink. Line 6: The number of letters lost by having been broken off at the beginning of the line is two, as is also assumed by Torczyner. It seems possible to decide with some amount of probability whether the reading Daleth Nu¯n or Mem Resh is to be preferred, i. e., whether the mutilated word was inda or irma. If we draw a line connecting the beginning of lines 1 to 4 and prolong it downwards to lines 5 and 6 we can be rather sure that we shall find the starting points of lines 5 and 6. The space available, as an experiment of mine showed, is too small for Mem and Resh which are nearly double as wide as Daleth and Nu¯n, and still more so for rbd. Thus it seems very probably that we must read inda, and that therefore the meaning of the letter is: “I received your letter. I am sorry I do not know what you asked”. This would unfortunately dispose of the interpretation proposed by Torczyner: “I never have known the thing I am accused of telling” – which protestation of innocence would connect it with Ostracon III . The establishment, however, of the reading inda has robbed us of this connection. In consequence of this reading rkb cannot be equated with rqb and its etymology is yet to be solved. If htydi, as is probable from the context, is in the first person, its scriptio defectiva seems to connect it with spellings like tta, rmal, abl, abh, bj, mbj, abn, mlw, tymw, mrmw as opposed to inda, ik, itbtk, il, uhlyiu, itarq. That might conceivably point to a rule that plene spelling is reserved for the end of a word.

000063

Birnbaum 01.04.11 18:17

63

The Lachish Ostraca

Transliteration of Ostracon II “ w “ iwuai . indala . ym

1

lwt – – wi“ ndatahuhi

2

‘ yimmi“ ktymik . ” ty . m db

3

“ i“ ndarkz . ikblkk . ta

4

ata huhi . ” rki . hdb

5

htydi . alrwa . rbd . i

6

Ostracon III Line 2: If the conjecture inda is correct, then the space between it and ymwi allows room for five normal letters so that the conjectured name – wuai – would not fill it. Hence this ostracon, in so far as it is legible, is not connected with the preceding one through the name of its addressee. Line 4: This line is not much bigger than line 1 and would allow room for only two or three more letters. Hence, from the start, we might be justified in regarding Torczyner’s four letters with suspicion. Further, a comparison of the facsimile with the hand copy shows that the Samekh of the original stands somewhat more to the right. That would mean that the conjecture kdbyxlwtyi is too long, and in fact, the letters of the hand copy are too small and crowded. itxlw might however be possible if we accept Torczyner’s reconstruction of the contents of this line. The Lamedh of the supposed la would have stood under the Mem of tymw in line 3, and would have started very near it, in order to attain at least the shortened and cramped form as proposed in the hand copy. But why is there then no trace of its top next to the distinct tail of Mem? There is no word divider at the end of the line, also the hand copy shows nothing. Torczyner seems unconsciously to have split up the Heth into He¯ + dot. ˙ The line is too effaced to allow more than the possibility of some connection between rps and a xÅQÈPÇ. But xqph might be some other word or form. Line 5: The conjecture hktbu before Zayin is too long. The hand copy made it feasible only through unusual overcrowding. The traces after Zayin do not seem to point to Mem, it would also be a rather small character in comparison with the other signs in the line. Torczyner’s reconstruction of this line, ingenious though it is, is not backed by graphic remains. Line 6: kdbyl is possible as a conjecture, of course, but apart from the Lamedh no letter is certain. The next trace is possibly Ayin; if the

000064

64

Birnbaum 01.04.11 18:17

[ 32 ]

next remainder represents Beth, it shows only the upper part of the shaft, but neither the lower top nor the loop which are shown in the hand copy correspond with such traces as are still visible. Line 7: Of the suggested hud I can see nothing.11 Line 9: The trace at the beginning of the line is very doubtful. The downstroke of Qoph would have stretched on to the edge of the sherd. Is it possible that a very fine splinter of the sherd broke off, forming the present corners at the beginnings of lines 4 and 12? That might also account for the following lines starting at the very edge. Line 10: There is no available space for Yodh and Shin. There might have been only the latter, and the spelling would be in consonance with the spelling system of our Letters. Line 12: The trace before the final ta of this line seems too near the Aleph for a Resh. This reading is also rendered doubtful by the spelling of the proposed tÇaÇrÄ which is not in agreement with the system of our Letters, which appears to prescribe scriptio plena at the end of words (cp. Ostracon II , end of our note). The signs after ta, at the end of the line, are very doubtful. The first one after Taw does not resemble any of the Mems of this sherd, with their heavy tops; the angle between the top and the tail also differs from their angle, and finally it does not, like them, have two top strokes. The letter in question looks like the Nu¯n in line 15 and this word would accordingly be ñta. The next letter could hardly be Nu¯n, neither in its more elaborate form as in line 15, nor in the more cursive one as in line 20. It seems to be most like a Beth, except for the difference in slant. The next sign, which is read as He¯ by Torczyner, lacks the long top stroke sticking out on the right characteristic of this sign, the cross strokes are longer than those of He¯ – they are of equal length, while those of He¯ gradually decrease. In comparing it with Heth, He¯ and He¯ + Waw in our sherd we can hardly doubt that the letter in˙ question is a Heth. ˙ Line 13: The space available at the beginning of the line would seem to imply the presence of a small letter. It might have been Yodh and we could then read lix, a word which would not sound out of place in the neighbourhood of abch rw in the next line. These words together, lixb ñta, do not look impossible. The three letters before them might have been the object of the sentence, supplemented by hta, and the space after them might have contained an adverbial clause. This sentence would be connected with the following and not with the preceding passage. 11

I have not seen the original of this ostracon, nor of nos. IV and VI .

000065

Birnbaum 01.04.11 18:17

The Lachish Ostraca

65

Line 14: Do the He¯ and Sadhe run into each other, or is there a black mark between them? Line 15: It is doubtful whether the empty space at the beginning of the line ever contained a letter. The neighbouring signs are so distinct that it does not seem likely that just here a letter should have been effaced. The second letter looks like Pe, cp. line 4. It is angular and has not a long shallow bow like Beth; the loop of Beth is not there; the angle of the little top stroke is wider than is characteristic for the exceptionally sharp loops of this scribe’s Beth. The next sign is very uncertain. To judge by the right top angle as well as by the rest of the traces, this might be Yodh (cp. this letter in line 9), but certainly not Resh. The space following, up to the supposed la, could take two letters but hardly Torczyner’s conjecture which consists of five letters, apart from two possible word dividers. The letters of the hand copy are too small and crowded. Thus if the line started with a name, it could not have been uhirbki. The traces which are supposed to be the tails of Beth and Nu¯n are so faint that we cannot rely on them; the loop of the Beth is not preserved and it is doubtful whether the mark under the Daleth of line 14 represents the top of Beth. The reading nb is therefore very doubtful. The next trace does not really look like the cross stroke of Aleph, and it is difficult to understand why there should not be a trace of the downstroke. The Alephs of this sherd are wider, too. So this trace is quite uncertain. It is equally uncertain as to whether the mark after it is the trace of a Lamedh. It does not look like the neighbouring ink marks; in structure and slant it looks unlike the Lamedhs of this sherd; there seems to be no room for a letter between the supposed Aleph and the following Nu¯n – the Nu¯n would have been written into the bottom part of the Lamedh: Why this crowding? The letters following are normally spaced; why should this bottom part only be effaced in a non-effaced neighbourhood?; the Lamedh would collide with the Resh of rw. Consequently the reading ñtnla is very doubtful. If we must abandon it, the only theophorous name with El would be removed from the Lachish documents. Our line most probably does not contain the name Yikhbaryahu ben Elnathan. But does it give a name at all? The letters ipk cannot be explained as the beginning of a name, and it might be remarked – although this is not conclusive – that it seems unnecessary to give, in a letter to a high officer, the title and name of his commander-in-chief, for such would be the position of the general in question, as implied by the He¯ of abch in conformity with Biblical usage. If we have no name here,

000066

66

Birnbaum 01.04.11 18:17

[ 32 ]

then what part of the sentence which starts with the predicate dri and the subject abch rw must we expect here? Probably an adverbial clause. Could it be Ólmh ipk ‘according to the command of the king’? The word klmh of the reverse, with dividing dots, would just fill up the available space. Or should this be the end of the sentence, leaving dri without an object? In that case ñtn would be a verb, meaning ‘to permit’, cp. ñti alu abl tixwmh (Ex. 12: 23), buwl Õhillym unti al (Hos. 5: 4) etc., but where would be its object? Is there a possibility of taking ñtn as a name, belonging to the adverbial clause? Whatever the way in which these difficulties will be solved, the non-existence of Yikhbaryahu ben Elnathan seems practically certain. This would rob Torczyner’s equation of the abn with Uriah, of an important argument in its support. Line 16: The last sign of the obverse cannot be Lamedh, there is no room for it and no trace of it. We have, like Torczyner, to read a Taw, the downstroke of which is visible. The reverse of this ostracon, lines 17 to 21, is the only example of a wholly preserved text. Line 19, however, proved to be a stumbling block. Line 19: Torczyner reads the second word as uhibdn but he notes “the difference in the writing of the n as seen on the obverse, from the reverse”. Yet one need not turn to the obverse with its indistinct signs. The reverse itself contains five clear Nu¯ns: they are angular, the upper stroke at an angle of nearly 90 degrees with the tail; the tail is long and under the line. But the sign in question is a bow and does not go under the line. The next sign cannot be a Daleth either. Daleth is a distinct triangle, rather wide, has a sharp angle on the left hand and the right stroke is straight. Our sign shows nothing of these characteristics. In fact, both signs do not look like any others on these ostraca. The little stroke between them gives the clue to the puzzle: We have here not two signs but one only: A bow right, a bow left, a cross in the middle – it is a Teth. All the examples of Teth in the Lachish documents have this form, nowhere do we find a real circle. The name is consequently uhibj ‘Tobias’. Torczyner’s reading of the next word, too, is erroneous. The first sign is even less likely to be Nu¯n than in the preceding word – it is hardly curved at all but nearly a straight line. In height and slant it corresponds to Ayin, as a comparison with examples of this letter will show. They are all rather narrow and this is narrower still. The subsequent sign has neither the fairly straight shaft of Kaph nor its great length and is without its left top stroke. Should just this little stroke have been forgotten by the scribe or effaced on this perfectly preserved ostracon?

000067

Birnbaum 01.04.11 18:17

67

The Lachish Ostraca

There is no reason for not taking it at its face value, i. e., as a Beth. It has the curve characteristic of Beth and belongs to the smaller variety of the letter. The right reading is therefore dby, and Torczyner’s first opinion was justified. This is regrettable as thereby we are not only robbed of a word of rare occurrence for our dictionary, but also of an interesting personal relationship between our documents and a king of Judah. Through the non-existence in our ostraca of “Nedabiah, the ‘nepos’ of the king”, we also lose textual evidence as to their date, which would have been provided through I Chron. 3: 18. Transliteration of Ostracon III Obverse ” . ” kd ‘ b “ y “ “ l . ” – lw“ . uhiywuh

1

. ymwi – ‘ – – – – – – i ’ n ’ – – ‘ l – – h“

2

’ ’ – – – h“ u h ” “i’ mlw . tymw in’ d

3

xqph – – – r‘ – s‘ – – – – – – – u’

4

“ b “ y “ “–z–––––– rwa“ . rpsl . ” kd

5

‘ . ik . wmak’ – by ‘ l ‘ – – da “ x ‘ lw bl

6

“ lw . zam – – – – db ‘ – “ y“ . la . kx db

7

“ aiku . ” k ’ “ . ” rm . htydi” . al . ” – nd” a

8

“ “ – r“ – a . hsn ’ . ma . huhix – – ps

9

” rpsilarq ‘ l ‘ . ” w‘ mgu” . ” xcnl

10

“ ila . abirwa “ rp ‘ s ‘ l “ k 11 . ma ” a – – – ht “ a . itarq“ xb‘ n t

12

” m . l – 13 . dgh . kdbyl – – m“ u a abchrwdr i ’ . ” rmal

14

” n” – a‘ – – – ip ” k “ “ 15 . ab l ’ . nt m

’ c’ · · 16 tau“ . hm i r

Reverse uuhixanbuhiuduh

17

.hzm . txqlxlwuwna 18 abh . klmh . dbyuhibjrpsu

19

mal . abnh . tam . ydinb . mlwla

20

“ . la . kby . hxlw . rmwh . r 21 . inda

000068

Birnbaum 01.04.11 18:17

68

[ 32 ]

Ostracon IV Line 4: There is no room for the proposed ila inda. The inda of line 2 transferred to this line goes up to the Lamedh, so the pronoun would not be ila but il; the inda of line 12, transferred to this line, even goes up to the Yodh. There are three possibilities: (a) il inda, (b) a personal name of three letters followed by ila, (c) a personal name of four letters followed by il. (a) seems the most likely. Line 5: Would it follow from the suggestion at the end of our notes on Ostracon II that the constructs tib, ñia are not plene spellings but that the Yodh indicates a diphthongic pronunciation (bayt, ayin)? Line 7: The sign after kdby might be Aleph but the next is a distinct Yodh, not Daleth with the entirely different angle of its downstroke. The subsequent trace is illegible. Line 8: The first three signs are illegible but the following Lamedh with the remains of Heth make a Shin before them probable. The sign ˙ or Aleph, the next perhaps Taw. after hmw might be Waw Transliteration of Ostracon IV Obverse ’ ty i ’ n ’ – – – – – – hi . ymwi mik

1

” axlwr “ walkktyu . bjtymw ind

2

‘ yitbtkkdb “ yhwy . nk“ lkktldhl

3

“ – xlwiku . i l ” – – – – xl‘ wrwa

4

“ t“ i ” b r “ bd . ly . ind amw“ . nia . dpr

5

‘ umd ‘ “ – uhiymw‘ . hxqluhikms

6

“ u“ hlyi – iak‘ dbyuhriyh

7

’ ‘uh ’ mwx l ’ – – – –‘ ht

8

rqbhtbstb“ . maik

9

a

Reverse “ m ‘ . laikyd “ iu 10 – n . wk“ l‘ taw “ . ttah . lkkmrmw un ntn rwa

11

“ ‘ n ’ . alik . inda zytahar

12

hq

13

000069

Birnbaum 01.04.11 18:17

The Lachish Ostraca

69

Ostracon V Lines 1 to 5 are beyond hope. The few letters which are still legible or imaginable, convey only the ideas of ‘I’ and ‘you’. Although not visible on the sherd, the photo might contain, as second and third signs, the letters mw. Line 5: Harding’s readings h in the middle of the line, and uz at the end, although still very uncertain, seem to be more probable than Torczyner’s s and uhi. The latter’s conjecture uhu[ – – – irp]s with kms as the probable completion, is too long. We have drawn these words from the letters of this ostracon and although we crowded them together there was no room left for ta. If this word were sacrificed, the traces of the first sign could hardly be made to tally with the form the clear Samekh has in the next line, and the traces in the middle of the line would hardly fit in with Yodh and its immediate surroundings. Line 7: The word at the end of the line is illegible. Its first letter is Yodh. The next one might conceivably be He¯, Daleth seems to be ruled out, on comparison with that letter in the same and the preceding lines, and Resh has a heavier downstroke in lines 6 and 10. Could the subsequent traces represent a Waw? If there should have been a He¯ after that it would have been unusually small. Similarly the last trace, if of a Yodh, would be undersized. Line 8: It is not impossible that the first sign is Heth. The He¯s of this hand have a rather heavy top stroke, which shows˙ a slight tendency to being curved downwards and sticks out boldly on top, while the sign in question does not show these characteristics. The cross strokes seem to be nearer each other, and the right downstroke appears to be smaller than in He¯. There seem to be two little strokes on the right hand top which would fit in neither with He¯ nor with Heth; perhaps they are not ink marks. The stroke on the left side has no ˙real continuation over the crack, the trace on top is too far left. Could this trace represent a word divider? The little top stroke of the next trace appears to be at too wide an angle for He¯ or Daleth, but perhaps these are not ruled out. Line 9: The second sign seems to be Yodh, cp. the slant of the Yodh in line 7 and near the end of this line. The top part of the third sign seems to show the remains of the small middle stroke of Mem. The subsequent character, a He¯, stands too far off – if we compare those parts of the ostracon where the writing is preserved – to be the immediate neighbour, but too near in view of the amount of intermediate space available to contain anything except a word divider or the bottom of a Lamedh. There is, however, no trace of this long letter. Thus we

000070

70

Birnbaum 01.04.11 18:17

[ 32 ]

arrive at the reading mih, i. e., ÕuÊIhÅ. The last letter of the line, though probably Aleph, is not beyond doubt. It looks as if the cross strokes were diverging, while in line 7 they are practically parallel. Line 10: The traces of the first sign might point to Teth, and the hand copy gives something similar; the Resh of the Reconstruction is out of the question – the traces do not fit in with the form the letter has in the next word, and there is not enough space between the alleged dot and the edge of the sherd. The adjoining mark visible on the facsimile is only a scratch. Of the last four signs of the line only two can be accepted as certain, the Lamedhs. On the photo there is something dark between the first Lamedh and the next sign. If this is ink it is difficult to see how this letter could be Mem. But even if the dark spot does not belong to the letter, the sign does not have the appearance of a Mem. Nu¯n does not seem to be any the likelier. Another possibility would be Kaph if the difference in slant could be ascribed to the cramping in of letters at the edge of the ostracon. The last letter is illegible. Apart from the greeting and the mention of correspondence, nothing is intelligible in this ostracon. What really does remain of lines 7 to 10 does not allow of a theory that this is a letter of exculpation written by a man who had cursed the king. Transliteration of Ostracon V – – – – – – – – – – – – – – i’

1 212

‘ “ – im ” “–––––––– – db“ y

3

‘ “ – – – – – – – – – – blk“ by

4

u’ z ’ – – – – – – – – h‘ – a‘ – –

5

“ h . kdby ‘ ‘ . ” – wh – – rps

6

i ’ – – – – i . indala – m“

7

– – – – – – – – – h‘ . ’ x‘

8

d ‘ – h . mih ” ai‘ – – – yl ‘ ’ – l – lyrzr – i – j‘

12

9 10

The transliteration of line 2 was omitted inadvertently in the original edition.

000071

Birnbaum 01.04.11 18:17

The Lachish Ostraca

71

Ostracon VI Line 3: Experiment shows that at the end of the line there is no room for the proposed letters psta. Line 4: Is the shaft of the third sign as seen on the facsimile – supposed to be Mem – really produced by ink? It does not tally with the forms of Mem in lines 2 and 6. It is doubtful whether at the end of the line there is sufficient room for the conjecture malmrw, the hand being rather wide and the letters not crowded together. Line 5: The second He¯ in the word read by Torczyner as hnhu, does not seem to have the form presented in the hand copy. The facsimile shows the top stroke of normal length going down to the base line; the impression of a ‘vertically curved stroke’ at the left seems – judging by facsimile – to be caused by a darker part of the sherd surface, combined with the effect of the nearness of the parallel strokes. Line 6: The third sign of the word read by Torczyner as tprl is much bigger than the distinct Pe in line 4 and the legible Pe in line 14. Pe is a small or middle sized letter in the ostraca. It is therefore doubtful whether that big sign here is really Pe. It appears to be Mem. tmrl ‘to deceive, to betray, to leave in the lurch?’ The next three signs are very uncertain, but the reading proposed in the edition, idi, may be possible. The next one, however, can hardly be a Mem if we compare it with that letter in the same line and in line 2. The sharp angle of the top, on the left, would rather remind us of the Beth in the same line and in lines 3 and 5. At the end of the line there seems to be more room than would be necessary for the conjectured whl. Line 7: The seventh sign is very doubtful, the facsimile is suggestive of Beth. This would not favour Torczyner’s conjecture riyhu Ñrah, nor does the second part of the phrase tally with the traces: the supposed Yodh is not certain, and the next sign is Daleth, unmistakable by its sharp top angle, as against the blunt top region of Resh. The readings tprl, Õidi, riyhi Ñrah in lines 6 and 7 being improbable, there is nothing to remind us of a prophet nor to make us infer that line 5 ends with the word abn. Line 11: An Ayin or Qoph seems to be visible in the middle of the line. Line 12: There appears to be a sign between ix and the next word. Line 13: The third sign looks like Kaph but the hand copy gives only a shaft which would imply that the mark on the photo is not caused by a trace of ink. Line 14: The sign under the Aleph of arq in line 13 seem too high up to be He¯. It probably is Samekh.

000072

Birnbaum 01.04.11 18:17

72

[ 32 ]

Transliteration of Ostracon VI ’ “i” i’nd ” a “ la ahuhi – r – w“ u a

1

‘ a ‘ i’nd ’ a“ – – – – mlw – hzht – ht

2

– – – – – n d” axlwikbl‘ kbdb“ –

3

“ s “ – – – kl “ –h– – – – – – hirp

4

’ – a‘ n a ’ r ‘ q “ ‘– – – – hirbd“ h nh ‘ l ‘ – bjal ” m “ r – – – – b‘ i d” ” it

5 6

‘ – dijq –‘ – – d i ’ – – – – – bh

7

n

a

–‘ – al‘ h – – – – – – – – – – – – – 8 “ t ‘ – – – – – – – – – ab “ t ‘ ‘ uwy 9 ‘ w – – – – – – tazk 10 lh – – kl m

’ r ‘ ‘ – q‘ – – – – – pl ‘ a‘ 11 – ih

b hlahuhiix – – – – – – m‘ – arq – a‘ – – – – ki – – h‘ – – – rps – – – k –

12 13 14 – – – – – – – – – – – l 15

Ostracon VII Of this nearly effaced sherd only a few letters of lines 4 and 5 are partly legible. The other readings given in the edition seem to be unacceptable. wÄ BÂ as ÕLÈÏ wÇ bÂ[u´j] and so it Line 6: The group mlwb might as easily be Õu´lÏ cannot serve as evidence to connect this sherd with Ostracon XVIII . Transliteration of Ostracon VII – – – – – – – u” h“ –

5

– – – – – – – – ml‘ wb‘

6

Ostracon VIII Line 3: The first legible sign might be a Nu¯n. The second and third traces are puzzling. The second might be an Ayin although it probably is not. But it is hard to imagine what the third trace represents. Could the two traces perhaps belong together, forming the upper part of a Qoph, the lower part of which is effaced? Or is it a Teth that we have here? Beth, however, is impossible. Thus we cannot read abn, quite apart from the fact that between the alleged Beth and Aleph there would be that doubtful sign as well as an empty space (a – – bn) which, of course,

000073

Birnbaum 01.04.11 18:17

73

The Lachish Ostraca

cannot be abn, the spelling of our ostraca. This sherd does not, therefore, effect a connection with Ostracon III . Reverse, line 1: The trace of the third sign is very faint, but it certainly cannot represent Taw. Most of what appears to be the base of the sign is just a shadow due to the surface. Transliteration of Ostracon VIII Obverse ‘ i‘ ” – – – – mw ‘ ‘–– – h‘ – – – – – – – bj – m‘ – – – – – qn‘ ” – – h ’ ‘ – – – – – – – r · b‘ uh

1 2 3 4

‘ a ‘ ‘ – – – – rt

5

’ a“ bz‘ k

1

inda

2

d

d

Reverse

Ostracon IX Line 1: It looks as if the Yodh had a little hook at the bottom, the same characteristic as in I, VIII and XVII ,2. Line 2: It seems unlikely that after mlw five letters – dbyim – could have been crowded in, especially in this bold hand. The big mark under the line after mlw cannot be the tail of Mem – compare the preceding character; in fact, there is a scratch on the sherd here, the shadow of which shows on the facsimile. Line 3: Also the six letters ikblkk would probably not fit in the space available. Heth after the Lamedh is a possible reading. Torczyner seems ˙ to have counted its left downstroke doubly: first as that and then as the downstroke of a Taw. The subsequent trace is given as Mem in the hand copy, but a glance at the Mem of mlw above it in line 2 will show that this is not likely. Could it be a Kaph? Line 7: The third sign cannot be the Ayin given by Torczyner in his transliteration, nor the upper part of a Qoph which he suggests as graphically possible. The hand copy shows a complete Qoph but the shaft is not visible in the photograph and on the ostracon. Nor would we be inclined to take only the right hand part as a narrow Ayin, like the fourth sign of line 1. Even then, the next trace, touching it, is not the beginning of a Nu¯n, which would make the suggested Ónydi possible.

000074

Birnbaum 01.04.11 18:17

74

[ 32 ]

Might the whole perhaps be a partly effaced Shin? Between it and the fairly legible uhi there might be room for two signs. But the traces visible give no hint as to what they might have been and do not favour a name like uhiymw, for example. Line 8: Of this line only the beginning – Shin, Resh – is certain. Of the next sign only the trace on the line is left. What looks like the trace of a tail is something in the material of the sherd. The last two signs of the line do not look like He¯ and Mem, of which we have fairly distinct examples in line 1. The incorrectness of the readings ydi, here and in V,7, makes it impossible to parallel these two sherds. Transliteration of Ostracon IX Obverse – – a huhiymwi

1

– – – mlw – – – – –

2

‘ ‘–– – – – – kx‘ l‘ – ma

3

–––h––––––

4

‘ ‘ – b‘ – – – dk

5

– r‘ –

6

‘ i auhi’ – – wd

7

– – – – – rw

8

– x‘

9

Reverse

Ostracon XI Of this sherd nothing is certain except the Samekh of line 5. Transliteration of Ostracon XI – l’ – ” m ’ ‘– hi‘ k ” ”ik ’ ‘–s uh

3 4 5

Ostracon XII It seems that nothing except the notions ‘you’ and ‘I’ can be gathered from this sherd.

000075

Birnbaum 01.04.11 18:17

75

The Lachish Ostraca

Ostracon XIII Line 1: The second letter might rather be Nu¯n than Mem as also the hand copy seems to convey. This is probably followed by an effaced sign, before the Lamedh. The reading of the third sign after the Lamedh as Taw is not probable. The subsequent trace is beyond hope and even the Lamedh which appears to follow is by no means certain. Line 2: Of the first half only a Samekh seems probable. The continuation uhikm is highly conjectural. This deprives us of a possible connection with Ostracon IV . Line 3: The first visible sign does not at all resemble the third which is probably Aleph. It differs in slant, the supposed beginning of the cross strokes is low down, nothing of cross strokes is to be seen right of the downstroke. The trace read Pe by Torczyner differs considerably from the probable sign in line 2, it might therefore rather represent Beth. Transliteration of Ostracon XIII “ y ‘ l ‘ ’ – mq ‘ ‘ · hk‘ – l’ – tw

n “ x “ ‘ – – – – – – – s“ – – – hrp tb‘ – a“ · t –

1 2 3

Ostracon XVI Line 1: The last sign is too big to be a Taw, even if we leave out of account that part of the sign which is broken off. It might be the rest of an Aleph (or Samekh?). Line 2: The first sign might be an indistinct Daleth. Line 3: At the end of the line there appears to be the trace of half a Shin, on the photo, but I did not see it on the sherd. Line 4: The first letter is possibly Beth. Line 5: The first two signs are illegible. He¯ + Waw would hardly fit in with the traces. They are rather low and are separated – also in the hand copy – while He¯ + Waw wherever they occur in these ostraca, run into each other. It is therefore improbable that Torczyner’s reading uh is correct. Hence, the conjecture built upon these traces, supplementing them to uhirua, unfortunately falls away with it. The two signs between He¯ and Aleph are very doubtful. Both stand higher than is usual for Beth, Nu¯n, Mem, Pe, etc., and do not remind us of any letter. The first cannot be a misshapen Mem, as a comparison with the distinct Mem of line 1 would show. It does not appear to be a misshapen Nu¯n either, in view of the heaviness of the top and the

000076

Birnbaum 01.04.11 18:17

76

[ 32 ]

shortness and thinness of the tail. The subsequent sign, alleged to be Beth, disagrees in slant and curvature completely with the Beth over it in line 4, quite apart from the missing loop which would have to be regarded as having been effaced, although the strokes of the next near letters, the Aleph after it and the tail over it, are well preserved. Thus, since it is highly improbable that this word is abn, and as it is even more unlikely that the conjecture “Uriyahu” is correct, this ostracon is thereby robbed of its great role in Torczyner’s theory, where the prophet Uriah is the centre of the Lachish Letters. Reverse: On the original hardly anything can be seen; it looks as though the marks were merely accidental surface formations, as most of them do not resemble letters. It would also appear suspicious that the writing on both the obverse and reverse has the same top side, a position which is in contrast with general usage. The photo, however, seems to decide in favour of writing on the reverse, too. Line 4: The last sign seems to be too high for the left stroke of Heth. Accordingly, we would have to read He¯ + Waw. The preceding sign˙ can hardly be a Waw, as in the edition. The two preceding letters might be Beth and Resh or Daleth, and db[y] or rb[d] would be possible. One is tempted to read the whole line as uhidby but Yodh would not tally very well with the trace. Transliteration of Ostracon XVI Obverse ‘ x“ am

1

“ ’ ihd

2

– hxl‘ –

3

” rb‘ inb a––h––

4 5

m‘ – – y‘ a–l– ”u h“ – db ‘ ‘– rd

2 3 4

Reverse

Ostracon XVII Line 1: If the second sign is Beth, then it has the clearest base of all the examples of that letter occurring in our ostraca. Line 2: The first letter is doubtful and the rest not certain. The last sign seems to be Yodh. If it is, then it would be of the older type, in keeping

000077

Birnbaum 01.04.11 18:17

77

The Lachish Ostraca

with the type of the Beth in line 1. There is, however, the third letter of line 3, which is distinctly Yodh. Unless we assume that the scribe used both types together (which is nothing very unusual and often occurs in our own writing) – we should have to decide against the letter in line 2 being Yodh. Line 3: There is some similarity between the first letter and between the second in the preceding line, which might tempt us to read both as either Daleths or Reshs. In the sign in line 3, the triangle seems to be more pointed and to have a longer downstroke; hence it might rather be a Resh; on the other hand, the top stroke sticking out would seem to hint at a Daleth. The next sign like the very similar one in line 2 is apparently a Nu¯n, of the angular three stroke type. This would be in keeping with the form of Beth in line 1 (and of Yodh in line 2?). It seems impossible to reconcile the letter with a Mem, as Torczyner would have it, because there is no trace of the stroke present even in the most cursive writing of the other ostraca, representing the union of the two original ticks which formed the top part of the letter. The last letter of the line is doubtful, but He¯ is improbable: the slant of the top line is very slight, its length very small, and one of the left strokes would be missing. We are inclined to think it is a Resh. It might be the beginning of a new word, the first being ind[a]; the Aleph may have been at the end of the preceding line. This ostracon does not afford any material for connecting it with Ostracon I as the name Jeremiah does not occur in it. Transliteration of Ostracon XVII x‘ b‘ y‘

1

i’ n’ d‘ a‘

2

r’ i n” r“

3

Ostracon XVIII Line 1: On the photograph, the word rwa clearly stands out at the end of the line, but I was unable to see it on the sherd. In placing it in the same position towards my eyes as it stands on the facsimile the word became visible – the letters seemed to be the result of light and shade, especially so in the case of the Aleph. The group mlw is hardly a connecting link with Ostraca I and VII . Line 2: The Aleph of inda seems to be visible. Transliteration of Ostracon XVIII ‘ w ‘ s ‘ ‘ –‘ – – b‘ – – lw ’ ‘ – ml ’ ‘ – – – – – br ‘ yh“ .” dy rp

1

‘ ‘ – – – – i” n” da “ “ . xlw ’ – – – – yh

2

000078

Birnbaum 01.04.11 18:17

000079

33.

Birnbaum 01.04.11 18:17

The Leviticus Fragments from the Cave*

Every Bible student will feel a thrill of excitement at the sight of the first relic of the text in its original script. Four of the five fragments shown in the photographic reproduction1 belong to three passages, all of them from neighboring sections in Leviticus: Ch. 19 : 31–34 20 : 20–23 21 : 24–22 : 5

Fragment A Fragment B Fragments C and D.2

On the photo we can see the horizontal and vertical lines which the scribe ruled on the parchment – before he started writing – to mark out the lines and columns. In the Cave Scrolls this was done ‘with a semisharp instrument,’ and the same method was, no doubt, employed in our manuscript. In accordance with traditional Jewish usage the writing is just below the ruled lines, the tops of the letters touching them. The scribe presumably marked out both sides of the column, as in the Cave Scrolls, although in our Fragments only the right-hand side is in evidence, the left-hand half of the text being missing. Our first question is: How wide were the columns? To find out the answer we shall have to count how many letters go to a line. But first we shall have to decide whether we have sufficient grounds for assuming that the text of the Leviticus Fragments (LevFrgs) had the same number of letters as the corresponding passage in the MT . As a matter of fact, the very first line of the LevFrgs contains a mater lectionis which is absent from the MT : Õinuydih. What if the LevFrgs are like the * First published in: Bulletin of the American Schools of Oriental Research [BASOR] 118 (1950), pp. 20–27. 1 The present writer wishes to repeat here his thanks to P. de Vaux for having sent him a photograph. P. de Vaux is publishing it in the October number of the Revue biblique which will be in the hands of his readers long before the present lines are printed. {R. de Vaux (O. P.), “La grotte des manuscrits He´breux”, in: Revue biblique 56 (1949), pp. 686– 609.} 2 In my photo (fig. 1) I have placed C and D in what must have been their position in the Column. (It is, of course, possible, – but not likely, – that C was at the bottom of a column and D on top of the subsequent one.) In the photo sent me by P. de Vaux he has separated C and D. He has placed C and E in juxtaposition, with E to the right of the lower third of C. I do not, however, see how the preserved letters could be fitted into the text. {See also S. A. Birnbaum, The Hebrew Scripts, vol. 2, plates 28–31, and 38.}

000080

80

Birnbaum 01.04.11 18:17

[ 33 ]

Cave Scrolls in their general employment of the scriptio plena? If that should be the case then our figures might be appreciably affected.

Fig. 1. The five fragments of the Book of Leviticus in early Hebrew script

Upon comparison we find, however, that apart from the above example (Õinuydih) all the other words preserved in the LevFrgs which might have had a mater lectionis are without one, i. e., they coincide with the forms in the MT : [u] ta (A: 4), lk (B: 3,3, C: 3,9), itqx (B: 3). It therefore does not seem rash to assume that the missing portions contained, if any, only a negligible number of matres lectionis not in the MT . In Fragment A we can easily ascertain the number of letters constituting a line, because the beginnings of the lines are preserved. But for the other fragments we cannot, to begin with, expect to get more than approximate results, as we have to start from a letter in the middle of a line and count to the letter below it in the subsequent line. The resulting figures are: Fragment A – 32, 34, 37, 34, 32; Fragment B – 32, 37, 35, 31, 34, 34, 38, 37, 36; Fragment C – 34, 38, 37, 36, 40; Fragment D – 38.

000081

Birnbaum 01.04.11 18:17

The Leviticus Fragments from the Cave

81

The considerable divergence between these figures suggests that the length of the lines must have varied quite a lot, thus complicating the task of building them up round the bits that have been preserved. We might, for instance, put a medium-sized word at the end of a line instead of at the beginning of the subsequent line, or vice versa, without our being forced to see from the resulting length of the line that we had made it too long or too short, as the case might be. But there is only one correct solution to the problem of each line, and it requires a lot of experimentation in order to arrive at it. The reconstruction has to be done in the original alphabet, of course, as transliteration into the Square script would render the spacing inaccurate. When trying out various possibilities, the present writer for a time worked on the assumption that the extent to which there was this variation indicated that some of the lines contained gaps – similar to those in A: 2 and C: 1 – caused by space being left empty after each verse to mark it off from the next. But he later came to the conclusion that our material does not favour that assumption. This does not mean that none of the other verses in our Fragments were divided in that way. If the two empty spaces in the LevFrgs are not verse dividers then they are obviously section dividers, as in the MT . The gap in Fragment C actually corresponds to the MT division between Chapters 21 and 22, but with the difference that the MT has a hxutp there, while our Fragment has a hmuts. In MT 19: 31, however, there is no corresponding division to the one in Fragment A. Perhaps our Scroll contained more sections than the MT . This would be in keeping with conditions in the Isaiah Scroll – the column reproduced in BASOR 111, p. 6,3 containing a section more than the MT . It should be observed that in both cases the last word of the section is the only one which is not followed by a word divider – a little slanting stroke next to the top of the final letter – in other words, the greater divider obviates the need for a lesser one. There was presumably no word divider after the last word of a line. Let us now work out the average length of a line in the three passages A, B, C plus D, respectively. The number of letters preserved (including the width of empty spaces) is 41, 45 and 30. They cover 135 mm., 158 mm. and 94 mm. (the photo is natural size). The corresponding passages of the MT have 173, 207, 331 letters. Hence the same passages in LevFrgs must have covered 570 mm., 727 mm. and 1037 mm. The 3

{John C. Trever, “Preliminary Observations on the Jerusalem Scrolls”, in: BASOR 111 (1948), pp. 3–16.}

000082

82

Birnbaum 01.04.11 18:17

[ 33 ]

average size of a line must, therefore, have been 114 mm., 121 mm. and 115 mm. These figures confirm either that there was considerable variation in the line length4 or they mean that the columns were not all of the same width. On the other hand, there is just a possibility that even in those early times – as in later periods – some device was employed to make each line as long as the preceding one, so that the edge on the left-hand side of the column might be as straight as that on the right. Whether this was done in our Fragments we cannot tell, but as we know that no such artificial method was employed in the Isaiah Scroll we would be rather inclined to think it was not used in the LevFrgs either, in which case the average length of a line and the approximate width of a column were identical, i. e., 114–121 mm. We have no means of deciding whether all the columns of the Scroll were of the same width. In the Isaiah and the Covenant Scrolls this is not the case, in the former the figures vary between 93 and 166 mm., in the latter between 102 and 187 mm. While we can establish – to within a few millimeters – the width of the columns of which our Fragments formed part, we can hardly do more than venture to estimate their height. One of the factors governing it is the size of the script. On it depends the distance between line and line, and, ultimately, the height of the column. The medium letters are themselves about 2 mm. high but when we include the ascenders and descenders, the actual line height is about 5 mm. The distance between the bottom of the medium letters of the one line, and the top of the medium letters in the subsequent line is also about 5 mm. This conforms to the regulation for the writing of Torah scrolls, according to which there has to be a space of line height between each line and the next. In our Fragments the distance from line ruling to line ruling is about 8 mm. According to the traditional regulations, a Torah scroll column may contain between 42 and 98 lines. Taking the minimum figure, the height of our column would have been about 32 cm. In view of the smallness of the script, that would have made a rather large book. It is possible, of course, that the regulations were different then and that the requisite minimum was less than 42. Perhaps the number of lines per column did not differ much from that of the Isaiah, Covenant and Habakkuk Scrolls. They have 29, 26–27 4

The first line of Fragment A, for example, could hardly have exceeded 106 mm. ( The reason is obvious: The word lhykm was too long for the end of the line and had to go on to the next, thus leaving the line with only 32 letters.)

The Leviticus Fragments from the Cave

83

and probably 19–20 lines respectively. If we assume that the scrolls containing the LevFrgs had round about the same number of lines per column then its height would have been between 15 and 22 cm. Another estimate could be based on the ratio between the height of the columns and their width in the Isaiah and Habakkuk Scrolls:5 in the former, the height is about 1.65, and in the latter probably about 1.35 times the width. The same ratio, applied to our own document, would give us about 16–20 cm. as the height. This tallies with the result obtained in the preceding paragraph. Although we cannot of course regard this as a proof, we should like to think it supports our assumption that the dimensions in the fifth century were similar to those in the second and first centuries. Now as to the height of the Scroll itself. Following the same procedure, we shall measure the margins in the photo of the Isaiah Scroll (Biblical Archaeologist 11, p. 48).6 The upper and lower ones together equal about seven line distances.7 In our Scroll this would amount to about 54 mm.; the height of the Scroll would, therefore, have been about 21–25 cm. (Isaiah Scroll: 26 cm.). We shall now give our reconstruction of the lines (see fig. 2). {A small circle above a letter indicates an unclear reading.} The reader will have noticed another spelling which differs from that of the MT – in B: 2 we have aih instead of auh (with its Qere perpetuum aih). Our form is both of palaeographical and textual importance. It seems to indicate that all the auh-forms of the MT are secondary, i. e., the result of something which took place after the introduction of the Aramaic script, at a stage of the Square development when Waw and Yodh had become sufficiently similar to make confusion possible. Palaeographically perhaps the most important aspect of this find is that it is a most welcome addition to our scanty stock of handwritten documents in Palaeo-Hebrew8 script. It is also noteworthy in that the LevFrgs are the first example of Palaeo-Hebrew handwriting on parchment, all our previous specimens being on ostraca.

5 6 7

8

The columns of the Covenant Scroll are not of uniform width. {John C. Trever, “The Discovery of the Scrolls”, in: BA (1948), pp. 46– 68.} The corresponding margins of the Covenant Scrolls are in a bad state of preservation; the Habakkuk Scroll has lost its lower margin, the upper equals four line distances. ‘Phoenician’, ‘Hebrew-Phoenician’, ‘Phoenician-Hebrew’, etc.

000084

Birnbaum 01.04.11 18:17

84

[ 33 ]

A [huhi'ina'Õhb'hamjl'uwqbt'la'] Õiënuëydiëhë'lëaë [u] [tariu'ñqz'inp'trdhu'Õuqt'hbiw'in]pëm

1

Õkihla

2

[unut'al'Õkcrab'rg'Óta'rugi'iku'h]uëhi'ina'Óihlam

3

[ul'tbhau'Õkta'rgh'rgh'Õkl'hihi']Õëkm'xrzak ['u] ta

4

[huhi'ina'Õircm'Ñrab'Õtiih'Õi ]rë [g'ik'Óumk]

5

B [utmi'Õir]ië [ry']uëawië ['Õajx'hlg'udd'tury'utdd'ta] [uixa'] turë [y]'aih'hdn'uix [a'twa'ta'xqi'rwa'wiau]

1 2

[ijpw]më'lk'tau'iëtqëxë'lk'ta'Õë [trmwu'uihi'Õiriry'hlg]

3

[aibm]'ina'rëwëaë'Ñrah'Õktaë ['aiqt'alu'Õta'Õtiwyu]

4

[rwa'iugh'tqxb']uëklt ['alu

hb'tbwl'hmw'Õkta]

5

[rmau'Õb'Ñqau'uwy'h] la ['lk'ta'ik'Õkinpm'xlwm'ina]

6

C–D [rmal'hwm'la'huhi'rbdiu]

larwi'[inb'lk'lau'uinb]

1

[alu'larwi'inb'iwdqm'urzniu]'uinb'lë [au'ñrha'la'rbd]

2

[rma'huhi'ina'il'Õiwdqm'Õh]'rwa'[iwdq'Õw'ta'ullxi ]

3

[Õiwdqh'la'Õkyrz'lkm'brqi ]'rwa'wia'lëkë [']Õëkë [itrdl'Õhla]

4

[htrknu'uily'itamju'huhil'larw]ië'inb'[iwidqi'rwa]

5

[ñrha'yrzm'wia'wia

huhi'ina'i] n [plm'aihh'wpnh]

6

[ygnhu'rhji'rwa'dy'lkai'al'Õiwdqb'bz'ua'yurc'auhu]

7

[wia'ua'yrz'tbkw'unmm'act'rwa'wia'ua']wëpn'aë [mj'lkb]

8

[ul'amji'rwa'Õdab'ua'ul'amji'rwa'Ñrw'] lkb ['ygi'rwa]

9

E ]Õ[

1 ] rw [

2

]n[

3

Fig. 2. Reconstruction of the text of the Leviticus fragments

000085

Birnbaum 01.04.11 18:17

The Leviticus Fragments from the Cave

85

Fig. 3. The Script of the Leviticus fragments

Let us now examine the script (fig. 3). For comparison we shall take documents from between the eighth and first pre-Christian centuries. This is a wider span of time than is really needed, but material in the Palaeo-Hebrew script being scarce, it is perhaps advisable to err on the side of caution. SamO Lmlk Lach JerSt

Samaria Ostraca9 Lmlk Stamps11 Lachish Ostraca12 ‘Jerusalem’ Stamp14

EF HabT Has

El Fragment10 Tetragrammaton in the Habakkuk Scroll13 Hasmonaean coins of the first cent. B. C. E.

Aleph: The cross strokes are parallel, and the lower one does not cross over to the left-hand side of the main stroke. These two features show that our form is younger than that of SamO or of Lach. In SamO, the cross strokes meet on the left-hand side to form a very acute angle; in Lach, this angle is so acute that the strokes are well nigh parallel; in EF , the upper cross stroke reaches the line bottom, which it does not do in either the Ostraca or LevFrgs. The letter on Has does not show this feature but its main stroke is without a top. – In LevFrgs the main stroke is somewhat curved; in all the other documents utilized for comparison it is straight. This feature, therefore, gives no clue to the chronology of LevFrgs and is presumably to be explained by the greater ornamentality of a formal book hand. Beth: There is a base, as in SamO and Lmlk whilst Lach has just a curved downstroke. Does the base in LevFrgs simply indicate that this 9 10 11 12 13 14

First half of eighth century B. C. E. First century B. C. E. First half of seventh century B. C. E. Early sixth century B. C. E. First century B. C. E. Fourth century B. C. E.

000086

86

Birnbaum 01.04.11 18:17

[ 33 ]

document is older than Lach? Or does it point to a difference in style (from which no chronological inference can be drawn)? Lach is written in a cursive, LevFrgs in a formal hand, and a formal hand in general maintains a more archaic character than its contemporaneous or even earlier cursive counterpart. The fact that the Samaritan script has a base favours the second explanation. Daleth tallies with the one in Lach. Heˆ is somewhat less wide than in SamO and Lach. As to the curving of the main stroke, see above, Aleph, second paragraph. Waw: In SamO we still have the old form which has been preserved in Roman Y. The form in Lmlk shows that the top part is not written as a tick but as two separate strokes, each starting from above. In Lach the left top stroke has become a tick – presumably via a curve. In LevFrgs this tick seems to have been dissolved: the top part of the letter now consists of two decidedly parallel lines, connected by an inner stroke, which runs from near the top of the left-hand stroke to the top of the right-hand stroke. The latter lies flat over the downstroke. These parts are quite differently proportioned in HabT. What in LevFrgs is the main stroke is in HabT just an appendage, in the way familiar to us from the Samaritan script. Thus LevFrgs is older than HabT. That the main stroke appears in Has cannot undo the fact that a more developed form has already come on the scene. As to the curving of the main stroke, see above, Aleph, second paragraph. Heth: As to the particularly ornamental shape of the two down˙ strokes, see above, Aleph, second paragraph. Yodh: The end of the main stroke bends to the left, at the same time merging into what was previously the bottom stroke, and slanting up to the line ceiling, generally, as a wavy line. From our material it is not possible to say whether there is a connection between this form and that of Lach, where the lower left-hand stroke actually crosses the main stroke. But it resembles the form of JerSt, where however the end of the main stroke has been completely fused with the bottom stroke into one straight line, which thus forms a real – although oblique – base. In HabT the downstroke is so short that the cross stroke and the base are nearly in a line. In Has the main stroke bends over to the left, and in the Samaritan script this has gone so far that a practically horizontal position has been reached. Kaph: The top part is an angle or curve, evidently a simplification of the two strokes in SamO and Lach. – The downstroke is a distinctly wavy line, unlike the straight or curved forms of SamO and Lach. This might signify that our document is written in a formal style. While in

000087

Birnbaum 01.04.11 18:17

The Leviticus Fragments from the Cave

87

SamO and Lach the end of the downstroke swings with a curve to the left, in LevFrgs this line is broken up into an angle so that a base is created, and this is heavy, straight and horizontal, corresponding to Has, the Samaritan Script, and some forms in Lmlk. We might feel tempted to explain the angular form on the stamps as being conditioned by the material, but then we have some nicely drawn curves on some other stamps. Thus it seems more likely that the explanation for this form is to be found in the formal style. While such a base is obviously a late development its presence does not signify that LevFrgs is younger than all the documents which have no such base. Lamedh: The horizontal is almost as long as the downstroke. JerSt has the same form, whereas the bottom stroke in the Ostraca is short – representing an older stage – and Lmlk has even yet a round form. Mem: The top part is in a horizontal position. In the Ostraca it slants rightwards to the downstroke. Thus the two parts form an acute angle in LevFrgs but a right angle in the Ostraca. The type in LevFrgs is the same as that in JerSt and Has, and is therefore the younger form – which is confirmed by the fact that the documents preceding SamO all have the same type, or even a more different one. Lach does not however constitute the terminus a quo for the form of LevFrgs, for it has already been employed in Lmlk, although the other form, too, can be found on these stamps. All we can say is that, to judge by the Mem, LevFrgs is not earlier than the seventh century B. C. E. There is however this complication: The horizontal position may originally have developed in the lapidary style or in the formal book hand, while the cursive still adhered to the oblique position. As to downstroke and base, see above, under Kaph, second paragraph. Nuˆn: The top part corresponds to the archaic forms in SamO and Has; these are very rare in Lach. That might indicate that LevFrgs is in a formal book hand, this being by nature more archaic than even a somewhat earlier stage of cursive. Ayin tallies with the one in Lach. Pe: The top stroke is an angle, rounded off at the tip. This corresponds to the curve of SamO, not to the straight stroke of Lach. Sadhe: The right-hand strokes are arranged side by side as in Lach, not on top of each other as in SamO. The left-hand stroke is short, evidently a reduced form of the corresponding long stroke of the Ostraca. Resh: The left-hand part rather resembles the flatter and wider form of Lach than the triangular, narrower shape of SamO. Likewise, the shortness of the downstroke places it nearer to the later Ostraca.

000088

88

Birnbaum 01.04.11 18:17

[ 33 ]

Shin: Whilst SamO, Lmlk, Lach and JerSt have the W-form, in LevFrgs the outer strokes are parallel, as in the Samaritan script, indicating that our document is later than the Ostraca, although not necessarily younger than the Stamps, because lapidary style is more archaic. Taw: As to the curving of the main stroke, see above, Aleph, second paragraph. Our examination has shown us that LevFrgs is clearly younger than SamO, Lmlk and Lach, and clearly older than JerSt, HabT and Has. Hence it must have been written between the beginning of the sixth pre-Christian century and some date within the fourth century B. C. E. If we assume the middle of this period, i. e., about the fifth century, we may reasonably expect to be not too far out. The evidence furnished by JerSt is very restricted. We shall, therefore, widen our range of comparison by taking the subsequent document, HabT, as a terminus ad quem. It is evident that the forms of LevFrgs are much further from HabT than they are from Lach. Consequently the position occupied by LevFrgs on the time graph between these two limits – a distance of over five centuries – must be considerably nearer the upper limit. This again brings us to about the fifth century B. C. E. To arrive at a slightly preciser date it might be of help to look at the palaeographical position of Lach between SamO and LevFrgs: we might say that Lach stands just about midway between the two others. If we assumed that this relationship would be roughly reflected also in their chronology, a date in the second half of the fifth century B. C. E. would appear to be likely. That is the period when the great change in the script of the Torah took place. Thus the scroll of which the Leviticus Fragments formed a part belonged to the final period when Palaeo-Hebrew was in use. To sum up: The Leviticus Fragments were part of a scroll written in a beautiful professional hand, about the second half of the fifth century B. C. E. They constitute our oldest specimen of the Bible text, being more ancient than any Targum that we have, and antedating the Septuagint by nearly two centuries. The words preserved in these Fragments are identical with those in the MT , and, from their position on the parchment, there is every indication that this identity extended also to the missing portions. The average line was 118 mm. long and contained 35 letters. The words were divided by little strokes. The final word of a section was followed by an empty space, the width of about four letters. The height of the columns may perhaps be estimated at 16–21 cm., that of the parchment at 21–26 cm. The lines and columns were marked out by ruling.

B. The Hebrew Script

000090

Birnbaum 01.04.11 18:17

000091

34.

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts*

Background I wonder how far it is generally known that the Hebrew script we use today is, in spite of its name, not that which the ancient Hebrews themselves employed. We all realise that things such as language and script change, so that the forms of our letters cannot be the same as theirs were thousands of years ago. However, in the present instance we are not speaking of a gradual development. There was a complete break, a change of the script. If we were to use the picture of a family tree, we would have to call our present Hebrew script not the daughter but the niece of the alphabet used by the Hebrews. I hope in this lecture to indicate – although in the roughest outline only – the development followed by these scripts. They cover a span of almost three thousand years, and so we are reduced to looking at a very small selection of specimens.

Palaeo-Hebrew Scripts Here (1) is the oldest document we have of the script of the Hebrews – the Palaeo-Hebrew script. This little slab is called the Gezer Tablet or Gezer Calendar, from the Biblical site where it was found. It dates from ca. 875 B. C. E. Although it was incised into the stone by an unskilled hand, the structure of the letters is perfectly clear. Let us have a look at them in my alphabetical chart (2). The first and second lines here represent the forms used in the Gezer Tablet (2, i), but, as some letters don’t occur there, I give in the third and fourth lines the writing of the famous Moabite Stone (2, ii) which is of slightly later date (ca. 830 B. C. E.), and has only one letter missing; this inscription is the work of an excellent mason. Slide (3) shows the Moabite stone itself. As you know, the Roman alphabet is derived from this North Semitic one. Some letters are unchanged. Waw, for instance, is Roman Y, ayin is * {This is the text, previously unpublished, of a lecture with slides given by Dr. S. A. Birnbaum in 1966, during a lecture tour of North American universities. The illustrations here are those shown at that lecture. We have slightly edited the text to fit the format of this volume, and have also appended a new section giving more detailed information about the illustrations. – D. J. B., E. B.}

000092

Birnbaum 01.04.11 18:17

92

[ 34 ]

Roman O, and qoph is Roman Q. The great majority have only to be swivelled round a hundred and eighty degrees, so that they face in the opposite direction, and then they are the familiar Roman letters: heˆ is Roman E, kaph is Roman K, nuˆn is Roman N, pe is Roman P, and so on. Let us go back to the Hebrews. The development of the Palaeo-Hebrew script can be followed for about four centuries. Here we have three ostraca from Samaria, from the eighth century B. C. E., (4), (5) and (6). Next is a pen-written specimen from the very time when the Kingdom of Judah was coming to an end (597–588 B. C. E.). It is one of the ostraca excavated on the site of ancient Lachish (7). The script survived the national catastrophe, and went on being employed by the Jews until about the middle of the fifth pre-Christian century. Then came the break I mentioned before. But among the Samaritans, the Palaeo-Hebrew script continued to be used – and still is used by them up to this very day. During these twenty five centuries the letter forms have changed considerably, of course. Already by the early Middle Ages they had become highly stylised, many so far removed from the originals that, without a knowledge of the development, one would not recognise them. Here (8) we have a Samaritan specimen, from the fourteenth century (1339/40 C. E.). As a matter of fact, we do find some Palaeo-Hebrew used amongst the Jews also after the fifth century before the Christian Era, but only in material of two special kinds – both of an official nature – while the people themselves now employed the Aramaic script which later developed into the Jewish, so-called Hebrew script. From the fourth century B. C. E. we have a few jar handle stamps and coins in Palaeo-Hebrew letters, and from the Hasmonaean and Bar Kokhba periods – that is, the first century B. C. E. and the second century of the Christian Era – we have a great number of coins inscribed with Palaeo-Hebrew legends.

Hebrew Scripts Palestinian Already before the fall of Judah, and even more so in the time following it, the Aramaic language and script became increasingly important amongst the peoples of western Asia. In the Persian period it became, also, the official vehicle of administration. Like other groups, the Jews also adopted Aramaic. By the third to second century B. C. E. the script had split into a number of branches, not very different from each other but nonetheless distinctive. One of these was developed by the Jews. Here (9) is a little

000093

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

93

chart showing the development of the letter mem from the end of the fourth century B. C. E. to about the beginning of the Christian Era. The next chart (10) shows the development of heˆ and samekh from the sixth century B. C. E. to the ninth Christian century. Here (11) we have some fragments from about 228 B. C. E., from the book of Samuel. They were found in a cave at Qumran.1 By the early second century B. C. E. the forms have already become rather close to what is generally called the Square script, as we can see from this slide (12), which contains a column from the famous complete Isaiah scroll found in a Qumran cave and dating from the middle of the second pre-Christian century. It is a very regular and beautiful book-hand. About two and a half centuries later, from the first third of the second century C. E., we have this book-hand (13). Structurally, the forms hardly differ from modern ones but in style they differ very much. The forms we are used to seeing were reached only some centuries later. Here (14) is a specimen from the tenth Christian century (930). Until that time there was no difference – or hardly any – between the thickness of horizontal and vertical strokes. Now the horizontals have become appreciably thicker, and almost dominate the picture. In the next few centuries the difference went on increasing until the top bar and bottom bar each took up a third of the line height. Up to now we have been following a straight line into the Middle Ages. However, our specimens have illustrated only one sector out of a considerable number, that of the script of Palestine. But already centuries before the fall of the Second Temple, Jewish settlements existed in various countries, and after the fall the Diaspora grew extensively. Aramaic: Egypt Of these early seats of Jewish settlement outside Palestine we shall first take Egypt, which furnishes our earliest material. Here (15) is a papyrus in the Aramaic script, from the Persian period. It dates from the beginning of the fifth century B. C. E., having been written in the twenty seventh year of King Darius (495 B. C. E.). Although at first glance there does not seem to be much similarity with the Square script, some of the letters are rather close to the eventual form, for instance: aleph, waw, yodh, shin / sin. By two centuries later, in Hellenistic times, the development had advanced somewhat further: this sherd (16) was written a generation after the time of Alexander the Great (ca. 300 B. C. E.). 1

Refer also to drawing 80* in The Hebrew Scripts.

000094

Birnbaum 01.04.11 18:17

94

[ 34 ]

Next (17) we have the fully developed writing, like that we saw in our tenth century example from Palestine. The manuscript to which this belongs was recently found in America, in the University of Michigan in Ann Arbor by my son Professor Eleazar Birnbaum, now of the University of Toronto. It dates palaeographically to the tenth / eleventh century C. E.2 Babylonian The other early Jewish settlement outside Palestine was in Mesopotamia, but we have no examples antedating the fifth or sixth Christian century. This example (18) comes from about the tenth century (916 C. E.). You will notice that the vowel signs are not those we use, and also that their position is different. They are all above the letters, in contrast to the mixed arrangement of the system that we now use, the Tiberian system. Not long after the time of our specimen, the Babylonian system was finally ousted even in its own home by the Tiberian one, but survived among the Yemenite Jews who had adopted it, until almost the present day. From the beginning the script was used not only for books but of course also for ephemeral purposes, such as for the writing of letters, notes, legal documents, etc. Here a development of its own took place: for the sake of greater convenience and speed, strokes were combined in a new way, so that finally special forms, which we call cursive writing, came into being. Thus there occurred a functional division in the script: the Square style was used for books, that is, Biblical manuscripts, and the cursive style was used elsewhere. The two diverged more and more, as we shall soon see from an example of the cursive which I shall show you. Sometimes cursive was used for non-Biblical literary texts, but in such cases it was written carefully and regularly. This finally led in its turn to the rise of a special book-hand, which we call Mashait. There was also a fourth functional group – the special Square script – employed in synagogue scrolls. Let us go back to our review of the regional groups and continue our travels in the Orient.

2

See: Eleazar Birnbaum: “The Michigan Codex”, in: Vetus Testamentum 17 (1967), pp. 373– 415.

000095

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

95

Parsic Leaving Babylonia and turning eastwards we come to Persia. This example of Square script (19) dates from 1571 C. E. Next (20) is a Mashait manuscript from the beginning of the seventeenth century (1601). Temanic We now change direction, crossing the Arabian Peninsula southwestwards to its opposite corner, the Yemen. This (21) is a specimen of the Yemenite or Temanic Mashait of the late fifteenth century C. E. (1490); it contains the Mishna with the Commentary of Maimonides. The general impression of Square Style given by this writing is deceptive. That becomes clear if one looks a little more closely at the forms, and even more striking if we compare them with the Square style of the same time and of the same group. Maaravic From the Yemen we move northwestwards, across Africa’s vast spaces – empty of Hebrew writing – until we arrive at the Maghreb, the southwestern coast of the Mediterranean. This example of Square script is from Qayrawan (Kairouan), and dates from 978 C. E. (22). The next slide (23) shows the cursive style of the Maghrebinian or Maaravic Type of the fifteenth century. It was written in 1479 in Marrakesh in Morocco. This is the most highly stylised of the Hebrew scripts. It is also the most difficult to read, with its overwhelming parallelism. It has its counterpart in the Mashait of the same group, as you can see in this specimen (24) from 1401. Sephardic From the Maghreb we go northwards into Europe, to the Iberian Peninsula and Provence. Everybody is familiar with the Type of this region. Most people regard it as the standard Hebrew script, not knowing that it is only one amongst many Types – and by no means the most important in Hebrew palaeography. It is in fact the Sephardic Type. Here (25) we have a Square manuscript written at Burgos in Spain in 1207 C. E. It is the same as the type face that we have in our printed books and the press, and this of course is the reason why people regard it as the

000096

Birnbaum 01.04.11 18:17

96

[ 34 ]

standard Hebrew script. But the fact that the printing press used only the Sephardic type face was just an historical accident. Originally the type face differed according to the group to which the printer belonged; it was either Ashkenazic, Italian or Sephardic. The same holds good for the Mashait style of this Type. Here (26) we have a halakhic work, copied in Mashait script in 1471 at Catalayud in Spain. This writing nowadays goes under the name of Rashi script, which is highly misleading. The designation is, of course, derived from the fact that Rashi’s Commentaries on the Bible and the Talmud are printed in it. However, he did not use this Sephardic script, but wrote in the script of Northern France. Italkian From Spain we turn eastwards, to Italy. This example of Mashait script is from 1383 C. E. (27). This Cursive hand (28) was written in the eighteenth century (1734). It is not easy to read. Yevanic Next, we go to the Byzantine lands, which include Asia Minor (Anatolia) and Greece. This slide (29) in Square script is from Anatolia ( Turkey), and dates from about the first quarter of the second century C. E. The next one, in Mashait, dates from 1451 C. E. (30). Zarphatic Our journey now takes us northwestwards into Northern France. Here we have a square manuscript of the French or Zarphatic Type, from 1285 C. E. (31). And here (32) a Mashait manuscript of the beginning of the fourteenth century (1301). This is a beautiful script but the letters are written so close to each other that it is not easily legible. Ashkenazic From France we turn east again and enter Germany. The Ashkenazic Type was originally identical with the Zarphatic one, because the Ashkenazim are mainly descendants of the Zarphatim who settled in the Rhineland and spread eastwards from there. These manuscripts were written in Square script 1236 C. E. (33), and 1432 (34). This (35) is a Cursive script, from 1532. As for Mashait, this manuscript (36) dates

000097

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

97

from 1264. It is aesthetically pleasing. Although less calligraphic than the Zarphatic counterpart that we saw, it is easier to read. As a matter of fact, the next step in the development of this script led to a Mashait style, shown in slide (37) (1383), that was very similar to the Zarphatic Mashait. The final slide (38) shows alphabets drawn from four Ashkenazic Cursive manuscripts: from 1382/83 C. E. (38, i); 1396/97 (38, ii); 1484 (38, iii); and 1574 (38, iv). We have now come to the end of our very brief review of the Hebrew scripts. A really representative picture could not be given even in a whole course of lectures. It would need a series of books, comprising, I estimate, hardly less than two thousand five hundred specimens.3 Perhaps this figure will give you some indication of how great the task is that still awaits the palaeographer of the Hebrew scripts.

3

{Note: The Author’s The Hebrew Scripts contains 547 facsimiles and alphabetic charts. – D. J. B.}

000098

Birnbaum 01.04.11 18:17

98

[ 34 ]

Details of Illustrations to Chapters 34 and 35 Compiled by D. J. B., E. B. Key: a: Reference number in The Hebrew Scripts (= HS) by S. A. Birnbaum, where applicable. b: Date. c: Group/Type. d: Where document / artifact found. e: Description and / or content of document / artifact. f: Present location of document / artifact. Slide number: (1)

a: HS 2. b: ca. 875 B. C. E. c: Palaeo-Hebrew. d: Tel-el-Gezer. e: Gezer Tablet / Calendar of agricultural seasons. f: Istanbul: Arkeoloji Müzesi, No. 2089.

(2, i)

a: HS 1. b: ca. 875 B. C. E. c: Palaeo-Hebrew. drawn by Author based on Gezer Tablet.

(2, ii)

a: HS 022. b: ca. 830 B. C. E. c: Moabite. e: Alphabetic chart drawn by Author based on Mesha Inscription (“Moabite Stone”).

(3)

a: HS 013. b: ca. 830 B. C. E. c: Moabite. d: Dibon. e: The Mesha Inscription (“Moabite Stone”). f: Istanbul: Arkeoloji Müzesi.

(4)

a: HS 5. b: ca. 780–773 B. C. E. c: Palaeo-Hebrew. d: Samaria (Sebastiye). e: Ostracon: Inventory/record for “bottle of old wine”. f: Istanbul: Arkeoloji Müzesi, No. 4608 (Ostracon 14).

(5)

b: 8th century B. C. E. c: Palaeo-Hebrew. d: Hazor, Area A, 1956, ˙ Stratum VI (Hebrew University Expedition). e: Potsherd: personal name yrb m bn lm . . . [See also HS 11B].

(6)

a: HS 13. b: ca. 740 B. C. E. c: Palaeo-Hebrew. d: Samaria (Sebastiye). e: Ostracon: business document. f: Jerusalem: Rockefeller Museum of Antiquities (formerly Palestine Archaeological Museum), No. 33 3119.

(7)

a: HS 23. b: ca. 597–588 B. C. E. c: Palaeo-Hebrew. d: Lachish = Tel elDuweir. e: Ostracon: letter. f: Jerusalem: Rockefeller Museum of Antiquities (formerly Palestine Archaeological Museum), No. 38 127, verso.

(8)

b: 1339/40 C. E. c: Samaritan. e: Deut. 14: 15–29. f: London: British Library, Or. 6461.

(9)

b: End of 4th century B. C. E. to beginning of 1st century C. E. e: Drawing by Author, based on six examples, of development of letter mem (in: Solomon A. Birnbaum, The Qumraˆn (Dead Sea) Scrolls and Palaeography. New Haven 1952 (= BASOR Supplementary Studies, Nos. 13– 14), p. 22.

(10)

a: HS 92*. b: 6th century B. C. E. to 9th century C. E. e: Drawing by Author showing development of Square Script from Aramaic. f: Various locations (noted in a above).

(11)

b: ca. 228 B. C. E. c: Palestinian Square. d: Qumran, Cave IV . I. Sam. 23: 9 ff. 4Q Samb. [See also HS 80*].

e: Alphabetic chart

e:

000099

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

99

(12)

a: HS 81. b: ca. 150 B. C. E. c: Palestinian Square. d: Qumran, Cave I. e: Isaiah Scroll A. Is. 38. f: Jerusalem: Hebrew University, Shrine of the Book, 1Q Isa.

(13)

a: HS 89. b: 1st third of 2nd century C. E. c: Palestinian Square. d: Murabba at Cave. e: Ex. 6: 7–9. f: Jerusalem: Rockefeller Museum of Antiquities (formerly Palestine Archaeological Museum), Mur. 1, part of Fragment 5.

(14)

b: ca. 930 C. E. c: Palestinian Square. e: Deut. 32: 24 ff. f: St. Petersburg: Publichnaya Biblioteka, II Firkovich 17, Tshufut Kaleh No. 36.

(15)

a: HS 122*. b: 495 B. C. E. c: Aramaic Cursive: Egypt. d: Elephantine, Egypt. e: Papyrus: legal document. f: Berlin: Ägyptisches Museum, P. 13 489.

(16)

b: ca. 300 B. C. E. c: Aramaic Cursive: Egypt. d: Egypt. e: Potsherd. f: Berlin: Ägyptisches Museum, P. 10 678.

(17)

b: 10th – 11th century C. E. c: Egyptian Square. d: Egypt. e: Michigan Codex f. 39a. Gen. 49: 3–23. f: Ann Arbor: University of Michigan Library, MS 88.

(18)

a: HS 189. b: 916 C. E. c: Babylonian Square. e: Ez. 35: 1–14. f: St. Petersburg: Publichnaya Biblioteka, Firkovich B3 (Prophets).

(19)

a: HS 205. b: 1571 C. E. c: Parsic Square. d: Isfahan, Iran. e: Gen. 47: 29– 48: 5. f: New York: Jewish Theological Seminary, MS (ENA 313; tyqwn s˙wprym).

(20)

b: 1601 C. E. c: Parsic Mashait. d: Lar, Iran. e: I. Sam. 1: 1 ff. Translation into Parsic. f: Paris: Bibliothe`que Nationale, MS He´b. 91, f. 1v.

(21)

b: 1490. c: Temanic Mashait. e: Mishna: Bekhoroth (bkrwt ) Chap. 9. Commentary by Maimonides, in Arvic. f: London: British Library, Or. 2224, f. 98r.

(22)

b: 978. c: Maaravic Square. d: Qayrawan (Kairouan), Tunisia. e: Legal document. f: Cambridge: University Library, T.-S. 12 468.

(23)

b: 1479. c: Maaravic Cursive. d: Marrakesh, Morocco. e: Arabic: Philosophical treatise. f: Paris: Bibliothe`que Nationale, MS He´b. 758, f. 112.

(24)

b: 1401. c: Maaravic Mashait. e: Liturgical poetry from prayer book (Hebrew). f: Paris: Bibliothe`que Nationale, MS He´b. 657, f. 81.

(25)

b: 1207. c: Sephardic Square. d: Burgos, Spain. e: Judges 5: 11–23 (Song of Deborah) f: Paris: Bibliothe`que Nationale, MS He´b. 82, f. 34.

(26)

b: 1471. c: Sephardic Mashait. d: Catalayud, Spain. e: Halakhic work (Responsa). f: Cambridge: University Library, Add. 499.

(27)

b: 1383. c: Italkian Mashait. d: “Har Albodo” = Ostra, Province of Ancona, Italy. e: Italkian translation of prayer book. f: London: British Library, Or. 2443.

000100

100

Birnbaum 01.04.11 18:17

[ 34 ]

(28)

b: 1734. c: Italkian Cursive. e: Moshe Chaim Luzzatto (autograph). f: London: British Library, Add. 27 085.

(29)

a: HS 285. b: ca. 1st quarter of 2nd century C. E. c: Yevanic Square. d: Akmoneia (Ahat), Anatolia, Turkey. e: Epitaph. f: Afyon Karahisar Müzesi, Turkey.

(30)

b: 1451. c: Yevanic Mashait. e: Torah commentary: from Minhath ˙ Ye˘hudha¯, supercommentary by Judah b. Eliezer on Rashi’s Torah commentary on Deut. 21: 16 ff. f: London: British Library, Harl. Or. 269, f. 129v.

(31)

b: 1285. c: Zarphatic Square. e: Deut. 34 (end), with Targum Onkelos. f: Breslau ( Wroclaw): Jüdisch-theologisches Seminar, Hs. 9.

(32)

b: 1301. c: Zarphatic Mashait. e: Hebrew commentary on liturgical poetry for Rosh Hashanah (New Year) service. f: Vienna: Israelitische Kultusgemeinde, Bibliothek, Hs II ,32.

(33)

b: 1236. c: Ashkenazic Square. e: Pentateuch and Targum Onkelos: Ex. 36: 11–24. f: Milan: Biblioteca Ambrosiana, B 30, f. 96r.

(34)

b: 1432. c: Ashkenazic Square. e: Mishna Aboth, Chap. 4 (end), Chap. 5 (beginning). f: Karlsruhe: Landesbibliothek, Hs. Reuchlin 4, f. 293r.

(35)

b: 1532. c: Ashkenazic Cursive. d: copied in Prague. e: Yiddish translation of Ps. 6: 7. f: Hamburg: Staats- u. Universitätsbibliothek, Hs. Hebr. 181.

(36)

b: 1264. c: Ashkenazic Mashait. d: probably Rhineland. e: Halakhic work. f: Paris: Bibliothe`que Nationale, MS He´b. 644. [See also HS 370.]

(37)

b: 1383. c: Ashkenazic Mashait. d: “Andernach beMedinat Qolonya” (Köln), Germany. e: Halakhic work. f: Oxford: Bodleian Library, MS Opp. 337.

(38, i) a: HS 356*. b: 1382/83. c: Ashkenazic Cursive. e: Alphabetic chart drawn by Author from Yiddish epic poem Avrohom Ovinu. f: Cambridge: University Library, MS T.-S. 10. K. 22. (38, ii) a: HS 356**. b: 1396/97. c: Ashkenazic Cursive. e: Alphabetic chart drawn by Author from Yiddish medical text. Archiv der Stadt Köln, Hebr. 4.

f: Köln: Historisches

(38, iii) a: HS 357*. b: 1484. c: Ashkenazic Cursive. e: Alphabetic chart drawn by Author from Hebrew legal document. f: Wiener Neustadt: Stadtarchiv (Magistrat, Abt. 10). (38, iv) a: HS 359*. b: 1574. c: Ashkenazic Cursive. d: Writer from “Byrswl”, Saxony, Germany. e: Alphabetic chart drawn by Author from Yiddish: Jewish customs and practices. f: Hamburg: Staats- u. Universitätsbibliothek, Hs. Hebr. 250.

000101

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 1

101

000102

102

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 2

000103

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 3

103

000104

104

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 4

Fig. 5

000105

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 6

105

000106

106

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 7

000107

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 8

107

000108

Birnbaum 01.04.11 18:17

108

[ 34 ]

Fig. 9

Fig. 10

000109

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 11

109

000110

110

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 12

000111

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 13

111

000112

112

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 14

000113

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 15

113

000114

114

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 16

000115

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 17

115

000116

116

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 18

000117

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 19

117

000118

118

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 20

000119

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 21

119

000120

120

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 22

000121

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 23

121

000122

122

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 24

000123

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 25

123

000124

124

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 26

000125

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 27

125

000126

126

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 28

000127

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 29

127

000128

128

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 30

000129

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 31

129

000130

130

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 32

000131

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 33

131

000132

132

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 34

000133

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 35

133

000134

134

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 36

000135

Birnbaum 01.04.11 18:17

The Development of the Hebrew Scripts

Fig. 37

135

000136

136

Birnbaum 01.04.11 18:17

[ 34 ]

Fig. 38

000137

35.

Birnbaum 01.04.11 18:17

Methodology in Hebrew Palaeography*

Background Let us start by defining the term palaeography. It is sometimes used to include everything connected with manuscripts, from the material written upon, to abbreviations, and to the arrangement of the contents. This does not seem to be right. The term (which literally means ‘oldwriting’) should be reserved for the central core of those things, that is, the study of the script itself. Our title does not imply that Hebrew palaeography has a method different from that of other palaeographies. One method is applicable to all scripts – exact investigation of the forms. Differences of method between the palaeographies of, say, Hebrew and Greek are of a nonessential nature. Our work begins with the assembly of the material. Script is to be found on the most diverse surfaces. It is therefore the basic element for various branches of research: the study of manuscripts, epigraphy, diplomatics, papyrology, and to a great extent also numismatics and sphragistics. It is a mistake to place these subjects into watertight compartments as far as their writing is concerned. All their scripts are within the domain of the palaeographer. It is, of course, obvious that the shape of certain letters which are written, carved or inscribed on a particular material may be influenced by the physical characteristics of that material. (For instance, rounded shapes or curves are difficult to incise in stone, and tend to appear in a more angular form.) Some scholars insist that palaeography must be limited to the study of documents written only by pen, and that inscriptions on other materials may not be used for the purpose of comparison with those written by pen. But palaeography must not be limited to the study of pen-written documents. As a matter of fact, purely epigraphical works such as the Corpus Inscriptionum Semiticarum 1 or Cooke’s North Semitic Inscriptions 2 * {This is the text, previously unpublished, of a lecture with slides given by Dr. S. A. Birnbaum in 1966, during a lecture-tour of North American universities. For illustrations of the slides shown at that lecture, and details about them, refer to Chapter 34 of this volume. The text has been slightly edited to fit the format of this volume. – D. J. B., E. B.} 1 {Paris 1881 ff.} 2 {George A. Cooke, A Text-book of North-Semitic Inscriptions: Moabite, Hebrew, Phoenician, Aramaic, Nabatean, Palmyrene, Jewish. Oxford 1903.}

000138

138

Birnbaum 01.04.11 18:17

[ 35 ]

deal not only with inscriptions but also with coins, seals and gems, and even pen-written documents, including ostraca and those written on soft material (e. g. papyrus). It need hardly be mentioned that the task of assembling the material is not an easy one. Very few libraries have good printed catalogues of their Hebrew manuscript collections; quite a number have only partial ones, or even none at all. Palaeographical information there is either non-existent, or – even in good catalogues – is restricted to a name. As the palaeographical classification is by no means always correct, we shall often meet with disappointment when we receive a photograph ordered on the strength of information given in the catalogue. It goes without saying that our photos should be made, as far as possible, from documents which bear a date and are of known provenance, so that we can, from the very beginning, establish regional patterns. The study of a script is based – so to speak – on measuring. In giving a description of a curve, what we are really doing is enumerating a series of measurements. While these measurements may not be precisely mathematical, they will be accurate enough to determine objectively the proportion between the height and width of the letters, the positions where the individual strokes of a letter touch each other or intersect, the comparative length and thickness of the strokes, the relation between them and the line ceiling and / or line bottom, and so on. By working out a comprehensive system of such ‘measurements’, letter by letter, and age by age, it is possible to build an unassailable basis for establishing order out of chaos, and for arranging the letter forms of the Hebrew scripts – about 100,000 of them, according to my very rough estimate – into groups: regional, functional and chronological. There is to my mind no other way of achieving reliable results. To make, for instance, style the basis of classification would generally be of little help in practice. It is too easy to succumb to vague, subjective and superficial impressions. Unless our impressions are related to a carefully worked out system and can stand precise analysis, they may mislead us, especially when our eyes have not undergone years of training in this field. Not that I would dream of denying the existence and importance of style in script. Very far from it – style is the great and dominant factor here, as elsewhere within the whole order of creation. However, it is accessible to objective analysis since it is built up on the basis of measurable details. The field of Hebrew palaeography is still in many ways virgin territory. We are still very far from having achieved an accurate delimitation of the

000139

Birnbaum 01.04.11 18:17

Methodology in Hebrew Palaeography

139

regional and functional groups, although it is to be hoped that we have got beyond the stage of the scholar, who, not so very long ago, could call a script ‘Greek Ashkenazi’. This name makes about as much sense as, let’s say, ‘Maghrebi Kufic’ would in Arabic palaeography or a term like ‘Merovingian Irish’ would in Western palaeography. However, one cannot be very optimistic if one sees what is happening to the term Sephardim outside the palaeographical domain. The Sephardim are the Jews of the Iberian Peninsula and their descendants, wherever they might have later settled. But at present we very often find that the term is used of any Jews who do not happen to be Ashkenazim. It is hard to imagine anything more absurd than to speak of ‘Yemenite Sephardim’, when what is meant is simply ‘Yemenites’, that is, Yemenite Jews. To go back to the script: How many groups are there? To give a precise figure is not possible but we might perhaps say that there are about thirty regional groups, apart from Palaeo-Hebrew and Samaritan, and these may each be subdivided into four functional groups. We would then arrive at a total of about one hundred and twenty groups. However, as our material is in some cases very deficient, the actual figure is much smaller, amounting to about seventy. This is not excessive if we consider the vast expanse of time and area involved. I shall not list the regional groups here. It goes without saying that the borderlines between some of them are not rigid. The same applies to the functional groups. These are: first, the Square script, employed in synagogue scrolls; second, the Square script used for Biblical, liturgical and occasionally for other manuscripts; third, the Mashait script, employed in commentaries, halakhic works, philosophical works, and so on; and finally we have the cursive used for writing letters, notes, legal documents and suchlike.

Review of Types and Styles Let us now take a quick glance at just a few specimens of various Types and Styles.

Palaeo-Hebrew Script Our first few slides show the script of the Hebrews. We cannot, however, call it the Hebrew script because that term is reserved for the script of the Jews. It is, of course, desirable to have a term which does, somehow, contain the word Hebrew. I have, therefore, coined the term PalaeoHebrew. Here (1), apart from a few letters of the alphabet, is the oldest

000140

Birnbaum 01.04.11 18:17

140

[ 35 ]

Palaeo-Hebrew document that we have. This little slab is the well-known Gezer Tablet or Gezer Calendar. As you see, the letters were incised into the stone by an unskilled hand, but the structure of the letters is perfectly clear. Next (7) we have one of the ostraca excavated on the site of ancient Lachish. It is written in a cursive script, by a skilled hand, at the very time when the Kingdom of Judah was coming to an end (597–588 B. C. E.). The national catastrophe did not mean the end of the script. It remained in use for about a century and a half after that. Then the Jews gave it up, while the Samaritans went on employing it, and continue to do so until the present. Here (8) is an example from a Samaritan Bible manuscript of the fourteenth century (1339/40 C. E.). The forms are highly stylised, and some of the letters are so far removed from those we saw in the last slide that without a knowledge of the development we would not be able to recognise them. Look at the zayin, for instance, or the samekh. As you know, the Jews, like other peoples of western Asia, adopted the Aramaic script, and, like them, developed a type of their own.

The Hebrew Script The third pre-Christian century may be regarded as the watershed between the Aramaic and what is generally known as the Hebrew script. Its correct name should really be the Jewish script. Palestinian Here (11) we have some fragments written in about 228 B. C. E. The forms are already well advanced towards becoming what is now called the Square script. In the next example (13), from the early second century of the Christian Era, the forms are already comparatively close to modern ones. The modern forms were reached by the ninth century C. E. You see (14) in this manuscript from the tenth century (930 C. E.) that the horizontals have become considerably thicker than the verticals. This contrast went on increasing so that the page became strongly dominated by the horizontals. Aramaic: Egypt Let us now leave Palestine and turn to the Egyptian group. The writing on this sherd (16) comes from about the year 300 before the Christian Era. Here (17) is a Bible manuscript which was found in 1962 in the University of Michigan Library at Ann Arbor by my son Professor

000141

Birnbaum 01.04.11 18:17

Methodology in Hebrew Palaeography

141

Eleazar Birnbaum, now of the University of Toronto. It dates palaeographically to the tenth / eleventh century C. E.3 Parsic Persia offers a style differing very much from that in the preceding slide. The Square writing of this specimen from 1571 C. E. (19) is less regular and formal, less calligraphic, than in the two specimens from Palestine and Egypt. Temanic This (21) Mashait manuscript comes from the Yemen of the late fifteenth century C. E. (1490). At first sight one might think it is in Square script, but comparison with a Square hand will show that it is not. Maaravic Here (24) we have a Mashait manuscript of 1401 from the Maghreb. There is no ‘Squarish’ atmosphere about this one! This (23) specimen of the cursive of the same group (1479) shows that cursive and Mashait are closely related. Neither of these – and especially the cursive – is easy to read. Sephardic The next two examples were written in Spain. This manuscript (25), written in Square script, is from 1207 C. E., and this (26), written in Mashait, is from 1471. Italkian This (28) eighteenth century cursive from Italy (1734) could also not be described as a model of clarity! Zarphatic Here (31) we have a French or Zarphatic Square script of 1285 C. E. It 3

See: Eleazar Birnbaum: “The Michigan Codex”, in: Vetus Testamentum 17 (1967), pp. 373– 415.

000142

Birnbaum 01.04.11 18:17

142

[ 35 ]

is very regular and calligraphic. The same applies to the corresponding Mashait, as you can see from this specimen (32). Its date is 1301. Ashkenazic Finally, we have two Ashkenazic manuscripts from the thirteenth century. This Type branched out from the Zarphatic, since the Ashkenazim are mainly the descendants of the Zarphatim. Here (33) is a manuscript written in Square script (1236 C. E.), and here (36) is one in Mashait (1264).

The Palaeographical Method One of the main practical applications of palaeography, and the principal one that outsiders are interested in, is determining the date and provenance of a document. This is done by comparing its script with that of other material for which these things are known. At this point we might ask ourselves whether it can be taken for granted that the script of any dated document is characteristic of its time, so that it may be utilised for the purpose. Here we meet with the problem of the overlapping generations. We know that at any given moment people of most unequal ages are writing. Should not this fact render it impossible to speak of the script of any particular period? The answer is that the various generations writing at the same time do not live in isolation, but overlap – they learn from each other and exert a mutual influence. Moreover, the number of extreme cases – that is, very old and very young people writing at the same time – is naturally rather small in comparison with those in the medium ranges, so that the medium ranges have, after all, a decisive influence. We are therefore justified in applying the law of averages, and do not run much risk if, as a general rule, we regard dated documents as representative of their time. When we have many dated documents of some particular time the risk can be regarded as negligible.

An Example We shall now show, by means of an example, how the date of an undated document is determined. If the forms of its letters are consistent, it is useful to draw charts like the ones that you see here (38). All four alphabets are from Ashkenazic Cursive manuscripts.4 4

{See also Chapter 56 in this volume.}

000143

Birnbaum 01.04.11 18:17

Methodology in Hebrew Palaeography

143

The first line (i) contains the script we are examining. One needs only a little experience to recognise where it comes from and what its style is: one can tell that it is Ashkenazic, and cursive in style. Before we start comparing it in detail with dated specimens, we will wish to narrow down our field of comparison in order to save unnecessary effort. Our material for the Ashkenazic cursive starts in the thirteenth century. It stands to reason that we need not work through all the eight centuries from the thirteenth to the twentieth in order to place it. Previous experience or a cursory comparison with a few dated Ashkenazic documents from different times will show the approximate general period to which the document belongs. In the present case we would say that it could hardly be as early as the thirteenth or as late as the seventeenth. Hence we need take into account only the intervening three centuries. I have drawn an alphabet from a dated manuscript from each century. You see them here in the lines (ii) to (iv). The one in the second line (ii) is from 1396, that in the third line (iii) from 1484 and the one in the last line (iv) is from 1574.

Detailed Comparison Let us now compare line (i), letter by letter, with the three others. Aleph: In lines (i) and (ii) the lower stroke on the right has only a slight downward bend, while in lines (iii) and (iv) it reaches the line bottom. Beth: In lines (i) and (ii) the base projects slightly rightwards, in contrast with lines (iii) and (iv). In line (i) it is of the same length as the top bar, in line (ii) not much longer; in lines (iii) and (iv) the difference is much greater. Gimel: In lines (i) and (ii) the left stroke is fairly straight, in line (iii) it is arched, in line (iv) it tends to be so. In lines (i) and (ii), it is shorter than the main stroke, in lines (iii) and (iv) of equal length or somewhat longer. Daleth: In lines (i) and (ii) the downstroke starts at an acute angle, in lines (iii) and (iv) the two strokes are fused into a curve. Heˆ: In lines (i) and (ii) the top bar is straight and the downstroke starts at an acute angle, while in lines (iii) and (iv) the two strokes fuse to form a curve. Zayin: In lines (i) and (ii), it consists of two parts, the head and the downstroke, the latter issuing from the centre of the head. In line (iii), the head and the beginning of the downstroke form an angle, which

000144

144

Birnbaum 01.04.11 18:17

[ 35 ]

then, in line (iv), becomes a curve, and the same happens to the rest of the downstroke. Teth: In lines (i), (ii) and (iii) the right stroke meets the left one somewhat above the bottom, in line (iv) they meet at a very acute angle. Final kaph: In lines (i) and (ii) the downstroke starts at an acute angle, in lines (iii) and (iv) the two strokes are fused into a curve. Mem: In lines (i) and (ii) the direction of the middle stroke is from the left below to the right above. In line (iii) it issues from the centre of the left stroke and is horizontal. In line (iv) the point of issue has actually moved up to the top, and the stroke now runs from above left to below right, that is, in the opposite direction from that in lines (i) and (ii). Final mem: In lines (i) and (ii) it is quadrilateral, in lines (iii) and (iv) it is bilateral: two curves, roughly symmetrical to an oblique axis, meet at the top left and at the bottom right. The upper curve corresponds to the original upper and right sides of the quadrilateral. The left half of the lower curve, however, does not correspond to the left side of the quadrilateral but is a secondary formation. This is what must have happened: when the writer had reached the left end of the base, instead of lifting his pen and starting afresh, he continued in an upward direction in order to begin the left downstroke from its correct position at the top. The bilateral outline now came to be regarded as the essential shape of the letter, the left downstroke being reduced to the role of a sort of appendage. Samekh: In lines (i) and (ii) the downstroke starts at an acute angle, in lines (iii) and (iv) the two strokes are fused into a curve. In lines (i) and (ii) the left side is a curve, in lines (iii) and (iv) it is straight. Pe: In line (i) the left stroke issues from the left end of the upper part, in line (ii) it is severed, in line (iii) it has become a dot. In addition, it is situated somewhat more to right, that is, inside the letter; in line (iv) it is again a stroke, linking the top bar with the base; however, it does not issue from the left end of the top bar but from somewhat more to the right, in conformity with the position of the dot in line (iii). Final pe: In lines (i) and (iv) the left part is severed from the right one, in contrast with lines (ii) and (iii). In lines (i), (ii) and (iii) the left part is oblique, in line (iv) it is vertical. Sadhe and final sadhe: In lines (i) and (ii) the right stroke is situated in the upper half of the line height, in line (ii) it juts out slightly above, in lines (iii) and (iv) it projects to almost an additional line height. Qoph: In lines (i) and (ii) the downstroke starts at an acute angle, in lines (iii) and (iv) the two strokes are fused into a curve. In line (iv) the upper part is stunted. It is continued right into the downstroke. The latter has developed, at the end, a large curve which sweeps rightwards.

000145

Birnbaum 01.04.11 18:17

Methodology in Hebrew Palaeography

145

Shin / Sin: In lines (i) and (ii) the middle stroke is straight and detached from the others. In lines (iii) and (iv) it forms almost a loop and joins the top of the left stroke.

Conclusion If we count the features in which line (i) agrees with one or other of the three alphabets, we find that in one instance the agreement is exclusively with line (iv), in one instance there is disagreement with line (ii), while in sixteen instances the agreement is exclusively with line (ii). This obviously means that line (i) must be nearer in time to line (ii) than to the other two lines (documents), and very much so: it must actually belong to the same period. Thus we may assign the document on which line (i) is based to the end of the fourteenth century. Is it possible to check this result? In the present case it certainly is. For in order to provide those of my listeners who are not palaeographers themselves with a means of checking the result, line (i) is from a manuscript, bearing the date of 5143 of the Aera Mundi, that is, 1382/83 of the Christian Era. In other words, our result is confirmed, and shown to be of a high degree of accuracy – as a matter of fact, the highest possible on purely palaeographical grounds, that being when the date arrived at falls within the limits of the correct generation. In our example the characteristic features of the script to be dated happen to point almost exclusively to one of the documents with which we compared it. This, of course, does not happen very frequently. In general, certain features point to this document and certain others to another document. We then have to take into account the ratio between the number of features which tally with this document or that, and thus work out the approximate date. If, for instance, 24 features point to script X and twelve features to script Y, then the unknown date of script Z will be double as close to X as to Y. For instance, if X is from the year 1400, and Y from 1700, then Z would be very roughly from 1500. I need hardly add that we are dealing here not with mathematics but with approximations. When necessary we not only have to count the numbers of features but also to weigh the quality of the evidence for each feature. When investigating cursive writing, a far greater degree of accuracy can be achieved than in the case of Square or Mashait, because the development of cursive script was comparatively quick. In Mashait the degree of approximation will be considerably smaller, and when we come to Square script we have to content ourselves with very much

000146

146

Birnbaum 01.04.11 18:17

[ 35 ]

less, because here changes took place very slowly indeed. Our greatest difficulty is in determining the date of synagogue scrolls, the script of which changed at an even slower rate. Perhaps the future will bring an expanding study of Hebrew palaeography and then even this most difficult chapter might yield up some of its secrets. And on this hopeful note I shall conclude.

000147

36.

Birnbaum 01.04.11 18:17

Notes on the Internal and Archaeological Evidence Concerning the Cave Scrolls*

Internal Evidence The Cave Scrolls are of two kinds. First we have the Biblical ones: Isaiah Scroll A and Isaiah Scroll B. Then there are the five non-Biblical ones: ‘The War Between the Children of Light and the Children of Darkness’, the ‘Book of the Covenant’,1 the Thanksgiving Hymns, the Lamech Scroll and the Habakkuk Commentary. Thus the internal evidence provided by this material is of two sorts: it is literary as far as the non-Biblical manuscripts are concerned, and purely textual in regard to the Isaiah Scrolls and to the Habakkuk passages embedded in the Commentary. In the War Scroll mention is made of the “Kittim of Misrayim” and ˙ is to the the “Kittim of Asˇsˇur”. It has been suggested2 that the reference Ptolemies and Seleucids respectively. This is an attractive interpretation. If correct, it would mean that the War Scroll was composed after the carving up of Alexander’s Empire and before the end of the Seleucid Kingdom, i. e., between 323 and 63 B. C. E. And not the War Scroll only. Certain internal similarities between it on the one hand, and the Book of the Covenant and the Habakkuk Commentary on the other, indicate that all three belong to the same period. * First published in: Journal of Biblical Literature 70 (1951), pp. 227–232. – The following lines constituted part of a lecture read before the Palestine Exploration Fund in London on March 22, 1950. 1 I use this name rather than ‘Sectarian Document’ because the whole group of scrolls is sectarian. In the published column the word covenant is frequently repeated and seems best to characterize the work. – In the Damascus Document (see Solomon Schechter, Documents of Jewish Sectaries. Vol 1, Cambridge 1910; Robert H. Charles, Fragments of a Zadokite Work. Translated from the Hebrew Text and Edited with Introduction, Notes and Indexes. Oxford 1912; and in {R. H. Charles [ed.],} The Apocrypha and Pseudepigrapha of the Old Testament in English. Vol. 2, Oxford 1913, p. 825), ch. XI , we find the words bspr hhgw wbyswdy hbryt. Does yswdy hbryt perhaps denote the title of another book and, if so, could it be our Book of the Covenant? (Charles, in his translation of the Damascus Document, spells the words with capitals, which looks as if he took them to be the name of a book: “learned in the Book of ‘the Hagu’ and in the Ordinances of the Covenant”.) 2 Eleazar L. Sukenik, Mgylwt gnwzwt {Megillot genuzot}; mtwk gnyzh qdwmh s´nms h bmdbr yhwdh; sqyrh r s´wnh. Vol. 1, Jerusalem 1948, p. 18. ˙

000148

148

Birnbaum 01.04.11 18:17

[ 36 ]

Opposition to this Kittim theory has been expressed on two grounds. First, that during the period of the Second Commonwealth Syria and Assyria were not confused, the Seleucids never being called the kings of Assyria. This, however, is an argumentum e silentio and is not a proof. Secondly, “Asshur, Assyria, and Mitzrayim, Egypt, during the Middle Ages referred to Rome”.3 This might possibly weaken the case for the Kittim theory if there were any hint in the War Scroll that these two terms c o u l d mean Rome. But there is not. On the contrary, there are obvious reasons against this interpretation. When the author of the work employs two different names the natural assumption is that he is simply referring to two different nations. Add to that – and it is a point of some importance – he does not merely speak about Misrayim and Asˇsˇur but about the “Kittim of ˙ of Asˇsˇur”. The stress is on the ethnic rather Misrayim” and the “Kittim ˙ than on the geographical aspect: the reference is to the same kind of people from two different countries. This would be fully in accord with the interpretation as Ptolemies and Seleucids, both being Macedonian dynasties. As a matter of fact, shortly before the discovery of the Scrolls it had been shown that the Jews of the Greek age used the name Kittim for the Macedonians.4 That there is a connection between the Damascus Document and the War Scroll, the Book of the Covenant and the Habakkuk Commentary was at once recognized.5 From this it should follow that our three works belong to the period of that famous fragment. Although no general agreement has been reached as to the time of its composition, it would 3

4

5

{Solomon Zeitlin, “Scholarship and the Hoax of the Recent Discoveries”, in:} Jewish Quarterly Review [JQR] 39 (1948/49), pp. 337–363. In support of this statement two sources are quoted. In Genesis Rabba, section 16, there is this passage: “That is it which Goeth toward the east (qdmt ) of Assyria [Gen. 2: 14] – R. Huna said: ‘With three things the kingdom of Greece preceded (qdmt ) this wicked kingdom’ [i. e., Rome] . . . R. Huna said in the name of R. Aha: ‘All kingdoms are called by the name Asˇsˇur [ SˇSˇuR] because they ˙ themselves [Mit aSˇSˇəRot] at the expense of Israel’ . . . R. Yose son of enrich R. Hanina said: ‘All kingdoms are called by the name Misrayim [MiSRayim] ˙ because they oppress [MəSiRot] Israel’.” Targum Jonathan˙ interprets ˙Asˇsˇur in ˙ Num. 24: 24 as Rome. A more detailed analysis of these passages is called for than the simple equation of Assyria and Egypt with Rome. By the way, the Amoraim here quoted lived in the third century, which can hardly be called “the Middle Ages”. H. L. Ginsberg, {“The Hebrew University Scrolls from the Sectarian Cache”}, in: Bulletin of the American Schools of Oriental Research [BASOR], 112 (1948), {pp. 19–23, especially} p. 20. E. L. Sukenik (above, n. 2), p. 18; H. L. Ginsberg (above, n. 4), p. 21.

000149

Birnbaum 01.04.11 18:17

Notes on the Evidence Concerning the Cave Scrolls

149

be safe to say that a date in the second or first century B. C. E. has gained most acceptance. If that is correct then it would support the date for the Cave Scrolls which follows from the Kittim theory. Conversely, if we can establish the age of the three Scrolls by external evidence we shall be able to form an opinion as to the date of the Damascus Document. The fact that one of the Cave Scrolls is a commentary on a book of the Prophets has been said to speak against an early date, the contention being that the Jews at the time of the Second Commonwealth were not in any need of commentaries and therefore did not write any.6 That they did not need any is a mere assumption. That they did not write any is again merely an argumentum e silentio. The other reasons which have been adduced against a pre-Christian date, and in favour of a late one, belong to this same category of negative argumentation, e. g., the alleged connection between the War Scroll and the Messianic speculations on the Oracles of Balaam, current during the Middle Ages.7 As a matter of fact, it had been shown before the Cave Scrolls had come on to the scene, that the Jews of the Hellenistic period were interested in these Oracles.8 Other arguments founded on the absence of evidence to the contrary are: There are certain terms in the non-Biblical Scrolls which have hitherto been known to us only from Talmudical or even later times.9 Their presence in the newly found documents is, we are told, an indication of the lateness of the Scrolls. Here are the examples on which this argument is based: Neither the Tetragrammaton nor the term dny is used, we have instead the word l; there is an expression mwrh hsdq ‘teacher of righteousness’; mrwdd is used for a certain sound of the˙ trumpet; the name ‘Israel’ is employed for the Jews. Let us consider this last example. Not only is it another piece of negative argumentation but it is unconvincing for an additional reason: the use of the word Israel here does not exclude either the possibility or the likelihood that the word ‘Jews’ was also employed at the period in question. That in this passage the word ‘Israel’ was given the preference over the word ‘Jews’ seems to have been simply a matter of style. The 6 7 8 9

{S. Zeitlin, “‘A Commentary on the Book of Habakkuk’: Important Discovery or Hoax?”, in:} JQR 39 (1948/49), {pp. 235–247, especially} p. 236 f. {Zeitlin}, (above, n. 3), p. 345. Ginsberg (above, n. 4), p. 20. {Zeitlin}, (above, n. 3), p. 347.

000150

150

Birnbaum 01.04.11 18:17

[ 36 ]

War Scroll, where the word occurs, is far from being a piece of dry narrative or prosaic description. How would it sound if the prayer offered up when the Jewish army is about to go into battle began not with the words “Blessed be the God of Israel” but with “Blessed be the God of the Jews”? Let us turn now to the Biblical Scrolls and see whether internal evidence can give us any help towards establishing their date. The opinion has been advanced,10 on the basis of the first specimens published from Isaiah Scroll A, that its text “agrees almost always with the later versions (Vulgate, Targum and Peshitta) against the earlier versions, particularly against the Septuagint”; and that the Scroll could not antedate the Christian Era because it represents the Masoretic Text “which seems to have grown up in the early years of the Christian Era”. This date for the Masoretic Text cannot, however, be regarded as an established fact. Hence, the date of the new text cannot be deduced from it. On the contrary, when we have dated Isaiah Scroll A by means of satisfactory external evidence, new light will be thrown on the history of the Masoretic Text. We shall now deal with another part of the textual evidence: that provided by the spelling. What differentiates the orthography of the Cave Scrolls – with the exception of Isaiah Scroll B – from that of the Masoretic Bible, is their general use of the scriptio plena for o and u,11 a final Aleph12 and a final Heˆ.13 This has been regarded14 as a proof that they are younger than the Septuagint. For it is obvious from the text of the Septuagint that in the Hebrew original from which it was made, this use of the scriptio plena was not general. The reason given for the scriptio plena is that it was only when Hebrew ceased to be commonly used that the matres lectionis were required as an aid to pronunciation. If that were so then we might expect to meet with evidence of such a general scriptio plena already in the Septuagint. For at the time of the Septuagint – the third century B. C. E. – the Jews no longer spoke Hebrew. Accordingly, the plene spelling of the Cave Scrolls does not prove a late age. To discover the relation between this orthography and 10 11 12 13 14

{G. R. Driver, “New Hebrew Manuscripts”, in:} JQR 40 (1949/50), {pp. 127– 134, especially} p. 129. Holem: qwl wmr; kwh; yw mr. Qames Hatuph: wkl hr, bsˇw lw. Qibbus: ˙ ˙ ˙ ˙ ˙ ˙ lkwlm. ky , py , my . At the end of the second and third persons: kmh and hmh, – indicating an earlier stage of pronunciation. {Driver}, (above, n. 10), p. 130.

000151

Birnbaum 01.04.11 18:17

Notes on the Evidence Concerning the Cave Scrolls

151

that of the Masoretic Text is an important piece of research which has yet to be done, particularly in view of the fact that a fragment of Isaiah Scroll B has in the meantime been exhibited and its spelling has proved to be practically identical with that of the Masoretic Text, although B is not much younger than A. Thus none of the arguments based on internal evidence15 are tenable, with the possible exception of the Kittim theory.

Archaeological Evidence The cave contained two earthenware jars, and a great number of fragments of what had once been jars and their lids. There had been in all 49 to 51 jars with lids. We know of course from literary sources – which have been confirmed by finds16 – that jars were used as containers for books. It is therefore only natural to assume that the Cave Scrolls were preserved throughout the centuries in these jars. That there should be no connection between these jars and the Scrolls seems to be out of the question. Still, in this, as in other respects, it is highly unfortunate that the find was made not by archaeologists but by treasure hunters, so that valuable evidence has been lost. Our first step must be to try to establish the date of the jars. The sherds are remarkably homogeneous as to material and technique.17 The archaeologists18 are unanimous in declaring that the ceramic evidence proves the vessels to be late Hellenistic. According to some scholars this period closes with the end of the second century B. C. E. Others add about another half century. But before we can utilize this evidence we must first consider how the ages of the jars and the Scrolls are related to each other. And that, in turn, is connected with the question as to why they were deposited there. 15

16 17 18

Recently it was reported that in 1947 a Haphtaroth Scroll had been seen together with Isaiah Scroll A. The former was described from memory and in quite a vague way, by a layman who had inspected it for a few minutes only, two and a half years previously ( T. Vkslr {Tuviah Wechsler}, “Hgnyzh hglwyh whgnyzh hgnwzh”, in: H wlm {Ha- olam} 37 [1949], p. 156). His description has now been brought forward in proof of the lateness of all the Cave Scrolls. This, of course, is no proof. No opinion can be formed until the manuscript in question is forthcoming and has been made accessible to experts. See R. de Vaux’s list {“La grotte des manuscrits he´breux”}, in: Revue biblique 56 (1949), {pp. 586– 609,} p. 591. R. de Vaux (above, n. 16), p. 587. W. F. Albright, G. L. Harding, O. Sellers, R. de Vaux, L. H. Vincent.

000152

152

Birnbaum 01.04.11 18:17

[ 36 ]

The jars are all identical in type, except for two, which are only very slightly different. That would favor the assumption that they are all of the same age and if so, that they were deposited in the Cave at one and the same time.19 For if they had been brought there on various occasions over some length of time, they would certainly not have all been alike. Hence the generally accepted view that they were hidden in the Cave in some sort of emergency. For in an emergency we would not expect the people who put the books into the vessels to have had at their disposal a collection of ancient jars, which were centuries – or even only several generations – old.20 It seems likely, therefore, that the jars were all brought in one lot and perhaps actually made to order for the purpose, which could have been done within a couple of days. And this possibility is favoured by the fact that they are – almost without exception – identical. Following this line of thought we arrive at the conclusion that the Scrolls were older than the jars, or, at best, contemporaneous but certainly not younger. And that the Manuscripts could not have been written – nor, of course, the works composed – later than the middle of the first century B. C. E. This conclusion from the archaeological evidence goes well together with that from the internal evidence, based on the Kittim theory, which places the Scrolls somewhere between 323 and 63 B. C. E. But neither conclusion helps us to limit down this span of 260 years or to discover whether the Scrolls are all of the same age, and if they are not, to decide what would be their chronological relationship. For that purpose we must turn to palaeography.21

19 20

21

R. de Vaux (above, n. 16), p. 595. J. Leveen, {“Hebrew Scrolls’ Age”, in:} The Times (London), August 26, 1949, p. 5; T. C . Lethbridge, {“Hebrew Scrolls’ Age”, in:} The Times (London), August 31, 1949; R. de Vaux (above, n. 16), p. 595. W. F. Albright, {“Editorial Note on the Jerusalem Scrolls”, in:} BASOR, 111 (1948), p. 2; idem, {“On the Date of the Scrolls from Ain Feshkha and the Nash Papyrus”, in: BASOR} 115 (1949), pp. 10–19; S. A. Birnbaum, in: BASOR 113 (1949), p. 33, 115 (1949), p. 20, JBL 68 (1949), p. 161, PEQ 81 (1949), p. 140. {These references to S. A. Birnbaum appear in the present book respectively as Chapters 40, 37, 41 and 42.}

000153

37.

Birnbaum 01.04.11 18:17

The Dates of the Cave Scrolls*

Since the writer’s articles on the Cave Scrolls were composed and the dates of their script suggested1 – on the basis of the facsimiles in BASOR 111 (1948) and Biblical Archeologist 11 (1948)2 – certain archaeological evidence has come to light. The excavation of the Cave in the vicinity of Ain Fashkha has shown that all the jars in which the Scrolls had been preserved belong to the late Hellenistic period, “towards the end of the second century B. C.” (Harding), in keeping with our terminus ad quem. It is to be hoped that scholars who believed they could see internal evidence proving the medieval origin of the Scrolls will now be convinced that this is impossible. But for those to whom it was at all times beyond question that the handwriting of none of the Scrolls is later than about the middle of the first pre-Christian century, this archaeological evidence has brought no additional help towards establishing a preciser date. An attempt at doing so – again by purely paleographical means – is made in the following lines. To that end we shall compare the writing of the Isaiah Scroll (Is) with that of the Nash Papyrus (Na). In our articles on the Cave Scrolls the Nash Papyrus did not figure among the material utilised for comparison. There is no archaeological evidence for the date of Na and we must as far as possible avoid deducing a palaeographical y from a palaeographical x. But there can be no objection to using Na for establishing its chronological relation to Is. The result should then enable us to determine the dates of both with greater precision since they belong to roughly the same period. The writer has examined3 Na in the same detail as Is and arrived at the first half of the second * First published in: Bulletin of the American Schools of Oriental Research (BASOR), 115 (1949), pp. 20–22. 1 Isaiah Scroll: 200–150 B. C. E. (BASOR 113 [1949], p. 33 {see Chapter 40 in this volume}); Covenant Scroll: 150–100 B. C. E. (PEQ 81 [1949], p. 140 ff. {see Chapter 42}); Habakkuk Commentary: 100–50 B. C. E. (JBL 68 [1949], 161 ff. {see Chapter 41}). 2 {John C. Trever, “Preliminary Observations on the Jerusalem Scrolls”, in: BASOR 111 (1948), pp. 3–16, facsimiles p. 6, p. 10, and p. 15; idem, “The Discovery of the Scrolls”, in: Biblical Archaeologist 11 (1948), pp. 45– 68, facsimiles p. 48 and p. 59.} 3 In his forthcoming book {The Hebrew Scripts, 1954–1971, vol. 1, no. 151}.

000154

154

Birnbaum 01.04.11 18:17

[ 37 ]

Comparative Table of Alphabets from Nash Papyrus and Isaiah Scroll. (The first four horizontal rows are put together from the Nash Papyrus {S. A. B., The Hebrew Scripts, plate 151}, the fifth and sixth from the reproductions in BASOR 111 (1948).

000155

Birnbaum 01.04.11 18:17

The Dates of the Cave Scrolls

155

pre-Christian century as its approximate date. Let us compare the letters which differ in the two manuscripts. Aleph: In Is the left hand stroke as a rule extends to the line bottom, as it frequently does in Na while in Eg (the third century B. C. E. documents from Egypt) it is small and sometimes high up. Daleth: The letter tends to be slightly wider in Is than in Na and Eg. Heˆ: The top stroke is horizontal in Is while in Na it is more often oblique, as in Eg. In Is the left hand downstroke is less oblique than in Na and Eg, occasionally it is almost parallel with the right downstroke. It issues from the middle of the top stroke. In Na it frequently goes forth from the right top corner, which it always does in Eg. Waw: The top part in Is is a right angle; in Na and Eg the whole upper half forms the head part. Teth: The left stroke is often not much above the line ceiling, in contrast with Na and Eg where it is tall. Yodh: Is has only the type with the strokes joining on the line ceiling, while in Na the right hand stroke issues from the middle of the left hand one, which is also the normal form in Eg. Mem: In Is the right top corner is sharply pointed. The Na form is intermediate between Is and Eg. At the line bottom the right downstroke turns leftwards to form a sharply pointed angle, while Na like Eg, has a curve. This bottom stroke is a fully developed base which Na does not yet have. In contrast with the curves of Na and Eg, the lines of Is are practically always straight. Final Mem: In Is the left stroke is vertical, in Na and Eg however it is oblique. Pe consists of straight lines with a right or acute angle at the base. In Na the letter is a curve. The base in Is is long, in Na the bottom part is small. Qoph: In Is the downstroke is mostly fairly long, while Na has only the old, short form. In Is it is often severed from the top stroke, Na however has no open form. Taw is of line height in Is. In Na and Eg it is a tall letter. In Is both strokes are somewhat straighter than in Na. Our investigation has shown that Aleph, Daleth, Heˆ, Waw, Teth, Yodh, Mem, final Mem, Pe, Qoph and Taw, i. e. nearly half the letters of Is are of a somewhat later type than those in Na. Thus it would appear that Is belongs to the later and Na to the earlier part of the period in question. The Isaiah Scroll belongs say to about the second quarter of the second pre-Christian century, and the Nash Papyrus to approximately the beginning of the century. This date is very near the maximum

000156

156

Birnbaum 01.04.11 18:17

[ 37 ]

figure – 165 B. C. E. – suggested by Albright in his masterly examination in the JBL for 1937.4 Let us see how our result affects our dates for the other two Scrolls. A detailed comparison between Is and the Covenant Scroll (Cov) reveals that Aleph, He, final Mem, Samekh and Qoph indicate Is to be the more archaic of the two.5 Thus, in assigning Cov to the second half of the second century B. C. E. we must not place it early in that period, otherwise it would be to close to Is. The quarter 125–100 is the more likely one. As to the Habakkuk Commentary (Hab), which we ascribed to the first half of the first century B. C. E., a detailed comparison with Cov shows a considerable difference between the two. Hence we must assume that there was an appreciable interval between Cov and Hab and place the latter at the end of the period in question – about the middle of the first century B. C. E.6

4 5 6

{W. F. Albright, “A Biblical Fragment from the Maccabaean Age: the Nash Papyrus”, in: JBL 56 (1937), p. 145–176.} Only Teth would seem to point the other way. This article was written and sent in to the editor before the writer had seen the July issue of the Jewish Quarterly Review. [It should be added that Dr. Birnbaum’s article was written without knowledge of the editor’s plans for this number and that it was received after the editor had written his own paper in this number. Thus there has been neither collaboration nor collusion. – W. F. Albright].

000157

38.

Birnbaum 01.04.11 18:17

How Old Are the Cave Manuscripts? – A Palaeographical Discussion*

I. The Weight of Palaeographical Evidence1 When the Cave Scrolls were discovered we were confronted with documents whose date was unknown and whose contents did not clearly place them within some obvious historical context. In a case like that the natural method of procedure is to begin by examining all the external evidence, and thus create a basis upon which to continue further investigations. The most important part of the external evidence is that provided by the script of a document. Here we have something tangible to work on, something which can actually be measured. Measuring is the basic method of palaeography. We determine the proportions between the height and width of the letters, the positions where the individual strokes of a letter touch each other or intersect, the relation between the letter and the line. In giving a description of a curve, what we are really doing is to enumerate a series of measurements. By working out a comprehensive system of measurements, letter by letter and age by age, it is possible to establish an unassailable palaeographical basis. To do this, all that is needed is a good eye, precision, patience and, last but not least, training. Once we have a reliable palaeographical basis to start with, it is easy to relate to it the script of any newly discovered document.

* First published in: Vetus Testamentum 1 (Leiden 1951), pp. 91–109. – The following study constituted part of a lecture read before the Palestine Exploration Fund in London on March 22, 1950. It is unaltered (except that the notes have been added and the palaeographical discussion of the Leviticus Fragments has here been shortened). The illustrations are the same as those then shown as slides but it has unfortunately not been possible to include here more than half of the original number, for reasons of space and expense. For the other half the reader is given the location of the documents and bibliographical references. The drawings in this article are by me. – S. A. B. 1 For abbreviations see the list at the end of this article.

000158

158

Birnbaum 01.04.11 18:17

[ 38 ]

II. Palaeographical References in Literary Sources Palaeography deals strictly speaking only with primary matter, with the raw material of research – the script itself. Literary references to script are decidedly of secondary importance. If however they are to be brought as arguments extreme caution must be exercised and only palaeographers should do it. It is perhaps worth quoting a few recent examples to show how even perfectly obvious things can give erroneous results if handled by scholars without the necessary palaeographical expertise. In Genesis Rabba, section 12, we have a description of the letter Heˆ, not, unfortunately, for the guidance of us poor palaeographers, but for Midrashic reasons. There R. Johanan bar Nappaha says: “The letter Heˆ is closed on all sides but open below – implying that the dead go down into Sheol. [ The letter Heˆ] has a little stroke sticking out on top – implying that the dead will rise. [ The letter Heˆ] has a window on its side – an allusion to those who [repent and] return to God”. This passage has been quoted2 as depicting the shape of the letter Heˆ in the Cave Scrolls, which would then date from Christian times, as R. Johanan lived in the third century. Let us take a look at the Heˆ of the Scrolls (fig. 1) and compare it with the description in the Midrash (fig. 2).

Fig. 1. The letter Heˆ from: 1. Habakkuk Commentary, 2. Covenant Scroll, 3. Isaiah Scroll A.

Fig. 2. R. Johanan’s Heˆ

In fig. 1 we see the opening below and the little stroke sticking out all right on top. But as far as the third characteristic is concerned, the window on the side, nothing here could be described as such – the lefthand wall stretches unbroken from top to bottom. In other words, the letter Heˆ in the Cave Scrolls is not the same as R. Johanan’s Heˆ, and the date of the Scrolls cannot be inferred from the Midrash. The form described by R. Johanan can without difficulty be found elsewhere in our palaeographical material: Here (fig. 2) we have the 2

{Solomon Zeitlin, “The Alleged Antiquity of the Scrolls”, in:} JQR 40 (1949/50), pp. 57–77, especially p. 59.

000159

Birnbaum 01.04.11 18:17

How Old Are the Cave Manuscripts?

159

stroke sticking out on top; the letter is open down below, and there is the window on the side. This corresponds to R. Johanan’s description but it is very unlike the Cave Scroll forms. Here is another example of how a literary reference should not be used: From Mishna Shab. 12 the deduction has been drawn3 that in the middle of the second century C. E. a Medial Mem could still be written in the final position, and from the Gemara (103b) discussing this Mishna it was inferred that even in the second half of the third century C. E. the final forms of Kaph, Mem, Nuˆn, Pe and Sadhe were not yet well established. It is not difficult to demonstrate that these inferences are fallacious and that, in fact, the very opposite is proved by the relevant passages4. But there is no need to do so. From fig. 3 it is obvious that already in the fifth century B. C. E.5 the trend towards the developing of Final letters had started. 3 4

5

JQR 40 (above, n. 2), p. 64. In Mishna Shab. 12: 3 we read: “R. Judah said: ‘We find a short word [made] from a long word: sˇm from sˇm wn or sˇmw l, nh from nhwr, dn from dny l, gd ˙ is being ˙ discussed: If on the from gdy l’.” In this passage a hypothetical case Sabbath a person intended to write the word ‘Simeon’ (sˇm wn) but stopped after the first two letters, sˇm, he would still have produced a word – sˇem, ‘name’, – and in that particular Mishna it is the production of a written word that is the Halakhic problem under discussion. What the Mishna means is simply that a Mem is always a Mem, whether in its medial or final form, and that this is all that matters here. There is not the slightest reason for assuming that the Tannaim thought sˇm with a Medial Mem a possible spelling for the word sˇem. – Nor did it occur to Rashi or to any of the other Commentators that the Mishna implied that the Mem in question was a medial form in a final position. Maimonides says: “If, in writing sˇm [with Medial Mem] it was his intention to write [the word] Simeon [sˇm wn] then he is guilty [of having desecrated the Sabbath by writing] because he has completed [the word] sˇm [i. e., sˇem ‘name’], although the Mem of [the word] sˇm [when written with the intention of writing sˇem ‘name’] is a closed [i. e., a Final] one, while that [which he has actually written] is an open [i. e., a Medial] one. What we must take into account is only [the fact that these two letters, when read as a word, do constitute a word in] the language”. Maimonides then expresses his agreement with R. Judah and adduces a parallel from another subject in support. – To R. Judah’s words the Gemara rejoins that “the Mem of [the word] sˇm [i. e., sˇem ‘name’] is closed, the Mem of [the two initial letters of] sˇm wn [i. e., sˇim on ‘Simeon’] is open”. This rejoinder makes it perfectly clear that the Final Letters were at that time already well established. The same holds good of the subsequent passage by R. Hisda: [R. Judah’s words] “prove that if a person has changed a Final Mem ˙into a Medial one, [the document in question] is still fit for use.” The dates represented on the chart are: (1) 495, (2) 484, (3) 471, (4) 465, (5) 460, (6) 460, (7) 456, (8) 456, (9) 447, (10) 441, (11) 420, (12) 419, (13) 419,

000160

160

Birnbaum 01.04.11 18:17

[ 38 ]

Fig. 3. Alphabets drawn by the author from dated papyri of the 5th century B. C. E. 6

In fig. 4 we have some Finals many centuries older than the second and third centuries C. E. – they are from a papyrus of the third century B. C. E.7

Fig. 4. Final letters in the Edfu Papyrus, 3rd century B. C. E.

6 7

(14) 416. Facsimiles in: A. H. Sayce and A. Cowley, Aramaic Papyri Discovered at Assuan. London 1906; and: Eduard Sachau, Aramäische Papyrus und Ostraka aus einer jüdischen Militär-Kolonie zu Elephantine, Leipzig 1911. {Cf. S. A. B., The Hebrew Scripts, plates 124–137.} A. H. Sayce and A. Cowley, An Aramaic Papyrus of the Ptolemaic Age from Egypt, in: PSBA 29 (1907), p. 260.

000161

Birnbaum 01.04.11 18:17

How Old Are the Cave Manuscripts?

161

Other documents preserved from earlier times than the second and third centuries C. E. containing Finals are the Ossuaries, from about the hundred years preceding the Fall of the Temple. Fig. 5 gives some Finals from this group.

Fig. 5. Final letters from Ossuaries, from the hundred years preceding the Fall of the Temple.

Now what about the argument that the Cave Scrolls must be late because they contain Finals?

III. Palaeographical Evidence A) The Scrolls in Square Script 1. External Features We shall first deal with three points of an external nature: the material written upon, the ruling, and the ink. a) The Material Written Upon The Cave Scrolls are written on leather. This does not help us with the dating. In Western Asia skin had been used for writing from time immemorial. The Palestinian Talmud8 gives it as a ha˘la¯ka¯ leˇ-mo¯sˇe¯ mis¯ on Sinai, that sı¯nay, i. e., as ordained by the Oral Law given to Moses all Torah scrolls must be written on skin. We need not deal here with the question as to what the term ha˘la¯ka¯ leˇ-mo¯sˇe¯ mis-sı¯nay really means in ¯ the history of the Halakha. In the present instance all that matters is that in the Palestinian Talmud the opinion is expressed that skin was so ancient as writing material that its use must go back to at least the time of Moses. It is only in Western palaeography that parchment is something late. The oldest dated Greek parchment preserved – and that, by the way, is 8

Meg., ed. Krotoschin, 71d.

000162

Birnbaum 01.04.11 18:17

162

[ 38 ]

of Syrian provenance – bears the date 195 B. C. E.9 This, though late in the history of skin as a writing material, is early enough to show that the use of skin for the Cave Scrolls is perfectly compatible with a preChristian date. b) Ruling The Cave Scrolls are ruled, in order to keep the lines horizontal and the columns vertical. What bearing has this fact on their age? A few scholars have seen in the ruling an indication of a late date, two reasons being given.10 First, that even in Greek documents ruling was not normally practised before the third century C. E., although “there are indeed papyri from Herculaneum with traces of ruled lines.” The implication is that the Jews learnt ruling from the West. But there is no basis in tradition or in known facts for such an assumption. We know that parchment was imported to Rome from the Orient, that it was not common in Rome even as late as under the first Emperors, and that no specimens were found at Herculaneum and Pompeii. Thus it seems improbable that the Greeks or Romans should have taught the new technique of ruling to the Orient, where skin had been used for ages. We would expect the opposite. What do we know from Jewish tradition about ruling? In the above-mentioned passage from the Palestinian Talmud ruling, too, is stated to be a ha˘la¯ka¯ leˇ-mo¯sˇe¯ mis-sı¯nay, and we may infer that ¯ ruling must antedate the time of these Amoraim – which is about the third century C. E. – by at least a good few centuries. It has been suggested in support of a later date for the Cave Scrolls, that at the time of that statement ruling was “a practice which was just beginning to take place in following non-Jewish usage”11 and that the statement was only an attempt to justify this practice by surrounding it with the aura of age – of Sinaitic origin.12 That, however, is a completely unhistorical explanation and shows quite a wrong appreciation of the stature and personality of the men in question.

9 10 11 12

F. Cumont, “Le plus ancien parchemin Grec”, in: Revue de philologie 48 (1924), pp. 97–111, especially p. 98. {G. R. Driver, “New Hebrew Manuscripts”, in:} JQR 40 (1949/50), pp. 127– 135, especially p. 129. {E. R. Lacheman, “A Matter of Method in Hebrew Paleography”, in:} JQR 40 (1949/50), p. 35. Ti August 30, 1949.

000163

Birnbaum 01.04.11 18:17

How Old Are the Cave Manuscripts?

163

As a matter of fact, this literary reference to ruling is not our earliest one. In the second century C. E. we find the following passage in Genesis Rabba:13 “Rabh said: ‘Even the ruling of the Book [goes back] to Adam’s teaching’.” This implies that at the time of the great Abba – generally called Rabh – ruling was considered to be “as old as the hills”, to date back far beyond the reach of living memory. On the strength of these two passages we may say that, even at a conservative estimate, ruling was used by the Jewish scribes long before the Christian Era. Finally, a point of comparative palaeography, which I think deserves attention. Samaritan scrolls are ruled. Now, it is completely out of the question that the Samaritans should have adopted a Jewish practice – or vice versa – after the Great Schism, and least of all, in connection with so sacred a matter as the writing of the Torah. Whatever the Jews and Samaritans have in common in this respect must go back to the time before the Schism. Thus ruling must have been in use by the fifth century B. C. E.14 The foregoing conclusions, based on literary and historical data, can now be confirmed by visual evidence: We have now at our disposal a very ancient document which is ruled: the Leviticus Fragments of the fifth century B. C. E. (see p. 176). In short, the argument in support of a late date on the score that the Cave Scrolls are ruled is also untenable. The Cave Scroll rulings are impressed with a “semi-sharp instrument which tended to make a slight crease in the material, discernible on the back of the Scroll almost throughout”.15 This, of course, is the traditional method, and is in accordance with the Talmudic regulation: me˘sarge˘lı¯n be˘-qa¯ne  , ‘Ruling is done with the reed’.16 When we think of ruling in relation to writing, we think of the letters as standing on the ruled line. That is what we are accustomed to from 13 14

15 16

Section 74. Even if the final separation is placed one or two centuries later, ruling would still be more than five or six centuries earlier than the third century C. E. – Since this paper was delivered, Dr. Plenderleith has mentioned in the course of a discussion that he recently examined a Ptolemaic papyrus and found it to be ruled. J. C. Trever, {“Preliminary Observations on the Jerusalem Scrolls”, in:} BASOR 111 (1948), p. 5. It was pointed out – JQR 40 (above, n. 10), p. 133 – that in the published photographs the ruling had an appearance of having been made with a lead point and it was suggested that if an examination of the originals were to show this to be so, then that of itself would establish the recent origin of the Scrolls, as lead was not used for ruling before the Middle Ages. The above description of the ruling in the Isaiah Scroll excludes such a possibility.

000164

Birnbaum 01.04.11 18:17

164

[ 38 ]

the Roman script. In the Cave Scrolls, however, the letters are below the ruling. This has been pointed out as noteworthy, and even advanced in proof of the medieval origin of the Scrolls.17 Suffice it to say that there has never been any other relationship between ruling and writing in Jewish script: the letters are always below, not above the ruled line. c) The Ink The statement has been made18 that if upon examination the ink used in the Cave Scrolls proved to be metallic, then they could not be early, because such ink had not come into use before the time of the Talmud. The opinion that until then metallic ink was unknown is based on certain passages in the Mishna and Gemara. They cannot, however, be interpreted in this way. Indeed we know that metallic inks were widely used in antiquity. Qanqantum ‘copperas’ is repeatedly mentioned as a writing material in the Mishna.19 To go back still earlier, the Lachish Ostraca have been found to be written with a mixed iron and carbon ink – and that was six centuries before the Christian Era. However, the ink of the Cave Scrolls is non-metallic. It was examined by Dr. Plenderleith in the British Museum Laboratory and he found it to be a carbon ink. This does not, of course, constitute a proof that the manuscripts are pre-Medieval. But if anything, it tips the scale a little on this side rather than on that, because in medieval times the use of metallic inks became general. 2. Palaeographical Evidence Proper What we have to do is to place the writing of the Cave Manuscripts side by side with that of documents whose dates are known to us, to mark the similarities and dissimilarities down to the minutest detail and so establish their date. We shall take into consideration documents covering a much wider span of time than would be really needed from the palaeographer’s point of view. But we shall do this because the conviction has been expressed that the script of the Cave Scrolls is medieval. Palaeographically such a statement is nothing short of fantastic, but we shall none the less begin our comparison by placing a column 17 18 19

JQR 40 (above, n. 10), p. 129. JQR 40 (above, n. 10), p. 134. Gittin 2: 3; Sota 2: 4.

000165

Birnbaum 01.04.11 18:17

How Old Are the Cave Manuscripts?

165

from Isaiah Scroll A next to a document20 of 105021 whose script it has been said to resemble.22 I do not think it requires much palaeographical experience to see that there is a big difference between the two. The same holds good for the Damascus Document B of the tenth century.23 whose writing, we have been told,24 also resembles that of the Cave Scrolls. Next we shall examine a few individual letters from these documents (fig. 6). First we shall take Aleph, Daleth, Lamedh and Shıˆn because their forms in the Cave Scrolls on the one hand, and in the two medieval manuscripts on the other, are supposed to be particularly similar.25

Fig. 6. The letters Aleph, Daleth, Lamedh, and Shıˆn from: 1. Damascus Document B (10th cent.), 2. Document of 1050, 3. Isaiah Scroll A.

Let us look at the first letter. In Isaiah Scroll A, a main stroke runs from the top left-hand corner to the bottom right-hand corner. In the medieval manuscripts the structure of Aleph is completely different. It consists of 20 21

22 23 24 25

Cambridge, University Library, MS. T.-S. 16: 79. Not of 750, as stated. The numbers are written in words, and – what is by no means an isolated occurence – the scribe has accidentally omitted a part of the date, jumping from “a thousand” (of the Seleucid Era) to “sixty two”, leaving out the intervening “three hundred”. Cf. J. L. Teicher, {“The Oldest Dated Document in the Genizah?”, in:} JJS 1 (1948/49), p. 156 f. {S. Zeitlin, “The Scholarship and the Hoax of the Recent Discoveries”, in:} JQR 39 (1948/49), {pp. 337–363, plates after p. 363, especially} p. 362. Cambridge, University Library, MS. T.-S. 16: 311. Facsimile in S. Schechter, Documents of Jewish Sectaries. Cambridge 1910, vol. 1, plate B. JQR 39 (above, n. 22), p. 362. Ibid.

000166

166

Birnbaum 01.04.11 18:17

[ 38 ]

a right-hand half and a left-hand half. What in the Isaiah Scroll is a leg, is, in the medieval documents, a tall main stroke, curving in an opposite direction from that in the Isaiah Scroll. If these two types of Aleph could be called similar – then what does “similar” mean? The differences in the case of Daleth, Lamedh and Shıˆn are, as the chart will show us, not less sweeping. As a matter of fact, these four letters happen to be among the less characteristic ones in the alphabet and so the dissimilarity between the two sets will be even greater in the more characteristic letters. Let us take, e. g., Heˆ, Zayin, and Samekh (fig. 7). By what stretch of imagination could the forms on the left and those on the right be called similar?

Fig. 7. The letters Heˆ, Zayin, and Samek – from the MSS. listed in fig. 6.

We had to compare the script of the Cave Scrolls and that of those two medieval manuscripts because specimens from both sets had been published on one page26 to show that they belonged to one and the same period. Actually, it would have been preferable to have avoided such a comparison because the manuscripts in question are not written in the same style of script: the Cave Scrolls are written in a careful, formal hand, corresponding to what we call Square script, while the two medieval manuscripts are written in cursive. The right procedure would have been to compare the writing of the Cave Scrolls with that of medieval documents in Square script, e. g. the Isaiah Scroll with a 9th century codex.27 We shall compare a few characteristic letters from some manuscripts written between the seventh and tenth centuries with their counterparts in the Cave Scrolls. 26 27

JQR 39 (above, n. 22), plate 2 after p. 363. British Library, MS . Or. 4445.

000167

Birnbaum 01.04.11 18:17

How Old Are the Cave Manuscripts?

167

Fig. 8. The letter Heˆ from: 1. a 9th cent. cod., 28 2. Is. Scroll A, 3. a 7th cent. fragm., 29 4. Covenant Scroll, 5. a 10th cent. codex, 30 6. Habakkuk Scroll.

Fig. 8 shows Heˆ. We have discussed this letter before, so that it is enough for us just to take a look at the two sets, to be convinced that they cannot possibly have been written during the same period. Fig. 9 gives Zayin from the same six documents. In the column on the right we can observe what seems to be a development: from a plain line to one with a head hardly more than a bulge to the right. In the left-hand column we see a letter consisting of two distinct parts, a body and a head, and this head looks to the left. Next, let us compare the Samekh (fig. 10). The idea that the two sets of documents belong to the same age, or even to periods not far apart, is absurd.

Fig. 9. The letter Zayin from the documents listed in Fig. 8.

28 29 30

Fig. 10. The letter Samekh from the documents listed in fig. 8.

Ibid. Cambridge, University Library, M.-S. T.-S. 12: 195. St. Petersburg: Gosudarstveˇnnaya Publichnaya Biblioteˇka: II Firkovich, No. 159.

000168

168

Birnbaum 01.04.11 18:17

[ 38 ]

We shall, therefore, have to go back much further than the period from the seventh to the tenth century. The difference between the Piyut Papyrus of the fourth century C. E.31 and the Cave Scrolls is very great indeed. Let us look at another example, a century or so further back, and compare the liturgical fragment of Dura-Europos32 with the Cave Scrolls. Let us examine a few individual letters (fig. 11).

Fig. 11. The letters Heˆ, Zayin, Mem, and Samekh from: 1. Piyut Fragm., 4th cent C. E.; 2. Dura Fragm., 1st half of 3rd cent C. E.; 3. Covenant Scroll; 4. Isaiah Scroll A.

We can see that the difference between the two sides is very great. In other words, in spite of the many centuries we have gone back we have not found a similar script. Let us therefore continue to move in the same direction until we reach the period of hundred years before the Fall of the Temple. This is the time of the Ossuaries.33 From the same period we have the famous 31 32

33

British Library MS . Or. 9180, No. A. – Cf. P. A. H. de Boer, {“Notes on an Oxyrhynchus Papyrus in Hebrew”, in:} VT 1 (1951), p. 49. New Haven, Yale University, School of the Fine Art, Dura Parchment 25. Facsimile in: The Excavations at Dura-Europos, vol. 6, ed. M. I. Rostovtzeff [et al.], New Haven 1936, plate 36. 1. {Reproduced in: S. A. Birnbaum, The Hebrew Scripts, no. 103*}. Q 1,1 (1921), p. 57. RA N. S. 25 (1873), p. 398. RA N. S. 36 (1878), p 305. RA Se´r. 3, 1 (1883), p. 257. Charles Clermont-Ganneau, Archaeological Researches in Palestine during the Years 1873–1874, London 1899, vol. 1, p. 381. RAO vol. 5, p. 334. RAO vol. 6, p. 211. RAO vol. 8, p. 59. MPJ, nos. 16, 51. S 4 (1923), p. 241. SBB 1885, p. 669. OLZ 15 (1912), p. 529. Samuel Klein,

000169

Birnbaum 01.04.11 18:17

How Old Are the Cave Manuscripts?

169

Uzziah Tablet.34 Let us now compare some letters from Isaiah Scroll A with their counterparts on the Ossuaries (fig. 12).

Fig. 12. The letters Aleph, Beth, Heˆ, Zayin, and Samekh from: 1. Ossuaries, 2. Isaiah Scroll A.

It is not the difference between the curves of the Isaiah Scroll and the straight lines of the Ossuaries which counts here. For it might perhaps be thought that these were conditioned by the material: straight lines might be said to be natural on stone. ( The Samekh, however, shows that the drawing of curves did not present great difficulties). But let us concentrate on differences of essential importance.

34

Jüdisch-palästinensisches Corpus Inscriptionum, Wien 1920. JPOS 9, (1929), p. 100. ESE vol. 1, p. 186, 312; 2, p. 72, 191; 3, p. 49, 302. HNE, p. 485. Robert A. S. Macalister, The Excavation of Gezer, London 1912, vol. 1, p. 347, 382. QS 63 (1931), p. 171. Q 1,2 (1925), p. 71. BBSAJ 5 (1924), p. 58 f., no. 3. BAMP 1 (1867), p. 21. RB 34 (1925), p. 253. RB 1 (1892). p. 262. JAOS 28 (1907), p. 355. ZDPV 4 (1882), p. 9. JPOS 4 (1924), p. 171. JPOS 8 (1928), p. 113. JPOS 9 (1929), p. 45. JPOS 12 (1932), p. 25. Spr hywbl lprwpyswr s´. qrwys {Festschrift Samuel Krauss. Jerusalem 1936/37}, p. 87. PEQ 69 (1937), p. 126. Spr zkrwn l s´r gwlq wls´mw l qlyyn {Memorial Volume Asher Gulak and Samuel Klein. Jerusalem 1942}, p. 133. RB 11 (1902), p. 276. RB 43 (1934), p. 564. JPOS 9 (1929), p. 41. JPOS 10 (1930), p. 21. QS 32 (1900), p. 75. RB 9 (1900), p. 106. BASOR 44 (1931), p. 8. T 2 (1930), p. 288; p. 295. QS 63 (1931), p. 217; 64 (1932), p. 106.

000170

170

Birnbaum 01.04.11 18:17

[ 38 ]

Aleph: In Isaiah the left stroke starts near the middle of the main stroke, on the Ossuaries it issues from its top. Beth: In Isaiah, the top part is a tick, on the Ossuaries it is a long stroke. It here runs to the top of the downstroke, but in Isaiah to a point lower down. In Isaiah the angle between the downstroke and the base is rounded off; on the Ossuaries this takes the shape of a base sticking out on the right. Heˆ: In Isaiah, the left downstroke is near the right one, on the Ossuaries it is on the left edge of the letter. In Isaiah, it issues from the top stroke, on the Ossuaries it is separated from it. In Isaiah it slants to the left, on the Ossuaries it is vertical. Zayin in Isaiah is a plain stroke, on the Ossuaries there is a little stroke added on top. Samekh in Isaiah consists of a long curve on the right and a shorter, or short one, on the left, connected on top. On the Ossuaries we have a closed figure, almost a triangle. Thus the two columns still clearly represent two different stages of development, so that the Isaiah Scroll must antedate the final decades of the first century B. C. E. We shall, therefore, take our comparison farther back and examine the script of a papyrus of the third century B. C. E.,35 by placing a few individual letters next to those of Isaiah Scroll A (fig. 13).

Fig. 13. The letters Heˆ, Zayin, Kaph, Ayin, and Taw, from: 1. Legal Papyrus, 3rd cent. B. C. E.; 2. Isaiah Scroll A. 35

Bodleian Library, MS. Aram. e. 2. (P).

000171

Birnbaum 01.04.11 18:17

How Old Are the Cave Manuscripts?

171

Here, for the first time in our quest, we gain an impression of similarity. Let us see how they compare in detail: Heˆ: In Isaiah, the lower stroke on the left issues from the upper one, but in the Papyrus it goes forth from the main stroke. Zayin: The form is identical in both documents, but there is a difference in the size. The same holds good for the top part of Kaph. Both letters have a base but in Isaiah it is straight, in the Papyrus curved. Ayin in Isaiah has our modern form but is very small. In the Papyrus it consists of two equal arms. Taw: In Isaiah, there is not much difference in the size of the strokes, in the Papyrus the left one is much taller. Here it is also very wavy while in Isaiah it is rather straight, apart from the bend at the bottom. In short, although the forms in the two columns are more similar than in any of the previous charts, they are not identical. If we went still further back and examined the forms of the fifth century papyri (see fig. 3) we would at once receive a warning: the differences between the two columns would again be much greater than here – but in quite a different way from those in the earlier charts. That would mean that we had gone too far in a different direction, i. e., too far back. In fact, already in this chart we have gone too far back. It is obvious, e. g., that the Heˆ in Isaiah (figures 12 and 13) stands half way between that of the Ossuaries (fig. 12) and that of the Papyrus (fig. 13), and that the Ayin in Isaiah is in between our modern form and that of the Papyrus. In other words, the forms in Isaiah represent a development away from those of the Papyrus, i. e., the script of Isaiah Scroll A must be younger than the third century B. C. E. but earlier than the final decades of the first century B. C. E. This leaves us with about a century and a half to limit down. Before going into this, it might be useful to take a quick glance at the development which took place in the shape of the letters during the fifteen centuries from the sixth B. C. E. to the ninth C. E. Let us take the letters Heˆ and Samekh (fig. 14). I wonder whether anybody examining this chart can doubt that the correct position for the forms in the Cave Scrolls is where they are in fig. 14, between lines 2 and 5.

000172

172

Birnbaum 01.04.11 18:17

[ 38 ]

Fig. 14. The letters Heˆ and Samekh from: 1. a papyrus of 515 B. C. E.; 36 2. Heˆ from a 3rd century B. C. E. papyrus (cf. fig. 13), Samekh from an ostracon of the same period; 37 3. Isaiah Scroll; 4. Habakkuk Commentary; 5. Ossuaries; 6. Piyut Papyrus, 4th century C. E. (see above, n. 31); 7. Cairo Codex, 895. 38

In order to establish the dates of the Cave Scrolls within those 150 years, we must compare in the utmost detail each letter of their alphabets with the corresponding forms of the adjoining periods. When we do so, we shall find certain features connecting the Cave Scroll forms with the lower limit, and others with the upper limit. Counting and weighing up this evidence – separately, of course, for each manuscript – 36 37 38

Hans Bauer / Bruno Meissner, “Ein aramäischer Pachtvertrag aus dem 7. Jahre Darius’ I.”, in: SBB (1936), p. 414. E. Sachau (see above, n. 5), pl. 62, 1, II {cf. Chapter 34, fig. 16}. William Wickes, A Treatise on the Accentuation of the Poetical Book. Oxford 1881; A. Neubauer, “The Introduction of the Square Characters in Biblical MSS . and an Account of the Earliest MSS . of the Old Testament”, in: SBE 3 (1891), p. 1; Richard Gottheil, “Some Hebrew Manuscripts in Cairo”, in: JQR 17 (1904/5), p. 609.

000173

Birnbaum 01.04.11 18:17

How Old Are the Cave Manuscripts?

173

we decide whether its writing is closer to that of the earlier or of the later period, and what is the degree of that closeness, so that we are finally in a position to estimate approximately at what point on the time graph between the upper and the lower limits each manuscript is situated. Such a detailed comparison is, naturally, a laborious affair. I have elsewhere39 supplied the details on which my datings are based.40 The results are these: Isaiah Scroll A and the Covenant Scroll were written in the second century B. C. E. – Isaiah approximately in its second quarter, the Covenant Scroll in its last quarter; the Habakkuk Scroll about the middle of the first century B. C. E. Thus the Book of the Covenant and the Habakkuk Commentary must have been composed before the Herodian and Roman periods. The War Between the Children of Light and the Children of Darkness, which internal evidence has shown to be connected with them, must belong to the same date, and this appears also to hold good for the Thanksgiving Hymns.

B) The Fragments in Palaeo-Hebrew Script 1. The Age of the Leviticus Fragments Among the fragments of manuscripts found by the archaeologists in the Cave there were a few41 which may from some points of view equal in importance all the large Scrolls put together. Although each of them contains only a few words, these belong to adjoining chapters (19, 20, 21) of Leviticus and there is sufficient material to enable us to reconstruct the whole of those passages. The words on our Fragments coincide with the Massoretic Text. An accurate dating does not seem possible from this piece of internal evidence, first because the material is too scanty, and second because the date of the Massoretic Text itself is controversial. As regards archaeological evidence, we are no better off. We cannot take it for granted that these bits came from the jars which contained the 39 40

41

BASOR 113 (1949), p. 33; 115 (1949), p. 20; JBL 68 (1949), p. 161; PEQ 81 (1949), p. 140 {= Chapters 40, 37, 41 and 42 in the present volume}. It may be mentioned that the dates of the Scrolls had been fixed by palaeography (BASOR 111 [1948], p. 2; 113 [1949], p. 6; 113, p. 33; JBL 68 [1949], p. 161) months before the pottery had even been taken out of the Cave, and still longer before the findings of the archaeologists had been made known. R. de Vaux, “La grotte des manuscrits he´breux”, in: RB 56 (1949), p. 597; S. A. Birnbaum, “The Leviticus Fragments from the Cave”, in: BASOR 118 (1950), p. 20 {= Chapter 33 in the present volume}.

000174

174

Birnbaum 01.04.11 18:17

[ 38 ]

documents in Square script, although we may assume such a connection for the fragments in Square script, because some of them actually form part of the Scrolls. Hence it is safer not to take the date of the jars as the terminus ad quem for our Fragments. Let us now turn to the palaeographical evidence. In our two final charts (figures 15 and 16) we have compared five characteristic letters from the Leviticus Fragments with their counterparts in a number of other Palaeo-Hebrew documents, ranging from the eighth century B. C. E. to the second C. E., and with Samaritan ones from the 2nd-8th cent. C. E.

Fig. 15. The letters Aleph, Waw, Yodh, Mem, and Shıˆn from: 1. Samaria Ostraca, 42 780–767 B. C. E.; 43 2. Lmlk Stamps, 1st half of 7th century B. C. E.; 44 3. Lachish Ostraca, 45 594–588 B. C. E.; 4. Leviticus Fragments.

42 43 44 45

George A. Reisner [et al.], Harvard Excavations at Samaria. Cambridge (Mass.) 1924. S. A. Birnbaum, “The Dates of the Gezer Tablet and of the Samaria Ostraca”, in: PEQ 74 (1942), p. 107 {= Chapter 30 in the present volume}. Cf. D. Diringer, “The Royal Jar Handle Stamps of Ancient Judah”, in: BA 12 (1949), p. 70. Harry Torczyner [et.al.], Lachish. Vol. I: The Lachish Letters. London 1938. S. A. Birnbaum, “The Lachish Ostraca”, in: PEQ 71 (1939), p. 20, p. 91 {= Chapter 32 in the present volume}; idem, Review of The Lachish Letters, in: BSOS 9 (1937/38), p. 807.

000175

Birnbaum 01.04.11 18:17

How Old Are the Cave Manuscripts?

175

Fig. 16. The letters Aleph, Waw, Yodh, Mem, and Shıˆn from: 1. Leviticus Fragments; 2. “Jerusalem” Jar Handle Stamps, 46 4th cent. B. C. E.; 3. Habakkuk Tetragrammaton 47 ( Waw, Yodh) and El Fragment, 48 about middle of 1st century B. C. E.; 4. Hasmonaean coins of 1st cent. B. C. E.; 5. Amwaˆs Capital 49 (Shıˆn), 2nd cent. C. E., Salbit Mosaic (Mem – Yodh – Waw), 4th cent. C. E.; Decalogue Tablet 50 (Aleph), 8th – 9th cent. C. E.

The detailed examination which we have made elsewhere51 could perhaps be summarised as follows: In fig. 15, going from top to bottom, a line of development may be discerned. Hence the document in line 4 – the Leviticus Fragments – is later than that from the sixth pre-Christian century shown on line 3.

46 47 48 49 50 51

W. F. Albright, in: BASOR 53 (1934), p. 20; E. L. Sukenik, in: JPOS 15 (1935), p. 341. M. Burrows (ed.), The Dead Sea Scrolls of St. Mark’s Monastery. Vol. 1, New Haven 1950, pll. 57, col. 6, 59. col. 10, 60, col. 11. ILN October 1, 1950, p. 494, fig. 6. Ch. Clermont-Ganneau, “Rapport sur une mission en Palestine et en Phe´nicie”. in: AMSL se´r. 3, vol. 11 (1885), p. 211. James A. Montgomery, The Samaritans. Philadelphia 1907, pl. I. Birnbaum (above, n. 41), p. 20.

000176

176

Birnbaum 01.04.11 18:17

[ 38 ]

In fig. 16, a line of development is also apparent between the forms of the Leviticus Fragments (line 1) and those of the latest documents (line 5). Our text must therefore have been written before the third century B. C. E. (line 2). Our final result was a date in the second half of the fifth century B. C. E. According to both Talmuds,52 the Samaritan53 and patristic54 literatures, the Palaeo-Hebrew alphabet was discarded for Torah writing by the Jews “in the days of Ezra”. As this presumably includes also the period of Nehemiah, we might place the great change somewhere between 458 and 532. Taking this into account we would be inclined to ascribe the Leviticus Fragments to about the middle of the century or slightly later.55

2. The Significance of the Leviticus Fragments Thus the Leviticus fragments give us our earliest glimpse of the Biblical text. They are about two and a half centuries older than the Nash Papyrus,56 Isaiah Scroll A, and our most ancient relics of the Greek version. And the words in our document coincide with the Massoretic text. The Leviticus Fragments are also our first specimen of the Scriptures written in the Palaeo-Hebrew alphabet. Hitherto this script was known to us only from ostraca, inscriptions, seals and stamps. Then some of the Cave Scrolls provided a few more examples of it, being used for the Tetragrammaton and the word l. Furthermore, our Fragments are the oldest relic of a Hebrew book. In spite of their smallness we are able to reconstruct its layout.57 52 53 54 55

56

57

Sanh. 21b, J. Meg. 1: 9, Zeb. 62a. Abu ‘l Fath, Chronicle. Origenes to˙ Ps. 2: 2; Jerome, Prologus galeatus. A date in the last quarter of the century would mean that they are from the pen of a Samaritan scribe. As it is not likely that the Cave was used by two hostile groups it is reasonable to suppose the Leviticus Fragments to be remnants of a Jewish scroll, and therefore closer to the middle of the century. (At the time of the Schism there was presumably no difference yet between Jewish and Samaritan script). W. F. Albright, in: JBL 56 (1937), p. 145; idem, in: BASOR 115 (1949), p. 10; S. A. Birnbaum, in: BASOR 115 (1949), p. 20 {= Chapter 37 in the present volume}. S. A. Birnbaum, in: BASOR 118 (1950), pp. 21–24 {= Chapter 33 in the present volume}.

000177

Birnbaum 01.04.11 18:17

How Old Are the Cave Manuscripts?

177

In addition, this is our earliest example of a formal book hand in the Palaeo-Hebrew handwriting, the letters on the ostraca being, of course, cursive. The Leviticus Fragments contain our earliest example of ruling on skin. Finally, they are also one of the oldest extant documents in any language written on skin. Abbreviations AMSL: BA: BAMP: BASOR: BBSAJ: BSOS: ESE: HNE: ILN: JAOS: JBL: JJS: JPOS: JQR: MPJ: OLZ: PEQ: PSBA: Q: QS: RA: RAO: RB: S: SBB: SBE: T: Ti: VT: ZDPV:

Archives des missions scientifiques et litte´raires. The Biblical Archaeologist. Bulletin arche´ologique du Muse´e Parent. Bulletin of the American Schools of Oriental Research. Bulletin of the British School of Archaeology in Jerusalem. Bulletin of the School of Oriental Studies (London). Mark Lidzbarski, Ephemeris für semitische Epigraphik. 3 Bde., Giessen 1902–1915. Mark Lidzbarski, Handbuch der nordsemitischen Epigraphik. Weimar 1898. The Illustrated London News. Journal of the American Oriental Sociaty Journal of Biblical Literature Journal of Jewish Studies Journal of the Palestine Oriental Society Jewish Quarterly Revue Rene´ Dussaud, Les monuments palestiniens et judaiques [. . .] Muse´e du Louvre. Paris 1912. Orientalistische Literaturzeitung. Palestine Exploration Quarterly Proceedings of the Society of Biblical Archaeology. qwbs hhbrh h brit lhqyrt rs ys´r l w tyqwtyh {= Journal of the Jewish ˙ ˙ Exploration ˙ Society}. ˙ Palestine Quarterly Statement of the Palestine Exploration Fund. Revue arche´ologique Charles S. Clermont-Ganneau, Recueil d’arche´ologie oriental. 8 vol., Paris 1888–1924. Revue biblique. Syria. Sitzungsberichte der preussischen Akademie der Wissenschaften, phil.hist. Klasse (Berlin). Studia Biblica et Ecclesiastica. Tarbis. ˙ The Times (London). Vetus Testamentum. Zeitschrift des deutschen Palästina-Vereins.

000178

Birnbaum 01.04.11 18:17

39.

The Qumran Scrolls and Palaeography*

A Matter of Presentation In the course of a report on the palaeographical work done about the Cave Scrolls, O. Eissfeldt expressed the opinion that it is ‘on the right track’.1 In his book on the Cave Scrolls,2 P. Kahle refers to Eissfeldt’s article but omits this essential statement. His mention of the article is embedded in an abundance of negatives: ‘[. . .] papyri and ostraca [. . .]. We have seen how uncertain their dating is. About the attempt [. . .] to reach an objective date for the scrolls found in the cave, with the help of these documents which are not datable with certainty, Eissfeldt, for example, has reported at such length, that I can refer the reader to that. I am convinced that Lacheman is right when he points out that in this way one cannot arrive at a safe result, at least not for the time being’.3 Anyone who has not read Eissfeldt’s article would conclude from Kahle’s words that Eissfeldt had condemned the palaeographical method, when, on the contrary, what he said was that the palaeographical investigation regarding the Scrolls is ‘on the right track’. A rather different matter.4 * Extracted from: Solomon A. Birnbaum, The Qumraˆn (Dead Sea) Scrolls and Palaeography, New Haven 1952 (= Offprint from: BASOR Supplementary Studies, Nos. 13–14), pp. 9–17. 1 “auf dem rechten Wege”; see O. Eissfeldt, “Ansetzung der Rollen nach paläographischen Kriterien”, in: Theologische Literaturzeitung 74 (1949), pp. 226– 228). 2 Die hebräischen Handschriften aus der Höhle. Stuttgart 1951. 3 ”[. . .] Papyri und Ostraka [. . .]. Wir haben gesehen, wie unsicher ihre Datierung ist. Über den Versuch [. . .] mit Hilfe dieser nicht sicher datierbaren Dokumente zu einem objektiven Datum für die in der Höhle gefundenen Rollen zu kommen, hat z. B. Eissfeldt so ausführlich berichtet, daß ich darauf verweisen kann. Ich bin davon überzeugt, daß Lacheman {“A Matter of Method in Hebrew Palaeography”, in: Jewish Quarterly Review 40 (1949/50), pp. 15–39} recht hat, wenn er darauf hinweist, daß man auf diesem Wege zu einem sichern Resultat nicht kommen kann, jedenfalls einstweilen nicht” (p. 22). 4 Since writing these words I have come across a passage in Einar Brønno’s article “Zu den Theorien Paul Kahles von der Entstehung der tiberischen Grammatik” (Zeitschrift der Deutschen Morgenländischen Gesellschaft 100 [1950], pp. 521–565, which is a reply to Kahle’s criticism of his book Studien über hebräische Morphologie und Vokalismus auf Grund der Mercatischen

000180

Birnbaum 01.04.11 18:17

180

[ 39 ]

Accepting Dupont-Sommer’s date for the composition of the Habakkuk Commentary – shortly before 40 B. C. E. – Kahle expresses the view, based on textual evidence, that the actual writing of the Scroll must have taken place before the destruction of Jerusalem (p. 59), i. e., according to him our copy originated somewhere within a period of approximately 110 years. As to Isaiah Scroll A – which would naturally be of very much greater importance for his particular studies – he allows himself a margin double as great, i. e., practically two and a half centuries. His terminus a quo, based on differences in text and spelling, is the Maccabaean Revolt, and his terminus ad quem the Fall of the Temple (p. 76). A somewhat more precise date would be of great value for his own particular field of research. As he is not able by means of the instruments at his disposal – that is, in the main, textual evidence – to arrive at one, it is a pity that he refuses the help of palaeography.

The Weight of Evidence Kahle brackets the names of Trever and myself with that of Albright, thus reducing the weight of the palaeographical results arrived at by the three of us to one piece of evidence: ‘About the attempt made by Fragmente der Zweiten Kolumne der Hexapla des Origenes, Leipzig 1943). Kahle had said: “Brønno quite arbitrarily decides not only which of the Greek vowels have to be pronounced as short or long, but also where we have to pronounce them as indefinite vowels (Murmelvokale ), corresponding to Tiberian Shwa!” (The Cairo Geniza. London 1947, p. 232). To this Brønno rejoined: ‘This statement of Kahle’s does not accord with the facts. If an authority like Kahle writes in such a way, then all those whose knowledge of my book is confined to Kahle’s mention of it, must come away with the impression that I have made arbitrary assertions. The position, however, is that I have brought detailed evidence in support of all I have said, and this cannot be demolished by means of an exclamation mark’. (“Diese Äußerung Kahles stimmt nicht mit den Tatsachen überein. Wenn eine Autorität wie Kahle so schreibt, muß jeder, der das Buch nur aus der Erwähnung Kahles kennt, die Auffassung erhalten, es handele sich um willkürliche Behauptungen meinerseits, während in Wirklichkeit von einer ausführlichen Dokumentation, die sich jedenfalls nicht durch ein Ausrufungszeichen vernichten läßt, die Rede ist” [loc. cit., p. 534].) Another passage: ‘Instead of checking up carefully whether my results regarding Codex B are correct, Kahle tries to turn down the whole piece of work as worthless, by polemizing against views which I never put forward’. (“Anstatt nachzuprüfen, ob meine Resultate hinsichtlich des Codex B zutreffen, versucht Kahle die ganze Arbeit als nichtig abzulehnen, indem er gegen Gesichtspunkte polemisiert, die ich niemals vorgebracht habe” [ibid., p. 564].)

000181

Birnbaum 01.04.11 18:17

The Qumran Scrolls and Palaeography

181

Albright, and, following him, Trever and Birnbaum, to reach an objective date [. . .]’.5 This cannot be done, however. That my examination was quite independent of those made by the other two scholars should at least have been clear to him from the fact that I did not utilise the Nash Papyrus. But he does not appear to have realised it even after I actually pointed it out.6 If he had, the circumstance that, independently of Albright and Trever, I arrived at similar results should have given him food for thought: Could such a thing have been a mere coincidence?

Reliability of Basis Kahle’s main criticism of the dating of the Cave Scrolls by way of palaeography is that the documents utilized for comparison are not ‘datable with certainty’. Of Albright’s material he says ‘that none of the texts utilised [by him] is datable with certainty’,7 and of Trever’s discussion, that ‘on the whole, [it] follows in the footsteps of Albright; but it betrays that he has not sufficient experience of his own in the field of palaeography’.8 As to the criticism contained in the last half of this sentence, one wonders whether a scholar who in the course of his work has handled many MSS . is ipso facto in a position to pass judgment on palaeographical work. Experience with MSS . is not a substitute for specialised, systematic palaeographical research. It is significant that in his whole book Kahle does not base a single argument on the forms of letters, which after all, constitute the subject of palaeography proper. Finally he says: ‘But although Dr. Birnbaum’s attempts at reaching a more precise dating of the Isaiah and Habakkuk Scrolls are made with greater knowledge of detail, one must remain suspicious because, of the parallel texts utilised, not a single one is datable with certainty’.9 From 5

6 7 8

9

“Über den Versuch, den Albright, und in seinem Gefolge Trever und Birnbaum unternommen haben [. . .] zu einem objektiven Datum [. . .] zu kommen [. . .]” (p. 22). BASOR 115 (1949), p. 20 {= Chapter 37 in the present volume, p. 153}. “daß keiner der beigezogenen Texte mit Sicherheit zu datieren ist” (p. 5). “Trevers Ausführungen bewegen sich im großen und ganzen auf den Bahnen Albrights; aber man merkt es ihnen an, daß er keine genügende eigene Erfahrung auf dem Gebiete der Paläographie hat” (p. 6). “Aber auch gegenüber den mit größerer Detailkenntnis ausgeführten Versuchen Dr. Birnbaums, auf paläographischem Wege zu einer genaueren Datierung der

000182

182

Birnbaum 01.04.11 18:17

[ 39 ]

which it would follow that the results arrived at on such a basis are useless. The documents utilised by me are, however, ‘datable with certainty’. To see that, it is only necessary to look up what has been published about them.10 It is generally accepted that the Egyptian papyri and ostraca which I utilised in my comparison are of the 3rd cent. B. C. E., the Tobias inscription of between 183 and 175 B. C. E. (Vincent: c. 260 B. C. E.), the Gezer Boundaries of the first third of the 1st cent. B. C. E., the Ossuaries, including the Uzziah Tablet, of the hundred years ending with the fall of the Temple, some of them of the late part of the 1st cent. B. C. E., some of the early decades of the 1st cent. C. E., and the Queen Helena Graffito of 50– 60 C. E. This is the evidence with which Kahle seems to have omitted to acquaint himself. If, however, he did study it and arrived at the conclusion that the results of Clermont-Ganneau, Cowley, Lidzbarski, Littmann, Sachau, Albright, Dussaud, Sukenik, Vincent, etc. are wrong or very doubtful, he should have explained why he thought so and made some attempt at disproving them. So much as to the ‘certainty’ of the dates. And as to their precision: Is a date which is precise to within a decade (Helena) not very precise? Are not those dates which go to within a third of a century (Gezer, some ossuaries) or half a century (some ossuaries) rather precise? And cannot even those documents (Egyptian papyri and ostraca, some ossuaries) which are datable only to within a century be nonetheless used to very great advantage? How does Kahle explain the fact that I arrived at dates that go very well with those he himself endorses or suggests on other grounds? After all, my results were not influenced by extra-palaeographical evidence. They were printed before the archaeologists had found the jar fragments and announced that these were of the late Hellenistic period, and before the various theories based on internal evidence and speaking for a pre-Christian date had been published.

10

Jesaja-Rolle und der Habakkuk-Rolle zu kommen, muß man mißtrauisch bleiben, weil von den herangezogenen Paralleltexten kein einziger mit Sicherheit zu datieren ist” (p. 6). See my bibliography in JBL 68 (1949), pp. 161–162 {= Chapter 41 in the present volume, pp. 197–198}.

000183

Birnbaum 01.04.11 18:17

The Qumran Scrolls and Palaeography

183

The Square Reconstruction Discussing the Leviticus Fragments, Kahle speaks of my ‘attempt at reconstructing the column but in Square script (!)’.11 His exclamation mark seems to imply: to use Square letters in reconstructing a PalaeoHebrew text is methodologically wrong. He fails to tell his readers what I expressly said about this point: “The reconstruction has to be done in the original alphabet, of course”.12 In the face of these unambiguous words of mine it is hard to understand how he can imply that I failed to make my own reconstruction in the original characters. The printed text is obviously not my original reconstruction in PalaeoHebrew characters but a transliteration of it, made for the convenience of the readers, many of whom would not have great fluency in reading Palaeo-Hebrew script.13 Kahle himself substitutes Square letters for Samaritan ones.14

Palaeo-Hebrew as a Display Script After saying that in the 1st cent. B. C. E. Bible MSS . were written in Square script, he continues: ‘But there can be no doubt that the OldHebrew script remained [after it had been discarded for the writing of Tora scrolls] also among the Jews as a kind of display script’.15 Kahle’s words ‘There can be no doubt’ show his statement to be based on an inference and not on actual evidence. Is he by any chance referring to the script on the coins? Whatever might be the reason for the lettering used on them, there is a complete absence of evidence to suggest that it was employed also for ornamental purposes in MSS . written in Square letters. And MSS . are what we are concerned with here.16 11 12 13

14 15 16

“Rekonstruktionsversuch der Kolumne, aber in Quadratschrift (!)” (p. 62). BASOR 118 (1950), p. 22 {= Chapter 33 in the present volume, p. 81}. Perhaps it would not have occurred to Kahle to make this criticism if my reconstruction had been published in printed type. As a matter of fact, the use of a drawing was an economy measure on the part of the Editor, owing to the expense involved in Hebrew type-setting. Op. cit., p. 20. “Es ist aber außer Frage, daß die althebräische Schrift auch bei den Juden noch lange Zeit als eine Art Antiqua im Gebrauche war [. . .]” (p. 20). The use of the Palaeo-Hebrew letters for the Tetragrammaton in the Habakkuk Scroll and for the word l in the Hodayot Scroll and El Fragment disproves the display script assertion.

000184

184

Birnbaum 01.04.11 18:17

[ 39 ]

Palaeo-Hebrew Bible Scrolls in the Christian Era The last-quoted sentence goes on: ‘and that Bible MSS . in Old-Hebrew script existed among Jews still in the Christian era’.17 To prove this he quotes mishna 5 of Yadayim, ch. 4. Here we read: ‘The Aramaic [texts] in Ezra and in Daniel make the hands impure. Aramaic [Biblical text] which was written in Hebrew [language], or Hebrew [Biblical text] which was written in Aramaic [language], or in [Palaeo-]Hebrew script, do not make the hands impure. Impurity is caused only by writing [a Bible MS .] in Assyrian [script], on skin, and with ink’.18 He believes this mishna furnishes him with an actual date: ‘This statement, coming from the 2nd cent. C. E., shows that at the time the Jews had, apart from Bible MSS . in Square script [. . .] also Bible MSS . in Old-Hebrew script’.19 The chronological inference is, however, wrong. The fact that the Mishna was codified in the 2nd cent. does not mean that every mishna in it belongs to this period – on the contrary, only a few would. Various collections of mishnas go back to the 1st cent. C. E., the Mishna Rishona to the 1st cent. B. C. E. And the mishnas themselves antedate the collections, of course. On the basis of his fallacious inference Kahle goes on to say that the Leviticus Fragments need not necessarily be several centuries earlier than the Christian era, and might therefore be contemporaneous with the fragments in Square script.20 In order to reach that conclusion, his reasoning would presumably have been somewhat on these lines: Our mishna 5 shows that Bible scrolls in Palaeo-Hebrew script were still in existence in the 2nd cent. C. E. It may therefore be assumed that these scrolls were written not 17 18

19

20

“und daß Bibelhandschriften in althebräischer Schrift bei den Juden noch in nachchristlicher Zeit vorhanden gewesen sind” (p. 20). Trgwm sˇb zr wsˇbdny l mtm t hydym trgwm sˇktbw bryt w bryt sˇktbw trgwm wktb bry ynw mtm t ˙hydym l wlm ynw mtm d sˇyktbnw sˇwryt l h wr ˙ ˙ wbdyw. “Diese aus dem zweiten nachchristlichen Jahrhundert stammende Feststellung zeigt, daß die Juden damals außer Bibelhandschriften in Quadratschrift [. . .] auch solche in althebräischer Schrift [. . .] gehabt haben” (p. 21). “Daraus ergibt sich [. . .], daß von Juden stammende Fragmente der Tora in alter hebräischer Schrift nicht unbedingt mehrere Jahrhunderte vor unserer Zeitrechnung geschrieben sein müssen [. . .]. Sie mögen älter sein als die in Quadratschrift geschriebenen Fragmente; aber ein Zwang zu dieser Annahme besteht nicht. Wir müssen mit der Möglichkeit rechnen, daß sie aus derselben Zeit stammen wie jene” (p. 21).

000185

Birnbaum 01.04.11 18:17

The Qumran Scrolls and Palaeography

185

very long before that time, i. e., in the centuries immediately preceding the 2nd cent. C. E. In other words, the Jewish scribes between the 2nd cent. B. C. E. and the 1st cent. C. E. were producing not only MSS . in Square script but also some in Palaeo-Hebrew. That being the period of the Cave Scrolls, the presence of the Leviticus Fragments among the Square script material would thus be accounted for. Hence there is no need to ascribe the Leviticus Fragments to a much earlier century. The assumption on which all this hinges is, however, an arbitrary one. Even if some Palaeo-Hebrew scrolls did survive into the 2nd cent. C. E. it would be no more reasonable to ascribe them to the 2nd cent. B. C. E. – 1st cent. C. E. (which would be in figures the time indicated by Kahle in general terms) than to the 4th or 5th cent. B. C. E. On the contrary, a very simple reason militates against that assumption. At a time when, in Kahle’s own words about Yadayim 4,5, ‘only texts written in Square script were considered sacred’ (p. 21), new Tora scrolls would not have been written in a discarded script which would ipso facto have rendered them unusable. That the 2nd cent. B. C. E. already belonged to the time when only Bible scrolls in Square letters were considered sacred seems to be disputed by no one. The Cave Scrolls, according to Kahle, date from between the 2nd cent. B. C. E. and the 1st cent. C. E. Hence it is impossible that during this period any Bible scrolls could have been newly written in PalaeoHebrew characters, and the Leviticus Fragments cannot therefore be contemporaneous with the Cave Scrolls. They must be older. It is not even possible to place the Leviticus Fragments in the Cave Scroll period by ascribing them to the Samaritans of that time. For – quite apart from the question as to whether the letter forms in the Fragments could belong to that period – their text, as Kahle himself has demonstrated, is not the one known to us as Samaritan.21 We have no reason to doubt that from the time when, centuries before the Christian era, Square became the script for Jewish Tora scrolls, no more were written in the discarded Palaeo-Hebrew alphabet. The historical argument has been used here22 only in order to show that Kahle has applied it incorrectly. But the safest way of showing that the Leviticus Fragments cannot have been written at the same time as the Cave Scrolls is by means of palaeography. We need only glance at 21

22

A kind fate has given us in the few words on our scraps a passage where the Jewish and Samaritan texts differ. The Leviticus Fragments have the Jewish form tkm h rs while the Samaritans read here h rs tkm. ˙ at all in establishing the date in my ˙ article on the Leviticus It was not used Fragments.

000186

186

Birnbaum 01.04.11 18:17

[ 39 ]

fig. 1 to see how much younger the forms in the Scrolls are than those in the Leviticus Fragments. The detailed discussion in BASOR 23 and the charts in Vetus Testamentum 24 demonstrate that the time interval between the two sets is approximately four centuries.

Fig. 1. 25 The Palaeo-Hebrew script in the Square script Scrolls contrasted with that of the Leviticus Fragments. Column a, line 1: Hodayot Scroll. 26 Lines 2–3: Habakkuk Scroll. 27 Line 4: Hodayot Scroll. 28 Line 5: El Fragment. 29 Column b, lines 1–5: Leviticus Fragments. 30 23 24 25 26 27 28 29 30

118, pp. 25–27 {above Chapter 33, pp. 85–88}. 1 (1951), pp. 107–108 {above Chapter 38, pp. 174–176}. The following notes (26–30) contain information as to where facsimiles from the MSS . utilized in the illustrations can be found. E. L. Sukenik, Mgylwt Gnwzwt {Megillot Genuzot}. Vol. 2, Jerusalem 1950, pl. 7. The Dead Sea Scrolls of St. Mark’s Monastery, ed. by M. Burrows. Vol. 1, New Haven 1950, coll. 6, 10, 11. See above, n. 26. Illustrated London News, October 1, 1950, p. 494, fig. 6. – J. van der Ploeg, in: Jaarbericht Ex Oriente Lux 11 (1949/50), fig. 12. R. de Vaux, “La Grotte des manuscrits he´breux”, in: Revue biblique 56 (1949), pp. 586– 609, pl. 18. – S. A. Birnbaum, in: BASOR 118, p. 21 {= Chapter 33, p. 80}.

000187

Birnbaum 01.04.11 18:17

The Qumran Scrolls and Palaeography

187

‘Strange Precision’ In a sort of summary of my article on the Leviticus Fragments, Kahle quotes what he calls my conclusion: ‘The Leviticus Fragments were part of a scroll written in a beautiful professional hand, about the second half of the fifth century B. C. E.’. This moves him to exclaim, ‘We ask ourselves in some surprise how he can arrive at this strangely precise date’.31 The explanation is to be found, he thinks, in this passage of mine: ‘That is the period when the great change in the script of the Tora took place. Thus the scroll of which the Leviticus Fragments formed a part belonged to the final period when Palaeo-Hebrew was in use’. In other words, what Kahle suggests is that my date for the Leviticus Fragments is based not on palaeography, but on an historical argument. This is more than strange in the light of pp. 25–27 of my article, where my examination32 of the script culminates in a conclusion as to its date – the result being reached by three lines of approach (paragraphs 2 to 4 of p. 27). To his explanation he adds the following remark: ‘All the learning of the article has essentially the purpose of proving that a certain Talmudic opinion, as he understands it, is correct’.33 These words imply that I had my result ready before starting to work, and ‘cooked’ my long, detailed and intricate palaeographical examination to fit a predetermined date!34

31 32

33 34

“Man fragt sich etwas erstaunt, wie er zu dieser seltsam genauen Datierung kommen kann” (p. 62). ‘which looks very learned’ (“sehr gelehrt aussehend”), he says (p. 62). Cf. with this expression what he writes about Zeitlin’s ‘tone’, ‘presumptuous tone’ (“überheblich”), the ‘high and mighty way’ (“von oben herab”), Zeitlin speaks to other scholars (pp. 8 and 10). Cf. also Kahle’s quotation marks for Wissenschaft in n. 34 below. “Die ganze Gelehrsamkeit des Artikels dient also im wesentlichen dazu, eine talmudische Anschauung, wie er sie versteht, als korrekt zu erweisen” (p. 63). At the end of this section, Kahle says that in view of such “Wissenschaft” on my part, it would suffice to point to Lacheman’s criticism of ‘analogous attempts at dating’. (“Es genügt gegenüber derartiger ‘Wissenschaft’ auf die Kritik hinzuweisen, die Ernest Rene´ Lacheman gegenüber analogen Datierungsversuchen geübt hat in seinem Artikel ‘A Matter of Method in Hebrew Palaeography’” {see above, n. 3}).

000188

188

Birnbaum 01.04.11 18:17

[ 39 ]

Codices Kahle constructs a terminus a quo for the depositing of the Scrolls in the Cave from the fact that some of the papyrus fragments discovered there had writing on both sides. He declares: ‘Texts written on both sides are parts of codices’35 – and in this simple way gets his terminus. For ‘manuscripts in codex form in the East can hardly be shown to have existed before the 2nd cent. C. E.’36 It is not correct to say that writing on both sides occurs only in codices. Long before the time of codices the backs of discarded scrolls were written upon. If it could be shown that the fragments in question contain a consecutive text on both sides, written by one and the same hand, and in the same direction, then it would be possible to say the fragments were parts of codices. What, however, are the facts? There are about 580 skin fragments from the Cave, and not one of them has writing on both sides. There are about 30 papyrus fragments, six of which have writing on both sides: Three are indistinct; the writing of the fourth is not identical on both sides: one side Greek, the other indistinct; the writing of the fifth is not identical on both sides, its script is indeterminate. These cases, although not conclusive, by no means favour the codex theory. The sixth fragment is definitely not from a codex because the writing runs in opposite directions on the two sides. According to this evidence,37 the likelihood that these are the remains of papyrus codices seems to be exceedingly small.

35 36 37

“Doppelseitig beschriebene Texte sind Teile von Codices [. . .]” (p. 58). “Handschriften in Codexform sind vor dem 2. nachchristlichen Jahrhundert im Osten kaum nachzuweisen” (ibid.). R. de Vaux (above, n. 30), p. 597: “Nous en avons recueilli environ six cents.” There are also some fragments in the Hebrew University. “Nous avons recueilli une trentaine de fragments sur papyrus; six sont e´crits des deux coˆte´s: l’un a de l’he´breu sur les deux faces, mais l’e´criture du ‘revers’ est inverse´e; ce n’est donc suˆrement pas un codex. Sur deux autres, la graphie n’est pas la meˆme des deux coˆte´s; l’un est une e´criture inde´termine´e, l’autre est d’un coˆte´ en grec, de l’autre coˆte´ indistinct. Les trois fragments restant sont indistincts [. . .]. Les fragments sont d’ailleurs minuscules et ne contiennent que quelques lettres”. R. de Vaux, in a letter to me, of February, 1951.

000189

Birnbaum 01.04.11 18:17

The Qumran Scrolls and Palaeography

189

Parchment In the chapter on the Leviticus Fragments Kahle points to what he imagines is an inherent contradiction between my date and my description of the Fragments as parchment: how could they be so old as I say, seeing that parchment is generally known to have been invented only in the Christian era? He asserts that I three times stress the fact of its being parchment, and underlines this by his manner of formulation.38 I have nowhere in my article discussed the question of the writing material (beyond mentioning it), much less stressed what the material is. Kahle’s words convey a quite misleading picture. As to the parchment question: Kahle did not know our terminus AD quem for the invention of parchment is 195 B. C. E., from which we have a dated document.39 The quality of this piece is practically as good as that of any later first class Western vellum. (“[. . .] une membrane si fine et si lisse qu’elle le ce`de de peu au velin des manuscrits les plus soigne´s”).40 This fact would seem to suggest, or at least to make it appear not unlikely, that the development culminating in such a quality had been going on a considerable time before 195 B. C. E.41 The whole question is to a considerable extent one of terminology. As far as I know the various ancient kinds of writing skin have not yet been systematically and comprehensively examined and classified. As long as that is so, it might be preferable to use the general term skin (as I did, e. g., in my article in Vetus Testamentum). The material of Isaiah Scroll A was thus described: “somewhat coarse parchment, or carefully prepared skins approaching the refinement of parchment. (Among the scrolls there is a considerable difference in the degree of refinement and quality of the skins making it somewhat 38

39 40 41

“Daß diese so alt sein sollenden Fragmente auf Pergament geschrieben sind, das bekanntlich erst in nachchristlicher Zeit erfunden ist, wird von Birnbaum in seinem Artikel dreimal betont” (p. 63). F. Cumont, “Le plus ancien parchemin grec”, in: Revue de philologie 48 (1924), p. 97 ff., idem, Fouilles de Doura-Europos. Paris 1926, p. 281 ff. Idem, Fouilles, p. 283. The pre-Christian origin of parchment disposes also of the basis of Kahle’s argument that the 4th cent. C. E. is the terminus ad quem for the depositing of the Scrolls in the Cave: ‘If the manuscripts had been deposited in the 4th cent. or later we would have to expect to find in the Cave, texts written on parchment’. (“Wären die Handschriften im 4. Jahrhundert oder später in der Höhle geborgen worden, so müßten wir erwarten, auf Pergament geschriebene Texte in der Höhle zu finden” p. 58).

000190

190

Birnbaum 01.04.11 18:17

[ 39 ]

difficult to classify the material)”.42 That of the Habakkuk Scroll is given as “soft leather”, “the hair side, on which the writing was done, was carefully dressed and is smooth to the touch”.43 The Leviticus Fragments, too, are written on the hair side.44 Hence, in the terminology of the Talmud,45 the material of the Cave MSS . is wr ‘hide, skin’ or gwyl (? ‘primitive kind of parchment’).46

42 43 44 45 46

Dead Sea Scrolls, vol. 1 (above, n. 27), p. XIV . Ibid., p. XX . R. de Vaux (above, n. 30), p. 597. J. Meg. ed. Krotoschin, fol. 71d. Since writing these lines I have come across two learned articles by G. Zuntz, devoted to rectifying Kahle’s statements about parchment and codices, and the chronological conclusions he drew from them regarding the Cave Scrolls. See Theologische Literaturzeitung 76 (1951), coll. 161–162 and coll. 533–534. – Cf. also the article coll. 533–536, where an expert like W. Schubart shows that the writing materials of the Cave MSS . cannot help with the dating. That must be done, he says, by means of palaeography and textual criticism.

000191

40.

Birnbaum 01.04.11 18:17

The Date of the Isaiah Scroll*

In the following lines an examination is made of the script of the Isaiah Scroll {1Q Isa}. Detailed comparison with documents hitherto known to us should make it possible to suggest a fairly reliable date. Our study is based on the reproduction of col. 33, published in BASOR 111 (1948), p. 6 {see HS 1 plate 81 (shown in part above in Chapter 34, fig. 12, p. 110) and HS 85 (here as fig. 1)}. Even a superficial comparison with the earliest specimens of the Square script hitherto available makes it clear that the Isaiah Scroll (Is) belongs to the same group. We have not sufficient material to decide whether there was a regional difference at that period between the script of Palestine and that of Egypt. Assuming that if there was it could not yet have been great, we shall utilize the writing of both countries. It does not require a detailed investigation to see that an earlier date than the third pre-Christian century is out of the question. It is equally obvious that the writing of Is is considerably older than anything that we have after the first Christian century. Thus we have limited the area requiring a detailed examination to about four centuries, the three final pre-Christian ones and the first Christian century. The following material has been used for comparison: the third century B. C. E. documents from Jewish colonists settled in Egypt (Eg),2 the Tobias Inscription ( Tob) {HS 80, 87 F(1)} of 180–175 B. C. E., the Gezer Boundary Stones (Gez) {HS 83, 87 F(2)} of 100– 63 B. C. E., the Ossuaries (Oss)3 of between the late first century B. C. E. and the first half of the first century C. E., the Uzziah Plaque (Uz) {HS 88, 87 F(7–8)} * First published in: Bulletin of the American Schools of Oriental Research {BASOR} 113 (1949), pp. 33–35. 1 {The Hebrew Scripts (HS) by Solomon A. Birnbaum, 2 vols., London 1954– 1957 / Leiden 1971.} 2 The Berlin Ostraca {Ägyptisches Museum}, P. 10678 {HS 155}, P. 10964 {HS 157}, P. 10974 {HS 156}, Weill Ostracon (Muse´e Guimet) {HS 161} and the Bodleian papyri Aram. a. 1 (P.), Aram. e. 2 (P.), and Heb. e. 121. (P.) {HS 159}. {Cf. also HS 158, 160, 162}. 3 Jerusalem, Palestine Archaeological Museum: 34. 7754; 36. 1867, 1868, 1871, 2177, 2181 {HS 95*}, 36. 2182 {HS 95}; Jerusalem, Pal. Arch. Mus.: Ossuaries from En Rogel {HS 96, 97, 98, 99}, Nahalat Ahˆım, University ˙ ˙Maria Ossuary; Hypogeum, Syrian Orphanage, Abu Ghosh; Kallon Ossuaries, British Museum: Nicanor Oss., Greek Colony, Zikhron Moshe; Dipinto {HS 100*}.

000192

192

Birnbaum 01.04.11 18:17

[ 40 ]

Fig. 1. Alphabetical chart drawn (by the author) from Isaiah Scroll A.

of 37 B. C. E. – 44 C. E., and the Queen Helena Inscription (Hel) {HS 100} of the second half of the first century C. E. Aleph: Only rarely does the Aleph of Is resemble the form used in Oss and Uz, where the left stroke begins at the top and is often straight. Generally the left stroke is nearer that of Eg, starting lower down – anywhere up to the middle of the main stroke – and curving towards the left. It is bigger than the corresponding stroke in Eg, reaching the bottom of the line. Beth: The top stroke is a curve as in Eg; in Oss it has been formalized into a top bar beginning with a small vertical stroke (sometimes this initial stroke is already missing). In Oss the base is, as a rule, produced towards the right. In Is there is no indication of this trend. The letter is of medium width as in Eg. In Oss it tends to become very wide. Gimel: The left stroke issues from the middle of the right one. This construction also occurs in Eg and Oss but in Is as well as in an Ossuary the strokes are curved. Daleth: The top of the letter is a narrow tick, corresponding to the narrow curve of Eg. A formalised top bar (cf. Beth) is characteristic of Oss. Here the letter is generally very wide. He: The right hand vertical is now and then slightly curved and occasionally starts just above the line ceiling. The horizontal is under

000193

Birnbaum 01.04.11 18:17

The Date of the Isaiah Scroll

193

the ceiling and rarely straight; as a rule it is an angular counterpart of the curve in Eg. The left vertical issues from the top, in the middle or near it. This form occurs in Tob and some Ossuaries. In Eg the left downstroke slants down from the right corner and the top stroke is often in a slanting position. On the other hand, while in the majority of Ossuaries and in Uz formalization of the top (cf. Beth) has taken place, there is no trace of it yet in Is. Nor is there any development towards the severing of the left vertical from the top stroke. However, in Uz three times out of five, on some Ossuaries and in Hel the open form is used. Waw: The top is an acute angle or a small and narrow curve as in Oss and Uz, while Tob has a right angle and a long top stroke corresponding to the wide top curve of Eg. Zayin is a plain vertical as in Eg, while in Gez and Oss a small top stroke is added, slanting from right to left and joining the vertical either at the top or near the top; in Uz this is formalized with the effect that the letter has become a symmetrical counterpart of Waw. Heth: The left end of the cross bar slides up to the top of the vertical. ˙ Slanting positions are also to be found in Eg; in Gez, Oss and Uz, however, the cross bar is always horizontal. On some Ossuaries and in Uz the top has been formalized, the cross bar having become a top bar. Teth is somewhat wider than in Eg and Tob but much narrower than in Oss and Uz. It is thus in an intermediate stage between Eg and Oss. While in Eg and Tob the left part is curved, it is here, as in Oss and Uz, nearly always straight. In Eg and Tob it is almost double the height of the right part, but in Is and on some Ossuaries it is often not much above the ceiling of the line. The fully developed, slanting, sometimes practically horizontal base characteristic of Uz and some Ossuaries, is however still absent from Is. Yodh is an acute angle, the right side of which often ends by curving very slightly towards the left. The letter is sometimes of line height, when it is as tall as Waw but yet clearly differentiated from it: the left and right strokes are slanting and of practically even weight. Waw, however, is a vertical, slightly curved stroke with a small curve or angle on top. The Yodh in Is is sometimes to be found in Eg. In Oss and Uz the right hand stroke is vertical, the one on the left is small and insignificant. The letter is clearly smaller than Waw; it has continued to remain so ever since. Thus the form in Is is earlier. Kaph starts with a curve or tick on the line ceiling and extends far below the bottom of the line. The same narrow form is used in Eg. In Uz, Hel and Oss the formalization of the letter to line height has caused it to acquire greater width. In Uz the top bar is also formalized. There is

000194

194

Birnbaum 01.04.11 18:17

[ 40 ]

no final Kaph in Is. At the end of a word the letter tends sometimes to increase a little in size. In Uz the final form consists of a formalized top bar and a long4 vertical downstroke. Lamedh is in two distinct parts, an upper and a lower one. In Eg this form is rare; as a rule the upper part gradually curves into the lower part, the letter thus becoming one wavy line. The lower part in Is is either a sharp curve or an acute angle. This type never occurs in Eg but is characteristic of Oss and Uz. The lower part is small, it does not reach the line base, which shows that it is not long, since the downstroke developed from a little tail. Mem: The fifth century top stroke is still in existence but it has often shrunken in size. The top right hand corner, usually round in Eg, is sharply pointed in Is. This tallies with the form in Gez, Oss, Uz and Hel. The left hand downstroke no longer passes through the top but issues from it in a downward direction, often starting very near the top right hand corner; it is sometimes slightly curved. In Gez, Oss and Uz we have a similar form but the left downstroke then always starts from the left end of the top. This is obviously a conventionalized, i. e., later form. (Still later, but already on some Ossuaries and in Hel, the top stroke and the slanting stroke were joined and the vertical beginning of the top stroke became independent). The right hand stroke turns sharply to the left at the bottom, forming a slightly slanting or almost horizontal short base, which is hardly shorter than in Gez, Oss, Uz and Hel; it is generally parallel with the left hand stroke. Final Mem: The left hand stroke still crosses the top, often also the base. The letter extends considerably below the line. In Eg there is an example where the letter is almost closed at the bottom but the downstroke is slanting, not parallel to the right hand stroke as here. In Oss the downstroke does not cross the horizontals, in Uz it is even severed from the top bar, which is, of course, a further step in its development.5 Nun consists of an almost vertical, sometimes slightly curved top stroke which is perpendicular to a {nowadays rarely used} horizontal base. In Eg there is not yet a base, in Oss the base is very long, so that the Is form takes up an intermediate position. Final Nun has the same long wavy form as in Eg and Oss. 4

5

The example in Is, line 14 {see line 11 above in Chapter 34, fig. 12 (= line 14 of HS plate 81)}, with a long vertical downstroke and no base, seems to belong to a re-inked and rewritten passage. In certain cases, final Mem is not used. This may be due to some orthographic rule. It seems unlikely that the final and non-final forms should have been employed indiscriminately in the final position.

000195

Birnbaum 01.04.11 18:17

The Date of the Isaiah Scroll

195

Samekh consists of a very curved, occasionally somewhat angular right hand stroke and a more or less straight left hand stroke (similar to the cursive forms of Heth in use up to this day). Eg has the identical ˙ form, whilst Oss has already the closed type, the downstrokes joining to form an acute angle. The top parts have coalesced into a bar. Ayin: The right hand stroke is produced downwards to the bottom of the line, a form not yet present in Eg; it occurs in Uz and Oss (in the latter side by side with forms where the right hand stroke extends well beneath the line). Of this further development there is as yet no trace in Is. Pe: A small stroke slants up to the right. At the line ceiling it turns downwards to form an acute angle. On reaching the bottom of the line it turns leftwards in the shape of a slightly slanting or almost horizontal base. In Eg the base is not yet developed and there is no specimen with a right or acute angle at the bottom. In Is we have the Oss form. The left hand stroke is sometimes broken or curved as in Uz. Final Pe: There is none. At the end of a word the letter tends occasionally perhaps to be just slightly bigger. Sadhe: The sometimes almost vertical downstroke passes well beneath the line and then turns to the left, forming a slanting, occasionally almost horizontal, medium sized base. In its upper half, the downstroke is met on the right by a large, often rather angular curve. The same form occurs in Eg and Oss. Final Sadhe: There is none. At the end of a word the letter is occasionally bigger. Qoph: The shaft either issues from the top stroke or is severed from it. This corresponds to Oss, whilst in Eg we have only the closed form. Resh: The top stroke is a narrow or medium wide curve as in Eg, while Gez has a narrow top bar and Oss a very wide one. Shin tallies with Eg and Oss but there is no example of the form, frequent in Oss, where the left hand stroke is produced downwards. Taw: The right hand stroke is vertical and fills the line height. The left vertical starts only just above the line ceiling and does not extend beyond the base, where it turns sharply to the left, forming a foot. Both verticals are straight. A top bar has not yet developed. ( The transition form, where the top of the left vertical and the top curve of the right hand part are linked together by a loop, occurs on some Ossuaries.) Thus the letter is like that of Gez, Oss, Uz and Hel, only rather less formalized. In Eg the left hand stroke is wavy and much longer, the one on the right very curved. To sum up: There are very few letters in Is which tally completely with their counterparts in the documents utilized for comparison. Hence

000196

196

Birnbaum 01.04.11 18:17

[ 40 ]

it is evident that Is was not written during the third century B. C. E. or after the middle of the first pre-Christian century. This leaves us with 150 years to limit down. Our investigations have revealed that the following forms are closer to Eg and Tob than to Gez, Oss, Uz, and Hel and indicate an early date: – Aleph, Beth, Daleth, He, Zayin, Heth, Yodh, Kaph, Mem, Samekh, ˙ Final Sadhe. Moreover we found Resh, absence of Final Kaph, Final Pe, that there are only half as many forms indicating a late date: – Gimel, Waw, Teth, Lamedh. Accordingly6 the first half of the second century B. C. E. suggests itself as the time when the Isaiah Cave Scroll was written.

6

Assuming that the other columns of the Scroll do not contain very different forms than those of col. 33.

000197

41.

Birnbaum 01.04.11 18:17

The Date of the Habakkuk Cave Scroll*

Each of the scrolls from the cave near the north end of the Dead Sea justifies the sensation the find created. The Habakkuk Scroll reveals the existence of a systematic commentary on a biblical book very long indeed before the ninth century, when, to our previous knowledge, the first works of this kind had been written. When was our Habakkuk Commentary composed? By determining when the Scroll was written down, through a detailed comparison of its script with that of other early documents hitherto available, we can at least establish the terminus ad quem. The following examination of the writing is based on the reproductions in The Biblical Archaeologist 11 (1948), p. 59 and BASOR 111 (1948), p. 15, showing the fragmentary columns 1–3. {Also reproduced on pp. 204 f. here, and in The Hebrew Scripts (HS ) by S. A. Birnbaum. 2 vols., London 1954– 1957 / Leiden 1971, plate 84. See also HS pl. 87.}1 No detailed examination is required for us to see that the script cannot be older than that of our third century B. C. E. documents from Egypt,2 and that on the other hand it is not as late as the Dura-Europos Fragment3 from before 256 C. E. Hence we can limit our comparison to the following material: the third century Egyptian documents (Eg), the Tobias Inscription ( Tob)4 of 183–175 B. C. E., the Gezer Boundary Stones (Gez)5 of 100– 63 B. C. E., the Ossuaries (Oss)6 of between the * First published in: Journal of Biblical Literature 68 (1949), pp. 161–168. 1 For abbreviations see the list at the end of this article. 2 BFMA . PSBA 37 (1915), p. 217. ESE vol. 2, p. 243, vol. 3, p. 22. E. Sachau, Aramäische Papyrus und Ostraka aus einer jüdischen Militär-Kolonie zu Elephantine. Leipzig 1911, p. 243. PSBA 29 (1907), p. 260. REJ 65 (1913), p. 16. {See also HS 155–162, and footnote 2 to Chapter 40 of this volume.} 3 The Excavations at Dura-Europos, vol. 6, ed. by M. I. Rostovtzeff [et. al.], New Haven 1936, p. 417 {HS 103*}. 4 D. Chwolson, Corpus Inscriptionum Hebraicarum, St. Petersburg / Leipzig 1882, p. 1. PJB 16 (1920), p. 33. JA se´r. 6, 10 (1867), p. 188. ESE vol. 3, p. 49. Syria, Publications of the Princeton University Archeological Expeditions. Leiden 1907 ff., Div. 3, Sect. A, p. 1. CRA se´r. 2, 1 (1865), p. 79. RA N. S. 10 (1864), p. 52. {See also HS 80, 87 F(1).} 5 BFMA . CRA se ´ r. 4, 26 (1898), p. 686; RB 8 (1899), p. 109; RAO vol. 3, p. 116; QS 31 (1899), p. 118. HNE , p. 484. R. Macalister, The Excavation of Gezer. London 1912, vol. 1, p. 37, vol. 3, pl. 10 f. JPOS 10 (1930), p. 19. {See also HS 83, 87 F(2).}

000198

198

Birnbaum 01.04.11 18:17

[ 41 ]

late first century B. C. E. and the first half of the first century C. E., the Uzziah Plaque (Uz)7 of 37 B. C. E. – 44 C. E., and the Queen Helena Inscription (Hel)8 of the second half of the first century C. E. Aleph: The right hand stroke generally starts above the ceiling of the line and reaches the main stroke about half way or lower down. The left hand stroke starts near the top and is more or less curved. It is of varying size and only rarely long. The letter is more akin to Eg than to Oss, where the left hand stroke starts on top of the main stroke and is practically always straight. Beth: The top part is a rather narrow tick from the end of which a slightly slanting and curved downstroke issues forth. At the bottom of the line this stroke meets a horizontal or slightly slanting base which is considerably wider than the top. In place of the tick Eg has a curve, which as a rule does not end on the top but runs into the downstroke. This feature is present also in Tob which, being lapidary, has a rectangular tick. In Oss however, the top region has been formalized into a top bar situated on the ceiling of the line, while the horizontal part of the tick or the bottom of the curve was formerly below the ceiling level. Accordingly, the small downstroke which in Oss represents the original initial downstroke of the tick or the beginning of the curve, is now above the ceiling. In Oss the base projects to the right. This is evidently a further development of the pointed corner of the Habakkuk Scroll 6

7 8

BFMA. Q 1,1 (1921), p. 57. RA N. S. 25 (1873), p. 398. RA N. S. 36 (1878), p. 305. RA Ser. 3, 1 (1883), p. 257. Ch. Clermont-Ganneau, Archaeological Researches in Palestine. London 1899, vol. 1, p. 381. RAO vol. 5, p. 334. RAO vol. 6, p. 211. RAO vol. 8, p. 59. MPJ , nos. 16, 51. S 4 (1923), p. 241. SBB 1885, p. 669. OLZ 15 (1912), p. 529. S. Klein, Jüdisch-palästinisches Corpus Inscriptionum. Wien 1920. JPOS 9 (1929), p. 100. ESE vol. 1, pp. 186, 312; vol. 2, pp. 72, 191; vol. 3, pp. 49, 302. HNE, p. 485. Macalister (see above, n. 5), vol. 1, pp. 347, 382. QS 63 (1931), p. 171. Q 1,2 (1925), p. 71. BBSAJ 5 (1924), p. 56. BAMP 1 (1867), p. 21. RB 34 (1925), p. 253. RB 1 (1892), p. 262. JAOS 28 (1907), p. 355. ZDPV 4 (1882), p. 9. JPOS 4 (1924), p. 171. JPOS 8 (1928), p. 113. JPOS 9 (1929), p. 45. JPOS 12 (1932), p. 25. Sep¯er hayyobel lapprofesor S. Krauss {Festschrift Samuel Krauss. Jerusalem 1936/37},¯ p. 87. PEQ 69 (1937), p. 126. Sep¯er Zikkaron le Asˇer Gulak uliSˇemu el Klein {Memorial Volume A. G. and S. K., Jerusalem 1942}, p. 133. RB 11 (1902), p. 276. RB 43 (1934), p. 564. JPOS 9 (1929), p. 41. {See also footnote 3 to Chapter 40 of this volume.} BASOR 44 (1931), p. 8. T 2 (1930), p. 288; p. 295. QS 63 (1931), p. 217; 64 (1932), p. 106. {See also HS 88, 87 F(7–8).} BFMA. CIS , no. 156. MPJ, no. 28. A. de Longperier, Muse´e Napole´on III , Choix des monuments. Paris 1868–1874. JA se´r. 6, 6 (1865), p. 550. {See also HS 100.}

000199

Birnbaum 01.04.11 18:17

The Date of the Habakkuk Cave Scroll

199

(Hab), where the downstroke curves to the left, i. e., its end swings to the right. In Eg the downstroke curves into the base and in Tob we have the corresponding lapidary form, a right angle. Gimel: The left stroke issues from the middle of the right one, as in Oss and some of Eg. There is no trace of that other Eg form, where the left stroke issues from the top of the right one. The letter is written with one lift of the pen, i. e., after having arrived at the line bottom the pen again moves upwards till it reaches the middle of the downstroke, then to make a left hand turn. The effect of this movement is to cause the strokes to be slightly curved. Curves occur in Oss, and the little upstroke in Gez may perhaps be a lapidary reflection of a handwritten curve. Daleth: The top part is a narrow tick, joining the right hand vertical not on top but somewhat lower down. This corresponds to the Eg top part except that there we have a curve instead of a tick. In Oss and Uz however, we have a formalized top bar (see Beth) and the letter is much wider than in Hab and Eg. Heˆ: The top stroke, a black triangle, reaches the right hand vertical somewhat below the line ceiling. Hab shares this feature with Eg, some Ossuaries and Hel, but in Eg the stroke is occasionally still in a slanting position. The formalization of the top, characteristic of many Ossuaries and Uz, has not yet been achieved. The left vertical is almost under the left end of the top stroke as in Oss and Uz, while in Tob it issues from the middle of that stroke; in Eg it slants down from the right hand top corner. There is no trace yet of the open form of the letter (through the severing of the left vertical from the top stroke). It occurs however, on some Ossuaries, in Hel and three times out of five, in Uz. Waw: The top is an acute angle as in Oss and Uz. The downstroke is often rather short, slightly slanting, sometimes slightly curved. Tob has a right angle, a long top stroke (with a small initial stroke, evidently based by ‘false analogy’ on the formalized top bar of Beth) corresponding to the wide top curve of Eg. Zayin is a vertical of line height, its top slightly thickened to the right. This is the form of Uz and corresponds also to the forms on some Ossuaries where the top is not thickened but slanting. In Gez and on other Ossuaries, however, this little slanting stroke joins the vertical somewhat further down. In Eg the letter is a plain vertical, sometimes of less than line height. Heth: The cross stroke, always horizontal, is rather high up but, as in Eg,˙ Gez and on some Ossuaries, not yet formalized into a top bar, like that of other Ossuaries and Uz.

000200

200

Birnbaum 01.04.11 18:17

[ 41 ]

Teth: The right hand part consists of an upper semi-circle, from whose right end a straight stroke issues forth and slants down to the line bottom. Here it is met by the left hand part, a vertical or very slightly slanting stroke of line height. The shape of this letter is very unlike that of Eg and Tob where it resembles a figure 6; its left stroke is very curved and almost double the height of the right part. Oss and Uz, however, tally with Hab, except that there the right hand downstroke is too slanting to be called a base. In all three of them the letter is wide, unlike the narrow forms of Eg and Tob. Yodh is an acute angle, the right arm of which slants rightwards, only occasionally becoming curved. The letter is not small; it is sometimes even of line height, like Waw. It differs from it in that the angle is distinctly maintained, while the top of Waw is more of a head. But often – or, at least, so it appears in the reproduction – the two letters are indistinguishable. The type with the short-armed angle is to be found also in Eg, while in Oss and Uz the right hand stroke is vertical, and the one on the left much smaller. In Oss, Yodh is considerably smaller than Waw. Kaph: The top part is a small and shallow curve in a slightly slanting position. The vertical which runs down from its right end is somewhat curved. At the line bottom it turns to the left, forming a medium sized, slightly slanting base. In Eg the base is considerably below the line bottom; Hab tallies with Oss, Uz and Hel. The decrease in length has made the letter wider than in Eg, but the shortness of the top stroke still gives it a rather narrow appearance. In the two fragments published there does not happen to be a Kaph at the end of a word so that we do not know yet whether there was a Final Kaph. Lamedh is in three parts: a very tall slanting downstroke above the line ceiling, its end extending slightly beneath. From here a small stroke slants upwards to the ceiling, whence another small stroke or shallow curve slants to the left, stopping in the middle of the line height, which shows that it is not long since it developed from a little tail. The top stroke sometimes starts with a loop. In Eg the letter is a wavy line, the three parts gradually curving into each other; letters consisting of two parts are also beginning to appear. The Hab type is characteristic of Oss, Uz and Hel. Mem: The top part is a tick. From its point a straight, occasionally curved, stroke slants to the middle of the line height. From its right end a straight, rarely a wavy, stroke slants to the line bottom. Here it turns to the left forming a slanting, sometimes horizontal, base of medium length. In Eg the top stroke is wider and still a wavy line. The top right hand corner is therefore round, while in Hab it is sharply pointed. In Eg

000201

Birnbaum 01.04.11 18:17

The Date of the Habakkuk Cave Scroll

201

the left hand downstroke passes through the top, which it never does in Hab. The Hab Mem tallies with Gez, most Ossuaries and Uz. Final Mem: The top tick is the same as in Mem. The downstroke runs far beneath the line bottom finally turning to the left – the corner is usually rounded – forming a slightly slanting base. The left hand downstroke starts with a triangle on the left of the tick and runs towards the left end of the base, sometimes without reaching it. Occasionally the vertical part of the tick and the downstroke under the triangle happen to be in a line so that an impression of the earlier form is created where a single stroke passes through the top. Owing to its length, the letter has a somewhat narrow appearance. A corresponding form occurs in Oss, whilst in Eg the left hand downstroke slants to the left, thus leaving the letter wide open. There is also the difference in the top part (see above and in Mem). Nuˆn consists of an almost vertical, sometimes slightly curved top stroke which is perpendicular to a horizontal base. In Eg there is not yet a base, in Oss the base is very long, so that the Hab form takes up an intermediate position. Final Nuˆn is an S-shaped line. The top curve is rather small. The lower part, which extends a full line height below the line bottom, is practically straight. Both Eg and Oss have the long and wavy type. Samekh: At the right end of an horizontal top stroke it runs downwards in the form of a short vertical, then to slant leftwards till it reaches the line bottom. Here it is joined or nearly joined, by a vertical which started by crossing the left end of the top stroke. The construction in Eg is similar but the letter is still wide open, and the left vertical sometimes does not pass through the top stroke but issues from it. The Samekh of Oss is completely closed. Ayin: The right hand stroke is produced downwards but reaches the line bottom only occasionally. This form is not present in Eg. It is used on some Ossuaries; on others the downstroke extends well beneath the line. Pe: A small stroke slants up to the right. At the line ceiling it turns downwards to form an acute angle. On reaching the bottom of the line it turns leftwards in the shape of a slightly slanting or almost horizontal base. In Eg the base is not yet developed and there is no specimen with a right or acute angle at the bottom. In Hab we have the Oss form. The left hand stroke is sometimes broken or curved as in Uz. Final Pe: The right hand stroke is a wide curve, extending a full line height below the line bottom. The left hand stroke is practically straight; at its end there is a tiny bend to the right. The top of the letter is an acute angle. Eg and Oss have similar types.

000202

202

Birnbaum 01.04.11 18:17

[ 41 ]

Sadhe: A slightly slanting downstroke extends to just below the line bottom and then turns to the left, forming a horizontal, occasionally slightly slanting, medium sized base. In its upper half the downstroke is met on the right by an obtuse angle or straight stroke. The same form is employed in Eg and Oss. Final Sadhe has the same top part but the downstroke does not turn into a base, extending instead to nearly three times the line height below the line bottom. Final Sadhe is used in Eg and Oss. Qoph: The top stroke – starting as a tick – slants slightly upwards. On reaching the line ceiling it turns towards the left again forming an acute angle or sharp curve, according to whether the downstroke is straight or curved. It does not reach the shaft. The shaft either issues from the top stroke or is severed from it. The latter type is not present in Eg but occurs in Oss. Resh: The narrow top stroke is a tick which slants up to the top of the vertical. The corner where they join is slightly rounded. The letter differs clearly from Daleth. Eg and Gez have a narrow top stroke, Oss as a rule a wide or very wide one. Shıˆn: The right hand stroke is vertical to begin with. It then slants leftwards to the line bottom. Here it meets the left hand stroke which is only slightly slanting or almost vertical, with the result that the bottom of the letter is sharply pointed. The inner stroke – straight or with a small vertical beginning – joins the left one at its middle. Both Eg and Oss have this type. In Oss the left stroke is frequently produced downwards, a development of which there are already traces in Hab. Taw: A top stroke, slanting up rightwards to the line ceiling, or under it and horizontal, turns downwards and as a straight vertical continues to the line bottom. The top stroke is joined on the left by a straight vertical which comes from just above the line ceiling. At the line bottom it turns to the left. Occasionally this corner is still round, but as a rule it is a right or obtuse angle, so that the foot type is produced. In Eg the left hand stroke is wavy and much longer, the one on the right very curved. Some Ossuaries tally with Hab; others have already developed a formalized top bar; others again show a transition form, where the top of the left vertical and the top curve of the right hand part are linked together by a loop. Our examination has resulted in establishing that most letters in Hab differ, appreciably to considerably, from their counterparts in the documents utilized for comparison, although they contain elements from all of them. Hence Hab must have been written in the interval between the early and the late of these documents – that is somewhere between 200–50 B. C. E.

000203

Birnbaum 01.04.11 18:17

The Date of the Habakkuk Cave Scroll

203

The following letters show early characteristics: Aleph, Beth, Daleth, Resh. The following, on the other hand, have characteristics of the late period: Gimel, Waw, Zayin, Teth, Kaph, Lamedh, Mem, Final Mem, Nuˆn, Samekh, Ayin, Pe, Final Pe, Final Sadhe, Qoph, Shıˆn, Taw. Accordingly it appears that the Habakkuk Cave Scroll was written in the first half of the first century B. C. E., with the result that this find has enlarged the range of our knowledge of systematical Bible commentaries by nearly a thousand years. Abbreviations BAMP: BASOR: BBSAJ: BFMA: CIS : CRA : ESE: HNE : JA : JAOS: JBL : JPOS: MPJ : OLZ: PEQ: PJB : PSBA: Q: QS : RA : RAO: RB : REJ : S: SBB:

T: ZDPV:

Bulletin arche´ologique du Muse´e Parent. Bulletin of the American Schools of Oriental Research. Bulletin of the British School of Archaeology in Jerusalem. W. F. Albright, “A Biblical Fragment from the Maccabean Age: the Nash Papyrus”, in: JBL 56 (1937), pp. 145–176. Corpus Inscriptionum Semiticarum. Vol. 2, Paris 1889 ff. Comptes Rendus, Acade´mie des Inscriptions et Belles-Lettres. Mark Lidzbarski, Ephemeris für semitische Epigraphik. 3 Bde., Giessen 1902–1915. Mark Lidzbarski, Handbuch der nordsemitischen Epigraphik. Weimar 1898. Journal Asiatique. Journal of the American Oriental Society. Journal of Biblical Literature. Journal of the Palestine Oriental Society. Rene´ Dussaud, Les monuments palestiniens et judaiques [. . .] Muse´e du Louvre. Paris 1912. Orientalistische Literaturzeitung. Palestine Exploration Quarterly. Palästinajahrbuch des Deutschen Evangelischen Instituts für Altertumswissenschaft des Heiligen Landes zu Jerusalem. Proceedings of the Society for Biblical Archaeology. Qobes hahebra ha ibrit lahaqirat eres yisra el wa attiqoteha {= Journal ¯ ˙ Jewish ˙ ¯ Palestine ¯ ¯ Exploration ˙ ¯ ˙ ¯ ¯ of the Society}. Quarterly Statement of the Palestine Exploration Fund. Revue arche´ologique. Charles S. Clermont-Ganneau, Recueil d’arche´ologie oriental. 8 vols., Paris 1888–1924. Revue biblique. Revue des e´tudes juives. Syria. Sitzungsberichte der preussischen Akademie der Wissenschaften, phil.hist. Klasse (Berlin). Tarbis. ˙ Zeitschrift des deutschen Palästina-Vereins.

000204

204

Birnbaum 01.04.11 18:17

[ 41 ]

Fig. 1. Part of the Habakkuk Scroll

000205

Birnbaum 01.04.11 18:17

The Date of the Habakkuk Cave Scroll

Fig. 2. Alphabetical chart (drawn by the author) from the Habakkuk Scroll

205

000206

Birnbaum 01.04.11 18:17

000207

42.

Birnbaum 01.04.11 18:17

The Date of the Covenant Scroll*

From a literary point of view, one of the most important finds amongst the scrolls from the cave near the north end of the Dead Sea is the Covenant Scroll (‘Sectarian Document’). As a preliminary to any attempt at determining when the author composed his work we shall have to establish the date of the script. A detailed comparison with that of other early documents hitherto available will provide us with the terminus ad quem for the composition of the text. Our examination is based on the reproduction of column 1, published in BASOR 111 (1948), p. 10. {Also reproduced on pp. 214 f. here, and in The Hebrew Scripts (HS ) by Solomon A. Birnbaum, 2 vols., London 1954–1957 / Leiden 1971, plate 82. See also HS, pl. 86.}1 It is evident at first glance that the writing belongs to the oldest type of Square script known to us from the scanty relics that Egypt and Palestine have until now afforded us. Although we have not sufficient material for differentiating between the alphabets of the two countries in those times we may perhaps assume that if there were differences they were not yet great. Upon the basis of the Jewish documents of either provenance it is clear that the Covenant Scroll (Cov) cannot possibly have been written earlier than the third century documents preserved from Egypt2 on the one hand. On the other hand again, it is obviously earlier than the Dura-Europos Fragment3 of before 256 C. E. Hence we can limit our comparison to the following: the third century Egyptian documents (Eg), the Tobias Inscription ( Tob)4 of 183–175 B. C. E., the Gezer Boundary Stones (Gez)5 of 100– 63 B. C. E., the * First published in: Palestine Exploration Quarterly 81 ( July – Oct. 1949), pp. 140–147. 1 For abbreviations see the list at the end of this article. 2 BFMA . PSBA 37 (1915), p. 217. ESE vol. 2, p. 243, vol. 3, p. 22. E. Sachau, Aramäische Papyrus und Ostraka aus einer jüdischen Militär-Kolonie zu Elephantine. Leipzig 1911, p. 243. PSBA 29 (1907), p. 260. REJ 65 (1913), p. 16. 3 The Excavations at Dura-Europos, vol. 6, ed. by M. I. Rostovtzeff [et. al.], New Haven 1936, p. 417 {HS 103*}. 4 D. Chwolson, Corpus Inscriptionum Hebraicarum. St. Petersburg / Leipzig 1882, p. 1. PJB 16 (1920), p. 33. JA se´r. 6, 10 (1867), p. 188. ESE vol. 3, p. 49. Syria, Publications of the Princeton University Archeological Expeditions. Leiden 1907 ff., Div. 3, Sect. A, p. 1. CRA se´r. 2, 1 (1865), p. 79. RA N. S. 10 (1864), p. 52. 5 BFMA . CRA se ´ r. 4, 26 (1898), p. 686; RB 8 (1899), p. 109; RAO vol. 3, p. 116;

000208

208

Birnbaum 01.04.11 18:17

[ 42 ]

Ossuaries (Oss)6 of between the late first century B. C. E., the Uzziah Plaque (Uz)7 of the same period, the Queen Helena Inscription (Hel)8 of the second half of the first Christian century.9 Aleph: The left hand stroke is straight or curved. It generally starts at the top of the main stroke and extends to the bottom of the line. Such forms are characteristic of Oss und Uz but are not to be found in Eg. Beth: The top is generally a semi-circle, sometimes a tick with a right or acute angle. The letter thus corresponds to Eg and Tob, while in Oss this stroke is formalised into a short initial vertical stroke and a long bar on the ceiling of the line (sometimes the initial stroke is already missing). Here the base is generally produced on the right. In Cov there is no indication of this trend. Gimel: Both strokes are very curved. Curves occur in Oss, and the little upstroke in Gez may perhaps be a lapidary reflection of a handwritten curve. The left stroke issues from the middle of the right one, as in Oss and some of Eg. There is no trace of that other Eg form, where the left stroke issues from the top of the right one. Daleth: The left side of the top part is a narrow tick, the horizontal stroke of which generally joins the main vertical well beneath the line ceiling. This corresponds to the narrow curve in Eg. In Oss and Uz, however, the letter is very wide, and the tick has become a top bar.

6

7 8 9

QS 31 (1899), p. 118. HNE , p. 484. R. Macalister, The Excavation of Gezer. London 1912, vol. 1, p. 37, vol. 3, plate 10 f. JPOS 10 (1930), p. 19. BFMA. Q 1,1 (1921), p. 57. RA N. S. 25 (1873), p. 398. RA N. S. 36 (1878), p. 305. RA Ser. 3, 1 (1883), p. 257. Ch. Clermont-Ganneau, Archaeological Researches in Palestine. London 1899, vol. 1, p. 381. RAO vol. 5, p. 334. RAO vol 6, p. 211. RAO vol. 8, p. 59. MPJ , nos. 16, 51. S 4 (1923), p. 241. SBB 1885, p. 669. OLZ 15 (1912), p. 529. S. Klein, Jüdisch-palästinisches Corpus Inscriptionum. Wien 1920. JPOS 9 (1929), p. 100. ESE vol. 1, pp. 186, 312; vol. 2, pp. 72, 191; vol. 3, pp. 49, 302. HNE, p. 485. Macalister (see above, n. 5), vol. 1, pp. 347, 382. QS 63 (1931), p. 171. Q 1,2 (1925), p. 71. BBSAJ 5 (1924), p. 56. BAMP 1 (1867), p. 21. RB 34 (1925), p. 253. RB 1 (1892), p. 262. JAOS 28 (1907), p. 355. ZDPV 4 (1882), p. 9. JPOS 4 (1924), p. 171. JPOS 8 (1928), p. 113. JPOS 9 (1929), p. 45. JPOS 12 (1932), p. 25. Sep¯er hayyobel lapprofesor S. Krauss {Festschrift Samuel Krauss. Jerusalem 1936/37}, p.¯87. PEQ 69 (1937), p. 126. Sep¯er Zikkaron le Asˇer Gulak uliSˇemu el Klein {Memorial Volume A. G. and S. K., Jerusalem 1942}, p. 133. RB 11 (1902), p. 276. RB 43 (1934), p. 564. JPOS 9 (1929), p. 41. BASOR 44 (1931), p. 8. T 2 (1930), p. 288; p. 295. QS 63 (1931), p. 217; 64 (1932), p. 106. BFMA. CIS , no. 156. MPJ, no. 28. A. de Longperier, Muse´e Napole´on III , Choix des monuments. Paris 1868–1874. JA se´r. 6, 6 (1865), p. 550. {These documents are reproduced in HS under the following numbers: Eg 155– 162; Tob 80, 87F(1); Gez 83, 87F(2); Oss 95–99; Uz 88, 87F(7–8); Hel 100}.

000209

Birnbaum 01.04.11 18:17

The Date of the Covenant Scroll

209

Heˆ: The right hand vertical starts above the line ceiling. The horizontal is slightly wavy; from its middle, or a little more to the left, issues the left vertical. This form corresponds on the whole to Tob. In Eg the left downstroke is still slanting and issues from the right hand top corner. In Oss and Hel forms similar to Cov can still be found but even in these the left vertical is more to the left. In other forms it has moved to the left end of the new, formalised top bar which has swallowed up the top of the right vertical. In some Ossuaries, in three times out of five in Uz, and in Hel, the left stroke is severed from the top bar. Waw: The top is an acute angle or a small and narrow curve as in Oss and Uz, while Tob has a right angle and a long top stroke corresponding to the identical angle and the wide top curve of Eg. Zayin: As in Eg, it is a plain vertical of line height, while in Gez and Oss a small top stroke is added, slanting from right to left and joining the vertical either at the top or near the top; in Uz this is formalised with the effect that the letter has become a symmetrical counterpart of Waw. Heth: The downstrokes are as a rule somewhat curved. The cross ˙ – under the ceiling of the line – is generally horizontal; occasionstroke ally, however, one of its ends slants up to the top, as in Eg. In Oss the cross bar is always horizontal. On some Ossuaries and in Uz the top has been formalised, the cross bar having become a top bar. Teth: It looks like a medium figure 6, except that it is generally open. But it is not written in one go: both the right and left hand strokes are written downwards and meet at a very acute angle. Cov tallies with Eg and Tob. It is very unlike Oss and Uz, where a base has developed. The letter is wide, the left stroke has become straight and tends to decrease in size until it has finally reached only line height. Yodh: It is long, often of line height. The top is an acute angle, the right hand stroke is as a rule longer than the other, slanting towards the right. It differs from Waw not in size but in the pronounced width or angularity of the top, and generally, the straightness of the downstroke. It resembles Eg rather than Oss and Uz, where Yodh is on the whole smaller and the right hand stroke a vertical. Kaph: The top is a narrow curve or tick. The downstroke is straight, curved or wavy, turning at right angles – occasionally at obtuse angles – to the left, to form a base of equal or slightly greater width than the top part. The base is somewhat under the line bottom, at times considerably lower. This narrow form is used in Eg. In Oss, Uz and Hel the formalisation of the letter to line height has caused it to acquire greater width. In Uz the top bar is in addition formalised. There is no Final Kaph but it appears that at the end of a word the shortened form is not

000210

210

Birnbaum 01.04.11 18:17

[ 42 ]

employed. In Uz the final form consists of a formalised top bar and a long vertical downstroke. Lamedh: It is in two distinct parts, an upper and a lower one. The first is a very tall stroke, slightly slanting or nearly vertical, beginning with a small or biggish upstroke or loop and sometimes extending a little below the line ceiling. The lower part is a slightly curved downstroke, which sometimes does not reach the line bottom. In Eg the letter is a wavy line, the parts gradually curving into each other. A kind of two part letter is, however, also beginning to make its appearance. The type of Cov is like that of Oss, Uz and Hel. Mem: The top part is a narrow, deep curve or tick, corresponding to the wide, wavy top stroke of the fifth century in Egypt and the intermediate form of Eg. The right hand top corner of the letter which is round in Eg is sharply pointed in Cov. This tallies with Gez, Oss, Uz and Hel. The left hand downstroke which in Eg crosses the top, in Cov issues from it, starting as a rule from the lowest point of the top part. It often slants down to rather near the line bottom. In these respects, too, Mem tallies with Gez, Uz, Hel and the majority of Oss. The right hand downstroke – mostly straight, occasionally curved or wavy – slants to the right, turning at acute angles to the left at the line bottom, to form a horizontal base of medium or considerable length. There is no base in Eg. In Gez, Oss, Uz and Hel the base is well developed and wide. Final Mem: The top and right hand parts are identical with those of Mem, but the left and downstroke reaches the base, either by being vertical or by turning vertical half way. Both forms occur in Oss, while Uz has only the rectangle. When the left stroke starts near the right hand top corner, the letter is rather like a triangle in shape. Such forms and even wholly triangular ones are to be met with in Oss. The letter extends considerably below the line bottom. There is nothing resembling the severance of the left hand downstroke from the top bar such as we find in Uz. Occasionally in Cov the downstroke even passes through the top stroke. Nuˆn: It consists of a vertical curve which, at the line bottom, turns to the left to form a short or medium long base. The two parts correspond to the upper and lower parts of the original wavy line of Eg. The development of the base on the line bottom reduces the letter to line height. Thus Nuˆn tallies with Oss but in Cov there are no forms with a straight top stroke like those which are sometimes to be met with in Oss. Final Nuˆn: It has the same long, wavy form as in Eg and Oss. Samekh: The top is a curve from whose right end a vertical goes forth. Occasionally this downstroke does not continue vertically up to

000211

Birnbaum 01.04.11 18:17

The Date of the Covenant Scroll

211

the line bottom but starts slanting to the left about half way down. On reaching the line bottom, it turns at right angles towards the left to form a base. From the left part of the top curve a stroke slants to the left, stopping at various distances from the line bottom. The letter corresponds to the somewhat less developed form of Eg.10 Ayin: The right hand stroke is a rather deep curve of line height, met on the left, half way, by a smaller curve. In Eg the strokes are of equal size. In Oss and Uz the Cov type is frequent, but there is a tendency towards prolongation of the right hand stroke beneath the line bottom; sometimes it extends considerably beneath it. Pe: From a point on the line ceiling a small stroke slants to the left and a downstroke to the right. On reaching the line bottom it turns leftwards in the shape of a horizontal – exceptionally, slightly slanting – base. In Eg the base is not yet developed and there is no specimen with a right or acute angle at the bottom. In Cov we have the Oss form. The left hand stroke is occasionally broken, corresponding to the curve in Uz. Final Pe: There is none. Sadhe: The more or less straight downstroke is slanting, and turns at the line bottom towards the left, to form a slanting, occasionally horizontal, medium-sized base. In its upper half the downstroke is met on the right by a large curve or tick. The letter corresponds to Eg, Oss and Hel, but in Eg it extends further down. In col. 1 no Sadhe occurs at the end of a word, so that at present we do not know whether there was a Final Sadhe. Qoph: A horizontal top stroke, or one slanting slightly upwards, starts with a small curve or straight downstroke, turns downwards and bends to the left again until it meets the left hand shaft, thus forming a triangular outline. The shaft, slanting slightly to the right, extends – more or less straight, or else becoming a very shallow curve – far beneath the line. This closed form is to be found in Eg as well as in Oss. There is in Cov no trace of the open form already making its appearance in Oss, with the vertical severed from the top stroke either on top or on the side or in both places. The downstroke is short in Eg, long in Oss. Resh: The top stroke is a narrow, sometimes medium wide curve or tick, as in Eg. In Gez, we have a narrow top bar, in Oss a very wide one. Shıˆn: The outer strokes are slanting and only slightly curved, or else straight. They meet at a point on the line bottom so that the letter is triangular in outline. The inner stroke is straight and therefore ends rather low down on the left stroke. Shıˆn tallies with Eg and Oss. There 10

The specimens which have a closed form have been re-inked by a later hand.

000212

212

Birnbaum 01.04.11 18:17

[ 42 ]

is no example of the Oss form with the left hand stroke produced downwards, occasionally considerably so. (In Oss this type is more frequent than the other one.) Taw: A small curved or slanting upstroke turns downwards at the line ceiling and runs, slightly curved or almost straight, to the line bottom. This is the right hand part of the letter. On the left there is a generally rather wavy downstroke, starting sometimes above the line ceiling. It does not extend below the line bottom. Its small end curve is now and then flattened into a foot. In Eg the letter is less formalised: the right hand stroke is more curved, the left one more wavy and much longer. Cov corresponds to Gez, Oss, Uz and Hel, but in them the letter is more formalised: there is a top bar, the downstrokes are straight. Our comparison has made it clear that nearly all the letters differ distinctly from their counterparts at both ends of the period in question, so that the time of our MS . must lie between them, say approximately somewhere between 200–50 B. C. E. Some of the letters have greater affinity with the forms of Eg than of Oss: Beth, Daleth, Zayin, Heth, Teth, Yodh, Kaph, Samekh, Resh. ˙ Others are closer to Oss: Aleph, Gimel, Waw, Lamedh, Mem, Final Mem, Nuˆn, Ayin, Pe, Qoph, Taw. In considering whether there is any preponderance in favour of either the earlier or the later age, we find that the scales are equally balanced. Hence the second half of the second century B. C. E. suggests itself as the time when the Covenant Scroll was written. The writing is thus more recent than that of the Isaiah Scroll, which was written in the previous half-century11 but earlier than the Habakkuk12 and Hodhayoth Scrolls,13 which I ascribed to the first half of the first pre-Christian century.14 I have since compared in detail the script of the Isaiah Scroll with that of the Nash Papyrus and dated the former as belonging to the second quarter of the second pre-Christian century. To allow for the fact 11 12 13 14

See my dating in BASOR 113 (Febr. 1949) {= Chapter 40 in the present volume}. See “The Date of the Habakkuk Cave Scroll”, in: JBL 68 (1949) {= Chapter 41 in the present volume}. A study of the script of the Hodhayoth Scroll is ready for publication. {Later published in PEQ 84 (1952) = Chapter 44 in the present volume.} Long after the present article had been handed in for publication (March), it became known that the fragments of the jars which had contained the Cave Scrolls had been excavated and dated by pottery experts as belonging to the late Hellenistic period, the latest possible date given being the early first century B. C. E. This is in keeping with our palaeographical results.

000213

Birnbaum 01.04.11 18:17

The Date of the Covenant Scroll

213

that the script of the Covenant Scroll is distinctly younger than that of the Isaiah Scroll, the former must not be placed early in the period in question. Hence my date for the Covenant Scroll is about the last quarter of the second pre-Christian century.15 Abbreviations BAMP: BASOR: BBSAJ: BFMA: CIS : CRA : ESE: HNE : JA : JAOS: JBL : JPOS: MPJ : OLZ: PEQ: PJB : PSBA: Q: QS : RA : RAO: RB : REJ : S: SBB:

T: ZDPV:

15

Bulletin arche´ologique du Muse´e Parent. Bulletin of the American Schools of Oriental Research. Bulletin of the British School of Archaeology in Jerusalem. W. F. Albright, “A Biblical Fragment from the Maccabean Age: the Nash Papyrus”, in: JBL 56 (1937), pp. 145–176. Corpus Inscriptionum Semiticarum. Vol. 2, Paris 1889 ff. Comptes Rendus, Acade´mie des Inscriptions et Belles-Lettres. Mark Lidzbarski, Ephemeris für semitische Epigraphik. 3 Bde., Giessen 1902–1915. Mark Lidzbarski, Handbuch der nordsemitischen Epigraphik. Weimar 1898. Journal Asiatique. Journal of the American Oriental Society. Journal of Biblical Literature. Journal of the Palestine Oriental Society. Rene´ Dussaud, Les monuments palestiniens et judaiques [. . .] Muse´e du Louvre. Paris 1912. Orientalistische Literaturzeitung. Palestine Exploration Quarterly. Palästinajahrbuch des Deutschen Evangelischen Instituts für Altertumswissenschaft des Heiligen Landes zu Jerusalem. Proceedings of the Society for Biblical Archaeology. Qobes hahebra ha ibrit lahaqirat eres yisra el wa attiqoteha {= Journal ¯ ˙ Jewish ˙ ¯ Palestine ¯ ¯ Exploration ˙ ¯ ˙ ¯ ¯ of the Society}. Quarterly Statement of the Palestine Exploration Fund. Revue arche´ologique. Charles S. Clermont-Ganneau, Recueil d’arche´ologie oriental. 8 vols., Paris 1888–1924. Revue biblique. Revue des e´tudes juives. Syria. Sitzungsberichte der preussischen Akademie der Wissenschaften, phil.hist. Klasse (Berlin). Tarbis. ˙ Zeitschrift des deutschen Palästina-Vereins.

See my article “The Dates of the Cave Scrolls”, in: BASOR 115 (Oct. 1949), pp. 20–22 {= Chapter 37 in the present volume}.

000214

214

Birnbaum 01.04.11 18:17

[ 42 ]

Fig. 1. Part of the Covenant Scroll

000215

Birnbaum 01.04.11 18:17

The Date of the Covenant Scroll

Fig. 2. Three alphabets drawn (by the author) from the Covenant Scroll, col. 1

215

000216

Birnbaum 01.04.11 18:17

000217

43.

Birnbaum 01.04.11 18:17

The Age of the Cave Scrolls Two Letters to the London Times*

To the Editor of The Times, August 25, 1949. Sir, – In the recent discussion about the cave scrolls one important aspect has been neglected. Dating works by internal evidence alone is always a risky matter. In the case of the scrolls it has led to results some of which differ from others by about a thousand years. Those, however, who relied on the evidence of the script itself have all arrived at almost identical results. In October, 1948, Professor Albright assigned the earliest document to about 150–100 B. C., the youngest one to a century later. Dr. J. C. Trever made a detailed study of the writing, and the corresponding figures were 125–100 B. C. and 25 B. C. – 25 A. D. The same issue of the Bulletin of the American School of Oriental Research in which Trever’s article was printed contained one by the writer of the present lines, who subjected the script of the Isaiah fragment published in the previous issue to a minute investigation, while analogous studies of the other scrolls were published by him elsewhere. His dates were 200–150 B. C. and 100–50 B. C. (more precisely in a forthcoming publication: 175–150 B. C. and about 50 B. C.). In other words, where our dates do not overlap the maximum difference in time is one generation only. Anybody who has taken the trouble to check a detailed palaeographical examination of the writing of the scrolls will realize what a high degree of precision is possible in this field, and will not be surprised to learn that the archaeological evidence – which became available long after the palaeographical results had been published – has corroborated these. Yours faithfully, Solomon A. Birnbaum. School of Oriental and African Studies, University of London, W. C. 1, Aug. 22 [1949]. To the Editor of The Times, September 13, 1949. Sir, – In his letter about the cave scrolls, the Rev. John W. Wenham asks the palaeographer to show that there are adequate data on which to base scientific conclusions regarding the dates of the script, and he expresses * First published in: The London Times, 1949.

000218

218

Birnbaum 01.04.11 18:17

[ 43 ]

the wish to learn in simple terms how these conclusions have been reached. The amount of material – Jewish documents written in ink – from the centuries around the Christian era is by no means small. The ostraca and papyri contain so much lettering that the whole of the alphabet is well represented in a number of hands – a lot of the letters recur very frequently, of course. From the later part of our period we have, in the main, something which is intermediate between handwriting and inscriptions. I am speaking of the numerous short texts on the Ossuaries. They are not written in lapidary style but are in a cursive script scratched into the soft chalk. Their forms can safely be used for comparison with those written with ink. There are, in addition, a few inscriptions, and with due caution the experienced palaeographer can utilize these lapidary forms, too. Thus it is correct to say that we have very far from a “totally inadequate” amount of data. How did the palaeographer arrive at his results? On the basis of experience he recognized at first glance that the script of the finds belonged to a period somewhere near the Christian era. He set to work comparing the alphabets of the scrolls with those of the above-mentioned documents. He examined every bit of every letter of the alphabet in minute detail under the lens – every angle, curve, hook, loop, up or down stroke, bend, wave, and slant. He established that certain letters of the scrolls bore a closer resemblance to the forms in the earliest documents under consideration (third pre-Christian century), that others were nearer to those in the later documents (the two centuries around the Christian era), others again were clearly intermediate. Weighing up the evidence in the usual way he arrived at the approximate date of each scroll, and, in view of the accuracy of the method, the results may be expected to be completely reliable. There is no need for the reader to put a blind trust in the art or authority of the palaeographer. Every step and detail can be followed and checked up in the published studies (Bulletin of the American Schools of Oriental Research, Journal of Biblical Literature, Palestine Exploration Quarterly). If there is scepticism as to the correctness of these conclusions, then it will have to be expressed in terms of a detailed examination, not in general assertions, and criticism will have to deal with the validity or otherwise, of the details of those studies. Yours faithfully, Solomon A. Birnbaum. School of Oriental and African Studies, University of London, W. C. 1, September 7 [1949].

000219

44.

Birnbaum 01.04.11 18:17

The Date of the Hymns Scroll*

Like the earlier studies by the present writer on the script of some of the Qumraˆn Scrolls,1 this article is based purely on the script.2 It therefore constitutes completely independent evidence as to the date when the Hymns were composed – the date of the manuscript being the terminus ad quem for its text. The present study is independent, too, of my own earlier palaeographical articles. Each of these presents a separate examination, the manuscript in question being compared with material whose dates are known to us. If these studies arrange themselves in a clear sequence of dates, and this sequence is confirmed when we compare the manuscripts with each other, then we have good grounds for thinking that our palaeographical results are reliable. When attempting to date a manuscript the palaeographer will start by making a preliminary examination of the script. That will enable him to limit down the area to be covered by a detailed investigation and to concentrate on this. In the case of our Scroll a preliminary examination shows beyond doubt that its script is neither as old as the papyri of Elephantine (last dated one: 400 B. C. E.) {see HS 147}3 nor as young as the Dura documents (244 C. E. and preceding decades {see HS 103*}). Between these two limits we have sufficient datable and dated material to locate the Hymns Scroll (Hym) with a fair degree of precision. Our examination is based on the only legible facsimiles hitherto published, which we here reproduce.4 The material utilised for comparison is:5 * First published in: Palestine Exploration Quarterly 84 (May – Oct. 1952), pp. 94–103. 1 The name Dead Sea Scrolls does not adapt itself well to the situation created by the newest finds, which also come from the neighbourhood of the Dead Sea. The designation Cave Scrolls or Cave MSS ., used by the writer from the start, appears to present no difficulties: the earlier finds might be called the Qumraˆn Cave Scrolls (MSS .), or for short Qumraˆn Scrolls (MSS .), the new ones Murabba aˆt Cave Manuscripts or for short Murabba aˆt Manuscripts. 2 Cf. my study “The Qumra ˆ n (Dead Sea) Scrolls and Palaeography”, in: BASOR Supplementary Studies, nos. 13–14 (New Haven 1952), pp. 7–52 {Chapter I, reprinted in this volume as Chapter 39}. 3 {HS = plate nos. in The Hebrew Scripts by Solomon A. Birnbaum. 2 vols., London 1954–1957 / Leiden 1971.} 4 With Prof. E. L. Sukenik’s kind permission, from his tuzung tuligm {Megillot Genuzot}. Vol. 1, Jerusalem 1948, pll. 12 and 13 {see also HS 87D, 87E, 87F}. 5 For further bibliographical references see above in this volume, Chapters 40, 42.

000220

220

Birnbaum 01.04.11 18:17

[ 44 ]

E = Eg. G = GezB. H = Hym. He = Hel. O = Oss. T = Tob. U = Uz

Alphabetic Chart drawn by the author, comparing the script of the Hymns Scroll with the script of six other documents (Aleph – Heth) ˙

000221

Birnbaum 01.04.11 18:17

The Date of the Hymns Scroll

221

the 3rd cent. B. C. E. documents from Egypt (Eg) {HS 155–162}, the Tobias inscription ( Tob) {HS 80, 87F(1)} of probably 183–175 B. C. E. (Vincent: c. 260 B. C. E.), the Sabbath boundary stones of Gezer (GezB) of the first third of the 1st cent. B. C. E. {HS 83, 87 F(2)}, the ossuary graffiti (Oss) {HS 95, 95*, 96, 97, 98, 99, 100*} of about the hundred years ending with the fall of the Temple, the Uzziah plaque (Uz) of the same period {HS 88}, and the Queen Helena graffito (Hel) of the decade 50– 60 C. E. {HS 100}. Aleph: The main stroke starts just below the line ceiling. It is straight as in Oss and Uz. The right stroke is of almost triangular outline and of considerable length, as in Oss and Uz, in contrast with the shortish form of Eg. The left stroke is straight and slants down to the left. As in Eg, it generally starts somewhere in the upper half of the main stroke. It is straight, as in Oss. As in Eg, Uz and on most Oss, it does not reach the line bottom. B: The top stroke is a rather narrow tick which corresponds to the rectangular tick of Tob and the curve of Eg, not to the formalised top bar of Oss. The downstroke6 slants slightly rightwards and just before arriving at the line bottom, swings to the right; it then turns leftward as a long base, which is sometimes slightly oblique or curved. Thus the general form is reached – the most frequent one in Oss – where the base is produced rightwards. G: Fairly often the downstroke starts with a sort of triangular thickening on the right, which corresponds to the small upstroke in GezB. The left stroke issues from the middle of the right one, never from the top or near it, which it does in Eg. Now and then the letter appears to be written with one lift of the pen, i. e., after having arrived at the line bottom the pen again moves back upwards to about the middle of the downstroke, thence to continue into the left stroke. Both curved and straightish varieties are to be found. Curves occur in Oss. D: The top-part is a medium-sized acute angle and thus stands in between the small curve of Eg and the long formalised top bar of Oss and Uz. It joins the main stroke near the top, as in Eg, some Oss and Uz. H: The top stroke is just below the line ceiling. It is horizontal and straight and has thus almost achieved the formalisation of Oss, Uz and Hel. As it is made by a double stroke it is very thick. The left downstroke starts in the middle of the top stroke as in Tob, not in the righthand corner (Eg) nor at the left edge (Oss). It is parallel to the righthand stroke, as in Oss but not oblique as in Eg. There is as yet no 6

The downstrokes are not vertical, they slant about 15 degrees to the right.

000222

222

Birnbaum 01.04.11 18:17

[ 44 ]

E = Eg. G = GezB. H = Hym. He = Hel. O = Oss. T = Tob. U = Uz

Alphabetic Chart drawn by the author, comparing the script of the Hymns Scroll with the script of six other documents (Teth – Mem) ˙

000223

Birnbaum 01.04.11 18:17

The Date of the Hymns Scroll

223

trace of the open form of the letter (through the severing of the left vertical from the top stroke) which occurs on some Oss, in Hel, and three times out of five in Uz. W: The top is an acute angle as in Oss and Uz, not, as in Tob, rectangular. The left stroke is short as in Oss and Uz, while Tob has a long top part corresponding to the wide top of Eg. The downstroke is now and then very slightly curved. Z: It is a downstroke of line height, practically straight, with a very slight indication of waviness, and with or without a very slight thickening on the right side of the top. The latter corresponds to the forms of Uz and some Oss where the top is not thickened but oblique. On other Oss, however, and in GezB this little oblique stroke joins the vertical somewhat further down. In Eg the letter is a plain straight vertical, sometimes of less than line height. H: There is a distinct horizontal top stroke, just below the line ˙ ceiling. The shafts are straight. The letter tallies with the form of Oss and Uz. T: The right part consists of an upper semi-circle, from whose right ˙ in the middle of the line height, a straight stroke issues forth and end, slants down leftwards to the line bottom. Here it is met by the left part of the letter, a somewhat oblique and very slightly curved stroke of line height, starting with a thickened twist. The shape of Hym is very unlike that of Eg and Tob, where it resembles a figure 6: its left stroke is very curved and almost double the height of the right part. Hym tallies with Oss and Uz, except that in them the right downstroke is less oblique than here and thus suggests a base. In all three of them the letter is wide, unlike the narrow forms of Eg and Tob. Y: It is an oblique downstroke of line height, on the left of which, not far from the top, a short stroke issues at right angles. This type is in between Eg and Oss. The top part differentiates Y and W: In Y it is thicker and triangular. The whole letter is sometimes heavier – and, if curved, often more so – than W. K: As in Uz the top part is a short or medium-sized tick. Its left, vertical stroke tends to be stunted, or even to disappear altogether. The right stroke is horizontal; as in Uz it joins the top of the main stroke, forming with it a well-defined, almost rectangular corner. The straight downstroke has a slightly rightward slant and on reaching the line bottom swings at right angles leftwards to form an oblique, fairly long base. In Eg the base is a curve far down below the line bottom, in Uz and Hel it is a horizontal on the line bottom. The letter is very different from B.

000224

224

Birnbaum 01.04.11 18:17

[ 44 ]

E = Eg. G = GezB. H = Hym. He = Hel. O = Oss. T = Tob. U = Uz

Alphabetic Chart drawn by the author, comparing the script of the Hymns Scroll with the script of six other documents (Final Mem – Pe)

000225

Birnbaum 01.04.11 18:17

The Date of the Hymns Scroll

225

k: Here the horizontal portion of the top tick is just below the line ceiling, as in Uz. As in Eg and Uz the downstroke extends below the line bottom to a length almost double the line height. It is slightly wavy as in Eg while in Uz it is straight.7 L: It is in three parts: a very tall downstroke, with a thickened initial twist, slopes rightward to just below the line ceiling. From here a very small horizontal runs rightward and then slants down sharply leftwards, without however reaching the line bottom – sometimes stopping in the middle of the line height. In Eg the letter is a wavy line, the three parts gradually curving into each other, but forms consisting of two parts are also beginning to appear. The Hym type is characteristic of Oss, Uz and Hel. M: The top part is a small tick. From its point a straight stroke slants down leftwards – a continuation, as it were, of the right-hand stroke of the tick. It stops in the middle of the line height. From the right end of the top a straight downstroke slopes down rightwards to the line bottom. Here it turns leftwards at a slightly obtuse angle to form an oblique base of medium length. In Eg the top is wider and still a wavy line. The right top corner is therefore round, while in Hym it is sharply pointed. In Eg the left downstroke passes through the top, which it never does in Hym. Here the M tallies with GezB, most Oss and Uz. m: As far as one can tell from the unsatisfactory facsimiles, the letter consists of a right and a left part. On the right there is a small tick from which a downstroke slants rightward to considerably below the line bottom. Here it turns at a slightly obtuse angle to the left, to form an oblique base of medium length. On the left there is a top tick and a straight downstroke which runs to the left end of the base. The horizontal parts of the two ticks have an appearance of forming one single top bar. The left part of the right tick and the left downstroke together seem to form one single stroke too, crossing the top bar. That this part was not constructed in this way is betrayed by the fact that the short top of this apparent cross stroke differs considerably in slant from the main part, the downstroke. What reason could the writer have had for changing the direction of the cross stroke as soon as he reached the top bar? We seem here to have the same construction as in the Habakkuk Scroll where it is beyond doubt that the earlier form was dissolved into quite new elements which, however, aimed at reproducing the old form. The 7

In fig. 1, line 5, the final form occurs in a non-final position. It is obvious that the writer realised what he had done before he started the next word and that he added the H at once. The same word without H occurs at the end of line 11.

000226

226

Birnbaum 01.04.11 18:17

[ 44 ]

E = Eg. G = GezB. H = Hym. He = Hel. O = Oss. T = Tob. U = Uz

Alphabetic Chart drawn by the author, comparing the script of the Hymns Scroll with the script of six other documents (Final Pe – Taw)

000227

Birnbaum 01.04.11 18:17

The Date of the Hymns Scroll

227

letter, though rather narrow, is one of the largest in the alphabet. It differs from Eg in being closed while there the left downstroke forms so large an angle with the vertical that it cannot meet the bottom curve. On the other hand it differs from Oss and Uz in never being open at the left top corner, and in having an oblique base. N: It consists of a slightly oblique downstroke which, at the line bottom, turns – at right angles – leftward to form a medium-sized base. In Eg there is not yet a base, in Oss the base is very long so that the Hym form takes up an intermediate position, but somewhat closer to Oss than to Eg. The downstroke has an initial triangular thickening on the right-hand side; in other words, this form of N is an angularisation of the wavy line characteristic of Eg (although some straightening has already started there). A corresponding form is the one that is generally used in Oss, although a type with a plain, straight downstroke also occurs there. Whether there are traces of this in Hym cannot be decided on the basis of our material. n: This starts as a vertical downstroke which very soon curves to the right, and then, in the lower half of the line height, to the left, so that the letter consists of a wavy line which extends below the line bottom to the length of almost a full line height. Both Eg and Oss are of the long and wavy type. S: From just below the line ceiling the top stroke goes towards the right with a gentle upward slant till it reaches the line ceiling. From there – at a very acute angle – a downstroke runs leftward to the line bottom. The letter is closed on the left by a downstroke which starts at the line ceiling and slopes to the right. S thus forms a triangle standing on one of its angles. Hym differs greatly from the open type of Eg and tallies with Oss. Ayin: The right stroke is produced downwards to the line bottom. This form does not occur yet in Eg but tallies with the shapes of Oss and Uz. On some Oss the downstroke extends to well below the line bottom, a development of which there are as yet no traces in Hym. P: The main stroke, sometimes very slightly curved, slants rightward from the line ceiling to the line bottom. Here it turns at right angles leftwards to form a medium-sized base, at the end of which there is often a slight thickening of the upper edge. From the top of the main stroke, at an acute angle, a stroke slants down leftwards, stopping at – or before it reaches – the middle of the line height. The outline of the letter is thus a triangle whose left side has a wide gap in the middle. In Eg the base is not yet developed. Hym tallies with Oss but the left stroke never resembles the curve of Uz.

000228

228

Birnbaum 01.04.11 18:17

[ 44 ]

p: This starts like P, but while in the latter the downstroke, upon reaching the line bottom, turns leftwards at right angles to form a base, in p the angle in question is an obtuse one, with the result that no base is formed and the letter is prolonged to almost double a line height. The left stroke is somewhat curved. Similar forms are present in Eg and Oss. S: A slightly oblique downstroke extends to just below the line ˙ bottom and then turns leftward to form – at right angles – a short base, straight or sometimes curved. In its upper half the downstroke is met (or closely approached) on the right by a triangular stroke. Hym tallies with Eg and Oss. s: This has a corresponding upper part, but instead of turning left˙ wards to form a base the downstroke is prolonged below the line bottom, the letter thus being of double a line height. Q: The top stroke, starting as a tick, runs to just below the line ceiling. From its right end issues – sometimes after a short transitional stage – a short stroke which slants down leftwards to about the middle of the line height. It thus does not reach the shaft. The latter starts from the top stroke.8 In Eg we have only fully closed forms, in Oss open ones as well, so that Hym stands nearer to Oss. The same holds good in regard to the longness of the shaft, a feature which is absent in Eg. R: The top is a medium sized or narrow tick which joins the top of the downstroke. This and the fact that the right part of the tick becomes horizontal, distinguishes it from D. Eg and GezB have a narrow top, Oss as a rule a wide or very wide and completely horizontal one. Hym is thus intermediate between the two. As regards the structure of the corner, it tallies with Oss. Sˇ: The outline of the letter is a triangle standing on one of its angles. The right stroke, sometimes starting with a slight thickening or bend, slants to the left at an angle of 45 degrees and is somewhat curved. The left stroke is a normal downstroke, i. e., almost vertical, with the result that the angle at the bottom of the letter is very acute. The inner stroke slopes towards the left one but often does not reach it. The right stroke joins the left one just above the line bottom. This feature is characteristic of Oss, where the left stroke is sometimes produced too far below the line bottom. T: A short stroke, just below the line ceiling, slants up to it, thence a straight downstroke runs down to the line bottom. On the left the top stroke is passed and touched by a straight downstroke coming from the 8

From the facsimile we cannot quite tell whether there is a tick or whether what looks like the left stroke of the tick is really the top of the shaft.

000229

Birnbaum 01.04.11 18:17

The Date of the Hymns Scroll

229

line ceiling. At the line bottom it turns leftward at right angles, so that the type with a foot is produced. In Eg the left stroke is wavy and much taller, the one on the right very curved. Hym tallies with Hel and with some Oss, while others have already developed a formalised top bar, and others again show a transition form, where the top of the left vertical and the top curve of the right part are linked together by a loop. To sum up: Three features of aleph, B and k are indicative of an earlier rather than of a later date within the period in question. Twentyseven features of aleph B G H W Z H T K k L M N S ayin P Q R Sˇ T ˙ it seems reasonable to assign point to a late date within the period. ˙Thus our manuscript to the Oss period. The absence in the Scroll of certain late features – in H, m, ayin, P and Sˇ – might indicate that our script belongs to the earlier part of that period – say the late 1st cent. B. C. E. or the early 1st cent C. E.

Fig. 1. Specimen from the Hymns Scroll

000230

230

Birnbaum 01.04.11 18:17

[ 44 ]

Fig. 2. Specimen from the Hymns Scroll

000231

45.

Birnbaum 01.04.11 18:17

The Date of the Incomplete Isaiah Scroll from Qumraˆn*

Only a few of the Qumraˆn MSS . – and not even all the longer ones – have been subjected to detailed palaeographical examination. The present article, investigating the incomplete Isaiah Scroll (IsB), is intended as a further contribution to this field. A first, preliminary glance at the letter forms1 indicates that the script belongs to the centuries around the Christian era. For a detailed examination we may therefore restrict our comparative material to this period. The documents utilized are:2 the early third century B. C. E. documents from Egypt (Eg), the Tobias inscription ( Tob) from Araˆq el-Amıˆr, of about 260 B. C. E., the Sabbath boundary stones from Gezer (Gez), of the first third of the first century B. C. E., the graffiti on the ossuaries (Oss), from about the hundred years ending with 70 C. E., the plaque of King Uzziah (Uz), of the same period, and the Jewish graffito on the ossuary of Queen Helena (Hel) of Adiabene, of the decade 50– 60 C. E. In the interests of brevity, telegram style and brackets are used for the comparisons. The downstrokes have a slightly rightward slope, the transversals a leftward one; between them is an angle of about 90 degrees or just over. Aleph: The main stroke tends to be slightly wavy (Eg; Oss, Uz – straight). The left stroke is curved (Uz; Eg, Oss – occasionally; Oss – there are, in addition, specimens with a ‘foot’, evidently an angularization of the curved form). The left stroke issues from the upper half of the main stroke (Eg – centre; Oss – top). It generally reaches the line bottom (Oss – often; Uz, Eg – never). The right stroke is longish (Eg, Oss, Uz). Beth: The top part is a ‘tick’ ( Tob, Oss; Eg – generally a curve) with the right arm sometimes rather long ( Tob, Oss) and, occasionally, horizontal. The left arm is now and then vertical ( Tob). The downstroke is * First published in: Palestine Exploration Quarterly 92 ( Jan. – June 1960), pp. 19–26. 1 See fig. 1 here, previously published in The Dead Sea Scrolls of the Hebrew University, ed. by E. L. Sukenik, Jerusalem 1955, or my The Hebrew Scripts [HS], pl. 87A (which corresponds to plates 8–9 of Sukenik’s book). 2 For bibliographical references {including HS } see my article “The Date of the Covenant Scroll”, in: PEQ 81 (1949), pp. 140–142 {= Chapter 42 in this volume}.

000232

232

Birnbaum 01.04.11 18:17

[ 45 ]

Fig. 1. Part of the incomplete Isaiah Scroll (IsB)

000233

Birnbaum 01.04.11 18:17

The Date of the Incomplete Isaiah Scroll from Qumraˆn

233

straight ( Tob, Oss), occasionally vertical. The right bottom corner is pointed, rectangular or acute-angled ( Tob, Oss). The base sometimes juts out slightly to the right (Oss). These two features suggest that this line was generally drawn from left to right. Gimel: The right stroke tends to be slightly wavy. The left stroke, which is very curved, issues from the centre of the right one (Gez, Oss; Eg – top or near it). Both strokes are of equal length. Daleth: The top part is a tick (Oss; Eg – rounded), rather narrow (Eg – very narrow; Oss – rather narrow or widish; Uz – wide), large (Oss; Eg – small), acute-angled (Oss; Uz – rectangular). The right arm meets the downstroke considerably below the top (Eg, Oss; Uz – near the top). The downstroke is slightly wavy (Eg – rarely; Oss – straight). Heˆ: The top stroke is below the line ceiling (Eg, Oss, Uz; in Tob it still consists of two strokes). It is in horizontal position ( Tob, Oss, Uz; Eg – often) and straight (Oss, Uz; Eg – often; Tob – the left part). Its thickness is double, and even more than double, that of the other strokes (not in Eg, Tob, Oss, Uz, Hel; Oss (very often), Uz, Hel – the stroke starts as a tick which might possibly be the mason’s rendering of the thick pen-written stroke); the upper outline slopes down rightwards so that the stroke tapers somewhat. The right downstroke has the usual stance ( Tob, Oss, Uz, Hel; Eg – strong slant). It is either slightly wavy, bent or straight. The left stroke issues from the centre, or from slightly left of the centre, of the top stroke (Eg – from right end; Oss, Uz, Hel – from near left end). It is not parallel with the right stroke because it slopes gently leftwards or it is vertical (Oss, Uz, Hel; Eg – at an acute angle with the right stroke). Waw: The top is an acute angle (Oss, Uz; Eg – obtuse; Tob – rectangular, starting as rectangular tick). The downstroke is generally slightly curved (Eg; Oss, Uz – straight) and often hardly reaches the line bottom (Eg – sometimes; Tob, Oss, Uz – never); but in final position it even now and then extends slightly below the line bottom. The left stroke is much shorter than the right one (Oss, Uz; Eg, Tob – less difference, up to almost equal size); Tob – starts as a rectangular tick, representing the earlier curve). Zayin: It is of line height (Gez, Oss, Uz; Eg – generally), very slightly wavy (perhaps corresponding to those Oss forms which start with a bend) or nearly straight (Eg; Gez, Oss, Uz – straight). Heth: Sometimes the downstrokes are parallel (Eg, Oss, Uz) but ˙ generally the right stroke is bent inwards (Eg – sometimes), the bridge is straight (Oss, Uz; Eg – generally), short (Eg, Oss, Uz – long) and lies just below the line ceiling (Eg, Oss, Uz; Gez – somewhat lower down).

000234

234

Birnbaum 01.04.11 18:17

[ 45 ]

Teth: The left stroke is slightly curved (Eg, Tob – strongly) or straight (Oss, Uz). It is of line height (Oss, Uz; Tob – almost double). The right top part has generally the form of a narrow Gothic arch (Oss; Uz – in oblique position; Eg – left arm generally straight) with its apex not far from the right edge of the letter (Oss; Uz – slightly more to the left; Eg – generally on the edge; Tob – an arc issuing from the centre of the left stroke). From the bottom end of the right stroke a straight line slants down to meet the end of the left main stroke (Oss, Uz; Eg – the right arm of the top and the downstroke together form a continuous line, curved or straight). The letter is wide (Oss; Uz – very wide; Eg – narrow). Yodh is an acute angle (Eg, Oss, Uz), the left arm very oblique (Eg, Oss, Uz), the right one has the usual stance (Oss; Eg – sometimes strongly oblique), a right-curve (Eg; Oss – sometimes) or straight (Oss, Uz). The right arm is roughly of line height (Eg, Uz; Oss – sometimes), the left one about half as long (Eg; Oss, Uz – sometimes shorter). The difference between yodh and waw is not that yodh is smaller – on the contrary, the top of yodh is larger than that of waw because the angle is, as a rule, bigger and the left stroke longer. Cf., e. g., cases where waw and yodh occur as a cluster of three letters: the yodh is then characterized by straightness.3 ( Tob has the old form but the main stroke has become short, stopping in the middle of the line height, the second transversal issuing almost from its end.) Kaph: The top part is a tick (Oss) or curve (Eg), it is rather narrow (Uz; Eg – generally; Oss – varying). It meets the main stroke, just below the top (Hel; Eg – sometimes. At the top – Oss, Uz; Eg – generally). The main stroke has just a trace of waviness (Hel, Oss; Uz – slightly curved). It bends (Oss, Uz, Hel) or curves (Eg) into a straight base (Oss, Uz, Hel; Eg – occasionally) which sometimes lies at the line bottom (Oss, Uz, Hel) but generally below it, sometimes considerably so (Eg). Kaph differs clearly from beth where the top and base are wider and where the base is never below the line bottom. Final kaph has a top ‘tick’ (Uz; Oss – a straight stroke; Eg – a curve), the right arm of which is transversal (Oss, Uz). This arm is from medium long to long (Oss, Uz; Eg – medium long to short). The top stroke is very thick. The downstroke is roughly double the length of the line height (Eg, Oss, Uz), somewhat longer than the descenders of nuˆn and sadhe. It is wavy (Eg; Oss, Uz – straight or has a final leftward curve). 3

E.g., Sukenik, Dead Sea Scrolls (above, n. 1), pl. 6, line 5; pl. 12, line 11.

000235

Birnbaum 01.04.11 18:17

The Date of the Incomplete Isaiah Scroll from Qumraˆn

235

Fig. 2. Alphabetical chart (drawn by the author) from Isaiah Scroll B

Lamedh: The top stroke is about double the height of the lower part (Eg, Oss, Uz). It starts with a small upstroke (Uz; Oss – occasionally; Eg – never), the effect of this movement being to fill the top corner with ink, giving the stroke a wedge-like shape. The transversal slants down rightwards (Eg; Oss, Uz – horizontal) and is straight (Oss, Uz; Eg – beginning of this development); its length varies from short to long (Eg, Oss; Uz, Hel – medium). The bottom stroke extends to about the middle of the line height (Eg, Oss, Uz); it is straight (Uz, Hel; Oss – generally; Eg – beginning of development). The two strokes form a sharply pointed corner (Oss, Uz, Hel; Eg – occasionally). Mem: The top stroke is an acute-angled tick (Oss – often, otherwise rectangular or obtuse; Eg – sometimes; Hel – rectangular; Gez, Uz – obtuse) with short arms (Hel). The top corner is sharply pointed (Gez, Oss, Uz; Eg, Hel – round). The right downstroke is occasionally a little wavy (Eg; Gez – slightly curved; Oss, Uz, Hel – straight). At the line bottom (Gez, Oss, Uz, Hel; Eg – lower down) it bends at right angles (Gez, Oss, Uz, Hel; Eg – curves) leftward to form a straight base (Oss, Uz, Hel; Eg – down-curve; Gez – slightly down-curved) of varying length (Gez, Oss, Uz, Hel – considerable; Eg – short). The left downstroke issues from the meeting point of the two arms of the tick (Gez, Oss, Uz, Hel; Eg – from above the top stroke, and crossing it) and is

000236

236

Birnbaum 01.04.11 18:17

[ 45 ]

short (Gez, Uz, Hel; Oss – varying, sometimes very short; Eg – generally rather long); it sometimes runs at the same angle as the right arm of the tick (Hel; Oss – often; Gez, Uz – obtuse angle) but is an independent stroke, thinner and often slightly curved (Eg, Hel; Gez, Oss, Uz – straight). This, combined with the fact that in the majority of specimens the two strokes are at a different slant, makes it clear that also in cases where there seems to be one continuous stroke of even thickness, there are really two separate strokes. Final mem: The top part is an acute-angled ‘tick’ with a rather long right arm (Eg, Oss, Uz). The right corner is rectangular or slightly rounded (Oss, Uz; Eg – exceptionally). The right downstroke is straight (Eg, Oss). Running to appreciably below the line bottom (Eg, Oss; Uz – line bottom or just below) it bends, at about right angles (Oss, Uz) or it curves (Eg) leftward to form a straight base (Oss; Eg – down-curve) of considerable length (Oss – medium; Eg – short) so that it ends almost a line height below the line bottom. The left downstroke starts above the top stroke, crosses it obliquely (Eg; Oss – one of the types is a further development of this form, in the other type the stroke issues from the top stroke) and then bends downward, continuing at a very slight leftward slant until it meets the base at the left end (Oss, Uz; Eg – letter is open) or just before this; accordingly the two downstrokes are not parallel (Eg; Oss, Uz – parallel). Nuˆn: The downstroke is sometimes straight and generally almost straight, being either slightly wavy (Eg, Oss) or curving slightly leftwards (Oss). At the line bottom (Oss; Eg – generally further down) it bends or curves at about right angles into a straight base (Oss; Eg – down-curve) or medium length (Oss; Eg – generally narrow). Final nuˆn: Generally of double the line height but shorter forms also occur (Eg, Oss). It is usually a wavy line (Eg, Oss); the upper part is generally thickened; sometimes it has become straight (Eg, Oss). Samekh: The top stroke is straight (Oss; Eg – sometimes) or slightly curved (Eg – generally). On the left end it is crossed by the downstroke (Oss – often; Eg – never). As far as I can see the two strokes are not actually linked up here (Oss – occasionally, by a loop) although they sometimes look as if they were. (At the next stage that link would have become the left arm of the top part.) This appearance of being linked is created by the ink having filled in the corner between the two ends. The right top corner is usually bevelled (Oss – sometimes; Eg – rounded right angle or wide curve). The upper half of the right downstroke is vertical or more often oblique. The lower half slants leftward to land at the line bottom opposite the beginning of the top stroke (Oss; Eg – often

000237

Birnbaum 01.04.11 18:17

The Date of the Incomplete Isaiah Scroll from Qumraˆn

237

stops on the right). The left stroke runs down vertically to meet the end of the upper stroke so that the letter is closed (Oss; Eg – curves leftwards, so that the letter is open). Thus the outline of the letter is generally a triangle standing on one of its corners but somewhat quadrilateral forms are not rare. Ayin: The right stroke slants leftward and almost reaches the line bottom (Eg – stops at about half the line height; Oss, Uz – line height). It is somewhat bent (Eg, Oss – occasionally), sometimes it is quite straight (Oss, Uz). The left stroke is straight (Uz; Oss – generally; Eg – rarely) and meets it about the centre (Oss, Uz; Eg – at the end). Pe: The right downstroke is straight (Uz; Eg, Oss – occasionally) or just slightly right-curved (Eg, Oss). At the line bottom (Oss, Uz; Eg – much lower down) it bends at about right angles (Oss, Uz; Eg – exceptionally) leftwards, now and then it curves (Eg), to form a straight base (Oss, Uz; Eg – exceptionally) of considerable length (Oss, Uz; Eg – exceptionally). The left stroke is of medium length, extending to about the middle of the line height (Oss; Eg – short), slants down leftward and is slightly bent (Oss – often; Eg – rarely; Uz – strongly curved). Final pe: The main stroke is appreciably less than double the line height (Eg, Oss); it is a right-curve (Oss – sometimes; Eg – top slightly curved, rest straight). The left stroke is generally curved (Eg, Oss); it extends to about the middle of the line height but is considerably longer than in non-final pe because the angle between the two strokes is bigger than there. Sadhe: The main stroke is straight (Oss – sometimes; Hel – bent) or a shallow curve (Oss – occasionally). At the line bottom (Oss, Hel; Eg – much lower down) it bends at about right angles (Oss, Hel; Eg – curves) leftward to form a straight base (Oss, Hel; Eg – down-curve). The base is short (Eg; Oss – short or medium long; Hel – medium). The right stroke consists of two straight or straightish arms forming roughly a right angle (Oss – often; Eg – corner rounded). The lower, left, arm is often slightly longer (Eg – short; Oss – medium to long). Final sadhe: Its length varies, ranging from somewhat more than one line height to fully two. The main stroke is wavy, a left-curve or even straightish. As to the right part, cf. sadhe. Qoph: The top stroke is an obtuse-angled or rectangular ‘tick’ whose right arm is long and straight (Oss; Eg – wavy line or down-curve). From its end a very short stroke runs down at about right angles, bending half way down, at an obtuse angle, after which there is a leftward slant (Oss – sometimes; Eg – large semi-circle); this stroke is of varying length but it never reaches the left downstroke (Oss – generally

000238

238

Birnbaum 01.04.11 18:17

[ 45 ]

very short; Eg – very rarely). The left downstroke issues from the top stroke (Eg; Oss – generally); its length varies, ranging from somewhat more than one line height to fully two (Oss; Eg – line height or very little more); it is usually very slightly wavy (Eg; Oss – sometimes). Resh: The top part is an acute-angled tick (Oss – often rectangular; Hel – rectangular; Eg – curve); the right arm is usually of medium length (Oss – varying, generally long; Hel – long). The main stroke issues from its end (Oss, Hel; Eg – top curve meets downstroke below the top) so that the letter is clearly distinct from daleth where the meeting between the top stroke and the main one is well below the top. The downstroke is often a gentle right-curve (Eg – often; Oss, Hel – straight) which makes the corner appear rounded. Shıˆn: The right stroke starts vertically (Eg, Oss – now and then) or obliquely, bends about half way down at an obtuse angle. It meets the left stroke just above the end (Oss – generally; Eg – at end). Now and then the left stroke continues rather far down (Oss – occasionally). The middle stroke, straight or practically so, but not parallel with the right one (Eg; Oss – rarely), generally reaches the left one rather far down so that it is of considerable length (Oss – often). The left stroke has just a trace of waviness. Taw: The left stroke is of line height (Oss, Uz, Hel; Eg – taller and lower); it is just slightly wavy (Oss – rarely; Eg – very wavy) or is a shallow left-curve, almost a straight stroke (Oss – generally straight; Uz, Hel – straight). At the line bottom it bends leftward to form a straight, oblique, ‘foot’ (Oss, Uz, Hel; Eg – oblique, curved, sometimes curled). From just below its top issues the middle stroke (Eg; Oss – sometimes, but more often, as also in Gez, Uz, Hel, it crosses from the left; in Oss now and then a loop has developed); it is straight (Oss – generally; Uz, Hel – sloping; Eg – curved, not an independent stroke but the beginning of the right stroke) of medium length (Gez, Oss, Uz, Hel; Eg – short). From its end, forming a rather sharp corner (Oss – often; Uz, Hel – rounded; Eg – very rarely) issues the right downstroke which is of line height (Oss, Uz, Hel; Eg – same, or shorter); it is straight or a slight right-curve (Uz, Hel; Oss – generally; Eg – often, but pronounced curves preponderate).

000239

Birnbaum 01.04.11 18:17

The Date of the Incomplete Isaiah Scroll from Qumraˆn

239

Conclusions In looking through the foregoing examination we see that as might be expected the scripts of the Gezer Stones, the Uzziah Tablet and the Helena Graffito nearly always tally with that of the Ossuaries. We shall, therefore, not deal with them separately here but simply include them in Oss. On the other hand, the letter forms of the Tobias Inscription agree either with those of both the Ossuaries and the Egyptian documents, or only with the former, or they differ from both. They thus afford us little help towards arriving at a result. We have accordingly the straightforward task of establishing the relation between IsB on the one hand and Eg and Oss on the other. In examining the 27 letters of the alphabet we utilized 164 different features. Out of these, 48 were common to IsB, Eg and Oss. This is natural since the interval between the two periods is, after all, not more than two to three centuries. Seventeen features (in aleph, heth, kaph, ˙ taw) are final kaph, lamedh, mem, final mem, samekh, pe, resh, shıˆn, common to IsB and Eg but not to IsB and Oss. Eighty-six features (in aleph, beth, gimel, daleth, heˆ, waw, teth, yodh, kaph, final kaph, lamedh, mem, final mem, nuˆn, samekh, ayin, pe, final pe, sadhe, final sadhe, qoph, resh, shıˆn, taw) are common to IsB and Oss but not to IsB and Eg. If we transfer this numerical proportion to a time graph beginning in the early third pre-Christian century and ending at year 68 of the first Christian century (the terminus ad quem for Qumraˆn) we arrive at somewhere round about the first quarter of the first Christian century as the time when our MS . was penned. It thus belongs to the Oss period. Since there is much Qumraˆn material of the same time we might try to check our result by comparing IsB and such MSS ., and thence by working out a relative chronology on which an absolute one could be based. For purposes of palaeographical comparison the War, the Hymns and the Genesis Apocryphon Scrolls provide a rich source. To do this work here would however lead us much too far and I will give only my result: IsB is the earliest of the four scrolls. It may be assigned to the first half or third of our period. When we attempt now to turn this relative date into an absolute one we receive important help from some extraneous evidence which considerably narrows down the limits within which our date may be placed. We know that there was a break in the settlement at Qumraˆn during the last third of the first pre-Christian century and a revived settlement at

000240

240

Birnbaum 01.04.11 18:17

[ 45 ]

the beginning of the next century. In view of the fact that so many of the Qumraˆn MSS . belong to the Oss period it might not be unreasonable to assume that many – if not the majority – were not brought from elsewhere when the settlement was re-established, but were written at Qumraˆn itself, i. e., between the years 4 (approximately) and 68 C. E. Accordingly IsB would belong to the first third or quarter of the first century C. E. This date coincides with the one we arrived at above. It can further be supported by a palaeographical comparison with the pre-Christian MSS . (the Habakkuk, the Yahad {The ‘Rule of the Community Scroll’, 4Q255} and the complete Isaiah Scrolls) which shows that IsB cannot be contemporaneous with them but must be some generations later than the latest of the three (the Habakkuk Scroll, of about the second quarter of the first pre-Christian century).

000241

46.

Birnbaum 01.04.11 18:17

Bar Kosba (Kokhba) and Akiba*

The forms of the letters in the Bar Kosba (Kokhba) document (see illustration p. 252) are already rather close to what is termed fully developed Square script. However, the differences are still of sufficient importance to lead our modern eyes astray.1

The Formula of Address From the small amount of epistolary material that has come down from the earliest times, it is clear that the forms of address varied considerably. Even the Lachish Ostraca – few as they are – contain quite a number of different types. A historical examination of all the known formulae would be well worth the effort. The method of address employed in the Bar Kosba Letter – ‘From X to Y, greeting!’ – does not occur in our oldest material. Our letter is, however, not the earliest example preserved. My colleague, I. Wartski, has drawn my attention to a passage in Midrash Tannaim. I quote his translation: He (R. Joshua [ben Hananiah]) said to him (R. Nehuniah ben Haqanah): ‘I shall ˙ my eyes saw, not what my˙ears heard (i. e., not hearsay) tell you something that [. . .] They (Rabban Simeon ben Gamaliel and Rabban Johanan ben Zakkai) said to him ( Johanan the Scribe): “Write: ‘From Simeon ben ˙Gamaliel and from ˙ Johanan ben Zakkai to our brethren [. . .] Peace! Be it known to you that [. . .].”2 ˙

Another letter is mentioned here and it begins with exactly the same formula. Simeon ben Gamaliel and Johanan ben Zakkai wrote their letter ˙ 3 Min appears many centuries about half a century before Bar Kosba. * First published in: Palestine Exploration Quarterly 86 ( Jan. – Apr. 1954), pp. 23–32. 1 The following note, though based on my own original readings, takes into account the studies by J. T. Milik {“Une lettre de Sime´on Bar Kokheba”, in:} Revue Biblique 60 (1953), pp. 276–294 and H. L. Ginsberg {“Notes on the Two Published Letters to Jeshua Ben Galgolah”, in:} Bulletin of the American Schools of Oriental Research 131 (1953), pp. 25–27. Milik’s identification of Bar Kosba and Bar Kokhba and his admirable interpretation of Õimwt and Õilbkt seem to me to be beyond doubt. 2 Ed. D. Hoffmann {Midrasch Tannaı¨m zum Deuteronomium. Vol. 2, Berlin 1909}, pp. 175–176. 3 As Wartski points out, roughly the same interval separates the Bar Kosba Letter from our next one having this formula: ‘Our teacher (i. e., R. Judah the Prince) said to R. Efes: “Write a letter in my name to our lord the Emperor

000242

Birnbaum 01.04.11 18:17

242

[ 46 ]

before our document in the Arsham Letters.4 Following a suggestion by Professor Hooke, it might be well to mention them here. Although they are not in Hebrew and not written by Jews, a connection in epistolary tradition seems possible, perhaps even likely.

Jeshua The name at the end of line 1 is spelt ywi, in contrast with the form yuwi employed in the Beth Mashku Letter. There is no room for a waw between the w and y, and no trace of the downstroke of waw is visible, although the surface of the papyrus seems to be undisturbed in this place. The little stroke to be seen on top, to the right of the ayin, cannot be the top of a waw: that letter starts with a small upstroke and then runs down vertically or with a slight slant, while our stroke is horizontal. As a glance at all the other specimens will show, it is the head of the left stroke of w.

Jeshua’s Men Is the first letter of the last word in line 2 a heth or a heˆ? The document contains no heth – unless our sign is one ˙– so that we shall have to ˙ restrict our comparisons to heˆ. Our sign corresponds to all four specimens of that letter. As in these, the top stroke is straight, horizontal and starts with a tick, the right shaft issues from its end, the left one at quite a distance from the left end. There are no features pointing to any other reading. Hence we have no reason to doubt that it is a heˆ. This is still further confirmed when we compare our sign with the heth of the contemporaneous Beth Mashku Letter. The horizontal top ˙ of our sign starts with a tick, and the left downstroke issues from a bar

4

Antoninus.” Thereupon he (R. Efes) rose and wrote: ‘From Judah the Prince to our Sovereign the Emperor Antoninus.’ He (R. Judah) took and read it, tore it up and wrote: ‘To our Sovereign and Emperor from thy servant Judah [. . .]’ (Gen. Rabbah {Midrash Bereshit Rabba}, ed. J. Theodor / Ch. Albeck, {Berlin 1903 ff., 2nd ed., Jerusalem 1965}, ch. 75, par. 5). The excellent Vatican MS . has the version: ‘tore it up and said to him: “Write: ‘From thy servant Judah to our Sovereign the Emperor Antoninus.’” A. Cowley, Aramaic Papyri {of the Fifth Century B. C., Oxford 1923}, no. 26; G. R. Driver, Aramaic Documents {of the Fifth Century. Oxford 1954}, all letters. In them the formula is only: ‘From X to Y’, without the word ‘greetings’, presumably because the letters are official communications addressed to subordinates.

000243

Birnbaum 01.04.11 18:17

Bar Kosba (Kokhba) and Akiba

243

point considerably further to the right. In heth there is no tick and the left downstroke fits on to the edge of the˙ bar. Our sign is therefore clearly not a heth. What is the˙ next letter? It has one or the other feature in common with daleth, heˆ, nuˆn, pe, resh, at various periods in their history. But it is obviously superfluous to make more than a cursory comparison with the forms these letters have in our document to see that identification with any one of them is out of the question. Only beth and kaph remain to be considered. Examining all the specimens of beth, we notice that the top stroke is generally horizontal, occasionally it slants up to the right. In our sign it slants down. In practically all the specimens it starts with a tick. In our sign it does not. The base projects considerably to the right, which it does not do in our sign. According to all these features it is not a beth. In kaph the top stroke is horizontal or, as in our sign, slopes down to the right. It has only the merest trace of an initial tick. The same feature occurs in Õklgr; in Õilbk even that hint of a tick is absent. As in the kaphs, the downstroke does not start above the top bar and the base does not jut out to the right. No doubt, our sign is a kaph. Accordingly the word is Órkh ‘the fortress’, ‘the fortified place’, as was also recognised by M. Delcor (letter of July 12, 1953, and forthcoming article in Vetus Testamentum)5 and Ginsberg. It refers, presumably, to the cave where the papyrus was found. It does not seem likely that the fortress had originally been elsewhere and that the letters were brought to the cave during the final retreat, as we know the role these caves played in guerilla warfare and have continued to do up to our time.

Were the Galileans to Be Avoided Or to Be Protected? The first word of line 4 is the crucial one of the whole document – and it happens to be the most difficult: two of the letters are ambiguous and a third appears to be missing. Is the first letter waw or yodh? There are many Yodhs for comparison in our document. A characteristic feature is that they start by slanting down to the right – although sometimes it does happen that there is a preliminary upstroke – and then bend to the left. 5

{Delcor does not seem to have published the article in question. At any rate, a suitable article is to be found neither in Vetus Testamentum nor in Delcor’s bibliography contained in his Festschrift (Me´langes bibliques et orientaux en l’honneur de M. Mathias Delcor. Kevelaer 1985). – E. T.}

000244

244

Birnbaum 01.04.11 18:17

[ 46 ]

There are fewer waws. They all have a small initial upstroke; the downstroke slopes just a very little or is vertical; there is no essential bend – at the bottom there may be a gentle movement to the left. From this it seems that our letter is a Yodh. Is the third letter a samekh? If it were, the root of the word could be gsp, dsp, xsp, isp, lsp, ssp, ysp, qsp, and some of these might possibly make sense. E. g., gsp can have the meaning ‘to break through, escape’. That – in conjunction with Milik’s and Ginsberg’s suggestion that the sentence is a conditional one – would give: ‘[if] even a single one of the Galileans [. . .] should escape’. Or we might choose dsp and translate: ‘[if] from the (number of the) Galileans . . . there should be a loss of (as much as) a single one’. But is the third letter a samekh? In the samekh of line 1, the left stroke is curved and forms a corner with the bottom end of the right stroke. In our sign it is straight and passes the bottom end of the right stroke. In our sign the top stroke starts with a tick, in the samekh of line 1 it does not. These differences speak against our sign being a samekh. The only other possibility then is a qoph. It is true the shaft is short, but an earlier form of the letter had this characteristic, and it generally takes a long time for a form to die out completely. If what we have here is a qoph – and that seems certain to me – our list of possible roots would be: dqp, xqp, Õqp, sqp, yqp, qqp, rqp, and it should be feasible to derive our word from one or the other of these. The niphal of dqp, with the sense ‘to be missing’, at once comes to mind, so that our passage would mean: ‘[if] even a single one of the Galileans [. . .] should be missing’. Other verbs would suggest some kind of connection with the idea of getting away: yqp is ‘to split, burst, break forth, escape’; rqp in the hiphil ‘to declare free, renounce ownership’. The remaining roots are not incompatible with some idiomatic expression meaning escape: qqp ‘to split, drive into, force open, make a breach’; sqp ‘to tear open, open forcibly’; Õqp ‘to prick, perforate’; xqp ‘to open’, hiphil ‘to cause to escape’. However, an external reason seems to stand in the way of accepting these verbs (as well as those of the samekh group). There appears to be insufficient space on the papyrus for a third radical. Even the narrowest of the above-mentioned letters would have some part of the extremities on its left-hand side showing on the left of the gap in the papyrus. This difficulty could perhaps be overcome if we assumed that the gap does not exactly correspond to what is torn off, i. e., that the right and left parts of the papyrus are not in exactly the same position as before the tear occurred: that they are slightly closer together than they were in

000245

Birnbaum 01.04.11 18:17

Bar Kosba (Kokhba) and Akiba

245

the original condition. Between lines 4 and 5, just below the spot where the missing letter would have been, we see a tiny strip of papyrus bridging the gap; its left end is not, like the rest of the strip, horizontal, but is vertical. If it was originally horizontal and we put it back into this position, the gap becomes by so much wider; the missing bit of papyrus might then have been wide enough to contain a narrow letter like daleth or resh. There are additional indications that the present gap is narrower than the width of the strip of papyrus that has been torn out: at the beginning of line 1 the shıˆn and the mem seem to be too close to each other, and the same holds good for the three letters g[l]g in line 2; it is hard to see how, in the first word of line 3, even a narrow final mem could have been squeezed into the width of the gap. An advantage of this interpretation would be that we would not need to regard the qoph as the last letter of the word and then have to account for the width of the distance to the mem. On the other hand, hlglg and iwnalu (disregarding the tail of the nu ˆ n from the line above) and Õda and inaw are also very far apart, so that the possibility of qoph being the final letter cannot be excluded, in which case only one root – qpn – would be open to consideration. As its meaning is ‘to go out’, our passage would read: ‘[if] even a single one of the Galileans . . . were to get away’. Hence we have reduced the number of candidates to this short list: dqÈPÄiÇ, rqÅpÂÁi, qÊ PiÇ. The meanings of all three happen to fit equally well into our passage and, as a matter of fact, they mean the same in the present context. Which of them would appear to be preferable? qpn, as an Aramaism, would not be surprising, and that our earliest example of it occurs in the Midrashic literature would not speak against such a derivation. However, the word is not a frequent one in our PiÇ. Hebrew sources and this does not favour the reading qÊ rqp is common from Mishnaic times onwards, but it does not seem ever to have been used in the sense of setting free; it is, of course, possible to suggest the semantic development ‘set free’ from the meaning ‘abandon’6 but it seems rather unlikely that among an abundance of examples this sense alone should happen not to be represented anywhere. Thus we are left with dqp. Unless we see a difficulty in the fact that the meaning ‘to miss’ (qal), ‘to be missing’ (niphal) is Biblical usage, it seems a reasonable and attractive possibility. The reading would then be: ‘[if] (even) a single one of the Galileans [. . .] should be missing’.

6

Or one of its derivative meanings: ‘make ownerless – declare ownerless – renounce ownership of [. . .]’.

000246

Birnbaum 01.04.11 18:17

246

[ 46 ]

To sum up: Bar Kosba did not desire to keep Jeshua ben Galgola away from the Galileans, nor did he order him to do them no harm. What he appears to have had in mind was that Jeshua ben Galgola should guard them well so that none of them should get away.7

Who Rescued the Galileans? The final signs on line 4 are somewhat difficult to read. Is the one before the last – consisting of a semi-circle and a shaft – a taw? The shaft protrudes below the line bottom. None of the three taws in our document has this feature. At the bottom this stroke looks as if it turned horizontally to the left, but my impression from the photo is that this is only the shadow, seen through a tear at the edge of the papyrus. If it really were a stroke, it would be longer than the analogous strokes in the taws. The top of the downstroke is straight while in taw it is bent. The sign is wide, so that its inside has roughly a circular shape. The taws are narrower – the one in ñtn is very narrow – with the effect that the inside is more or less an oblong. The right stroke of our sign is almost a semi-circle. In the taws we have a generally straight, or slightly curved, downstroke, connected by a straight bridge with the left stroke, the two right ones forming a very acute angle. This comparison indicates that our sign is not a taw. Could it be a samekh? The right part of our figure is three-quarters of a circle, in samekh it forms, together with the left stroke, a triangle. This does not favour the reading samekh. Could the letter in question be a final mem? In that case it would be a quadrilateral, looking like a somewhat angular Roman D. Our sign, however, resembles a Roman P with a shortened shaft. Hence identification as final mem seems hardly possible. Before we try to solve this problem, we shall examine the subsequent sign. Is it a yodh? The ‘head’ is turned to the right. In the yodhs a head is rare; when there is one, it is turned in the opposite direction. So also is the plain bend which is more prevalent than the head. This speaks against the reading yodh. Could our sign be a nuˆn? If what appears to be a bottom stroke were only a shadow (see above), identification as nuˆn would be quite impossible, of course. If it were a real stroke it would slant upwards. In the nuˆns the base slopes down, and is longer, too. Hence, our sign does not seem to be a nuˆn. 7

This would fit in well with Milik’s suggestion that these Galileans were ( Jewish) Christians.

000247

Birnbaum 01.04.11 18:17

Bar Kosba (Kokhba) and Akiba

247

Could it be a zayin? Not if what looks like a stroke at the bottom should really be one. But if it were a shadow, then a similarity with the zayins of the Beth Mashku Letter would exist. However, the top is much too pronounced to make this identification likely. Having tried all the letters looking sufficiently similar to warrant examination, and having had to reject them, we can only conclude that we are on the wrong track. This would mean that the strokes which seemed to us to belong together must be combined differently if they are to make sense. It is not difficult to see why we have been misled: all the strokes run into each other because the writer had reached the edge of the papyrus and had to squeeze in the end of the word. There seems to be only one other possible way of connecting the strokes: the straight downstroke in the middle of this tangle does not belong to the curve on the right but to that on the left. What remains on the right is thus a kind of semi-circle, somewhat angularised. Looking through all the letters, we find one, of which most specimens tally with our sign. That letter is kaph. There is an additional detail which fits in with it. A magnifying glass reveals that what appeared to be the top of the big downstroke is slightly narrower than the rest of the stroke. That evidently means it does not form part of it all, but belongs to the kaph, i. e., to its initial tick, which is a feature of many of the kaphs. The last letter is a narrow triangle tapering downwards. Such a description would apply to samekh (apart from the narrowness, which would be due to the accidental lack of space), but in our sign the main downstroke is on the right, in samekh it is on the left. Thus it is not a samekh that we have here. There is only one other letter which is closed on all sides. That is final mem. But it is a quadrilateral and our sign is a triangle. Let us compare the strokes. All the specimens have a horizontal top stroke, apart from that in Õimw. The left stroke slants down to the right. Both features are present in our sign, too. In mem, the lower third of the right stroke turns left, while in our sign there appears to be only one straight, vertical line. Closer inspection, however, shows that this is not so. Lack of space forced the writer to make the angle of the bend even more obtuse than usual, so that it has been flattened out to almost 180 degrees – but the bend is none the less indicated, thus removing any doubt that our letter is final mem. Accordingly our word is Õklchw. To read this as ÕkÃlÂCÇhÇwà seems out of the question. lCÇhÇ is a singular, and as it is connected with Õiallg by the relative w, the subject contained in lCÇhÇ cannot be identical with the

248

[ 46 ]

plural Õiallg. What the writer presumably intended was ÕkÃlÂcÂhÃ, meaning ÕkÃlÂcÂaÃ. This was either a slip, pure and simple, or, more likely, a spelling mistake based on pronunciation. The tendency to confuse the glottal stop and /h/ is, of course, a common feature in old and modern languages. We find it represented in the Cave MSS ., too: Isaiah A has hknkh instead of hknka (MT Ónka, 45: 4) and conversely sda, ulilia instead of sdh (53: 13), ulilih (23: 1); the Covenant Scroll has iwnh (13: 8) instead of iwna, and conversely, yipua (2: 10) instead of yipuh. Thus the passage in question means: ‘the Galileans who are at your place’.

Who Put Ben Aflul Into Irons? Bar Kosba’s threat to do to Jeshua ben Galgola what had been done to Ben Aflul, seems to suggest that it was Bar Kosba himself who had meted out the punishment in question. This has led to the reading ‘as I did to Ben Aflul’. Let us, however, take a look at the last, and crucial, word of line 6. The third letter in it is not a samekh, for the reasons mentioned when we discussed the identical sign in line 4. Here, in line 6, an additional difference may be seen: in this specimen the top stroke still has an old feature – the upward slant of its left half. The subsequent letter is a yodh, like that of iwna (line 2) or diym (line 3). The last letter is clearly a beth, all the left part of whose base has been torn off, and that, possibly, together with one more letter – but with certainly not more than one. Our reading is thus [–]biqyw. We can hardly doubt that this is a reference to the great Rabbi Akiba. We had known that he looked upon Bar Kosba as the Messiah, applying to him the messianic passage bqyim bkuk Órd (Num 24: 17), and that he was possibly the author of the name ‘Bar Kokhba’. But from the three final words of this short Letter we now learn a great deal more. They give us information on a point that has long been the subject of controversy: Did Akiba actively participate in the War or not? Here he is revealed as having done so. The historical interest of this letter is therefore even greater than had been thought. lulpy ñbl [a] biqyw hmk sounds somewhat abrupt but in no wise surprising in a hastily written letter. On the other hand, although we have learnt by experience that the difficulties which exist for us are often not shared by a writer’s contemporaries, it may be permissible to attempt an emendation. Perhaps the present difficulty is only due to a haplography for [a] biqy hwyw hmk, since both words start with ayin.

000249

Birnbaum 01.04.11 18:17

Bar Kosba (Kokhba) and Akiba

249

Why was the end of line 6 left vacant and the word ñbl transferred to line 7? There would have been ample room for it. And why does line 7 not start below line 6 but slightly more to the right? And why is the distance between the two lines so much smaller than between the other lines? It is true that the line distances 2 – 3 – 4 are also small, but that is only because line 2, from the middle onwards, droops very considerably; at the beginnings of the lines the distances between them are wide and regular. These three questions would be answered if we assumed that the letter originally ended with the word [a] biqy, either through a slip, or because this had actually been the first wording – and that lulpy ñbl was added as a rectification or as an afterthought, when it was realised that (hwy) w hmk [a] biqy might not make the reference sufficiently clear.

Prince of Israel On the basis of other documents, as yet unpublished, Milik suggests that the signature might possibly have been followed by the title aiwn / aisn larwi. If our letter had contained these words they could hardly have been written in line 8, straight after Kosba. Even a very crowded arrangement, as in line 4, would have brought the final letters to the edge of the papyrus – and this edge is preserved and it is empty. Thus fifteen letters – even omitting the intervals between the four words – would have had to be compressed into the part which is torn off. And that is not possible. The corresponding space in lines 1– 6 contains 10–13 units (i. e., letters or intervals between the words) each. The title would therefore have been written in the subsequent line. What is left of the papyrus under the name makes it unlikely (though not impossible) that there was writing there, because we would expect the top of it to be just showing at the bottom edge of the papyrus. The title would thus have had to be more to the left (of the hypothetical line 9). On the whole, it seems rather unlikely that in this document the name was followed by the title.

Who Penned This Letter? If we had only the body of this letter we would have no means of deciding whether Bar Kosba himself wrote it or whether he had it written for him. Fortunately, we have a few letters of the signature. Let us compare them with their counterparts in the body of the document. Mem: Only part of the letter is preserved, and, add to that, the lower portion of the long stroke apparent on the photo seems to be just a tear

000250

Birnbaum 01.04.11 18:17

250

[ 46 ]

in the papyrus. But it is possible to compare the top left portion of the letters: they differ completely. Ayin: In the text, the strokes have heads, in the signature they have not. Waw resembles the yodh rather than the waw of the text. Final nuˆn is very short and almost straight. In the text it is very long and wavy. With these letters differing so essentially it seems hardly possible that the text and signature were written by the same hand. Hence I believe the body of the letter was not written by Bar Kosba himself. The writer drew his letters irregularly. In addition, his lines are not straight. Comparison with the writing of the Biblical fragment from about the same time8 makes it clear that Bar Kosba’s secretary could hardly have been a professional scribe, nor a man who did a great deal of writing. On the other hand, he is perfectly literate. Vacillation in the use of matres lectionis and the omission of aleph in ta are of no great importance. Even the rigidly standardised spelling systems of modern languages allow some variant spellings, and many more do actually occur in the writing of even educated people.

The Import of the Letter Let us summarise what the discovery of this document means to us. The letter emanates from Bar Kosba himself, the leader in the last struggle of Judaea against Rome, and is a direct piece of contemporary evidence in connection with that War, giving us a glimpse of an episode in which Jewish Christians possibly play a role. It contains a reference by a contemporary to Rabbi Akiba, the leading figure of his time, and one of those responsible for the shaping of historical Judaism. It tells us of his active participation in the War. It acquaints us with the original form of the name on which the Bar Kozba / Kozeba of our sources is based. It confirms that Simeon was his first name, which till now had only been inferred from the name Simeon on certain coins. It adds two new names to our onomasticon for the period. It shows that Hebrew was employed for at least official letters, and illustrates the influence of Aramaic on this Hebrew. It is one of our few first Hebrew letters in the original not written on ostraca.9 8 9

Revue Biblique 60 (1953), pl. XII , fig. a {also reproduced above in Chapter 34, fig. 13}. As to the Beth Mashku Letter, see PEQ 87 (1955), pp. 21–33 {= Chapter 48 in this volume}.

000251

Birnbaum 01.04.11 18:17

Bar Kosba (Kokhba) and Akiba

251

It demonstrates the use of papyrus for letter writing in second century Palestine, and gives us an initial formula employed at that time. It shows us Bar Kosba’s handwriting, albeit in a fragmentary way. It provides a specimen of a private – i. e., non-scribal – hand of certain date. It is a human document of extraordinary interest that has come to us, in the original handwriting, from antiquity. The Text ywil hbsuk ñb ñuymwm Órkh iwnalu hlg [l] g ñb Õimwtily ina diym [Õ] ulw Õklchw Õiallgh ñm [d] qpi Õilbktñtn inaw Õda lk [a] biqyw hmk Õklgrb lulpy ñbl [hbsuk ]bë ñuymë [w]

1 2 3 4 5 6 7 8

Tentative Translation 1 2 3 4 5 6 7 8

From Simeon ben Kosba to Jeshua ben Ga[l]gola and the men of the fortress, Gree[tings]. I call Heaven to witness against me 〈that if〉 of the Galileans, who are at your place, there should be missing (even) a single one, I will put fetters on your feet, as Akiba 〈did〉 to Ben Aflul. [Si]m eon b [en Kosba]

000252

252

Birnbaum 01.04.11 18:17

[ 46 ]

A letter from Simeon Bar Kosba

000253

47.

Birnbaum 01.04.11 18:17

Akiba and Bar Kosba* (Kokhba)

In my article1 about Mur 43 I read the last word of line 6 as [a]biqy. The facsimile accompanying the publication,2 being a half-tone and not enlarged, was not distinct, the word ending as a smudge. I therefore give here an enlargement of this word (fig. 1), and hope it will be distinct enough even in the half-tone facsimile. To make sure, for the purposes of comparison, accompanying chart (fig. 2) shows copies of all examples in the document of the letters under discussion, including a tracing – which I made from the original enlarged photo – of the final bit of the word (left side of the fourth line of the chart, marked ✜). Let us examine the individual characters of this word. The third letter is, at first glance, ambiguous – is it a samekh or a qoph? The top stroke starts as a tick, unlike the samekh of hbsuk (line 1). The downstroke issues slightly to the right of the tick, while in the samekh of line 1 the downstroke is completely outside the letter. The downstroke continues to a point slightly below the end of the right stroke, while in the samekh of line 1 the downstroke and the right stroke end by meeting. Hence, what we have here is a qoph. It is true the downstroke is very short but this feature is to be met with at various periods. ( The final kaph at the end of line 2 is also hardly longer than the preceding resh.) Accordingly, the third letter of line 4 is also a qoph. The next letter (of the final word) is a yodh, like those in ywi (line 1), ily ina diym (line 3). The last letter is a beth, tallying with all five other specimens in this papyrus. These two letters – yodh beth – have been read as a single one: taw. Analysis of the final complex, i. e. comparison with the undoubted specimens of taw in this text – and due regard for the little gap inside that complex, would have obviated the error. As a matter of fact there are many examples here of letters touching each other, or of almost doing so: the lamedh heˆ of hlglg, the shıˆn yodh of iwna, the resh final kaph of Órkh (line 2), the shıˆn lamedh waw of * First published in: Palestine Exploration Quarterly 100 ( July – Dec. 1968), pp. 137–138. 1 PEQ 86 (1954), pp. 23–32 {= Chapter 46 in the present volume. – See also The Hebrew Scripts by S. A. Birnbaum, 2 vols., London 1954–1957 / Leiden 1971, no. 102.} 2 {See p. 252 in the present volume.}

000254

254

Birnbaum 01.04.11 18:17

[ 47 ]

Õulw, the ayin yodh of diym, all the letters of Õimw (line 3), the pe qoph of qpi, all the letters of sklch (except between heˆ and sadhe) (line 4), the daleth final mem of Õda, the kaph beth lamedh of Õil˙bk (line 5), all the letters of hmk (line 6), the waw lamedh of lulpy (line 7). But even if we were to take what is the beginning of the yodh and the top stroke of the beth as a single, uninterrupted stroke, we would still not get a taw: (1) the little horn under it, protruding upwards to the right, could not be explained as part of a taw, (2) the alleged form (chart, line 5, left) differs essentially from that of the three taws occurring here: they have no initial tick – the piece on the left is the top bit of the downstroke, the top stroke runs upwards to the right, not downwards as in the alleged form, (3) the latter does not occur before the Middle Ages (and is not a frequent one) – moreover it is then restricted to certain Jewries. The fact that the spelling [i]tsy instead of [i]t[i]wy would be incorrect would not necessarily rule out the reading of the letter as samekh, since our text, as well as the other texts of these finds, contain many ‘irregular’ spellings. On the other hand, it certainly does not favour the reading hwy. It is unfortunate that the last letter happens to be missing. If a heˆ3 or an aleph had been preserved, the wrong reading might not have arisen.

Fig. 1. Enlarged word from the Bar Kosba document (Mur 43), line 6

3

More likely, in view of hbsuk and hlglg.

000255

Birnbaum 01.04.11 18:17

Akiba and Bar Kosba (Kokhba)

Fig. 2. Chart of selected letters from the Bar Kosba document {✽ ✽✽ ✜ ✜✜

Last letter of line 6: beth (left bottom broken off) Penultimate letter of line 6: yodh Last two letters of line 6: yodh beth Last two letters of line 6: misread by some scholars as a taw. }

255

000256

Birnbaum 01.04.11 18:17

000257

48.

Birnbaum 01.04.11 18:17

The Beth Mashku Document*

The initial sentence of Bar Kosba and Akiba 1 applies also to this document: the forms of the letters are already rather close to what is termed fully developed Square script. However, the differences are still of sufficient importance to lead our modern eyes astray.2

The Introductory Formula The first word occurs three times in the initial line (see illustration p. 272). Does it start with a K or M? In those mems of our document about whose identity there is no question, the top stroke is a tick. From its bottom end a stroke slants down to the left. When the base is horizontal and not too short, that slanting line runs rather near to the end of the base, the effect being that the letter is nearly closed. While the rest of the letter under examination more or less tallies with this description, one essential stroke is absent – the slanting line which comes down from the tick. This makes the reading M impossible. We cannot assume that this crucial stroke was accidentally effaced in all three identical words in this line and survived in ukwm, in the same line. Let us now look at the kaphs of our document. They consist of a top tick, a downstroke and a base, i. e., they are identical with our sign, which is thus an undoubted K. The suggested vowelling ka¯n – reported by Albright from the Jerusalem Press3 – is based on legal usage. On the other hand, our papyrus is not simply a legal document, it is a letter as well. It does not seem feasible to make the two lines independent of each other by taking line * First published in: Palestine Exploration Quarterly 87 (April 1955), pp. 21–33. 1 PEQ 86 (1954), pp. 23–32 {= Chapter 46 in the present volume}. 2 The present article is based on my own readings but reference is made to the excellent studies by R. de Vaux, “Quelques textes he´breux de Murabba at”, in: Revue Biblique 60 (1953), pp. 269–275, J. Rabinowitz, “A Hebrew Letter of the Second Century from Beth Mashko”, in: Bulletin of the American Schools of Oriental Research 131 (1953), pp. 21–24 and H. L. Ginsberg, “Notes on the Two Published Letters to Jeshua Ben Galgolah”, ibid., pp. 25–27, as well as to the joint article by O. H. Lehmann and S. M. Stern, “A Legal Certificate from Bar Kochba’s Days”, in: Vetus Testamentum 3 (1953), pp. 391–396. 3 BASOR 131 (1953), p. 24 {editorial note}.

000258

Birnbaum 01.04.11 18:17

258

[ 48 ]

1 as the beginning of the legal contents and line 2 as the initial formula for the epistolary section. Hence, in our document, line 1 is parallel to the first line of the Bar Kosba Letter which reads: hbsuk ñb ñuymw ñm. Having made up our mind on this point we are however still not in a position to vowel the word with certainty – it might be ka¯n or ken; in either case the word ‘say’, or something similar would I think be understood: ‘Here/Thus (say) the parna¯sı¯n’.

Parna¯sı¯n Is this word written with S as in the Talmud and Midrash or with a S˙? Enough of the letter in question has survived to enable us to decide. The left stroke is sufficient for the purpose. On top there is clearly a trace of a tick or a head. Not one of the s˙¯ıns has anything resembling it. On the contrary, they all bend or curl to the right. Our sign is therefore not S˙. On the other hand, comparing it with the S in line 3, we find that the surviving strokes are in conformity with the corresponding ones in this specimen. Hence we have no new spelling of the word here.

Aramaic or Hebrew? It has been suggested that the commander’s title should be read wir not wur. There is, however, throughout this document no difficulty in distinguishing between W and Y. W is always bigger, Y is never of line height. In our word the ink has flaked off from the lower part of the stroke but its bottom end is preserved. Hence what we have here is an undoubted W and the title is in Hebrew, not in Aramaic. By the way, if the whole title had been taken over bodily from Aramaic, then we would expect the second part of it to be in that language, too. It is, however, in Hebrew – hinxmh – like all the rest of the message.

The Men The end of the fifth word in line 2 is slightly effaced but enough is visible to enable us to establish the reading. There are many heˆs in the document and not one of them is a symmetrical figure, i. e., with a tick on both ends of the top bar. In the 13 distinct heˆs of the text and 8 clear heˆs of the signatures, the top stroke ends at the right shaft. The shaft generally starts a little above the top. What we see on the right of our sign does not, therefore, belong to it but must be something else. The

000259

Birnbaum 01.04.11 18:17

The Beth Mashku Document

259

small vacant space between the shaft and the dot on the right was always vacant, i. e., it is not a spot from which the ink has come off, connecting the dot with the shaft on the left and the little stroke on the right. The dot was originally connected with the top of the little stroke, being the end of an upstroke to it. In other words, what we have here is a Y, e. g., like the third letter from the end of the line. This spelling is an early example of seghol being rendered by Y as a mater lectionis. There is no reason why we should discard de Vaux’s reading, as has been suggested.

Imperfect or Perfect? Is the second letter after s˙hprh in line 3 a Y? The yodhs of our document consist of a little upstroke and a straight downstroke. The letter in question would have to be that downstroke. It is not, however, straight but curved. Hence the reading Y cannot be right. There is only one letter of which it can be a remnant: L. What is effaced is not the upstroke of Y but the top and middle strokes of L. Thus our word is xqlw ‘which he bought’. This makes the awkward rendering of the verb in the present tense (or in the future tense, according to another suggestion) unnecessary. The imperfect expresses incompleteness. From the context, however, there can be no doubt that the cow had already changed hands.

Attestation? In this document the main difference between W and Y is in their size: W is longer. Accordingly in line 4 the second letter of the third word is Y, followed by W. The last letter of this word cannot be R because of the small horizontal stroke issuing from it at the line bottom; this is clearly part of a base – and the only letter with a downstroke and a base projecting to the right is B. De Vaux’s reading as bwuiw ‘who dwells’ is thus confirmed.

byt Yalon was doubtless right when he equated tiba in the Habakkuk Scroll with the same form in Tosephta and Baraitha, pointing to J. N. Epstein’s hnwmh xsunl aubm,4 p. 1258 f. This de Vaux applied to the tiba of our document (line 4). What were the vowels of this word, and how did it 4

{Mavo le-nusah ha-Mishna. Jerusalem 1948.} ˙

000260

Birnbaum 01.04.11 18:17

260

[ 48 ]

develop from tibb? Perhaps the evolution ran on something like these lines: bəbe¯t > *bbe¯t > *bbe¯t. Now, gemination did not occur in an initial ¯ doubled ¯ ¯ ¯ position.¯ The consonant was therefore transferred to a medial position by the creation of a new first syllable, to which the first half of the long b would belong. As a vocalic beginning was not possible in Semitic, the syllable had to begin with an aleph: *bbe¯t > əbbe¯t or ¯ ¯ the general abbe¯t.5 This avoidance of initial gemination is in line with ¯ avoidance of an initial consonantal cluster: instead of two different consonants there is a doubled one here. Epstein drew attention to the Phoenician forms wdqmba and iba ixba, which occur side by side with wdqmb and iba ixb.6 He was perhaps thinking only of the prosthetic aleph before the preposition b, but this aleph is, of course, frequent before many consonants in Phoenician and in the other Semitic languages.7

One Word or Two? Do the last six letters of line 4 constitute one word or two? At first glance the interval between the second letter and the third suggests two, especially as the signs tend to get crowded at the end of the line. However, in case this interval happens to be accidental, we shall examine the individual letters. The first sign is almost closed by a slanting stroke which issues from the top tick, and is clearly a M. The other specimens of that letter confirm this identification. The base does not project to the right. The dark patch on the right does not belong to any of the strokes but to the texture of the papyrus. It has been suggested that the subsequent sign is a Z, superimposed onto the lower half of the long-tailed final N from the line above. However not all the downstrokes of this writer’s finals are of such extraordinary length. The nuˆns in line 1 and the kaph in line 3 do not run beyond the ceiling of the subsequent line. Careful inspection of what is supposed to be the lower half of our final nuˆn shows that it is slightly further to the left than the alleged upper half. In other words, the two do not form a continuous line. If they did, the stroke would unaccountably have changed its direction, and that just at the ceiling of the lower line. It is most unlikely 5

6 7

In looking through Epstein’s list, I find some examples of a spelling which seem to confirm this suggestion: tibba. This is presumably a transition form – the original tibb is preserved but the new pronunciation is indicated by the aleph. i. e., iixb. Carl Brockelmann, Grundriss {der vergleichenden Grammatik der semitischen Sprachen. Berlin 1908–1913}, § 82. An analogous device is found elsewhere, e. g., spiritus > espirito, esprit, etc.

000261

Birnbaum 01.04.11 18:17

The Beth Mashku Document

261

that the writer wrote a Z straight onto the tail of the final nuˆn. If he had done so it would have caused a thickening of the underlying stroke, but not the slightest indication of that is apparent. Elsewhere he clearly avoided doing such a thing: in line 7 the left stroke of the first T is coterminous with the downstroke of the final K above it, but it keeps its separate entity; at the end of the same line care is taken to write the aleph and L to the right and left of the tail of the final N from the line above. In addition, the head of the Z at the end of line 1 is horizontal, but the triangle we have here at the line ceiling, is caused by a narrow curve like that in all the specimens of final N. Our sign is therefore not a Z but a final N (also de Vaux’s reading). It is, in fact, a perfect replica of the letter above it. Our word is accordingly ñm, and it remains for us to deal with tunb. ulw follows straight after the name and description of Jacob ben Judah, the former owner of the cow. The relative clause tunb ñm ulw ihw is therefore to be regarded as part of the description, i. e., it refers to him, not to Jehoseph ben Ariston whose name is further away. This clause seems to specify how Jacob ben Judah originally came to own the cow, and that would be contained in the words tunb ñm. On the face of it, there does not perhaps seem to be much likelihood of this being simply the word ‘daughters’. If it referred to his own daughters, where is the suffix? Nor is it very likely to be a construct with the name of the father accidentally omitted. For we would expect some qualifying designation, as in the case of the other persons (apart from Jehoseph ben Ariston, about whom see below, p. 264): ‘the parna¯sı¯n of Beth Mashku’, ‘chief of the camp’, ‘resident of Beth Mashku’, ‘witness’, etc. Could ñm tunb perhaps be a specific term denoting a certain kind of inheritance through the female line? I am rather inclined to doubt it.8

Zayin or Nuˆn? The reading tuzb instead of tunb has been suggested. Z (line 1) is straight. Our sign, however, is a curve; it has moreover an unmistakable base. Hence Z is out of the question. One letter tallies with the nuˆns in the document. 8

My colleague, I. Wartski, suggests the reading min be¯no¯t. Adducing bici ñm (Dan 2: 8) and other adverbial expressions with ñm, he translates ‘meanwhile, for the time being’. He thinks the clause refers to Jehoseph ben Ariston. Be¯no¯t is an attractive reading but how is one to work it into the interpretation of the document as a whole?

000262

Birnbaum 01.04.11 18:17

262

[ 48 ]

The Heathens In the third word of line 5, the two signs before the last are identical. If therefore the first were to be read W, the second could not be Y (and we would have Õuug which is impossible). The signs are too short and small for W, and far too curved. What we have here is simply a defective spelling as far as the o is concerned and a plene one for i, as in Gen 25: 23 and Ps 79: 10, resulting in Õiig.

Azay or Ani ? In line 5 in the second word before the last the middle letter is not a straight line as in Z (line 1), but a curve as in N. It has not a horizontal head like Z but a pointed top as in some of the nuˆns. The base is effaced like so many other strokes, or parts of strokes, in this document. Our word is ina. It makes good sense: ‘I myself would have come up.’

Imperative or Not? It has been suggested that the last four letters in line 5 constitute an imperative. That cannot be the case for the simple reason that the S has ˙ not the final form (as de Vaux has already pointed out), being identical with the one in ñujcra. If it were a final, the downstroke would no doubt be as long as in the final N and K of our document. Nor can cphu be taken as an imperative on the assumption that it is in the plural; a W would then have to follow the S – for which there would be enough room; but the space is vacant. ˙ These four letters are the beginning of a word. The writer does not, by means of careful spacing, attempt to avoid a gap at the end of a line: when he arrives at the margin without having reached the end of his word, he abandons the incomplete word and starts it anew in the next line. This system is familiar to us from the Cave MSS . and from the Middle Ages. (Both de Vaux and H. L. Ginsberg realised the true nature of these redundant letters in the present manuscript.)

Ps, s or Psr ? ˙ ˙ ˙

The third letter in line 6 cannot be an ayin. Its base is short and horizontal while that of ayin is long and slanting. In all the specimens of this letter the left stroke slants to the right, in our sign it slants to the left and even crosses the base, a thing that does not happen in any of the ayins. That stroke is simply the end of the final P in the line above (as

000263

Birnbaum 01.04.11 18:17

The Beth Mashku Document

263

de Vaux has pointed out). Thus the meaning ‘advice’ – which would not make good sense, anyhow – is excluded. The base of the S apparently ˙ have been has a small oblique stroke on the left; in later times this could a flourish, but it would be too early for that here. Careful scrutiny shows that it touches the base but is not the continuation of it. The only possible connection can be with the little stroke somewhat further to the left, at the line ceiling; the link between them would then have been by means of a very shallow curve. Together with the shaft on the left, to which the curve would have been joined by the little top stroke, they would form a T. This is also de Vaux’s reading. The traces on the left of S ˙ can in no way be combined to form a R. There is simply no room here for any letter except Y – and the traces cannot form part of a Y. On the other hand, this shaft cannot, together with the stroke on its left, form a T: there is no bridge between the two; the left stroke would have to be taller, there is no room on the left of its bottom end for a longish horizontal foot, even if we assumed that it is only effaced here. De Vaux’s reading Óitc is beyond doubt, and the reading Óitrcph consequently out of the question. Ñip or icp?

The sense ‘libe´rer, de´gager’ does not seem to offer an attractive solution to the problem. The help of one villager – or even of several (if we assume that he included the others as a matter of course) – would hardly have meant much to an organised military force. If we knew whether hkk ly refers to the legal contents of the preceding text or to the words about the heathens, it would be easier to explain the verb Óitcphu. At first glance either is possible. However, let us consider the following. If hkk ly refers to the heathens, then Õibrq would probably have to be read qəre¯bı¯m ‘are approaching’ because the reference would obviously be to enemy movements, in a report to a military commander like Jeshua. Now, if the purpose of this sentence was to inform him of the approach of the enemy, two things would not be easy to understand: why would so vital a piece of intelligence have been put in the protasis of a conditional sentence, and why, in fact, would the whole sentence have been placed at the end of the document? It would surely have deserved pride of place. Finally, if it had been a question of giving a piece of intelligence, the root ydi with its precise meaning would presumably have been used, as in line 2. For these reasons it seems to me preferable to regard hkk ly as relating to lines 3– 4. There remains the question as to why it should have been necessary to inform Jeshua ben Galgola about this business transaction? We would

000264

264

Birnbaum 01.04.11 18:17

[ 48 ]

have to assume that he was not only in command of the garrison but that he was head of the civil administration for the district as well. De Vaux does think so, pointing out that such a thing would be normal at a time of armed revolt. Rabinowitz’s explanation – Jacob ben Judah was afraid the military authorities might confiscate the cow – is ingenious but would still not explain why this matter should dominate, as it does, a document which, according to his theory is in the first place a situation report to the military commander. From all that it seems that Óitcphu is unlikely to have anything to do with the idea of information. Can the root icp perhaps solve our problem? Its piel means ‘to compensate’, and assuming that its hiph il might – as in so many verbs – denote the same as the piel, the meaning might be ‘to compensate’. But why should it have been necessary to compensate the garrison for the sale of somebody else’s cow, and why should the writer have had to go in person and report the transaction to the commander? We shall discard this root. If we adopt Ginsberg’s etymology (Ñip < sip), which has independently been mentioned to me by Wartski, we might try another sense of the hiph il of sip – that of ‘arbitrating, deciding’. The meaning would then have been: ‘I would have left the decision to you’. But why should the military authorities be concerned with a private business transaction between some peasants? Looking for a possible reason we might perhaps think of an illegal action: Is xql by any chance not ‘to buy’ but ‘to take away illegally’? I do not think so. For that type of acquisition there are the legal and colloquial terms lzg and bng. Since xql need not necessarily mean ‘to buy’ could it here signify ‘to take away a thing without having paid’? Perhaps this was a case of requisitioning, Jacob ben Judah’s cow having been requisitioned by Jehoseph ben Ariston? As the matter is dealt with in a document-cumletter addressed to Jeshua ben Galgola, it might be inferred that Jehoseph ben Ariston was one of Jeshua ben Galgola’s men and had acted on his orders. That would also explain why his signature does not appear on the papyrus – at the time of writing he is again with Jeshua. How would this interpretation affect our derivation and rendering of Óitcph? If xql here means ‘to requisition’ it would be quite natural to connect Óitcph with the idea of compensation. We mentioned above that the pi el of the root icp means ‘to compensate’. We shall therefore go back to that root. Its hiph il should mean ‘to cause to compensate, ask for compensation, arrange or discuss compensation’. Our earliest occurrence of the sense ‘compensate’ seems to be in B. Kethub. fol. 91 b.

000265

Birnbaum 01.04.11 18:17

The Beth Mashku Document

265

The names Rami bar Hama and Raba place this passage into approxi˙ of the 4th century or slightly earlier. That is mately the second quarter only two centuries later than our document, and it does not seem out of the question that we should meet with that sense already here, in 132–135. If so, our derivation would be possible.

Omer or Amur? In line 6 it would be more convenient to read rima, the active participle (with plene writing of the second vowel), rather than ruma, the passive participle. But the sign between M and R is too big for a Y so that we have to decide in favour of the lectio difficilior, the passive voice.

Castra? In the last word of line 6, the first letter cannot be K. In those kaphs about which there is no doubt, the downstroke and base form an angle, written in one continuous movement. In our sign the base juts out slightly to the right. This form occurs in some specimens of B. What we have here is the root bsr suggested by Albright, and independently of him, by Wartski. The mater lectionis is a W. ( This was recognised also by Rabinowitz and Wartski.)

First or Second Person? According to the excellent suggestion of de Vaux the aleph of hiha is prosthetic. The following reasons seem to me to speak for this explanation. It would be rather striking for a writer to conclude his letter with good wishes for himself. It would be even more striking if he did not add any for the addressee to whom he has written in a deferential manner. Another consideration in favour of taking hiha as the imperative may be this. Could not this perhaps be a similar case to tiba, discussed above (p. 259 f.) – similar in that the aleph would have the function of preserving the consonantal cluster at the beginning of the word. For we know that in those times the /h/ was liable to disappear. Thus our imperative / hje / was prevented from becoming / je/ by a prosthetic aleph: / əhje / or / ahje /. (As in tiba, this device would presumably have been used instictively.) Although it is of course not impossible to regard hiha as in the first person, the claims of the second person seem to be stronger.

000266

Birnbaum 01.04.11 18:17

266

[ 48 ]

The Signatures Everyone of these six country folk could write although they were clearly not skilled in handling the pen. The latter does not, however, explain the many striking features in the forms of their letters. A detailed examination leads to an unexpected result. Some of the letters seem to be at a different stage of development from those in the text part of the document. Others are not what they seem to be. E. g., the second letter of line 13 is an ayin but looks like a D or R – and how could an ayin be confused with a D / R? Looking at these unusual letters, it struck me that they bore a striking resemblance to those in a fragment I had deciphered some months previously.9 Our chart10 – showing the alphabet of that Murabba at fragment11 – may enable the reader to compare the forms there with those written by the signatories. We shall go through all the non-Hebrew forms in the signatures. B: While in the body of the document (i. e., lines 1–7) the base projects only very little to the right, this is very pronounced in lines 8, 10, 13. In lines 8 and 13, to be precise one would have to speak of the whole of the base as projecting, because the leftward slant of the downstroke is so great that it finally joins the left end of the base. The letter tallies with that of the chart, except that in line 8 the top stroke starts as a tick12 as it does in the Hebrew shape: the result is a mixed form, possibly an attempt to Hebraize the ‘Murabba at’ form or a ‘Murabba at’ influence on the Hebrew form. D: Instead of the straight top stroke starting as a tick (lines 4, 11) we have here (line 10) a curve; the downstroke bends to the left, while in Hebrew it slopes to the right. The letter is intermediate between Hebrew and Murabba at. The daleths in lines 12, 13 are Hebrew, but uncharacteristic. H: In all specimens the right shaft clearly projects on top as in Murabba at. In Hebrew the top stroke and right shaft are joined to form a corner; now and then, when they have been carelessly joined, the shaft 9

10 11

12

“A Fragment in an Unknown Script” and “An Unknown Aramaic Cursive”, in: PEQ 84 (1952), pp. 118–120, and 85 (1953), pp. 23– 41 respectively {= chapters 49 and 50 in the present volume}. {See pp. 280–281 in the present volume.} The first two lines of the chart show variants of 18 letters, lines AA and BB contain 5 letters (and variants) about whose identity there was some doubt. (Cf. below n. 13.) The second specimen, somewhat incongruously, continues as a curve, as in the chart.

000267

Birnbaum 01.04.11 18:17

The Beth Mashku Document

267

projects a little. In one or two specimens only (line 8 and presumably in the second H of line 12) the top stroke has the Hebrew form: straight and starting with a tick. In line 9 and in the first specimen of line 12 we have the arch of Murabba at. The H of line 13 is so mutilated that it is hard to decide what it was really like. The top strokes on line 10 to some extent approach the Hebrew but consist of a very obtuse-angled tick only, not, as in Hebrew, of an initial, very acute-angled tick and a very long straight horizontal. The left downstroke in line 10 is not parallel with the right one, as it would be in Hebrew, where the left stroke issues from the middle of the top stroke; instead it slants to the left issuing from the top right-hand corner. This is the form we get in Murabba at. The right stroke is everywhere straight as in Hebrew. Thus all the heˆs in the signatures are of mixed Hebrew and Murabba at types. M: The Hebrew form has a base, usually long, at right angles to the downstroke. The specimen in line 13 differs essentially, instead of the base there is an angularisation of a deep curve, as in Murabba at. The left stroke is very long, reminiscent of the Murabba at form. Thus the letter is a mixture of Hebrew and Murabba at. Final N: The tops of the letters in lines 8, 11, turn to the left wherein they differ both from the Hebrew and the Murabba at forms. Ayin: In lines 1–7, the right stroke is bent into a practically vertical top part and a sloping downstroke which descends below the line bottom. The left stroke is more or less parallel with the other top stroke and runs to the middle of the slanting stroke. In lines 8, 13, the right stroke is straight and does not extend below the line bottom, while the left stroke is curved and runs horizontally to the top of the right stroke. That is the Murabba at form. The ayins of line 10 are mixed: the right stroke is slightly bent but does not run below the line bottom, and the top part is not vertical. The left stroke is curved but does not run to the top of the right stroke. Q: Instead of the big triangle of the Hebrew form we have here a very flat one (line 13) or actually two parallel horizontals close to each other (line 10). S˙ in line 8 is of the Murabba at type. T: In line 8 the bottom of the left stroke is looped with the top of the right one. This tallies with Murabba at as against Hebrew. In line 9 a tick is added to the bottom of the right stroke. It looks like an erroneous reflex of the little stroke or curve to be found in the Hebrew form at the bottom of the left stroke. To sum up: We have in the signatures three kinds of letters: (a) of the Hebrew type, (b) of mixed Hebrew and Murabba at types, (c) of pure Murabba at type.

000268

Birnbaum 01.04.11 18:17

268

[ 48 ]

Reason for the Mixed Types Such a mixture of types of different scripts presents an unusual state of affairs. How are we to account for it? Perhaps the explanation is that the script of the Murabba at Fragment13 was that ordinarily employed by the Beth Mashku Jews and that its imprint can be seen on their Hebrew alphabet, when they did use it. The more closely two forms of script (or language) are related the more easily do such things happen. The signatures would show that the extent of this influence varied with the individual. They obviously did not write much at all, and such people would be the very ones to be most easily influenced.

The Writer In the writing of the actual body of the document, however, there is no trace of ‘Murabba at’ influence. The man who penned it was educated enough to write Hebrew but he was neither a scholar nor a professional scribe; the latter is clear from the irregularity of his writing and from the fact that he found it difficult to keep to the lines – which is particularly striking as the texture of papyrus is an aid towards keeping straight. There is a roughly contemporaneous scribal hand available for comparison.14

The Date of the Murabba at Fragment The Beth Mashku document has also a bearing on the dating of the Murabba at Fragment. The identity or near-identity of certain letters in the two documents suggests that they are not very far apart in time. The date suggested by us on palaeographical grounds in PEQ 85, p. 41 {= p. 297 in the present volume} was: “about the second half of the 1st cent. C. E. [. . .] or fairly close to it.” This is now confirmed by the Beth Mashku Document, whose date, 132–135 C. E., is certain.15 13

14 15

And now of another papyrus (unknown to me when the present article was sent to the Editor, and published since by J. T. Milik, “Un contrat juif de l’an 134 apre`s J. – C.”, in: Revue Biblique 61 (1954), pp. 182–190). This new material, unlike that narrow, mutilated scrap, the Murabba at Fragment, consists of a complete text of considerable length and has made it possible to clear up the unavoidable uncertainties of my editio princeps. Revue Biblique 60 {see above, n. 2}, pl. 12, fig. a {also reproduced in The Hebrew Scripts by S. A. Birnbaum. London 1954–1957 / Leiden 1971, plate 89}. A further corroboration is now afforded by the second Murabba at deed, which actually bears a date – 134 C. E.

000269

Birnbaum 01.04.11 18:17

The Beth Mashku Document

269

Origin of the ‘Murabba at Script’ Perhaps we can learn something from the Beth Mashku Document about the origin of the ‘Murabba at script’. Mashku, as de Vaux has pointed out, is a Nabatean name, so that the village might have been a Nabatean settlement. Now the ‘Murabba at script’ is closely related to the Nabatean alphabet. (See my article in PEQ 85, p. 40 {= p. 296 in the present volume}.) That would indicate that the ‘Murabba at script’ grew up in places where Jews and Nabateans lived side by side. Incidentally, if that is so, the Beth Mashku Document confirms the result we reached, without a knowledge of these signatures, in our examination of the Murabba at Fragment, namely that the ‘Murabba at script’ stands in a special relationship to the Nabatean alphabet.

The Language Like the Bar Kosba Letter our document shows that in addition to what we know of its role as a literary language, written Hebrew was at this time used also for ephemeral purposes. A plain literate person in a village who was neither a scholar nor a scribe was able to write it. Does this document contain any indication as to what was the spoken language of those who sent it? The colloquial of the Nabateans was Arabic. No traces of that tongue are to be found in our papyrus. From what we know of the linguistic situation in those regions at that time it is a matter of course that we should expect to meet with indications of Aramaic influence. What we find is: (a) In line 12 there is the Aramaic rb (as pointed out correctly by Lehmann / Stern, l. c.). (b) The spelling with aleph in aluglg (line 2) is Aramaic (the Bar Kosba Letter has H: hlglg.16 (c) The ending ¯ın (line 1) is Aramaic. Whatever the real nature of this form in the Bible, it could hardly be an ancient Hebrew survival here in ñisnrp. (d) Whether the final H of line 10 is Hebrew or Aramaic depends on the meaning of the word. Everybody’s first reading surely is – like de Vaux’s – ‘for himself’. It has been objected that wpn as a reflexive pronoun is of only rare occurrence in Mishnaic Hebrew and that we 16

The ih (line 4) without aleph is presumably a phonetic spelling rather than a result of Aramaic influence, since the Judaeo-Aramaic spelling is also aih, as in Hebrew.

000270

270

Birnbaum 01.04.11 18:17

[ 48 ]

should equate it with the Mishnaic úug, meaning ‘essence’. However, this is on the assumption that the word is Hebrew – which is precisely what we do not know. But even if it were, it could have been employed here in the rare, reflexive sense, or its meaning might have undergone Aramaic influence. Rather than postulate a new meaning (‘essence’), it is preferable simply to read ‘self’. In that case the suffix would be masculine and the word, accordingly, Aramaic.17 (e) The root bsr (line 6) is known to us from Midrash Rabbah Exodus and Tanhuma and so an Aramaic derivation seems likely. ˙ ultimate source of snrp (line 1) is controversial but as it (f) The appears only late in our Hebrew sources, is present in Syriac, and came via Syriac into Arabic, it would be reasonable to suppose that it is an Aramaic loanword in Hebrew, too. Taken together, these points would indicate that the language of our document had an Aramaic background.18 It therefore seems reasonable to assume that Aramaic was the mother tongue of the Beth Mashku Jews.

17

18

The final heˆs in lines 8 and 9 might be Aramaic. If hbtk were to be a noun it would be Aramaic. Its spelling with H would then be in contradiction to that with aleph in aluglg (line 2). However, in view of the many orthographic irregularities in our papyrus, this would be of little account. If hbtk were a noun, it could hardly mean ‘writer’. That we have hitherto not come across such a word would not disprove that it existed, but the two people against whose names it is written could not both have been the actual writers of a document which is in one hand – a hand very unlike either of their signatures. One would therefore have to assume that hbtk meant ‘a man who causes something to be written’. Such a noun seems very unlikely; it is preferable to take hbtk as a verb with a pronominal suffix. ( The word used in the scribal formulae appended to the Elephantine letters and deeds is a verb [btk]). – For the reason just given hbtk must mean ‘he had it written’ – ‘it’ referring to the body of the document. ( This sense of btk is general in the Elephantine Papyri. The term in the scribal formulae for the act of writing is Õp ly btk or Õpk btk ‘as told by ....’). The suffix would then be feminine. As Hebrew and Aramaic have the identical form of the feminine we would not know which language it was. The root bsr and the word snrp might, however, at that time already have been part and parcel of Hebrew.

000271

Birnbaum 01.04.11 18:17

271

The Beth Mashku Document

The Text rzyla ñku yuwi ñk ukwm tib lw ñisnrph ñk ydiw Õulw hinxmh wur aluglg ñb yuwil bqyi ñm ñujcra ñb úsuhi xqlw hrphw Ól ihi tunb ñm ulw ihw ukwm tiba bwuiw hduhi ñb cphu itly ina unly Õibrq Õiighw illa úau ñurwb ñm ruma iht alw hkk ly Óitcphu larwi tib lku Õulw hiha Ólca itly al hbtk rzyla ñb yuwi hbtk úsuhi ñb rzyla hwpn ly hduhi ñb buqyi dy rzyla ñb luaw dy úsuhi rb úsuhi diym úsuhi ñb buqyi

1 2 3 4 5 6 7 8 9 10 11 12 13

Tentative Translation 1 2 3 4 5 6 7 8 9 10 11 12 13

Thus / Here (say) the parna¯sı¯n of Beth Mashku: thus / here (says) Jeshua and thus / here Eleazar To Jeshua ben Galgola, commander of the garrison, Greeting. Made known to you may it be that the cow which Jehoseph ben Ariston requisitioned from Jacob ben Judah, who is a resident of Beth Mashku, that it was his property ‘from daughters’ (?). Further: If the heathens were not (so) near to us, I myself would have come up and would have discussed compensation in the matter, so that you should not say, out of disrespect have I not come up to you. Peace to you and to the whole House of Israel. Jeshua ben Eleazar had it written Eleazar ben Jehoseph had it written Jacob ben Judah, for himself Saul ben Eleazar, witness Jehoseph bar Jehoseph, witness Jacob ben Jehoseph, confirming the (identity of the) witnesses.

000272

272

Birnbaum 01.04.11 18:17

[ 48 ]

The Beth Mashku Document

000273

49.

Birnbaum 01.04.11 18:17

A Fragment in an Unknown Script*

The first account that was published about the Wadi Murabba aˆt finds spoke of “manuscripts [.. .] apparently in Hebrew and Aramaic but written in a type of cursive hitherto unknown and therefore very difficult to read.” (Manchester Guardian, April 7, 1952.) When the Editor of PEQ sent me this photograph1 and suggested that I should let him have a note on it, I realised with a thrill that this must be the manuscript (or one of the manuscripts) in question, for I was not able to read it. The script was clearly of the Aramaic group, but to which branch could it belong? Examination revealed that it was not a form of Pahlavi, and that the northern Aramaic branch could also be eliminated. The writing of the fragment did not look like any of the southern branch either: Palmyrene, Hebrew, Nabatean, Arabic. However, we had known them – apart from Hebrew – from inscriptions and graffiti only. Could the new script perhaps be a handwritten cursive of one of them? In a separate article – ready for publication2 – I have attempted to answer this question by means of a detailed investigation. I shall confine myself here to the result: The writing is related to Nabatean but it would perhaps be safer to call it a parallel branch rather than simply Nabatean cursive. My date for the manuscript is roughly the first century C. E. To decipher the script has been by no means a simple task and certain letters may have to be identified differently. I have not arrived at a smooth, self-evident text – in view of the scantiness of the material and its state of preservation, that would have been surprising. A slightly better result might have been possible if the original had been available to me. It might then, e. g., have turned out that some of the black parts of the photograph, which looked as if they belonged to letters, were merely holes. Perhaps, too, it would have been possible to make out more of the mutilated writing on the right and left-hand edges, as well as on the creases which have broken the writing of lines 3, 10 and 12. A real improvement, however, could I think only have been achieved with * First published in: Palestine Exploration Quarterly 84 (May – Oct. 1952), pp. 118–120. 1 {See Chapter 53, p. 374, of the present volume. See also The Hebrew Scripts by S. A. Birnbaum, London 1954–1957 / Leiden 1971, plate 121(4).} 2 {See Chapter 50 in the present volume.}

000274

274

Birnbaum 01.04.11 18:17

[ 49 ]

the aid of additional material. It is to be hoped that there really are some additional fragments in this script (i. e., that the plural ‘manuscripts’ in the above-quoted passage was not a slip or misprint). Here is my provisional transliteration, in which  represents a in final position {and a dot above a letter indicates an unclear reading}: ]yhdx ¡ ¡ ¡ ¡ id ¡ ¡ ¡ ¡ h ¡ ¡ [ ]buw±iula cëbqëhwnm[ ]ñ±iu±iun hxal±iuluht[ ] . . . ±iubëbë ¡ ¡ ¡ ¡ nqë±iul ¡ ¡[ ] ¡ ¡ ¡ ¡ ¡ ¡ ¡ b yibjb ¡ ¡[ ]wët mam±iup±iuh[ ]Óht hëml yt±iubl[ ] ¡ ¡ ñnbma±iuharsë[ ]tblhnañhuañulbë ¡ ¡[ ]±iusëbm±iuhnitmi[ ] ¡ ¡ b qlidulmra[ ] ¡ ¡ ¡ ¡ ñiarxa ¡ ¡ ¡ ¡ Ó[ ] ¡ ¡ ¡ ¡ ¡ ¡ yëutëruhëdkëii[ ] ¡ ¡ ¡ ¡ rjwbulë[ ] ¡ ¡ ¡ ybwb[ ] ¡ ¡ ¡ ¡ ¡ ¡ ab[ ]ulacëbq[

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Line 1: id ‘which’. hdx ‘one’? Line 2: hwnm ‘Manasseh’? Or could the word possibly be connected with (a) ‘to forget’ or (b) ‘women’? Could lacëbq be a theophorous name? ( l is not necessarily Jewish.) Line 3: uht impf. sg. 3rd f. or 2nd of auh ‘to be’. If ul is the correct reading, could it perhaps be the same as Talmudical ual ‘is not’?. The next word cannot be al, otherwise the aleph would have the final form. Line 5: ybj ‘to sink, to coin’, etc.? The end of the line is taken up by something that looks like the figures in the Beth Phage graffiti,3 or in the Pahlavi papyri from Egypt.4 If it is a figure, might not the preceding sign, which I could not identify as a letter, also be a figure? It only vaguely 3 4

Syria 4 (1923), p. 241, Journal of the Palestine Oriental Society 4 (1924), p. 171, Revue Biblique 32 (1923), p. 352. Olaf Hansen, Die mittelpersischen Papyri der Papyrussammlung der Staatlichen Museen zu Berlin. Berlin 1938 (Abhandlungen der Preussischen Akademie der Wissenschaften, Phil.-Hist. Kl., 1937.9).

000275

Birnbaum 01.04.11 18:17

A Fragment in an Unknown Script

275

resembles the Nabatean ‘100’ or the ‘100’ of older Aramaic, but there is a chance that it represents the one or the other. If the beth in front of it were meant to function as a numeral – 2 × 100 – it would be an exceptional combination of a figure and an alphabetic numeral. Line 7: It looks as if something had happened to the letter after the mem. Could it be a non-final aleph connected with the mem by a blot? The letter seems to be too tall for a non-final aleph as compared with other specimens; these are decidedly short, apart from a medium-sized one (line 6). In addition, if it were a non-final aleph, it could not have been used at the end of a word. Nor can it be read as the beginning of the subsequent word. Could it perhaps be a crippled heˆ? The right downstroke would have been shortened in order to avoid crossing the downstroke of the mem, and the left-hand downstroke of the presumed heˆ would accordingly have been pushed to the left. Could we then read Óht hmly tibl ‘to her eternal home she will go’? The expression is known to us from Phoenician, Palmyrene, Midhrash, Talmud and Bible. Line 8: mau ‘and mother’? ñnb ‘daughters’? Line 9: ñul might be l with suff. 3rd pl., the heˆ having dropped out, as it frequently does in Palestinian Aramaic. Or is the preceding letter by any chance a beth? It would then be open to us to reconstruct the word as Óulb[qi] ‘they will / shall receive’. hna ñh ua ‘or if I’? (It seems very improbable that we have here scriptio defectiva for hnia ‘he / she is not’.) Line 10: hnitmi ‘he will / shall await him / her / it’. Line 11: l id ulmra ‘widowhood that not’. If the black mark after b denotes traces of a letter then we cannot restore ¡ ¡ l]bq. Line 12: ñiarxa ‘responsible’? Line 14: rjwb ‘in the deed’. Line 17: The writing starts somewhere in the middle of the original line. This, in conjunction with the apparently blank line preceding it, suggests that we have here the date of a document: ]ybwb ‘on the 17th’. The reading ‘17’ seems to be more likely than ‘7’, because there is no yodh between beth and ayin. In Nabatean it could not be 17, because there the tens precede the units. Line 18: If lacëbq were a name it might be the signature. All this together does not go very far towards elucidating the contents of our fragment. Words like rjw (line 14) and iarxa (line 12), the possible occurrence of a date in a special position (line 16), or of even a signature (line 18), might point to something of a legal nature. The mention of widowhood (line 11) and of future death (line 7; perhaps the contingency of future death) would suggest the possibility of a marriage

000276

276

Birnbaum 01.04.11 18:17

[ 49 ]

document. Other hints of a legal or business character might be found in line 5, if ‘coined’ is mentioned there, and if the signs are figures, and in line 1, if the reading ‘one’ is correct. Whatever the contents, the fragment is of unique interest from the palaeographical point of view.

50.

An Unknown Aramaic Cursive*

In the issue of the Quarterly for July – October 1952 {= the previous Chapter in this Volume} the present writer offered a provisional transliteration of one of the fragments recently discovered in the Wadi Murabba aˆt caves and written in an unknown cursive.1 The following lines contain the palaeographical examination referred to there, on which the decipherment and the date assigned to the document were based. In view of the scantiness of the material the number of absent or doubtful letters is surprisingly small. As more fragments in this script appear to exist,2 their eventual publication will presumably clear up these doubts and make it possible to fill the gaps. In the meantime here is our examination. For comparison with the Murabba aˆt script we have utilised the three alphabets which are obviously most closely related to it, those constituting the southern group of the Aramaic branch: Palmyrene, Nabatean and Hebrew. Extant Palmyrene and Nabatean material3 covers three to four centuries around the Christian Era and so, of course, the Hebrew script of the corresponding period has been used.4 * First published in: Palestine Exploration Quarterly 85 (1953), pp. 23– 41. 1 {See the facsimile p. 374 in the present volume; also reproduced in The Hebrew Scripts (HS) by S. A. Birnbaum, 2 vols., London 1954–1957 / Leiden 1971, plate 121(4).} 2 The Manchester Guardian report spoke of the discovery of “manuscripts” in this alphabet. 3 Re the comparison of pen-written and engraved scripts: The difference in medium must be taken into account, but that does not mean that comparison between scripts written on different media is inadmissible in principle. This holds good even for comparison between pen-written style and lapidary style (provided that due care and discretion is exercised). Not every writing on stone is however in lapidary style. As a matter of fact, in our material it is generally not. As far as Hebrew is concerned, a real lapidary style never developed. The few inscriptions we have are simply pen-written forms transferred to stone. Cf. also p. 18 of my monograph The Qumraˆn (Dead Sea) Scrolls and Palaeography. New Haven 1952. 4 From the letters in our Fragment which are unmistakable, a general idea as to their time will be so clear to anyone acquainted with the development of the Aramaic alphabet, that it would be quite irrelevant to encumber our study by assembling comparative material from so late a time as the Middle Ages. Cf. my remark in PEQ 84 (1952), p. 94, lines 14–17 {= Chapter 44 in this volume, second paragraph}.

278

[ 50 ]

In the Palmyrene and Nabatean sections of the Chart (see below, pp. 280–281) specimens from various inscriptions are assembled. The Hebrew section, taken from my forthcoming book The Hebrew Scripts 5 represents the following: 1 Tobias Inscription {HS 80}. 2 Gezer Sabbath Boundary Stones {HS 83}. 3a, 3b Hymns Scroll {HS 87E}. 4a, 4b Ossuaries {HS 95–99}. 5a, 5b Uzziah Tablet {HS 88}. 6 Queen Helena Graffito {HS 100}. I have drawn the alphabets of the Murabba aˆt Fragment and of the Hebrew inscriptions numbers 1, 4, 5, 6 from original photos, the other alphabets from photographic reproductions. I have given the most common forms, and occasionally, in addition, forms that have been specially mentioned in this article.

Aleph Aleph was difficult to decipher. It consists of two strokes only. A faintly arched top stroke, sloping a little to the right, is joined on the left to the top of a slightly curved downstroke, which slants gently leftwards. Let us first consider how this form could have arisen. We shall take as our starting point the form the Aramaic aleph had in the 3rd century B. C. E., i. e., the century preceding the period with which we are dealing. Its short left-hand stroke might have undergone a gradual lengthening while its point of issue from the main stroke was raised until it reached the top.6 At a subsequent stage, instead of adding the right- and left-hand strokes to the initial main stroke, the writer would gradually have made the whole letter without lifting his pen off the paper. Having started from the top of the right-hand stroke, at the end of it his pen would have moved upwards to the left, making only the upper half of the main stroke, and finally, still without lifting his pen, have added the vertical left-hand downstroke, altogether omitting the lower half of the main stroke.7 At the next stage the initial tick thus formed would have lost its angle and become a curve; then it would have 5 6 7

{See above, n. 1; vol. 2.} Such a form is known to us from Hebrew. This labour-saving form was independently evolved at various times and in various groups of Hebrew cursives.

000279

Birnbaum 01.04.11 18:17

An Unknown Aramaic Cursive

279

straightened out, subsequently even becoming very slightly arched. An alternative possible line of development, leading to our form, might have come about through the gradual stunting of the left-hand stroke of the 3rd century B. C. E. form, and its final disappearance into the main stroke. Simultaneously, or later, the stance of the main stroke would have changed until it became upright, while the right-hand stroke no longer ran to somewhere in the middle of the main stroke, but joined it on top. In our Palmyrene material the aleph consists of three strokes. Only in the graffito CIS 8 II , no. 4180 does one of the three different specimens of this letter resemble Mur.9 This graffito dates from 160 C. E. Hence Mur might date from approximately the 3rd century C. E. Hebrew has the three-stroke form, so that there would appear to be no close connection between Mur and Heb.10 A form similar to Mur however occurs in the Beth Phage graffito of the Louvre, with which we shall deal in the next section. If Mur is related to this form, a period shortly before the Christian Era would be the terminus a quo for Mur. Nabatean, as exemplified in Halasa, has originally a plain three˙ E.˙ we find this aleph has developed stroke form. In the 1st century B. C. into a number of curved strokes meeting at the centre. By the end of the 1st century B. C. E. the full form is replaced, when in initial and medial position, by a cursive one,11 and by the end of the 1st century C. E. the latter has ousted the full form from the final position too. Hence the divergence between Mur and Nab12 would have preceded the 1st century B. C. E.

Final Aleph Our Fragment has a final aleph. It seems to be closely related to the non-final form. Curiously enough, both facts did not become clear to me until after I had identified the two forms independently of each other. To decipher them was not easy, as, in spite of Nabatean, we do not usually think of aleph as a letter with a final form. 8 9

10 11

12

{Corpus Inscriptionum Semiticarum, vol. 2. Paris 1889 ff.} “Mur / Palm / Heb / Nab” denote: “the form the letter in question has in the Murabba aˆt Fragment / in the Palmyrene script / in the Hebrew script / in the Nabatean script”, respectively. Cf. above, n. 9. Its origin is clearly traceable in the inscription CIS II , no. 163 (E. Littmann, Nabataean Inscriptions. Leyden 1914, no. 100; J. Cantineau, Le Nabate´en. Paris 1930–1932, vol. 2, p. 12, Hauran I). Cf. above, n. 9.

000280

280

Birnbaum 01.04.11 18:17

[ 50 ]

Alphabets of the Four Scripts (drawn by the author):

000281

Birnbaum 01.04.11 18:17

An Unknown Aramaic Cursive

Murabba aˆt, Palmyrene, Hebrew, Nabatean.

281

000282

282

Birnbaum 01.04.11 18:17

[ 50 ]

Palmyrene has only one form and thus provides no clue to our problem. Our Hebrew material appears to contain an isolated example of the use of final aleph. In the Beth Phage graffito in the Louvre13 (1st century B. C. E. / C. E.) aleph occurs four times in final position (the one initial specimen is indistinct); the left-hand stroke is missing and the main stroke lengthened downwards. This lengthening clearly indicates its cursive character. The Beth Phage form and Mur bear a considerable resemblance and, if related, Mur would belong to approximately the 1st century B. C. E. / C. E. From shortly before the Christian Era onwards, Nabatean employs the full form at the end of a word, from the second half of the 1st century C. E. onwards the cursive form – formerly used only initially and medially – is transferred to the final position too. This cursive form differs essentially from Mur. The terminus ad quem for the origin of Mur would therefore be earlier than these dates. As a matter of fact, we find a similar form in the oldest Nabatean inscription (of Halasa, prob˙ ably between 120 and 96 B. C.E).14 Here the full form ˙is employed 15 initially and an abbreviated one at the end of a word. The latter corresponds to Mur, where however the strokes are differently joined and differently proportioned. Both in Nab and Mur the left end of the upper stroke is joined to the top of the downstroke, but in Mur it coalesces with it lower down; in Halasa the upper stroke is short, in Mur ˙ stroke or else of roughly equal it is either much longer than the ˙lower length. Nowhere after Halasa do we find a similar shape in Nabatean, ˙ C.˙ E. would be the terminus ad quem for the hence the 2nd century B. origin of Mur.

Beth A short straight vertical top stroke curves into a short downstroke which slants to the left. Top stroke and transition curve may also be combined to form a straight stroke which slopes rightward. The downstroke runs to the base but does not always reach it. The base is a short curve or angle. 13 14 15

R. Dussaud, “Comptes d’ouvriers d’une entreprise fune´raire juive”, in: Syria 4 (1923), pp. 241–249. A. Cowley, {“A note on a Nabatean inscription at Halasa”, in:} Palestine ˙ ˙ Exploration Fund Annual, 3 (1914), pp. 145–147. The letter is at the end of the line, near the edge of the stone, and may be mutilated, though the photo of the rubbing does not seem to favour such an assumption.

000283

Birnbaum 01.04.11 18:17

An Unknown Aramaic Cursive

283

In contrast with Mur, the top stroke of lapidary Palm16 is a downcurve. In cursive Palm, however, straight top bars are the rule – the earliest specimen known to me being of 113 C. E.17 In Heb the top part, from being a tick, becomes a straight bar in the Ossuaries period (1st century B. C. E. / C. E.), but on the left of this bar there is a remnant of the earlier structure. Nothing of that kind survives in Mur. However, this feature is occasionally lacking in Heb too. The letter in the Beth Phage graffito (see above, under final aleph) is almost identical with Mur, except that while the former has a straight base, the latter generally has a curved one. Thus, as far as Heb is concerned, Mur would be somewhat later, and the 1st cent. B. C. E. / C. E. would perhaps only be the terminus a quo for Mur. In Nabatean, from the Christian Era onwards, the full form seems to be hardly used outside the final position, and after the 1st cent. C. E. it disappears here too. Shortly before the Christian Era a cursive form which has lost the top stroke enters the scene. Hence Mur would have parted ways with Nab before the late 1st cent. B. C. E. And even for that period the idea of a close relationship between Mur and Nab is not attractive because a straight top stroke is of only exceptional occurrence in Nab. The bottom part of Mur might have come about in this way: The base was not written from right to left as a continuation of the downstroke, but was started independently on the left. Hence the writer’s hand easily went beyond the end of the downstroke. This development presupposes an approximately rectangular corner at the base, because a very obtuse angle or a wide curve offer no inducement for the hand to leave the paper and to start the base from the other side. This corner is to be found throughout in Palm, i. e., 1st cent. B. C. E. – 3rd cent. C. E., which period might then be the terminus a quo for Mur. Heb has this corner from the 2nd cent. B. C. E. onwards. In the 1st cent. C. E., or perhaps somewhat earlier, the base projects to the right. That time might then be the terminus a quo of Mur.

Daleth Daleth is a wavy line or zigzag. It proved rather troublesome to decipher. It looks like a nuˆn in 3rd cent. B. C. E. Hebrew and therefore not unlike the lapidary nuˆn. It also resembles a certain infrequent form of Nabatean yodh, which has a somewhat flattened appearance. In addition, 16 17

Cf. above, n. 9. CIS II , no. 4374.

000284

Birnbaum 01.04.11 18:17

284

[ 50 ]

the Murabba aˆt ayin looks temptingly like the daleth of Palmyrene, early Nabatean and Hebrew, prevalent during many centuries from pre-Christian times onwards, so that the decipherer is misled into trying to find other values for the real daleth. Furthermore, our sign does not resemble the Palmyrene, Hebrew and Nabatean daleth – its greatest similarity being with the daleth in the Nabatean inscription of 124 C. E.18 Our form very likely developed from a type with a formalised horizontal top bar which was joined to the top of the downstroke, because, if the top stroke had reached the downstroke at a point lower than its top, the two strokes could not have curved into each other. In Palm and Heb they never do. In Nab this process begins just before the Christian Era, which would therefore be the terminus a quo of Mur.

Heˆ This is one of the few letters in the Fragment about whose identity there could have been no doubt, revealing the script as belonging to the Aramaic group. The top stroke is short, slightly arched and in a horizontal position. A straight slanting specimen occurs in line 2. The top stroke joins the right-hand downstroke below the line ceiling. The downstroke has a rightward curvature, but there are also rather straight forms. The lefthand downstroke – straight or curved – generally slants from the juncture of the other two strokes but is sometimes severed from that point. While in Palm the top stroke is straight and long, in Mur it is generally curved and rather short. Mur could therefore hardly be later than the early 1st cent. B. C. E. By the beginning of the 1st cent. C. E. the difference is considerable and the divergence increases rapidly. Heb of the 2nd cent. B. C. E. is more advanced in its development than Mur – the left downstroke has moved somewhat to the left. Hence Mur would be older than that time. 3rd cent. B. C. E. Heb is differentiated from Mur in the same way as pre-Christian Palm is from Mur. The early 3rd cent. B. C. E. would therefore be the terminus a quo for Mur. The earliest Nabatean inscription contains two heˆs: both differ from Mur in that the top stroke is full-sized and slanting; in one the left-hand downstroke is short and slants from the inner corner; in the other it is very short, parallel with the right-hand downstroke and severed from the top stroke. Mur is considerably less advanced in its development; hence divergence from Nab would have started before the 2nd cent. 18

Littmann (above, n. 11), p. 25.

000285

Birnbaum 01.04.11 18:17

An Unknown Aramaic Cursive

285

B. C. E. Furthermore, it is without a final form, while Nabatean already has one in the early 1st cent. B. C. E. The existence of a final form is, of course, a cursive feature and this form might therefore be expected in a cursive document like Mur, if the script were Nabatean and younger than the 2nd cent. B. C. E. Its absence indicates that the two scripts must have diverged not later than in the 2nd cent. B. C. E.

Waw Waw is either a slightly rounded angle whose upper arm is somewhat shorter than the lower one, or a plain, gently curved stroke. From our earliest material onwards Palmyrene has an entirely different form. Even in the cursive inscriptions it is very remote from Mur. A terminus a quo later than the 2nd cent. B. C. E. would therefore seem to be unlikely. By the 4th / 3rd cent. B. C. E. Hebrew has already reached the form we see in Mur. As Heb is straightening out in the 1st cent. B. C. E., a connection between Heb and Mur would have to precede this date. From the 1st cent. B. C. E. onwards Nab differs so much from Mur that the latest possible connection must be put at the 2nd cent. B. C. E.

Heth ˙

The two specimens look somewhat different but are of identical construction: two vertical shafts are linked by a slanting and very short bridge. Palm is identical. In Hebrew a half-developed form of this type occurs in the 3rd cent. B. C. E., but it is rare. In the older type – with a horizontal bridge – by the 1st cent. B. C. E. the cross stroke has started moving towards the line ceiling. Hence a connection between Heb and Mur would have to precede that period. In Halasa the cross stroke is on top, and although less extreme forms occur,˙ this˙ remains the rule in Nabatean. Shortly before the Christian Era we come upon a cursive form where the left-hand stroke has lost its top. No trace of such a development is to be found in our document. Mur and Nab would therefore have started diverging by the early 1st cent. B. C. E. at latest.

Teth Teth looks like an angular, recumbent letter S. Palmyrene has nothing similar to Mur, so that any connection between Mur and Palm would have had to precede our earliest Palmyrene material (1st cent. B. C. E.).

000286

286

Birnbaum 01.04.11 18:17

[ 50 ]

Hebrew shows no trace of this development either. But a chronological inference cannot be drawn until we have discussed the derivation of our form. Nabatean has it shortly before and during the 1st cent. C. E.,19 although the older type does not go out of use. It is clearly of cursive origin: first the left stroke was made, then the writer’s hand, instead of leaving the paper to begin the right-hand part of the letter near the line ceiling, slid along the paper on its way to the top. Finally the curve stopped being run back to the left-hand stroke. Alternatively we might derive our shape from a teth whose right-hand stroke touched the left one to begin with.20 The principle is the same: the hand runs straight from the bottom of the left-hand stroke in an upward direction to form the right-hand part. It appears to be less likely that this development took place in Nab and Mur independently of each other, than that we are here dealing with one and the same event. Thus, according to teth, the relationship between the two scripts seems to be rather close. Hence the earliest possible date would be the 1st cent. B. C. E. We can now discuss the chronological aspect in relation to Heb. The form from which we derived Nab and Mur – where the right and left parts are of roughly the same height – begins to appear in Hebrew in the 2nd cent. B. C. E., although specimens with the tall left-hand stroke do not disappear before the 1st cent. B. C. E., it seems. Hence the terminus a quo for Mur would be the 2nd cent. B. C. E.

Yodh Yodh is a more or less straight vertical stroke, either practically of line height or appreciably shorter. Palm is at first an angle of about 90 degrees, open at the bottom; later it becomes rounded. At the same time its position gradually changes, clockwise, from horizontal to slanting. By the last quarter of the 2nd cent. C. E. it is vertical. The 3rd cent. C. E. would then be the terminus a quo, if we derive Mur from Palm by assuming a flattening or straightening out of the curve.

19

20

No dated examples from a later time seem to exist, but in view of the great frequency of this form at the Wadi Mukattab, there is every likelihood that it continued beyond the 1st cent. C. E. This type occurs at various times in Aramaic script.

000287

Birnbaum 01.04.11 18:17

An Unknown Aramaic Cursive

287

In Hebrew in the Ossuaries period (1st cent. B. C. E. / C. E.) we now and then meet with straight forms without a head, corresponding to Mur. Accordingly the terminus a quo for Mur would be about the 1st cent. B. C. E. / C. E. As to Nabatean, the yodh of Halasa, continuing one of the 3rd cent. ˙ be ˙ the source of our form. In the B. C. E. Aramaic types, can hardly early 1st cent. B. C. E. Nabatean begins to develop – from another of the 3rd cent. B. C. E. types – a very characteristic form which attains to full line height before the Christian Era. It is hardly probable that Mur is descended from a certain somewhat flattened type, approximating to a wavy line, which is occasionally to be met with. The same applies to the final form occurring in Nabatean from the beginning of the Christian Era onwards. Hence the only possibility of a connection between Mur and Nab is during the 2nd cent. B. C. E. The 3rd cent. B. C. E. Aramaic form, consisting of a slanting downstroke from the end of which a curve slants rightward, became in Nab fuller, rounder and bigger; in Mur, on the other hand, flatter, a wavy line, and finally short and straightish.

Kaph The identification of the second letter in line 13 as kaph is doubtful. One more example might have settled the question. (See qoph.) The top stroke is straight and horizontal. In all our examples of lapidary Palm, the top stroke is a down-curve, and even in the cursive of the 2nd and 3rd cent. C. E. it seems rarely to be straight. This would indicate that a connection with Mur could not antedate the 2nd cent. C. E. In Heb there are now and then some rather straight cursive top strokes from the 3rd cent. B. C. E. onwards, but it is only slowly that this feature becomes increasingly frequent. The 3rd – 1st cent. B. C. E. might be given as the terminus a quo for Mur. In Nabatean, from the 1st cent. B. C. E. onwards, we quite often meet with the straight top stroke. That century seems therefore to be the earliest possible providing a link with Mur. The downstroke of Mur is a wavy line. In Palmyrene this is always a lapidary form. In the cursive inscriptions straight downstrokes also occur, our earliest specimen belonging to the 2nd cent. C. E. Thus Mur might possibly be from round about this period. In Hebrew by the beginning of the Christian Era the wavy form has practically disappeared. Hence the connection between Mur and Heb could not be later than the 2nd / 1st cent. B. C. E.

000288

Birnbaum 01.04.11 18:17

288

[ 50 ]

In Nabatean the wavy downstroke is an exception. The Agrippa inscription21 has it. That would make a connection between Mur and Nab possible as late as in the 1st cent. C. E. The downstroke of Mur is long and seems to end vertically, or with a very slight bend leftwards. In lapidary Palmyrene there is always a big curve at the bottom, in cursive a straight base. Hence the terminus a quo for Mur would precede our earliest Palmyrene material. In Hebrew by the 3rd cent. B. C. E. there is already a big curve at the bottom, or a straightish base. Hence the point of divergence of Mur and Heb would have been earlier than that. By the 2nd cent. B. C. E., Nab has a straight base so that Mur would have to be derived from an earlier form.

Final Kaph Our Fragment has a final kaph. Palmyrene has not. Hence the connection between the two scripts would have to precede the beginnings of our Palmyrene material. Hebrew has a final form from the 3rd cent. B. C. E. onwards. This is accordingly the terminus a quo for Mur. Nabatean has a final kaph from the beginning of our material. Thus the 2nd cent. B. C. E. is the earliest possible time for a connection between Mur and Nab. In Murabba aˆt the letter consists of a horizontal top bar – with or without (?) the tick on the left – and a vertical downstroke on the right. In Heb the top tick or curve begins to straighten out in the 1st cent. B. C. E. / C. E., which is therefore the terminus a quo for Mur. In Nab, while the top curve is quite often rather shallow, it does not become straight. Forms with straight downstrokes are known to us in Hebrew from the 2nd / 1st cent. B. C. E. onwards, and this would accordingly be the terminus a quo for Mur. In Nabatean straight downstrokes occur in our material from about the 1st cent. B. C. E. onwards, which would therefore be the terminus a quo for Mur. In Heb, as in Mur, we find both features together – horizontal top bar and straight downstroke – from the 1st cent. B. C. E. onwards. In Nab they are not combined.

21

Littmann (above, n. 11), no. 102.

000289

Birnbaum 01.04.11 18:17

An Unknown Aramaic Cursive

289

Lamedh Lamedh is another of the letters identifiable at first sight. It is a long, or longish, straight or slightly curved stroke above the line ceiling, generally reaching well into the middle of the line height. At the end it turns to the left and – unless it is the final letter of a word – tends to run into the subsequent letter or even to be linked with the beginning of it. Both forms of Palm preclude a close connection with Mur. Nor is there a corresponding form in Hebrew. Hence Mur and Heb could not have diverged later than the 3rd / 2nd cent. B. C. E. In Nabatean a cursive form is already in use before the Christian Era, together with a full one. Hence the 1st cent. B. C. E. is the terminus a quo for Mur.

Mem Mem is a kind of oval, imperfectly closed on the top left; from here a stroke slants down leftwards. No similar form occurs in Palmyrene, where the letter is open at the bottom. Hence any connection between Mur and Palm would antedate the 1st cent. B. C. E. On the other hand Heb, like Mur is closed at the bottom. This form is known to us from the 2nd cent. B. C. E. onwards. As to the little gap in Mur, it recalls the gap we sometimes see from the Ossuaries period onwards. Two isolated specimens22 tally with Mur: they are round and have the downstroke. Nab is closed at the bottom from shortly before the Christian Era onwards, and in the final form already in the early 1st cent. B. C. E. There is frequently a gap on the left. This feature is already present in the final mem of the same inscription. Frequently Nab is similar to Mur or virtually identical with it. Hence the 1st cent. B. C. E. is the terminus a quo for Mur. Our Fragment is without a final mem. Palmyrene has none either. But its mem differs completely from Mur. Hence their connection would have to antedate our earliest Palmyrene material. Hebrew has the double form from the 2nd cent. B. C. E. onwards. Hence Mur and Hebrew forms would have parted ways before that period. Nabatean, too, has two forms. Our first example is in the early 1st cent. B. C. E.,23 indicating that Mur and Nab must have diverged at latest in the 2nd cent. B. C. E. 22 23

Hebrew University, ossuary graffiti, Judah b. Amram and Johanan b. Amram. J. Cantineau (above, n. 11), vol. 2, p. 1.

000290

Birnbaum 01.04.11 18:17

290

[ 50 ]

Nuˆn The nuˆn in line 9 is small, the specimen in line 2 medium-sized; in both the top is rather vertical. If the sign before the final nuˆn in line 8 is a nuˆn, it is of line height and the top is bent leftwards. Lapidary Palm is a wavy line with a big curve at the bottom. The cursive form, however, corresponds to Mur, though not in every detail. Our earliest specimen is of the 2nd cent. C. E. and this period might therefore be the terminus a quo for Mur. Hebrew has similar forms. The downstroke, especially its top part, begins to straighten out in the 3rd cent. B. C. E., which might then be a terminus a quo for Mur. Nab, too, is similar from the time of our earliest specimens onwards so that the 2nd cent. B. C. E. would be the terminus a quo for Mur.

Final Nuˆn Final nuˆn is a long downstroke the lower half of which slants to the left. Palm is very different. The lapidary form has a big initial curve, which later becomes a very acute angle. The latter type is the only one represented in the cursive. Hence any connection between Mur and Palm precedes our earliest Palmyrene material. In Hebrew, forms resembling Mur begin to emerge in the Ossuaries period. The 1st cent. B. C. E. / C. E. would therefore be the terminus a quo for Mur. In Nab, the top part retains traces of the earlier curve. Hence Mur would have originated before the time of our earliest Nabatean specimens, which are from just before the Christian Era (there happen to be no final nuˆns in the earlier inscriptions). The rest of the letter is always straight, indicating the same terminus.

Samekh The first letter of line 8 seems to be a samekh, mutilated at the bottom. The full form would be rather triangular. The mutilated tenth letter of line 10 may also be a samekh. A great deal of development has taken place in Palm since the 3rd cent. B. C. E. Aramaic form. But it is still a wide open letter. No later contact with Mur than well before our earliest Palmyrene material seems possible. Heb tallies with Mur. The triangular shape is reached by the 1st cent. B. C. E.

An Unknown Aramaic Cursive

291

A corresponding but practically quadrangular form appears in the early 1st cent. B. C. E. in Nabatean, rounded specimens shortly before the Christian Era, and mainly triangular outlines in the Christian centuries. In Nab the left stroke and the top one cross each other, often – as already in Halasa – forming a loop. As this feature is absent in Mur, ˙ divergence ˙between Nab and Mur is to be assumed before the time of our oldest Nabatean specimen (early 1st cent. B. C. E.).

Ayin This letter is one of those in the Fragment which appear to be quite unmistakable, once they have been deciphered, but are very misleading before. It might be a daleth or resh in any Aramaic cursive from the 6th cent. B. C. E. onwards, and even a gimel – of some late period – would not be impossible. Ayin consists of a straightish or slightly curved downstroke of line height, joined on the left, very near the top or somewhat lower down, by a short down-curve. Lapidary Palmyrene is of an earlier type, with the left-hand curve meeting the right one in the middle. Cursive Palm, on the other hand, is much more developed than Mur, the left stroke running to the top of the other, which is in a horizontal position. Our earliest dated cursive specimen is of 113 C. E.,24 giving us the 1st cent. C. E. as the latest possible date for Mur. Hebrew, from the 2nd cent. B. C. E. onwards, has the same form as lapidary Palm. In Halasa the right stroke projects only very little below its juncture ˙ left ˙ one. Hence it seems that in Nab of the 2nd cent. B. C. E. with the this development was still in its initial stages. But not much later – in the 1st cent. B. C. E. – in Nabatean, too, the form with the fully lengthened right-hand stroke is in use, and already at the end of that century we here meet with the forms which we see in Murabba aˆt.

Pe The fourth letter of line 6 is possibly a pe. Palm is very different, so that any connection between Mur and Palm would have to precede our earliest Palmyrene material (1st cent. B. C. E.). The top part of Mur is not too unlike the 3rd cent. B. C. E. Heb, where however not much later a pointed top had developed. Hence Mur 24

CIS II , no. 4374.

000292

292

Birnbaum 01.04.11 18:17

[ 50 ]

and Heb would have begun diverging at latest in the 3rd cent. B. C. E. During this same period a horizontal base begins to develop in Heb. Mur has only the merest trace of a curve, so that a connection between Mur and Heb would have to antedate the 3rd cent. B. C. E. In Nab the top part resembles Mur from the 1st to the 3rd cent. C. E. Nab has, however, also a curve at the bottom. Mur and Nab would therefore have parted ways before the time of our earliest Nabatean specimens (4 B. C. E.).25

Sadhe Sadhe, being a rare letter, may not be represented in our Fragment. There are, however, some possible candidates. Is the sign in the middle of line 2 a sadhe, the lower part of whose shaft has disappeared, or have we two letters here: a beth followed by a nuˆn? The same letter or letters can be seen in line 13, and possibly in line 3, near the end. If it were a sadhe, the direction of the right stroke – i. e., towards the shaft – might give us a possible connection between Mur and Heb and probably Palm, but not Nab. The fifth sign of line 2 (= the tenth of line 11 = the first of line 18 = possibly the third / fourth of line 4) may be an alternative possibility. The letter consists of a more or less oval top – which is closed or almost closed – and a long descender issuing from its left end. The right stroke runs towards the shaft. The letter would therefore tally with Palm and Heb but not with Nab. It however differs from all three in being closed, or nearly so, while they all have wide-open forms. Also the complete straightness of this shaft speaks against it being a sadhe, since in all three groups the downstroke has either a base (Heb, from the 3rd cent. B. C. E. onwards) or it at least curves to the left (Palm and Nab). If the eighth letter of line 13 is not a somewhat deformed taw, it may conceivably be a sadhe. Here, as in the first sign discussed in this section, the originally long shaft would have been shortened. However the right stroke is in the opposite direction from the first sign – it issues from the ‘shaft’. Hence a connection between Mur and Nab but not with Palm and Heb might be possible.

25

A. Jaussen / R. Savignac, Mission arche´ologique en Arabie {vol. 1, Paris 1909}, p. 142 f., pl. VIII .

000293

Birnbaum 01.04.11 18:17

An Unknown Aramaic Cursive

293

Qoph The fifth sign of line 2, etc. (see above, under sadhe) is perhaps rather a qoph than a sadhe. In Palm both sides are of equal length, i. e., the left stroke is no longer a descender, and the letter is wide open below. If the sign in question is qoph, any connection between Mur and Palm would accordingly have to antedate our Palmyrene material. In Heb the top part is not, as in our sign, restricted to the right side. A remnant of what was previously the left-hand semi-circle never wholly disappeared, but by the 1st cent. B. C. E. it had become sufficiently small for the letter to have been a possible transition form to Mur. The shaft becomes long in the 2nd cent. B. C. E. Thus the 2nd / 1st cent. B. C. E. would be the terminus a quo for Mur. In Nab the shaft is short in the 1st cent. B. C. E., and the right part very wide, but from the subsequent century onwards the shaft is long and turns leftwards at the bottom, or the shaft is altogether a wavy line while the right part is often practically circular. This would on the whole correspond to Mur, giving us the 1st cent. C. E. as its terminus a quo. Is the second letter of line 13 perhaps another candidate for qoph? If it were a qoph, then its form would preclude a derivation from Palmyrene. On the other hand, there might be a bridge to Hebrew, where, in the 1st cent. B. C. E. / C. E., forms with a plain straight top bar, which is severed from the shaft, are already in existence. The cursive hand might afterwards have run on from the end of the top bar to the beginning of the shaft,26 producing our form. Nabatean qoph differs so greatly from our sign that a connection between the two would have to be earlier than our oldest Nabatean inscription.

Resh Resh greatly resembles daleth but is distinguishable from it: while daleth is a zigzag or fairly regular wavy line, resh is less regular, its parts are less slanting and in its lower half it contains a vertical portion. Mur differs so greatly from Palm that a connection between the two would have to antedate the extant Palmyrene material. In 2nd and 3rd 26

Such a development occurred spontaneously in various medieval and later Hebrew cursives.

000294

294

Birnbaum 01.04.11 18:17

[ 50 ]

cent. C. E. Palm a diacritical dot is sometimes placed on top to differentiate it from daleth. In Mur there is none. However, as the dot is not a regular feature in Palmyrene, we should not draw a chronological conclusion from its absence in Mur. Mur also differs greatly from Heb. At earliest in the 2nd cent. B. C. E. do we find any forms in Hebrew from which Mur could possibly be derived. As regards Nabatean, the position is similar.

Shıˆn The second and third letters of line 2 run into each other, which made it difficult for me to decide how the cluster is to be divided. The right part somehow looked to me like an archaic shıˆn, and this may perhaps have helped me to recognise the letter itself. There are two types. One consists of a slanting Z, joined on the left to the middle of a straight, slanting stroke of line height. ( The letter looks like a 9th century sadhe.) The other type has a less slanting stroke on the left which ends by turning leftwards; on the right there are two more or less parallel down-curves, so placed that the left end of the right one runs into the left half of the other. It soon became clear to me that this was not the original, the W-form, but that the left shaft of Mur represented the left-hand stroke of the Aramaic form known to us from the 5th cent. B. C. E. onwards, lengthened downwards however; in writing the right part of Mur, the hand ran from the end of the right stroke – which did not reach the left one – to the top of the middle one, thus re-creating a four-stroke figure, not unlike the original shıˆn in appearance. Palmyrene provides no similar type. Hebrew has however a transition form. In the 1st cent. B. C. E. / C. E. the left stroke is considerably lengthened downwards; the first traces of this process are to be met with even earlier.27 Hence the 1st cent. B. C. E. might be the terminus a quo of Mur. In Nabatean the old form disappeared at the beginning of the Christian Era. The left-hand stroke had by then become a vertical shaft with a little bend at the bottom, the right strokes were straight and parallel, and virtually always on the upper half of the shaft. This form represents the same transition to Mur as in Heb and would indicate the 1st cent. B. C. E. as the terminus a quo for Mur. 27

Such a lengthening also developed independently in the Middle Ages in one of the non-Square scripts of Hebrew.

000295

Birnbaum 01.04.11 18:17

An Unknown Aramaic Cursive

295

Taw Taw is one of the few letters of our Fragment which are unmistakable at first glance. The earliest specimens of Palm are very similar but without the cursive feature whereby the bottom of the left stroke and the top of the right one are linked up. Instead, in Palm there is a leftward bend at the end of the left stroke. Later (2nd cent. C. E.) this bend is occasionally joined to the top of the right stroke, forming a loop, but by that time the left-hand stroke has become double as long as the right one, in sharp contrast with Mur. Hence the terminus a quo for the latter would be not later than the 1st cent. B. C. E. / C. E. Heb of the 2nd cent. B. C. E. corresponds to the earliest Palm and is thus very close to Mur. Moreover, a cursive form having a loop also occurs. Hence the terminus a quo for Mur would be the 2nd cent. B. C. E. The oldest Nab differs somewhat from Mur, but by the early 1st cent. B. C. E. it has the form of early Palm and 2nd cent. B. C. E. Heb. From the 1st cent. C. E. onward we now and then find looped forms. Thus the 1st cent. B. C. E. would be the terminus a quo for Mur.

A Note The forms in lines 14–18 differ slightly from those in the rest of the Fragment. One might be tempted to ascribe them to another hand. But there is not enough material to ascertain whether the differences are beyond the normal range of a single hand.

Relation to the Other Scripts Let us now tabulate the features which link our document with Palmyrene, Hebrew or Nabatean respectively. With Palmyrene: heth, nuˆn, samekh, ayin, resh, taw; non-existence of final mem. With˙ Hebrew: beth, heth, yodh, kaph, final kaph, nuˆn, final nuˆn, samekh, ayin, resh, taw (once ˙only: final aleph). With Nabatean: final aleph, daleth, teth, lamedh, mem, nuˆn, samekh, ayin, resh, shıˆn, taw; existence of final aleph. We see that the connection between Mur and Palm is only about half as strong as it is between Mur and each of the other two groups. Let us now see which features our Fragment shares with any one of the three groups, to the exclusion of the other two. With Palmyrene it has no such common features, apart from the absence of differentiation

000296

Birnbaum 01.04.11 18:17

296

[ 50 ]

between final and non-final mem. With Hebrew it is linked by such features in beth, yodh, final kaph, final nuˆn. With Nabatean by features in final aleph, daleth, teth, lamedh, mem, shıˆn and the existence of a final aleph. These summaries indicate that the relationship between the script of our Fragment and Palmyrene is very distant; between Murabba aˆt and Hebrew the ties are closer, but much less close than between Murabba aˆt and Nabatean. This does not mean that the script of our Fragment must be Nabatean. There appear to be too many differences – e. g., in beth, heˆ, waw, heth, yodh, final nuˆn; lack of final heˆ, lack of final yodh – for us to take it˙ for granted that it is simply the handwritten cursive form of Nabatean. Hence it might be safer to classify it for the time being as a parallel branch to the three groups but much closer to Nabatean than to Palmyrene and Hebrew.

The Date Our comparison between the Murabba aˆt and Nabatean forms yielded two kinds of termini. (I) Termini a quo, giving the earliest times at which the forms in question (or features in them) could have started diverging from a common ancestor: 2nd cent. B. C. E. – final kaph, nuˆn, ayin, resh; existence of final kaph. 1st cent. B. C. E. – daleth, teth, final kaph, lamedh, mem, shıˆn, taw. The fact that no century later than the first B. C. E. is represented in our table indicates this period to have been the time when the parting of the ways took place between Nabatean and Murabba aˆt.28 (II) Termini ad quem, giving the latest times at which the forms in question (or features in them) could have starting diverging: (a) 2nd cent. B. C. E. – aleph, final aleph, heˆ, waw, yodh, samekh; non-existence of final heˆ, non-existence of final mem. (b) 2nd / 1st cent. B. C. E. – beth, heth, final nuˆn. This table suggests that the divergence˙ between the two scripts dates from the 2nd cent. B. C. E. For (b) contains only a fraction as many letters as (a), and these might belong equally well to either of the two centuries in question. 28

Or between Nabatean Lapidary and Nabatean pen-written Cursive – if additional material in this script and / or other evidence should point to our document(s) having been written by Nabateans.

000297

Birnbaum 01.04.11 18:17

An Unknown Aramaic Cursive

297

Let us now integrate the results of (I) and (II). The fact that the two termini are actually contiguous, suggests that the branching off of Nabatean and Murabba aˆt29 from their common ancestor began round about the time where the two centuries meet – say during the period 150–50 B. C. E. approximately. While that is not the date of the Fragment itself, we may use it as if it were its terminus a quo. We have, however, no terminus ad quem, the document being our only example of the script. Whether a terminus ad quem based on archaeological or any other evidence exists is not known to me. We may hear about that when the report on this excavation is published. In the meantime we shall see what more we can discover by means of paleography. If we compare the forms the letters have in the Fragment with the forms that they have at the times of their termini a quo, we see that some hardly differ, or do so only very little. As we are dealing with a cursive – a type of writing which develops at a fairly rapid pace – the interval of time between the two stages must be correspondingly short. Since more than a third of the letters in question show very little development – heˆ, heth, teth, nuˆn, samekh, taw – it is likely that the ˙ document is not a great deal later than the terminus a quo. In a couple of letters the difference is slightly more pronounced: waw, final kaph. The remaining half of the letters, however, show medium to considerable development – aleph, final aleph, beth, yodh, final nuˆn, ayin, resh, shıˆn – and it is to these that we must give our main attention. There is, of course, no yardstick or slide-rule by which to calculate how long the process of development between the beginning of divergence and the writing of our Fragment could have taken. But – judging from experience with other scripts – it seems reasonable to assume that it required quite a few generations, say, six to eight, corresponding to about 150–200 years. That would bring us to round about the second half of the 1st cent. C. E. So perhaps we may hope to be not too far out if we estimate that the Murabba aˆt Fragment was penned during this period or fairly close to it.30

29 30

Cf. above, n. 28. Postscript (March, 1953). I hear from P. R. de Vaux that the script treated of in the above article is not the one referred to in the Manchester Guardian article. Thus we may look forward to yet another palaeographical thrill.

000298

Birnbaum 01.04.11 18:17

000299

51.

Birnbaum 01.04.11 18:17

The Negeb Script*

The manuscripts from the Wadi Qumraˆn caves are of such extraordinary importance for students of a number of subjects – the Bible, Judaism in the final centuries before the fall of the Second Temple, nascent Christianity, Hebrew palaeography – that these finds have overshadowed the later and less extensive ones from the Wadi Murabba aˆt.1 However, if the two groups of newly discovered material2 are compared from the purely palaeographical point of view it may be no exaggeration to say that Murabba aˆt has brought an equal surprise.

The Material3 Among the three photos accompanying the preliminary report on Murabba aˆt4 there was one showing a pen-written fragment in an unknown script {reproduced in Chapter 53, p. 374, in the present volume; see also HS 121(4)}. This was evidently one of the manuscripts described in an authoritative article5 as “apparently in Hebrew and Aramaic but written in a type of cursive hitherto unknown”. Soon after having dealt with this fragment6 I came across the same kind of characters in some of the signatures at the end of a Hebrew document – another find from the same place – the Beth Mashku Papyrus.7 This interesting document is addressed to a certain Jeshua ben Galgola, the commander of a camp. It can be dated because another letter addressed to him (discovered * First published in: Vetus Testamentum 6 (Leiden 1956), pp. 337–371. 1 The Wadi Qumra ˆ n runs into the north-western end of the Dead Sea, the Wadi Murabba aˆt is further to the south. 2 There is no connection between the material from the two wadis – the manuscripts were written at different times and by quite different sets of people. 3 For abbreviations see the list at the end of this article. 4 G. Lankester Harding, “Khirbet Qumran and Wady Muraba at”, in: PEQ 84 (1952 [Dec.]), pp. 104–109. 5 Manchester Guardian, April 7, 1952. 6 PEQ 84 (1952), pp. 118–120 and 85 (1953), pp. 23– 4l {reproduced as Chapters 49 and 50 in the present volume}. A few doubtful signs remained undeciphered. It was possible to iron out the difficulties when the far ampler and not much mutilated Kephar Bebhayu Conveyance was published (see below, n. 9). 7 They are discussed on pp. 28–31 of my article in PEQ 87 (1955) {= Chapter 48 in the present volume, pp. 266–269}.

000300

Birnbaum 01.04.11 18:17

300

[ 51 ]

together with the first) emanated from Bar Kokhba, the leader in the Second War against the Romans (132–135).8 Some time after the publication of these two documents, the KBeb Conveyance appeared in print.9 Here at last was a text of considerable length in our script. By that time I had realised that long before setting eyes on the first Murabba aˆt document I had actually met this script elsewhere – without being aware of it: among the names scratched on to some of the ossuaries found in the environs of Jerusalem there are a number in this type of writing. (Until now they have been read as Hebrew – see the effect of this in note 15, below.) Finally, when I was putting together the present article, a few single letters on tiny fragments of papyrus, found in the Qumraˆn caves, were published.10 This completes the list of material in this alphabet available for discussion.

The Name of the Script We do not know where Beth Mashku and Kephar Bebhayu were situated but it does not seem unreasonable to assume that these villages were not very distant from the cave where the papyri were found: there is a greater likelihood that the owners of these documents, fleeing for their lives during the Bar Kokhba War, escaped to a refuge nearby, rather than they had travelled to it from a great distance. For the sake of convenience we shall therefore call the new writing – for the time being – the Negeb script. That we find it in Jerusalem, too, on certain ossuaries, need not prevent us from doing so – people hailing from the Negeb may have died in Jerusalem or have been brought there for burial.

The Method We shall discuss each letter in three sections. In the first, this will be done without reference to other contemporary scripts. Comparison with these will be made in the second sections (Comparison). Only Hebrew and Nabatean will be taken into account, as 8

9 10

J. T. Milik, “Une lettre de Sime´on Bar Kokheba”, in: RB 60 (1953), pp. 276–294. – S. A. Birnbaum, PEQ 86 (1954), pp. 23–32 {= Chapter 46 in the present volume}. J. T. Milik, “Un contrat juif de l’an 134 apre`s J. – C.”, in: RB 61 (1954), pp. 182–190. D. Barthe´lemy / J. T. Milik, Qumran Cave I. Oxford 1955, no. 70.

000301

Birnbaum 01.04.11 18:17

The Negeb Script

301

already our first examination11 made it clear that Palmyrene – not to speak of the other branches of Aramaic – is very much farther removed from the new script than either Hebrew or Nabatean. The abbreviations Heb and Nab will be used to designate these two scripts – or any individual letter, form or feature of a letter – from the 2nd cent. B. C. E. to the 2nd cent. C. E. In the third sections (Derivation), the relation between the new script and the two other alphabets will be examined, together with the chronology of that relation. Since the early Nabatean material is older than the Negeb documents we shall assume that connection between Neg and Nab means derivation of Neg from Nab. Similarly we shall assume that connection between Neg and Heb means derivation of Neg from Heb.

The Charts The letters serving as material for our examination were written or incised by 37 hands, whose individual contributions to our chart range from a single character to the whole alphabet twice over. They are arranged in 25 lines, so that in some lines letters from two (in one case: three) hands are assembled, while, in the case of a few documents, it has been possible to devote two lines to a single hand. Here is a list naming the documents utilised for our Neg chart and giving the sources from which I have made the drawings. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

Qumraˆn fragments. – See above, n. 10 (nos. 1 and 15). Beth Phage G. – Photo. [HS 121 (1)]. OGi. – (a) OLZG 534, end. (b) RBS pl. X, no. 5. OGi. – Photo. (a) [HS 121 (3)]. (b) AR 397. OG . – Photo. [HS 121 (2)]. OGi. – (a) RBV 412 ff. (b) AR 431. OGi. – (a) RBS pl. X, no. 2. (b) ESE 2, p. 196. OGi. – RBS pl. X, no. 7. OGi. – (a) RBL 221. (b) RBS pl. X, no. 8. OG . – AR 402. OGi. – (a) JPOSS 12, pl. 4. (b) JPOSS 8, pl. 4. OG . – MQYQ 24. OG . – MQYS 67. OG . – RBS ˙pl. X, no. 6. OGi. – (a) OLZG 531, no. 2. (b) RBS pl. XI , no. 4. (c) OLZG 531, no. 1a. OG . – OLZG 531 / 532, no. 1. Papyrus. – Photo. [HS 121 (4)]. Papyrus. – Photo. (a) [HS 102*, line 8]. (b) [ibid., line 9].

11

See above, n. 6.

000302

Birnbaum 01.04.11 18:17

302 19. 20. 21. 22.

Papyrus. Papyrus. Papyrus. Papyrus.

[ 51 ] – – – –

Photo. Photo. Photo. Photo.

[ibid., lines 10 and 11]. [ibid., line 13]. [HS 121 (5), text]. [ibid., signatures].

Our examination is not based on this chart but on the sources of it. The same applies to the other charts.12 Here is a list naming the documents utilised for the Hebrew chart and giving the sources from which I have made the drawings: 1. Gezer Boundary Stone. – R. A. S. Macalister, Excavation of Gezer. London 1912, vol. 3, pl. 10, fig. 1. [HS 83]. – (c. 100– 63 B. C. E.). 2–3. Hodhayoth Cave Scroll. – Photo. [HS 85E]. – (c. 50–70 C. E.). 4. OGi. – Photos. [HS 95–99] a. o. – (c. 50 B. C. E. – 70 C. E.). 5. Uzziah Tablet. – Photo. [HS 88]. – (c. 50 B. C. E. – 70 C. E.). 6. Queen Helena Graffito. – Photo [HS 100]. – (50– 60 C. E.).

Further sources used for the present examination were photos of the Kedron Dipinto [HS 100*] (60–70 C. E.), the Murabba aˆt Tephillin [HS 101], the Bar Kokhba Letter [HS 102], the Beth Mashku Document [HS 102*], (the two latter of 132–135 C. E., the Tephillin of about the same time), the Kafr Bir im Inscription [HS 103] (150–200 C. E.). In the Nabatean chart I have given the most frequent forms of each letter, as well as some special forms. I have selected and drawn them from the photos, or, when necessary, hand copies, in the original publications.13 For each specimen mentioned in the discussion the exact reference to the document where it occurs is given.

The Dates Our Neg material belongs to two different periods, which are however separated by no more than about two generations: nos. 1–16 date from roughly the last hundred years before the fall of the Temple, say from about the middle of the 1st cent. B. C. E. to 70 C. E. Nos. 17–22 are of the Bar Kokhba period (no. 17: ca. beginning of the 2nd cent. C. E.,14 nos. 18–20: 132–135 C. E., nos. 21–22: 134 C. E.).

12

13 14

A chart can obviously not give all the specimens occurring in the sources, and must not, therefore, be used as the basis of research. But it can serve to illustrate the text to some extent. CIS, JaSa, Litt. a.o. This date is purely palaeographical. At the time of the discovery I dated the papyrus as in the 1st cent. C. E. (PEQ 84 [1952], p. 118 {= Chapter 49 in this

000303

Birnbaum 01.04.11 18:17

The Negeb Script

303

The Alphabet Aleph In no. 17, aleph has two different forms, one for the initial and medial, the other for the final position. We shall therefore have to devote two sections to aleph. It will be convenient to deal with final aleph first.

Final Aleph In no. 2, the letter consists of a straight stroke, sloping down rightwards, and joined, not far from the centre, by a smaller stroke that comes slanting from the right. In this document aleph occurs only in final position; we do not, therefore, know whether our sign is exclusively a final. It is obviously an abridged form, the usual left stroke being omitted. As it occurs in final position, its development is evidently due to the fact that aleph is a very common ending in Aramaic. The letter in no. 17 (second column) is a cursive counterpart to the one in no. 2: the strokes are curved, the right one tends to join the main stroke at the top. This development has been fully accomplished in no. 21 – the form here has become so unlike that in no. 2 that, at first glance, they hardly look related. The specimen in no. 1415 is, it seems, of a different type – the right stroke slants down, not to the left, but to the right, i. e., it corresponds to the non-final form of nos. 3, 9, 14, 16, 17.

15

volume, p. 273}, more precisely “about the second half of the 1st cent. C. E. [. . .] or fairly close to it” (PEQ 85 [1953], p. 41 {= Chapter 50 in this volume, p. 297}. In the Temporary List of Facsimiles in the second fascicle of The Hebrew Scripts I gave the first third of the 2nd cent. C. E., on the basis of the study of the new material as here presented. {This Temporary List no longer appears in the final form of the HS, being replaced by the definitive list in the Text Volume, pp. 317–336. – Eds.} – If J. T. Milik’s conjecture (RB 61, p. 189; see above, n. 9) of sr after the ayin at the end of line 1 is correct – as seems probable – the resulting year, the eleventh, could be taken as a year of the Era of Arabia. The date of the papyrus would then be 117 C. E. Less likely would appear to be the regnal year of the last Nabatean king, Rabbel II , which would make the date 81 C. E. The first date would completely tally with the palaeographical result, the second practically so. The inclusion of this document might require explanation. The name has been read 〈ús〉ui by one scholar, and Ñmaui by another. But the word is askla (Alexa). The same name (cf. Lidzbarski, ESE, vol. 2, p. 197) occurs on the British Museum ossuary of askla rnqn.

000304

304

Birnbaum 01.04.11 18:17

[ 51 ]

Fig. 1. Chart of the Negeb Script (drawn by the author)

000305

Birnbaum 01.04.11 18:17

The Negeb Script

Fig. 2. Chart of the Hebrew Script (drawn by the author)

Fig. 3. Chart of the Nabatean Script (drawn by the author)

305

000306

Birnbaum 01.04.11 18:17

306

[ 51 ]

Comparison In Heb there is no final aleph. Nab and Neg both have a final aleph, although it is not consistently employed. The forms themselves are, however, completely different in Nab and Neg. The two-stroke sign of Neg contrasts strongly with the four-stroke type of Nab – which latter is a very florid version of the Heb ‘Square’ form, i. e., while the final aleph of Neg is an abbreviation, that of Nab is formal, often even highly ornate. Yet once we have a specimen16 which, except for the shortness of its top stroke, tallies with Neg (cf. no. 17 and, particularly, no. 21). (Although the letter stands at the edge of the stone it does not seem to be just a mutilated specimen of the complete letter). This inscription may perhaps contain a clue to the development of our form. One of the examples is in non-final position. Here the left stroke is small and issues from the top of the main stroke (a common type in Heb). At the next stage these two strokes might have been written cursively, with one pen lift, starting from the bottom of the left stroke. It is easy to see how this might have led to the little stroke eventually being swallowed up. Derivation The existence of a final aleph in both Nab and Neg seems to indicate relationship, and the form in Khal17 would point in the same direction. On the other hand, this (Khal) is our only inscription with Nab and Neg tallying; everywhere else the forms are not only unlike but result from different – even opposing – trends: in Neg it is development towards cursiveness, in Nab, towards greater formality. If, in spite of this, we were to decide in favour of derivation from Nab, our terminus ad quem would be the 2nd cent. B. C. E.

16

17

Khalasa Inscription, PEFA, 3 (1914), p. 145 {see above, Chapter 50, n. 14}. – King Haretat, mentioned in this inscription, is taken to be Haretat II , who reigned approximately 120–96 B. C. E. However, the script differs very much from that of the two subsequent inscriptions, which are of about 70 and 62 B. C. E. respectively. A period of roughly one or two generations seems too short to account for the change (even allowing for the regional difference) – it is of a much earlier type. I wonder whether the inscription might not refer to Haretat I, instead – he is known to have been reigning in 169 B. C. E. See preceding footnote.

000307

Birnbaum 01.04.11 18:17

The Negeb Script

307

Non-final Aleph There are three types of this letter. In nos. 10 and 18 we have the ‘abridged Square’ form (identical with the final of no. 2). Was perhaps the frequency of the final form (due to the frequency of the ending aˆ in Aramaic) largely responsible for the ousting of the non-final aleph? This factor would presumably have combined with the general tendency for cursive shapes to replace formal ones.18 The second type is represented by nos. 21 and 22. It tallies with the final aleph in the same document. (Except that in the final position the lower stroke is often longer – a natural final feature). The same explanation as before would be applicable: the final form has ousted the other. – No. 7 seems also to belong to this second type. The third type appears in nos. 3, 9, 14, 16, 17. The right stroke is horizontal, or slants down very little to the right. The left end joins the top of the downstroke, or almost so. As to its origin, it might be derived from the ‘abridged Square’ form, or rather from its cursive counterpart. We might picture the development something like this: The right stroke moved increasingly upwards until it reached the top of the downstroke and thus ended up in a horizontal position. This pressed the main stroke more and more into a vertical stance, because an acute angle is not quite convenient to make (with one pen lift) and / or the writer was intent on preserving the customary angle – roughly one of 90 degrees. When the horizontal-vertical pattern had been reached, a change of direction in the writing of the top stroke took place: it was now drawn from left to right. This caused it to become slightly arched: it was now able to slope down rightwards and the downstroke leftwards – the two strokes having exactly interchanged directions: the letter was turned clockwise through roughly 90 degrees. If this reconstruction is correct the question arises as to why, in no. 17, the development took place in non-final position only, while in the final one, the earlier form remained static. We might go on to ask: 18

Cf. the ousting of the non-cursive final aleph in Nab by the cursive, non-final type – a movement already in progress towards the end of the 1st cent. B. C. E. Later, in final position, the lower part has very often just turned into a thick stroke (MahAg, see below, n. 25). This must have spread to the non-final position, too, because in Arabic, which developed from Nabatean, there is only one form (it is a stage further advanced: the thickness has disappeared, the whole letter is one continuous stroke).

000308

Birnbaum 01.04.11 18:17

308

[ 51 ]

why is there in nos. 21 and 22 no trace of such a development? It might perhaps be that these nos. represent a more developed stage than no. 17 – by now, as we mentioned in our description of the second type, the non-final form had been altogether ousted.19 Comparison Neither Heb nor Nab have any forms of non-final aleph which can be said to correspond to the non-finals of Neg. Derivation We cannot, therefore, derive non-final aleph of Neg from either of the two other scripts.

Beth There are two main types of beth. Type A is characterised by a long base, type B by the absence of the top stroke. In type A, we find quite a number of variants. Nos. 11, 13 and 21 have a more or less straight and horizontal top stroke, which curves into the roughly vertical downstroke. In nos. 7, 17, and 22 the top stroke slants down to the right. In nos. 2, 16, 17 (2), 18 and 20 it is no longer possible to speak of two separate strokes because they are fused to form one wide curve. The downstroke slants to the left, so that it reaches the base at the left end (the letter then resembles a Roman Z or a figure 2) or not far from it, never beyond the centre (nos. 11, 13, 16, 21[1] and 22). The base is straight, ends with a downward bend, or it is arched. In type B there is no top stroke: it has, as it were, been swallowed by the downstroke. The downstroke is usually oblique, slanting either to the right or left; it is sometimes short. The base more often than not slopes down to the left. Between the two strokes there is a right angle in no. 15, but an obtuse one in nos. 3, 5, 6 and 8 – in the latter the whole form would be better described as a curve in oblique position. Type B occurs only in ñb or rb. These words are, however, sometimes written in type A, too. Neither of the two types can have developed from the other. The parent form must have had a top stroke like type A, and the downstroke 19

This, by the way does not imply that there must have been a considerable interval between nos. 17 and 21. At every moment a minority goes on using a certain form which the majority have given up in favour of a more advanced one, and it takes a long time before a new form is in universal use.

000309

Birnbaum 01.04.11 18:17

The Negeb Script

309

have joined the right end of the base, as in type B. In other words, the letter must have been trilateral. Comparison In Heb the types of Neg do not occur but, in both Heb and in type A, after about the middle of the 1st cent. B. C. E., the downstroke and base do not together form a corner on the right. In Nab, type A does not occur. The early Nab form has the complete vertical-horizontal construction, as in no. 11, but the base runs from right to left. The cursive type of Nab, however, tallies with type B, our earliest specimen deriving from the end of the 1st cent. B. C. E.20 Nab develops a differentiation between final and non-final position. Of this there is no trace in Neg. Derivation Heb and Neg might possibly be connected in this way: From about the middle of the 1st cent. B. C. E. onwards the base in Heb juts out a little on the right and thus resembles Neg. The explanation for this development might be that the downstroke reached the line bottom at roughly right angles. In cursive writing it may have been a trifle inconvenient now to turn the pen abruptly to the left, in order to make the base. The hand, therefore, lifted the pen off the paper, moved leftwards and drew the base from left to right. It would be perfectly natural for this stroke to go a little beyond the precise end of the downstroke. This form might then have become the starting point for a development which resulted in Neg. No. 11 represents the first of the stages: the base is produced very far to the right. The writer clearly did not feel any need to make the downstroke run exactly to the right end of the base. In no. 13 we have a less extreme version of such a base, but, on the other hand, the top stroke is not properly straight – it forms, with the downstroke, a wide curve. This weakens the idea that the top stroke has to be horizontal, thus enabling the curve, on occasion, to be retranslated into an oblique top and an oblique downstroke (nos. 7, 17[1] and 22). That, in its turn, leads to the shallow curves of nos. 2(2), 16, 17(2) and 20. But elements from different phases can be combined: no. 21(1) has a top stroke – albeit a short one – and a curved downstroke; no. 21(2) retains the horizontal top stroke but has an oblique downstroke. (As all these forms occur side by side in our material, they cannot be utilised for chronological purposes.) 20

J. Cantineau, Le nabate´en. Paris 1930–1932, vol. 2, p. 7, no. V, after 9 B. C. E.

000310

Birnbaum 01.04.11 18:17

310

[ 51 ]

The obliqueness of the downstroke caused the base, which had already started projecting beyond the letter, by now to lie half outside it – or even completely so, when the downstroke met the base at its left end. A further consequence of the new method of writing the base (i. e., from left to right) was the arching of the stroke (nos. 20 and 21[1]) or the growth of a tail (nos. 17[l], 21[2] and 22). One specimen with the straight top stroke – no. 11 – still has the initial tick, and no. 18 even gives us a full-curve example of the letter. The plain straight top stroke does occur in Heb of the 1st cent. B. C. E. / C. E., but exceptionally so; the nature of its connection with Neg must, therefore, remain in doubt. If the connection with Heb as suggested at the beginning of this section is correct, the terminus a quo for type A is about the middle of the 1st cent. B. C. E. Type B is clearly of Nab origin, and its terminus a quo, therefore, is the end of the 1st cent. B. C. E.

Gimel Comparison Heb tallies with Neg. In Nab, during the first century for which we have material, it occurs only once.21 Here – in the first half of the 1st cent. B. C. E. – the left stroke issues, as far as can be made out from the reproduction, almost from the top of the right stroke, and in Savignac’s drawing22 the letter is a plain circumflex – a form which was already disappearing in the 3rd cent. B. C. E. But in the rest of the material Nab tallies with Neg. Sometimes, however, the left stroke is horizontal. Derivation The rareness of gimel in our material, and in the first century or so of our Nab material, makes it difficult to say whether Neg is more likely to be connected with Heb or with Nab. The same holds good for the terminus a quo. For Heb it would be the beginning of the 2nd cent. B. C. E., for Nab the end of the 1st cent. B. C. E.

21

22

G. Dalman, Neue Petraforschungen und der heilige Felsen von Jerusalem. Leipzig 1912, no. 90; the date of the inscription is about 62 B. C. E. (96–90 B. C. E. seems to me less probable). RB 22 (1913), p. 441.

The Negeb Script

311

Daleth The left stroke is straight, curved or angular. It joins the downstroke either below the top or at it. In no. 17 the stance of the letter differs by about 45 degrees, clockwise, from that of the other documents; the downstroke is shorter than the middle one. This form is evidently later and more cursive than that of the other documents. Frequently the corners are rounded – the letter then becomes practically a wavy line. In nos. 2(1) and 21 we have specimens of an earlier form, the top stroke joining the downstroke below the top. Comparison In Heb the top stroke is, as a rule, straight and horizontal – which tallies with nos. 2(1), 4, 6, 9 and 20. In Heb the top stroke joins the downstroke below the top, in Neg it generally joins it at the top. However, in nos. 2 and 21 we have the Heb construction. In Nab the top stroke is curved, as in nos. 2(2), 21 and 22; it seems always to run to the top of the downstroke, tallying in this respect with nos. 4, 6, 9, 17, 20 and 22. Only in this way could a form like no. 17 have come into being. A similar one occurs in a Nab inscription of 124 C. E.23 Derivation There are connections with both Heb and Nab. The most important of these, however, seems to be that the downstroke issues from the end of the top stroke, thus favouring a Nab derivation. All the same, the occurrence of Heb constructions must be taken into account. The Nab form is already present in the earliest documents. Hence, the 2nd cent. B. C. E. would be the terminus a quo for Neg. The terminus ad quem can be established from the width of the letter. Already in CIS II , no. 34924 it is very narrow, and medium-wide forms become rare; in MahAg25 the top stroke is tiny, sometimes being practically omitted. Hence the origin of Neg would hardly seem to go back farther than, say, the end of the 1st cent. B. C. E.

23 24 25

Litt, p. 25. Of ca. 70 B. C. E. J. Starcky, “Un contrat nabate´en sur papyrus”, in: RB 61 (1954), pp. 161–181; of 60– 69 C. E.

000312

Birnbaum 01.04.11 18:17

312

[ 51 ]

Heˆ The top stroke is short, or of medium length, horizontal or oblique (more often sloping up to the right than slanting down to it). In penwriting it is usually somewhat arched. It joins the right downstroke not too far from the top. This downstroke has generally a rightward curvature but straight forms are not rare. The left downstroke – straight or slightly curved – issues from the same point as the top one, and slopes to the left, more often than not stopping before the line bottom is reached. Comparison This type exists in Hebrew but becomes extinct at the beginning of the 2nd cent. B. C. E., i. e., just before the time of Heb and Nab. The form in no. 19, with a tick initiating the top stroke, develops in Heb during the second half of the 1st cent. B. C. E. Nab tallies with Heb, as developed after the beginning of the 2nd cent. B. C. E., but frequently retains the oblique position of the top stroke and / or the left one. The closed form of Nab which already appears in our early material as a final, and in the 1st cent. C. E. begins to invade the non-final position,26 is absent from Neg. Derivation The form in no. 17 would make the second half of the 1st cent. B. C. E. the terminus a quo for Heb. But as it is the only exception among a great number of specimens with a straight top stroke, it might be safer not to base any terminus on it. The obliqueness of the left and / or top strokes would favour connection with Nab. Its terminus a quo would then be the 2nd cent. B. C. E. (Khal).27 The absence of final heˆ might suggest the early part of the 1st cent. B. C. E. as the terminus ad quem.

Waw Waw is very often just a straight line. In nos. 2, 17 and 19 there is an obtuse-angled top. Comparison In Heb from about the middle of the 1st cent. B. C. E. onwards, the straight-line form is occasionally to be met with. None of the Neg forms 26 27

JaSa no. 22, of 75 / 76 C. E. ( The specimen JaSa no. 2 seems to me extremely doubtful). See above, n. 16.

000313

Birnbaum 01.04.11 18:17

The Negeb Script

313

are used in Nab, with the exception of Khal,28 where the obtuse-angled type occurs side by side with a right-angled one. By the early part of the 1st cent. B. C. E. Nab has already developed its round head.29 Derivation Neg is connected with Heb through the straight type. The terminus a quo is accordingly the middle of the 1st cent. B. C. E. There is also a connection with Nab, yielding 169 B. C. E. as the terminus ad quem.

Zayin Zayin starts by sloping down to the left but, immediately after, it changes its direction and either runs down vertically or slopes to the right. Near the line bottom it may curve leftwards, in which case the letter becomes a wavy line. Comparison In Heb the first of these forms is the usual one from the 1st cent. B. C. E. onwards. In Nab the earlier, straight form continues to be used right up to the end. Derivation Hence Neg is connected with Heb. The terminus a quo would be about the beginning or middle of the 1st cent. B. C. E. Its terminus ad quem would be the 2nd cent. C. E., when the acute-angled top begins to become prevalent.

Heth ˙ There are three types. (A) The middle stroke slopes from the top of the left stroke to about the centre of the right one: nos. 2(2), 17(2). (B) The middle stroke slopes down to the bottom end of the right stroke: nos. 2(1), 7, 17(1). (C) The two strokes are connected by an arch. The left stroke is shorter, with the left end of the arch resting on the top of it: no. 21. (No. 6 is presumably only a misshapen specimen of this type.) Both B and C are clearly developed from A.

28 29

Ibid. See above, n. 24.

000314

Birnbaum 01.04.11 18:17

314

[ 51 ]

Comparison In Heb, type A is very rare. It no longer occurs after the middle of the 1st cent. C. E. There are no traces of types B and C. In Nab, the earliest specimen of type A is of the third decade of the 1st cent. B. C. E.30 type B does not occur. Type C makes its appearance early in the 1st cent. C. E.31 It is rare in our material but was presumably prevalent; it may even have been the only form employed in pen-writing – MahAg32 has no other type. Derivation Thus while type A gives no clue, type C links Neg with Nab. As regards Heb, the terminus ad quem for type A would be the middle of the 1st cent. C. E.; as regards Nab, the terminus a quo would be the late 1st cent. B. C. E.

Teth In the earlier period (nos. 2 and 13) teth consists of two downstrokes: the left one ends by curving at the bottom towards the right, the angular or curved right stroke touches, in passing, the end of the bottom curve of the left stroke (type A). What happened here is evident: The right stroke was originally on the same level as the left one. The two sides of teth were each made separately: they had never been written with one pen lift but with two, both from above, with a downward movement – the right one clockwise, the left one anti-clockwise. Consequently, the two strokes did not always meet exactly at a point because, in drawing the right stroke, the hand easily runs on a little – the left stroke was written first, as the development of the letter proves. Soon this downward projection would have become the norm. No. 2 is at a later stage of evolution than no. 13: the angle on top of the right stroke has already disappeared. From this type a new cursive development takes its start, a one-lift form grows up through ligation33 between the end of the bottom curve and the top of 30 31 32 33

c. 48–30 B. C. E. – Ch. Clermont-Ganneau, Recueil d’arche´ologie orientale. Paris 1888–1924, vol. 8, p. 229. JaSa no. 2, of 4 C. E. See above, n. 25. For the sake of convenience I have coined the term ‘ligation’ (and ‘ligate’) to describe a very frequent development in cursive writing: At the end of a stroke the pen is not lifted off, as it should be, to start the next stroke but is slid along the paper instead, thus binding the end of the one to the beginning of the next, creating a new part of the letter.

000315

Birnbaum 01.04.11 18:17

The Negeb Script

315

the right stroke. The result of this evolution becomes apparent in the second period: nos. 17 and 21 (type B).34 Comparison In Heb there is no trace of these forms. In Nab, type A does not occur either, but type B does, although only seldom. In the graffiti of the Wadi Mukattab, on the contrary, it is frequent – evidently because it is a cursive form. In MahAg, where the majority of specimens are of the ‘Square’ type, there is a similar form: the right stroke finally runs back into the left part of the letter, not at the bottom, however, but crossing the middle stroke, producing something like an open and imperfect, recumbent figure 8. Our earliest dated example is from the end of the 1st cent. B. C. E.35 From after the 1st cent. C. E. no dated specimen seems to be extant, but in view of the great frequency of this form at the Wadi Mukattab, it presumably survived beyond that time. Derivation It is hardly likely that Nab and Neg would independently of each other have arrived at the same form. Our terminus a quo is accordingly the end of the 1st cent. B. C. E., and the terminus ad quem a century later.

Yodh Yodh is generally short but somewhat longer specimens occur in no. 21. It is usually straight and headless. Sometimes, near the end, it curves slightly leftwards. In no. 17(1) it consists of a bend or curve. No. 21 contains obtuse-angled examples as well. Comparison Heb has this cursive type, side by side with the formal one – i. e., with an acute-angled head – from the second half of the 1st cent. B. C. E. onwards. Nab has it in MahAg.36 Derivation Hebrew origin seems more likely than Nab, because on some ossuaries the formal Heb type appears side by side with the cursive Neg type, 34 35 36

Cf. the reconstruction of this development, on purely theoretical grounds, on p. 29 of my article in PEQ 85 (1953) {= Chapter 50 in this volume, p. 286}. CIS II , no. 163, of 2 / 1 B. C. E. See above, n. 25.

000316

Birnbaum 01.04.11 18:17

316

[ 51 ]

while there is no trace of the Nab type – two arches meeting at a point on the left – which is already fully developed by the early 1st cent. B. C. E.37 Our terminus a quo is accordingly the middle of the 1st cent. B. C. E. Furthermore, a negative feature speaks against Nab. Like Heb, Neg has no special form for yodh in final position, while Nab develops one during the 1st cent. B. C. E. Our terminus ad quem, in relation to Nab, would then be the end of the 1st cent. B. C. E.

Kaph The top stroke varies in length. It is usually straight and joins the downstroke below the top, except in no. 18. Also slightly curved forms occur. The downstroke and the base meet at right, or slightly obtuse, angles (nos. 14, 18, 21[1]); sometimes they curve into each other (nos. 17, 21[2], 22). The base is straight. The letter is of line height and rather wide. Comparison In Heb curved top strokes have disappeared by the middle of the 1st cent. B. C. E. In Nab, straight top strokes are not rare – our earliest example is from the end of the same century38 – but the curved type remains the rule. In Heb, the top stroke joins the downstroke either below the top, or at it, the latter form becoming increasingly prevalent from the middle of the 1st cent. B. C. E. onwards. In Nab, we have only the second type. In Heb, the letter is at first somewhat narrow, but becomes rather wide from the middle of the 1st cent. B. C. E. onwards. In Nab, the letter gradually becomes increasingly narrow. In MahAg39 the top stroke and base are tiny, or have partially disappeared. In Heb, the base is straight. In Nab, straight forms cease at the beginning of the C. E. Shortly before, curved bases appear. Derivation Most of the features of Neg might be derived from Heb or Nab. The width of the letter seems to favour Heb. As far as the top stroke is concerned, the terminus ad quem for Heb origin would be the middle of the 1st cent. B. C. E., for Nab origin the end of the century. 37 38 39

See above, n. 24. See above, n. 35. See above, n. 25.

000317

Birnbaum 01.04.11 18:17

The Negeb Script

317

With regard to the juncture of down and top strokes, the terminus ad quem would be about the middle of the century – only Heb origin needing to be taken into account. As to the width of the letter, the Heb terminus a quo would be the middle of the 1st cent. B. C. E. As regards the straightness of the base, the terminus a quo for Heb is about 200 B. C. E., the terminus ad quem for Nab the beginning of the C. E.

Final Kaph The top stroke varies in size. It is horizontal or oblique, straight or – apparently a rarer form – curved. It joins the downstroke somewhat below the top. The downstroke is straight, or near the end, curves gently to the left. Comparison In Heb, curvature of the top stroke has disappeared by the 1st cent. B. C. E., and the right stroke of the corresponding tick has become very long. No. 17 is of this type while no. 21(1) continues the curve tradition. In Nab, straight top strokes are exceptional, the curve persists until the end. It joins the downstroke to form a pointed top, which, by the first cent. B. C. E., has become rounded. This results in the development of a symmetrical curve, in the opposite direction from the original one, i. e., an arch appears between the original top stroke and the downstroke. The letter being narrow, its two halves are pressed together.40 Derivation Neg can be derived from Heb. The terminus a quo of the straight type would be about the middle of the 1st cent. B. C. E., and this time would also be the terminus ad quem for the curved type. Derivation of the latter type from Nab is unlikely, as in our earliest Nab document, the meeting point of the top stroke is already at the upper end of the downstroke, not somewhat below it, as in Neg.

40

CIS II , no. 196, of 37 C. E.

000318

Birnbaum 01.04.11 18:17

318

[ 51 ]

Lamedh Lamedh usually consists of a straight, long stroke above the line ceiling, reaching down well into the middle of the line height and finally curving leftwards. But there are considerable variations: some specimens are completely straight, others are straight and end by bending, at right angles, to the left, etc. Comparison Heb has none of these forms. Nab has them all (and others, too). Nab differs from Neg, however, in having a special form for the final position, although it is not consistently employed; MahAg41 has a final form. Derivation The derivation of Neg from Nab seems evident. It would have originated before the end of the 1st cent. B. C. E., when Nab begins to have a final lamedh.

Mem The outline of mem varies greatly. There is perhaps a preponderance of oval or elliptical forms, but triangular, quadrilateral and circular ones are not rare. This description does not include the downstroke that issues from the top left hand side of the letter. It is altogether absent, however, in one specimen of no. 5, and in nos. 15 and 16, while in nos. 10–14 it issues from the bottom. The length of the left stroke varies from very small (no. 9) to long, stretching considerably below the line bottom (no. 5). The letter is usually written clockwise, and, in making the left side, the pen often does not go right up to the top, a gap developing. That the open form is not the original one, from which the other developed by closing up, becomes clear when we try to visualise how this type could have developed: The outside stroke is obviously the original left stroke – as in no. 20. Ligation took place between the bottom end of the right stroke and the starting point of the left downstroke. When the whole letter is thus made with only one pen lift it can easily happen that the pen stops short of the top, a gap then being formed. Nos. 11–14, which have only one left stroke, are perhaps survivals of an earlier, closed stage, whose top part later became stunted, finally disappearing. 41

See above, n. 25.

000319

Birnbaum 01.04.11 18:17

The Negeb Script

319

Nos. 15 and 16 might either be derived from the same earlier form by simplification of the top plus a general rounding, or from the triangular type by rounding. Nowhere does the left stroke jut out on top. This seems to indicate that our form developed at a time when the left stroke no longer crossed the top one but issued from it. Comparison None of these forms occur in Heb. Here the letter is never written in one – clockwise – movement, beginning and ending on top left, but always remains open. This gap has of course nothing to do with that of Neg. In Neg the gap is at the top and represents a new development, while in Heb it is at the bottom and represents the preservation of the preceding stage. Nowhere in Neg is the top stroke a curve or a tick – developed from a curve – as in Heb. Only no. 20 appears to have a corresponding form. Nowhere in Heb is the top stroke stunted, or more or less fused with the downstroke, as it generally is in Neg – not to mention forms like nos. 11 and 14, where it has actually disappeared. Until about the end of the 1st cent. B. C. E., Nab tallies with Heb as against Neg. At that time, however, Nab begins to tally with Neg, in that the letter is closed. But there is not always a left outside stroke, the letter having the construction of a Heb final mem. As in Neg, there is frequently a gap at the top of that downstroke. The earliest occurrence in our material is in the final years of the 1st cent. B. C. E.42 Unlike Neg, the top stroke continues in existence during the remaining life of the script. In contrast with both Heb and Nab, Neg has no special form for final mem. The difference between the two forms in no. 5 has presumably nothing to do with the distinction between non-final and final – compare no. 4, which is from the same ossuary as no. 5, and refers to the brother of the man mentioned there: here, in no. 4, in the same name, that of their father Amram, the two mems are identical. However, already from the late years of the 1st cent. B. C. E. onwards, the two mems of Nab are not, as a rule, much differentiated, and it happens that they are practically identical. Derivation It is clear that Neg and Nab are closely connected. The forms of the letters, as well as the lack of differentiation between a final and a non42

See above, n. 24.

000320

Birnbaum 01.04.11 18:17

320

[ 51 ]

final form, point to the late years of the 1st cent. B. C. E. as the terminus a quo for Neg.

Nuˆn The top stroke is straight (in no. 17 bent), vertical or oblique. The bottom stroke is horizontal or oblique. The two link up rather angularly. In nos. 19 and 21 the top stroke is wavy. Comparison Both Heb and Nab tally with Neg. Hence our letter affords us no clue as to relations between the scripts.

Final Nuˆn The letter is either straight, vertical or oblique, or its lower half bends or curves leftwards (type A). In nos. 4(a), 16, 18(a), the stroke starts by sloping down rightwards but very soon bends to the left (type B). This is what remains of the wavy form of the preceding period. Comparison In Heb, type A occurs occasionally on ossuaries, but not in pen-written documents, which have only the wavy form. In Nab, type A is exceptional. Type B is dominant – the wavy type going out of use at the beginning of the C. E. – our earliest specimen dating from the end of the 1st cent. B. C. E.43 Derivation As regards type A, connection with Heb would be more likely than with Nab, but type B is definitely linked with Nab. For type A, the terminus a quo would be about the middle of the 1st cent. B. C. E., for type B the end of the century.

Samekh The letter varies a great deal between roughly triangular and approximately circular outlines. The left stroke, generally a straight one, closes the letter. It often juts out below the line bottom, rarely upwards (type A). There are also open types. One (type B: nos. 20 and 22), where the 43

See above, n. 31.

000321

Birnbaum 01.04.11 18:17

The Negeb Script

321

leftward slant of the left stroke prevents it from reaching the bottom stroke; the other (type C: nos. 14 and 15) has the gap on top, the left stroke issuing from the lower stroke – i. e., this form is identical with the usual mem of Neg (while the mem of nos. 14 and 15 is different). There are two possible relationships between these three types. Type B might have been the predecessor of the others in this way: if the left stroke became vertical it would meet the bottom stroke, and the letter would thus be closed, resulting in type A. Type C might have developed through ligation44 between the end of the bottom stroke and the top of the left downstroke, and through the downward movement starting before the line ceiling had been reached, so that a gap ensued. The other possibility is that type A formed the starting point: type B would have developed through the left stroke being written obliquely, instead of vertically, so that it no longer met the bottom stroke, the letter becoming open. Type C would have developed in the same way as described above, with type B as the starting point. Comparison In Heb we have, in the 2nd cent. B. C. E., a type in between types A and B: the left stroke does not curve leftwards but runs straight down. In the second half of the 1st cent. B. C. E. the right stroke reaches the left one, so that the letter is closed, as in type A of Neg – its outline, however, being more regular. The top stroke and the left one are often looped together. There is, in Heb, no trace of type C. In Nab we have all three types of Neg. The most frequent one corresponds to type A, except that the top stroke and the left one are generally looped together. This form possibly went out of use by the middle of the 2nd cent. C. E. Specimens corresponding to type B are very rare.45 Type C is fairly frequent. Our earliest example is from the end of the 1st cent. B. C. E.46 – the form survives type A. Nab differs from Neg in having a final samekh: the left stroke projects very far below the line bottom. We have examples in the 1st cent. C. E.47 (It might be mentioned that samekh is not a common letter, and that it would occur even less frequently in final position.) 44 45 46 47

See above, n. 33. RB 14 (1905), p. 592, 15 (1906), p. 131, of about 148 C. E. – Perhaps CIS II , no. 161, of 95 C. E. CIS II , no. 160, of 4 B. C. E. See above, n. 26. – CIS II , no. 209, of 31 C. E.

000322

Birnbaum 01.04.11 18:17

322

[ 51 ]

Derivation Heb might be the source of type A. The terminus a quo would then be the second half of the 1st cent. B. C. E. This type might on the other hand have its source in Nab, where it is already present early in the 1st cent. B. C. E.48 This derivation is possibly preferable, because both in Nab and Neg, the outline of the letter is more variable than in Heb. Heb might also be the origin of type B. In that case, the middle of the 1st cent. B. C. E. would be our terminus ad quem because in the second half of the century, the letter becomes closed. Nab as the source of type B is possible. The terminus a quo would then be the end of the 1st cent. C. E. Type C is connected with Nab only. Its terminus a quo is the end of the 1st cent. B. C. E.

Ayin Ayin consists of a straight, or almost straight, main stroke which slopes down from right to left, from line ceiling to line bottom. On the left there is a straight or curved, usually small, stroke which generally reaches the right one fairly near the top, or even at it, tending to be in a horizontal position. Comparison In Heb the right stroke usually starts with a vertical part. The left stroke has no tendency to occupy a horizontal position and reaches the right stroke at the centre, never near, or at, the top. Nab, from shortly before the C. E. onwards, tallies with Neg (although the other form does not disappear altogether). Derivation The link between Neg and Nab is evident, and the late 1st cent. B. C. E. is our terminus a quo.

Pe A short, oblique stroke on the left, together with the beginning of the main stroke, form a pointed top. The vertical, or oblique, downstroke is convex, straightish or wavy. The base is horizontal or slants down to the left. The angle between the two strokes is roughly one of 90 degrees; the corner thus formed is not, however, actually pointed. 48

See above, n. 24.

000323

Birnbaum 01.04.11 18:17

The Negeb Script

323

Comparison Heb has the same form. In Nab, the top is usually rounded and the base generally a horizontal curve. Derivation Hence Neg corresponds to Heb. Accordingly, the beginning of the 2nd cent. B. C. E. is the terminus a quo for Neg.

Final Pe The upper part is identical with that of non-final pe, but the downstroke, instead of turning left at the line bottom, continues downward, and in the lower half usually develops a slant to the left. Comparison Heb shares this form, too, with Neg. In Nab it is exceptional.49 Generally, the final differs from the non-final form in having a long and straight base. Derivation Neg corresponds to Heb, and the beginning of the 2nd cent. B. C. E. is the terminus a quo.

Sadhe The letter occurs only in two of our documents, and in one of them it is mutilated. The main stroke is straight or wavy. On reaching the line bottom it curves into a more or less straight horizontal base of medium length. The right stroke is slightly concave in no. 2 (type A), and in no. 21[1] (type B) seems to be an angle, open on top. No sadhe in final position occurs in our material. Comparison In Heb, type B is the norm; straight strokes are very rare, concave ones do not occur. From Nab we have, unfortunately, no examples from before the beginning of the C. E. Then the right stroke is a generally very narrow arch. This is unlikely to be the curved sequel to the old, gable-shaped or wide-arched form which had already disappeared during the 4th cent. 49

E. g., CIS II , no. 198.

000324

Birnbaum 01.04.11 18:17

324

[ 51 ]

B. C. E. It would more likely have developed from the convex type, via a straight stage, followed by a concave one. Derivation Accordingly, type A would be closer to Nab than to Heb while type B (when an angle) would tally with Heb.

Qoph The top stroke is straight, apart from no. 7. It is preceded by a small initial stroke – the remains of the left half of the original circle. Our two pen-written documents, however, no longer have it. The downstroke does not issue from the top stroke but starts from the end of the righthand stroke. In nos. 7, 17, 21 (type A), the right-hand stroke is straight or curved and slopes down to the left. In nos. 19 and 20 (type B) it is straight and horizontal, and runs along the top stroke, contiguous with it half, or all, the way. The history of type A is obvious. The downstroke originally issued from the top stroke. Later it was severed from it. Finally, the end of the right-hand stroke was ligated50 with the beginning of the down stroke. Type B might have developed from type A, by the downstroke again issuing from its original position, but the ligation being preserved. Comparison Heb has none of the Neg forms. In Nab from our earliest examples51 onwards until our last, the letter is generally closed. The top part is quadrangular or circular. However, there are occasionally open forms, and these tally with Neg, with the difference that in Nab the right top comer – except in one case – is not pointed. Our specimens date from the 1st cent. C. E.;52 from the middle of the next century we have a slightly open one.53 Nab has a final qoph, characterised by a very long downstroke which issues from the lowest point of the top circle. Our examples are from the

50 51 52

53

See above, n. 33. See above, n. 24. CIS II , no. 209, of 31 C. E., no. 182, of 50 C. E. – Litt no. 23, of prob. 70 C. E. – R. Dussaud / F. Macler, Voyage arche´ologique au Safaˆ et dans le Djebel ed-Druˆz Paris 1901, p. 187, of 96 C. E. J{ebel} Ramm Dipinto, of 147 (RB 1935, pl. X, 1a) {Cf. M. R. Savignac / G. Horsfield, “Le temple de Ramm”, in: RB 44 (1935), pp. 245–278}.

000325

Birnbaum 01.04.11 18:17

The Negeb Script

325

1st cent. C. E.54 In our Neg material there are two cases of the letter standing in final position, and, in one of them (no. 19), it seems that the downstroke is longer than in the other qophs of the document, running into the letter on the next line. Derivation A connection with Heb would be possible, there being Heb forms open on the left. But these are rare. The existence of an almost identical type in Nab links it with Neg. Its terminus a quo is about the beginning of the 1st cent. C. E.

Resh Except in no. 21, the downstroke starts from the end of the middle stroke. The strokes are either vertical-horizontal (type A) or the whole letter is turned clockwise by 45 degrees. In the latter the top stroke is either particularly long (nos. 7, 17[1]) (type B) or rather short (nos. 16, 17[2], 22) (type C); the angles tend to be obtuse. Thus the letter easily becomes rounded, resulting in a wavy line (type D). In no. 21, the top and middle strokes form a curve, which is generally stretched into a straight line. In either case this part does not run to the top of the downstroke but meets it lower down (type E). Comparison Heb, from about the middle of the 1st cent. B. C. E. onwards, is of the vertical-horizontal type (A). In Nab, the top stroke is generally a curve, joining the top of the downstroke – mostly in a horizontal position – but straight top strokes are also to be found. No. 15 possibly corresponds to the curved form; no. 21(1) has a curve, which joins the downstroke below the top. One of the characteristics that distinguish Nab from both Heb and Neg, is the excessive narrowness of the letter; in MahAg55 the top stroke has practically disappeared. Derivation Type A connects Neg with Heb. It is possible that type E links Neg with Nab. The terminus a quo for Heb would be about the middle of the 1st cent. B. C. E., for Nab the 2nd – 1st cent. B. C. E.

54 55

See above, n. 25 and CIS II , no. 209, of 31 C. E. See above, n. 25.

000326

Birnbaum 01.04.11 18:17

326

[ 51 ]

Shıˆn There are a number of forms but they may perhaps be said to be variants of one type. The main characteristic is that the two right strokes constitute a unit in relation to the left stroke. They meet at about the centre of the left stroke. In nos. 3 and 14 they meet before making contact with the left stroke, so that the letter falls apart. Sometimes it is only the middle stroke that reaches the left one, the right one then running into the middle one (nos. 17, 18, 22). The left stroke generally slants down rightwards but vertical stance is not uncommon. The strokes are straight or curved, the latter being more frequent, it seems. Comparison In Heb the middle stroke does not form a unit with the right one, as in Neg. The right stroke slopes to the bottom of the left stroke, the middle one is parallel with the right stroke, reaching the left at about the centre. Thus the letter is, unlike Neg, triangular in outline. In the 1st cent. B. C. E. the left stroke is generally produced a little in a downward direction, and, on the ossuaries, there are a few specimens where it is very long (e. g., in sˇlmsywn brt sˇm wn)56 as in Neg. However, there is this difference: in Heb˙ the left stroke extends beyond the line bottom, the three strokes filling the line, while in Neg the whole letter remains within the line height, the main part of the letter occupying the upper half of the line height. In Heb the left stroke is very seldom vertical. In Nab as in Heb the two right strokes do not form a unit, but run parallel. But, as in Neg, and unlike in Heb, the right stroke meets the left one at the centre, so that the main part of the letter is in the upper half of the line height. Our first example comes from early in the 1st cent. B. C. E.,57 which is also the date of our last Nab specimen having the Hebrew type. In Nab the vertical stance of the left stroke very early becomes prevalent, slanting specimens very soon almost disappear. Neg differs from Nab in that the right strokes are often curved – a very infrequent feature in Nab. Derivation Neg, although not identical with Nab, is closely related with it. The terminus a quo, in regard to the position of the right part is the end of the 1st cent. B. C. E. This is also the terminus for the stance of the left stroke. 56 57

RBS pl. X, no. 3. See above, n. 31.

000327

Birnbaum 01.04.11 18:17

The Negeb Script

327

Taw The left stroke is sometimes convex but generally straight – in either case it has a vertical stance. Its bottom end is ligated58 with the top of the right stroke, resulting in a loop. The right stroke is convex or straight. The left one, at times, starts somewhat above the line ceiling, the right one rather lower, passing below the line bottom. The size of the loop varies. It is sometimes only a blob or fat stroke, the pen practically running upwards along the left downstroke itself. Specimens written with two pen lifts are very rare indeed. No. 3 is a transition form between a loopless and a looped one. The letter is often narrow, so that the top of the right part is pointed. Comparison In Heb the loop is exceptional.59 In Nab, the loop is rare. Our earliest examples come from the first half of the 1st cent. C. E.60 Derivation It is perhaps safer not to take the loop as an indication that Neg is somewhat closer to Nab than to Heb.

Conclusions The Source of the Negeb Script Certain forms would seem to point neither to Heb nor to Nab as the source of Neg: non-final aleph, heth (B), teth (A), resh (B – D). In other ˙ internal, development. words, we have here independent, Some forms allow a connection with both Heb and Nab, and do not, therefore, help us: gimel, waw, heth (A), kaph (top stroke), kaph (base), nuˆn, samekh (A), samekh (B). ˙ Relationship with Heb is indicated by beth (A), zayin, yodh, kaph, final kaph, final nuˆn (A), pe, resh (A).

58 59 60

See above, n. 33. In the Nash Papyrus, it occurs in a third of the specimens. See above, n. 40. Here it appears once in final position but other examples in that position are not looped. – CIS II , no. 182. Here it appears twice in final position while one final and the non-final specimens are not looped. – J. Euting, Sinaitische Inschriften. Berlin 1891, no. 463, has it in final position.

328

[ 51 ]

Relationship with Nab is indicated by final aleph, beth (B), daleth, heth (C), teth (B), lamedh, mem, final nuˆn (B), samekh (C), ayin, sadhe ˙ (A), qoph, resh (E), shıˆn. From this tabulation it is clear that Neg is not just a form of either Heb or Nab but that, while being related to them, it is an independent script. What is the nature of this twofold connection? The first Murabba aˆt document was too fragmentary to provide a basis towards answering this question. But since then KBeb and MahAg have been published and they enable us to attempt it. These documents show that Jews and Nabateans were living side by side, in close contact. As the names of the places where the documents were written are Nabatean (Beth Mashku, Kephar Bebhayu) the Jews were presumably the later settlers, forming a minority. If, in such a situation, we find the Jews using a script which is linked with both Hebrew and Nabatean, we seem to be safe in assuming that it came into existence through the influence of Nabatean on Hebrew, i. e., the new writing is the result of a synthesis of two scripts. The Cause of the Development That the Negeb script was obviously a small, unimportant and ephemeral writing does not rob it of its twofold palaeographical interest as a hitherto unknown unit and as the product of a synthesis. There is another point of interest: script is part of the cultural – i. e., basically, religious – heritage of a group, and has a way of persisting through all historical vicissitudes – unlike language – so long as there is no religious change in the writers.61 When, therefore, in this particular instance we come upon a script other than Hebrew being used by a Jewish group, it strikes us that this development does not conform to the general pattern. And it seems especially striking because the Jews have furnished a good example of the principle that the use of a certain script is governed by the religion of the group in question: for twenty-three centuries they have been employing Hebrew characters for a considerable number of languages – each Jewry its own language, according to time and region. Are we then to conclude that the people who used the Negeb script had changed their religion, or at least modified it? We have no grounds 61

Cf. p. 65 of my article “Jewish Languages” in: Essays in Honour of the Very Referend Doctor J. H. Hertz [. . .]. On the occasion of his 70. birthday, Sept. 25, 1942 [5703]. London [1944], pp. 51– 67.

000329

Birnbaum 01.04.11 18:17

The Negeb Script

329

for presuming that this happened – it seems out of the question that at so late a time as this the Jews of those parts should have become involved in the polytheism of the Nabateans. There must have been special factors at work to account for the amalgamation of the scripts. What these factors were we cannot guess.62 One point must be taken into consideration in this connection: The writers of our documents did not give up the use of the Hebrew alphabet. That is proved by the fact that the body of the Beth Mashku Document and some of the signatures are in Hebrew script. If this is typical, then they employed Heb for the Hebrew language and Neg for the Aramaic tongue. At first glance it might be thought that the rise of the Negeb script is, as it were, analogous to the evolution of a Jewish language.63 But the analogy applies only to the general sociological principle of religion being a group-forming factor. The development in the two cases under consideration was in opposite directions: As regards Neg, it is the Jewish element (Heb) that constitutes the material basis of the synthesis, as regards the Jewish languages, it is the other way round. The Date of the Synthesis When did the amalgamation of the two elements take place? In summing up and integrating the individual results arrived at in the course of the foregoing discussion on the letters, we hope to throw some light on this question. The latest terminus a quo and the earliest terminus ad quem – which delimit the period within which an undated manuscript was written – would indicate the time of the amalgamation. For Heb, the latest terminus a quo is the middle of the 1st cent. B. C. E.: beth (A), yodh, kaph, final kaph, final nuˆn (A). This means that in the middle of the 1st cent. B. C. E. Neg was still under the influence of Heb. The earliest terminus ad quem is also the middle of the 1st cent. B. C. E.: kaph, final kaph. This would suggest that the influence of Heb ceased at about this time. The two letters form only a narrow basis for this conclusion, but it may be corroborated by the fact that there is no later terminus a quo than the middle of the 1st cent. B. C. E. For Nab, the latest terminus a quo is the first half of the 1st cent. C. E.: qoph. This being a single case only, it might be safer to assume 62

63

It seems evident that we are not dealing here with conscious adoption of certain signs from another writing, which is, of course, a well-known occurrence in the history of script. Cf. “Jewish Languages” (above, n. 61).

000330

330

Birnbaum 01.04.11 18:17

[ 51 ]

that the absence of earlier examples of this form of qoph – itself a rare letter – is accidental, and to pass instead to the next terminus, as it is based on numerous cases, i. e., the end of the 1st cent. B. C. E.: beth, heth (C), teth (B), mem, final nuˆn (B), samekh (C), ayin, sadhe (A), ˙ ˆn. This would mean that at the end of the pre-Christian era Neg was shı still under the influence of Nab. None of the letters of Neg which we regarded as possibly Nab in origin provide us with a terminus ad quem. But perhaps we may utilise for this purpose those forms in Nab which have no counterpart in Neg: For from their absence we might infer that the influence of Nab on Neg ceased before these forms had developed in Nab. Here we have, from the early part of the 1st cent. B. C. E. onwards, the florid aleph / final aleph, the closed final heˆ, the waw with a round or curled top, the 5-shaped yodh, the final yodh looking like an elongated S; from the end of the 1st cent. B. C. E. we have the retort-shaped aleph, the qoph in the shape of a reversed q. To the absence of these forms in Neg might be added the fact that Neg has no counterpart to the hyper-tall, narrow style of Nab which develops in the 1st cent. B. C. E. – no detailed examination is needed to recognise the striking difference between the styles of MahAg64 and KBeb.65 The first group of forms (those which suggest that the influence of Nab ceased before the early 1st cent. B. C. E.) is in contradiction to our terminus a quo: the latter is based on forms which could only have originated much later in the century. The second group of forms (those which suggest that the influence of Nab must have been over by the late 1st cent. B. C. E.) results in a terminus ad quem identical with the terminus a quo. This must surely mean that the influence of Nab was restricted to a short period. It was, therefore, only partial. Our date for the Hebrew element at the time when Neg was born was slightly earlier (about the middle of the 1st cent. B. C. E.). As our dates are only approximate, that interval of one to two generations is, in the circumstances, more apparent than real, both results referring to what is one and the same period in the second half of the 1st cent. B. C. E.

64 65

RB 61 (1954), plates I – III (above, n. 25). RB 61 (1954), pl. IV (above, n. 9).

000331

Birnbaum 01.04.11 18:17

The Negeb Script

331

Summary To sum up: From some time in the second half of the 1st cent. B. C. E. to the Bar Kokhba War (132–135) there existed a Jewish script which contained Hebrew and Nabatean elements. The Murabba aˆt finds show that there was close contact between Jews and Nabateans in, presumably, the north-eastern corner of the Negeb. Palaeographically, the documents in this alphabet appear to indicate that the new script came into being in the second half of the 1st cent. B. C. E., through the influence of the Nabatean on Hebrew script. The period of amalgamation seems to have been of short duration. As to the further fate of Neg, it seems reasonable to assume that the upheaval caused by the Bar Kokhba War resulted in the dispersal of those who used this writing, and that after the loss of its territorial base the Negeb script would have had no possibility of remaining alive.

332

[ 51 ]

Abbreviations AR:

Ch. Clermont-Ganneau, Archaeological Researches in Palestine. London 1899, vol. 1. CIS: Corpus Inscriptionum Semiticarum. Vol. 2, Paris 1889 ff. ESE: M. Lidzbarski, Ephemeris für semitische Epigraphik. 3 vols., Gießen 1902–1915. G: Graffito. Heb: The Hebrew script – or any individual letter, form or feature of a letter – from the 2nd cent. B. C. E. to the 2nd cent. C. E. HS: A reproduction of this photo is given in the writer’s The Hebrew Scripts. London 1954–1957 / Leiden 1971. JaSa: A. Jaussen and R. Savignac, Mission arche´ologique en Arabie. Vol. 1, Paris 1909. JPOSS: E. L. Sukenik, (1) “A Jewish Hypogeum near Jerusalem”, (2) “Two Jewish Hypogea”, in: Journal of the Palestine Oriental Society 8 (1928), pp. 113– 124 and 12 (1932), pp. 22–31. KBeb: Kephar Bebhayu Deed; see above, n. 9. Khal: Khalasa Inscription; see above, n. 16. Litt: E. Littmann, Nabataean Inscriptions from the Southern Hauraˆn. Leyden ˙ 1914 (Publications of the Princeton University Archaeological Expeditions to Syria, Div. IV , Sect. A). MahAg: Mahoz Aghalten Deed; see above, n. 25. ˙ Sukenik, “M rwt-qbrym yhwdywt bsbybwt nhl-qdrwn, I”, in: QeMQYQ: E. L. ˙ dem 2 (1945), pp. 23–31. MQYS: E. L. Sukenik, “M rt-qbrym yhwdyt bmwrdwt hr hswpym”, in: Qobes ˙ ˙ (Journal of the Jewish Palestine Exploration Society) 2˙ (1934), pp. 62–73. Nab: The Nabatean script – or any individual letter, form or feature of a letter – from the 2nd cent. B. C. E. to the 2nd cent. C. E. Neg: The Negeb script, or any individual letter, form or feature of a letter. OG: Ossuary graffito. OGi: Ossuary graffiti. OLZG: H. Grimme, “Inschriften auf Ossuarien aus Jerusalem”, in: Orientalische Literaturzeitung 15 (1912), pp. 529–534. PEFA: Palestine Exploration Fund Annual. PEQ: Palestine Exploration Quarterly. RB: Revue Biblique. RBL: M. J. Lagrange, “Epigraphie se´mitique”, in: RB 2 (1893), pp. 220–222. RBS: R. Savignac, “Nouveaux ossuaires juifs avec graffites”, in: RB 34 (1925), pp. 253–266. RBV: H. Vincent, “Ossuaires juifs”, in RB 16 (1907), pp. 409– 414.

000333

52.

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Conveyance*

I. The Letters1 This chapter is devoted solely to an examination of the writing, so many letters of which are mutilated and therefore difficult to establish. Our document {see illustration p. 345; see also HS 121(5)} is written in the Negeb script.2 In order to avoid being influenced by the contents and so arriving at erroneous readings of doubtful signs, we have confined our attention rigidly to the letters. This spade work has been shown in detail. Line 1 The third letter is a samekh – the right-hand half of the top stroke is effaced, only its right-hand edge being visible. The next letter, a daleth or resh, is followed by what might be either a shadow or the trace of a sign. If the latter, then it might have formed part of a gimel, heˆ, waw, zayin, heth, yodh or lamedh. After that there is ˙ a hole, followed by a black configuration which appears to consist of two parts: on top the lower end of a downstroke, and below this the lefthand end of an oblique stroke, the greater part of which would have occupied what is now a gap on the right. This stroke could thus have formed part of only a heˆ, mem or ayin. Between this letter and the preceding resh there would not have been room for an additional letter – as suggested above (gimel, heˆ, waw, zayin, heth, yodh or lamedh) – ˙ but the two traces might have belonged to a single letter. Our two lists of possibles both contain heˆ, and that must be the required letter. The trace above the left-hand bottom end of this heˆ might have been part of a daleth, zayin, yodh, lamedh or resh. Of these it would most easily fit into the shape of the two specimens of lamedh contained in the same line. Next comes the bottom end of a possible aleph or beth. This is followed by the bottom trace of what might have been a daleth, waw, * First published in: Palestine Exploration Quarterly 89, July – December 1957, pp. 108–132. 1 For abbreviations see the list at the end of this article. 2 See NS and my earlier papers in PEQ 84 (1952), pp. 118–120, 85 (1953), pp. 23– 41 {= Chapters 49 and 50 in the present volume; see also Chapter 53}.

000334

334

Birnbaum 01.04.11 18:17

[ 52 ]

yodh, lamedh or resh. The next trace (or traces, if what we see above it is ink) might belong to a daleth, heˆ or resh. If it was a heˆ then there is not enough room for the preceding letter to be a daleth or resh – it would have been a waw, yodh or lamedh. If, on the other hand, the second letter was a daleth or resh, then its distance from the preceding letter would have been too great for the first letter to be a waw, yodh or lamedh – it would have been a daleth or resh. The subsequent five letters are clear. At the right-hand bottom edge of the hole that follows we have the end of a stroke that might have formed part of a teth or taw. Following this is a slightly mutilated heth. ˙ After the next four letters there is a partly effaced sign that could have been a daleth, waw, zayin, yodh or resh. The sixth letter after this is a probable daleth or resh, followed by what is evidently the base of a beth. Next we get a mutilated beth. The last two letters of the line differ in length and seem to stand for the sequence yodh waw. Line 2 The first ten letters present no difficulties – apart from the daleth / resh ambiguity – in spite of the bad state of the papyrus after the fourth letter. The eleventh sign is unlike any letter of the Hebrew or the Nabatean alphabet. It would be possible to derive it from final sadhe or qoph. We meet the same sign again at the beginning of the next line, and this position precludes the reading as final sadhe, of course. Our sign might, on the other hand, be made up of two letters, the first running into the second. For the second, final nuˆn immediately suggests itself. To identify the first we have to look among the letters having a base or bottom stroke. The most likely seems to be a nuˆn. Other possibles – allowing for greater distortion – might be: beth,3 kaph, mem, samekh or pe. The third, fourth and fifth strokes from the end of the line, belong together, although the meeting point of the third and fourth strokes is effaced. The outer strokes (third and fifth) curve in opposite directions. At the point where they are nearest to each other they are linked by a short bridge (fourth stroke). This sign is the letter teth. The second stroke from the end, which touches the preceding teth and the succeeding aleph, seems to be a slightly mutilated waw or yodh. 3

Cf. NS p. 340–341 {= Chapter 51, p. 304 in the present volume}: lines 3, 5, 6, 8, 15.

000335

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Conveyance

335

Line 3 As to the first sign, see the discussion of the eleventh sign in line 2. The next five letters are followed by a trace that might belong to a gimel, heˆ, waw or zayin. Then comes – at the line bottom – a trace that is obviously the end of a downstroke with a leftward slant. It might have formed part of a gimel or heˆ. There being no room here for two letters, the two traces must belong to one and the same letter. Our two series of possibles have gimel and heˆ in common. The damaged letter would therefore be one of the two. In the middle of the line – exactly below the second beth of line 2 – there is a mutilated letter of uncertain identity. It looks as if the dot just under that beth were the top of a lamedh, and the surface of the papyrus between this dot and the main stroke actually is damaged. But that main stroke has not the typical slant of a lamedh and it ends by turning rightward instead of leftward. Thus the dot is presumably after all a shadow, like that below the daleth / resh – five letters to the left. Our letter might then be a nuˆn of the same type as the fourth sign on the left, the base practically disappearing into the subsequent heˆ. Further to the left there is another nuˆn whose base also touches a subsequent letter. Line 4 Above the sixth letter (heth) there appears to be a spot of ink that could hardly have formed part˙ of a letter. In the tear between the heth and the slightly mutilated sadhe there is room for a narrow letter: ˙waw, zayin, yodh or lamedh; there may, however, have been a blank space. Line 5 The seventh sign – the letter after the aleph – is a mutilated waw yodh. This sign is followed by a (not quite complete) waw or yodh. The third sign after this has an unnaturally straight top edge suggesting damage by flaking off. The blackness above it appears to be shadows although the stroke seems to have run on upwards into it. If so, the letter would be an aleph. The sign after the gimel in the middle of the line is straight. If it were a zayin, we would expect the characteristic rightward top; moreover, typical zayin slants in the opposite direction, or is at least vertical. Thus a waw or yodh is, on the whole, more likely. The last letter of the line is puzzling. The left-hand side looks like a final pe with its left-hand stroke curving back into the main stroke. The

000336

336

Birnbaum 01.04.11 18:17

[ 52 ]

right-hand part of the whole configuration would then be an independent letter, running into the final pe. It does not, however, look like any of the letters of this MS . The nearest resemblance is with the right half of a mem, samekh or pe. But where then would the left half be? Was it run into the pe and swallowed up by it? The letter could not, in that case, have been a mem. If the first letter were a pe it would be possible because only the bottom tip of its left-hand stroke would then he covered by the main stroke of the subsequent pe. But if the first letter were a samekh, its left-hand stroke would be too big to disappear into the final pe. In that case sp would be a monogram: the left-hand stroke of the samekh serving at the same time as the upper part of the long downstroke of the pe.4 It might be tempting to explain this as a cursive development, thus: the bottom stroke of the right-hand letter was produced to the left and then bent upwards to form the left-hand stroke of the final pe; on reaching the line ceiling the pen turned downwards and the resulting stroke served for both letters. Upon careful inspection this explanation is, however, not acceptable, for this reason: there is not one continuous bottom stroke but one on the right and one on the left of the main stroke: the bottom stroke of the right-hand letter and the bottom part of the left-hand stroke of the final pe. We must therefore assume that the left-hand stroke of the final pe came down from the top in the normal way. This monogram was accordingly made with four pen lifts. ( Was the writer perhaps rendering an existing one-lift type with four pen-lifts?) Line 6 The gap after the eighth letter (heˆ) is followed by an unusual curve. It does not look like part of any of the letters in this MS . Further inspection shows that the right-hand tip of this curve differs clearly in tone from the downstroke and might not belong to it at all. Perhaps this is corroborated by the appearance of the bottom end of the downstroke. The abruptness with which it stops might indicate damage through flaking off. The letter seems to have been a waw or yodh. The tiny right-hand tip mentioned might have belonged to a lamedh or mem. Below it, somewhat to the right, there is another spot of ink: it might have formed the extremity of a gimel, heˆ, heth, mem, nuˆn or ayin. ˙ There is no room here for two letters. Can both traces be fitted into any 4

See NS p. 340 {= Chapter 51, p. 304 in the present volume}, especially the samekh of line 6.

000337

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Conveyance

337

single letter, common to our two series of possibles? Yes, into the mem: the lower trace is the bottom end of the oval, the upper one is the end of the tail. The stroke after the taw might have been the right-hand half of a daleth, mem, samekh, pe or resh. On the photo it is impossible to make out how much of the black configuration following it is ink and how much shadow. The biggish triangle at the end seems to me to be a shadow and the two dots before it, ink. If they are, they might well belong to the tail of a mem. The next sign, though slightly effaced, is a nuˆn. The letter between ayin and final nuˆn, in the middle of the line, resembles the daleth or resh in the KBK (and resh elsewhere) but it differs markedly from the specimens of these letters contained in our MS . Perhaps it is closer to a zayin, except that the top is vertical, not oblique as in a zayin here. Can it be a slightly distorted waw or yodh? Before the shıˆn towards the end of the line, between the lamedh and daleth / resh, the writer has squeezed in on top a small mem which he had accidentally omitted (M). Line 7 The seventh letter is damaged on the right-hand side. There is room for only a plain-stroke letter here. It could be either a waw of the straight variety discussed in line 5 (if that was a waw) or a zayin. The trace before the large hole is hardly identifiable. It may have belonged to a beth like that of zb in the second half of line 5. Heˆ, teth, mem, samekh, pe and taw are other possibles – none of them attractive but not out of the question either. In the ensuing gap there would have been room for two or three letters. The little line which follows the hole at, and above, the line ceiling appears to be a shadow. The trace after the hole might represent the top of a daleth, heˆ or resh. The smallness of the left-hand stroke and the slant of the right-hand one speak against daleth and resh – the right stroke here slopes rightward while in daleth and resh it slants initially to the left. Furthermore, in our sign the left-hand stroke seems finally to turn downwards. It stops abruptly with an unnaturally straight bottom edge, suggesting that the stroke might not have ended there, i. e., that the rest has flaked off – the big tear at this point would have affected the writing on the edges by damaging the surface of the papyrus there. For these reasons it is more likely that the trace formed part of a heˆ.

000338

338

Birnbaum 01.04.11 18:17

[ 52 ]

Line 8 The traces of the first letter seem to fit in with the shape of an ayin. Next we have the top (mutilated) and the bottom of a pe, like that of line 4, letter 8. The third sign might be daleth, waw, yodh or resh. The fifth sign might be the lower part of a daleth, waw, yodh, lamedh or resh. Where the big hole begins there is a tiny trace which might have belonged to an aleph, daleth, zayin, heth, kaph, final kaph, nuˆn, final ˙ nuˆn, ayin, qoph or resh. The hole has obliterated seven to eight letters. Then there is a dot, just under line 7; at this point it touches the end of the gimel. This can only be the top of a lamedh. The next trace might be the top of a beth, gimel, daleth, heˆ, waw, zayin, yodh, nuˆn, final nuˆn, samekh, ayin, resh, shıˆn or taw. What follows is a daleth or resh. Line 9 The first sign must be one of the narrow letters: waw, zayin, yodh or lamedh. If it were a lamedh, the straight black line just above could not be the top stroke of that latter because its direction and position preclude the possibility of its being a continuation of the bottom stroke. It is a shadow. The second letter is clearly a heˆ. The third might be a waw, zayin, nuˆn or final nuˆn. As regards the fourth letter, the papyrus fragments do not seem to have been placed in quite the correct position here: the outline on the left has a break in the middle which seems to indicate that the top and bottom parts do not belong to each other. The bottom part ought perhaps to be moved slightly leftwards to form the bottom of the subsequent (fifth) letter. Our sign would then be an aleph, mutilated at the lower end. The next one is clearly a damaged shıˆn, followed by a stroke belonging to a possible daleth, waw, zayin, yodh, lamedh, nuˆn or resh; other letters which might have ranked as possibles are excluded by the narrowness of the space. The next letters are ayin and waw. The second sign under the hole does not look like any of the letters in this MS .: the strokes of none of the other letters cross. This might indicate that the two strokes do not form a single letter. The second stroke would be part of a final letter – kaph, nuˆn or pe. In the previous line, in discussing the trace above our strokes, one of the possibilities mentioned for that trace was a final kaph. Our stroke here in line 9 seems to confirm this. The letter would then be an exceedingly long specimen.

000339

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Conveyance

339

Of the two strokes that cross, the right-hand one cannot form part of a heˆ because it turns vertical at the top. It is too long and low for the tail of a mem. Perhaps it is an ayin; the top bit would then be the end of the left stroke of that ayin. Under the left-hand end of the hole, between the mem and beth, the lower extremity of a narrow letter is visible. It might have belonged to a waw, zayin or yodh; aleph or lamedh are not precluded. The sign after the beth seems to be a waw or yodh; a lamedh, too, is possible. Line 10 The eighth letter appears to be a mem rather than a samekh. Its left-hand side is mutilated, and we cannot tell whether it was followed by another letter. Next we have parts of a beth: the downstroke running to the lefthand end of the base; a trace of the top is faintly visible. After the hole, in the middle of the line, fragments of a letter are preserved: a downstroke with a leftward slant and a smaller stroke – apparently issuing from it – which runs down vertically. These traces fit in perfectly with a mem: the oblique stroke is the end of the tail, the lower trace is a bit of the left-hand side of the letter. After the ayin that follows we get a very long stroke running right up to the line above and touching a letter (aleph) there. In spite of its unusual length it looks like a single stroke and not as if it were partly a shadow caused by a break in the papyrus. It would be a lamedh, of course. The continuation of this line downwards – from about the upper third of the ayin to the line bottom of the next line, however, looks on the photo like a shadow cast onto the background by a break in the papyrus. The strokes after the next hole do not constitute a sadhe: there is a short top stroke on the left which precludes that reading. What we have here is a pe preceded by the end of a missing letter. This oblique little stroke must be the tail of a mem. Line 11 Enough is preserved of what, according to its position, is presumably the second letter to identify it as a waw or yodh. The sixth sign before the hole is a heˆ, carelessly written – the lefthand stroke does not issue from the right-hand top corner, as in the sequence drth in line 5, and it is not oblique. The long, oblique stroke that runs down from the left-hand top corner is puzzling. Is it just a slip? Or is only the upper half a slip, while what

000340

340

Birnbaum 01.04.11 18:17

[ 52 ]

appears to be the lower half is really the lower half of the subsequent stroke, i. e., have we a final nuˆn here? There are similar forms on this papyrus but none quite so curved as this specimen would be. It therefore seems, on the whole, more likely that no part of the long line belongs to a letter. At the beginning of the big hole there is a trace of a letter that might be a daleth, heˆ, waw, heth, yodh, kaph, lamedh, mem, samekh, pe, resh ˙ or taw. If it were a lamedh, the stroke immediately above (in line 10) might belong to it. After the hole there is the end of a letter base. As it is concave the sign is more likely to have been a beth than a kaph, nuˆn or pe. In the gap there would be room for one to three letters. Line 12 At the right-hand edge of the big hole there is a black configuration. The upper stroke could belong to an aleph, ayin or shıˆn. If it were an ayin, then, at that height, its left-hand stroke should already be visible; if the letter were a shıˆn then the stroke in question would, according to its height, be the inner stroke, and the parallel stroke on the right should be visible. Hence the upper stroke is likely to have formed part of an aleph. The lower part of the black configuration, at first glance, appears to be a shadow. If it were, its rectangular outline on the right – which would then coincide with the edge of the papyrus – would cast a rectangular shadow to the left. Instead of this we have a straight line on the left. The lower part is therefore a stroke (apart from the extreme lefthand bottom tip). It could form part of an aleph or, less likely, of a gimel or heˆ. As the upper part of the configuration belongs to an aleph there can be no doubt that the lower part is its continuation. After this, at the bottom edge of the hole, we have a very small oblique stroke. It might be a base – one that turned out somewhat too thin and slanting – or else it might be the end of a curving downstroke. Possible letters would accordingly be: beth, gimel, heˆ, teth, nuˆn, ayin, pe, sadhe, shıˆn, taw. The hole could accommodate three to five letters. The first letter after the hole is difficult. A shorter downstroke on the right and a longer one on the left point to a taw – but the strokes are straight, in contrast with the pronounced curves of a typical taw; nor is there a loop. Two parallel straightish strokes would be characteristic of heth, except that in heth the position of the shorter and the longer stroke ˙ the reverse of what ˙ we have here. is

000341

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Conveyance

341

Line 13 The third stroke might belong to a gimel, waw or yodh. The next trace is clearly the upper part of a lamedh. Left of it there is a tiny grey line slanting down to the left. Is it part of a letter? If so, it might have been the extremity of an ayin or shıˆn, either of which letters would then have filled the gap. Next we have the upper part of a lamedh again. The stance of the third letter after that would favour identification with a zayin rather than with a waw or yodh. The trace on top of the next hole presents difficulties. It does not really tally with any of the letter tops, and all we can do is to say that the following letters are not quite impossible: daleth, heˆ, waw, zayin, yodh, kaph, nuˆn, final nuˆn, resh. Diagonally opposite it on the left, on the lower edge of the hole, there is a trace which might conceivably form part of the same letter, giving us a possible heˆ or kaph. On the other hand, it might belong to the next letter – although this would then be rather wide – i. e., it would be the end of the right-hand stroke of a taw; the left-hand half of a taw is clearly preserved. After the taw comes the top of a letter which is hardly identifiable – perhaps it was a daleth, waw, zayin, yodh, nuˆn, sadhe or resh. The traces after the big hole might have belonged to a heˆ, mem, ayin or taw. Of these the first three are less likely than the fourth in view of the heavy oblique stroke. It fits in better with the loop of a taw. Towards the end of the line, the sequence hl seems to be followed by either waw or yodh. After this there appears to be another waw or yodh; it is also possible that both traces together formed a heth. On the other ˙ hand, the second trace might have constituted the right-hand side of a mem; the oblique stroke at the bottom would then be the tail of the mem. Alternatively, this oblique stroke might have formed part of the bottom of a beth. There is, however, no trace of its upper part. Nor is there any sign of the left-hand stroke of a mem. Line 14 At the bottom edge of the initial hole there seems to be the remnant of the base of what would have been the third letter of the line. Letters with a base are: beth, kaph, pe and sadhe. Of the fourth letter in the line, three traces are preserved: two bottom ends of downstrokes and a straight horizontal top stroke. All these traces would fit into a heˆ. The black formation at the next hole presents great difficulties. The top part looks like ink and the lower part like a shadow. What we have here

000342

342

Birnbaum 01.04.11 18:17

[ 52 ]

might then be the top of a beth, mem or pe. After sw / sy comes a teth like the one in line 2, but the lower part of the right-hand stroke has flaked off. What follows now is a mutilated taw: the loop has gone (the black line at the bottom is a shadow). The sign after it is very badly damaged. The right-hand downstroke and base suggest a mem like the one in line 13 (middle). The next letter (beth) is followed by a sign which can only be a mutilated gimel: the upper half of the main stroke has come off. The rest of the line is written by another hand, with a different pen. After the big hole, there is the left-hand part of a taw, and perhaps the bottom tip of the right-hand stroke. Line 15 The first trace might be the top extremity of an aleph, daleth, zayin, heth, kaph, nuˆn, ayin, resh or shıˆn. The second trace might be the ˙ bottom end of a daleth, waw, zayin, heth (one of the downstrokes), ˙ or resh. A third possibility is yodh, lamedh, samekh (left-hand stroke) that the two traces formed part of one letter. If so daleth, zayin, heth or resh, which occur in both series, are at our disposal. In view ˙of the position of the two traces, only heth would satisfy the requirements. ˙ trace were to be connected with the If, on the other hand, the second third and fourth ones, the sign could only be a heˆ. But if only the third and fourth traces belong together, they would constitute a daleth or resh. The letter after that is clearly a daleth or resh: above the hole the tips of the main stroke and of the top stroke are showing and under it the bottom end of the main stroke can be seen. If the preceding letter is also a daleth or resh, then the writer apparently took the trouble to differentiate between the two letters because they followed each other. The bottom stroke of the next – mutilated – letter reveals it to be a beth. The next stroke might belong to a daleth, waw, zayin, heth, yodh, lamedh or resh. The next trace might be the bottom end of˙ a daleth, waw, zayin, heth, yodh, lamedh or resh. Next, after ˙the little hole, are two downstrokes which might form part of a heˆ. Whether the oblique little stroke above could be the top stroke of such a heˆ is very doubtful. The trace after the next hole might belong to the top of a daleth, kaph, ayin or resh. The oblique stroke before the hole in the middle of the line might form part of a gimel or heˆ. The rest of the line is written by a different – a third – hand. After the second letter (lamedh) there is a tiny dot which might be an ink trace. It is, however, too high up to form part of a letter. The sign after the next

000343

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Conveyance

343

hole might be a daleth, heˆ (the left part), waw, yodh, ayin (the left half), resh or taw. Daleth, waw, yodh and resh are less likely because the downstroke is too oblique. The top stroke would be uncharacteristic for a heˆ and the lower stroke somewhat too low down. Our sign differs clearly from the undoubted heˆ at the end of the line. Ayin, therefore seems to fit in best here. The right part of the letter would just fill the space covered by the hole. The next sign might be a waw, zayin or yodh. It is followed by a daleth, waw, yodh or resh. At the left-hand corners of the hole that follows, there are traces of a letter. They are compatible with daleth, waw, yodh, kaph, samekh, ayin or resh. The letter after the first mem is a shıˆn, on the general pattern of NS 17a, 18, 22 {see p. 304 in this volume}. The next sign is difficult. The manner of construction suggests a teth but the main strokes are straight instead of curved; the right-hand stroke starts at the line ceiling instead of half-way down. Identification as a taw would be open to even stronger objections: the left-hand stroke stops less than half-way down, and there is no loop; the right-hand strokes are completely straight instead of curved; and some of the minor features are different, too. Nor is it feasible to split up the strokes of this sign: only the left-hand one could constitute a separate letter. On the other hand, neither of the strokes could form a letter with the strokes adjoining them on the right or left, respectively. The next letter is a final aleph, the lower half of its main stroke being almost effaced. The formation between the two mems that follow is misleading: the right-hand part is the end of the final nuˆn from the line above. The lefthand part seems to be non-final aleph, of the type NS 17a. But if it were, then the first letter of this second half of the line could not also be an aleph – and its identity seems beyond doubt. Could our sign be a waw or yodh? Or is its left-hand part by any chance the curled end of the final nuˆn? That seems very improbable. Or is there perhaps no letter at all between the two mems but just a smudge and / or shadow? The need for avoiding the end of the final nuˆn would explain the unusual distance between the two. Line 16 Here we have another hand again – the fourth. The first letter is a shıˆn of the type NS 17, 18, 22 {see p. 304 in this volume}. The eighth sign might be a waw, yodh, nuˆn (with base effaced) or pe (the same). Then, after the heˆ, there might be a waw, zayin or yodh.

000344

344

Birnbaum 01.04.11 18:17

[ 52 ]

The general outline of the next letter would only be compatible with mem or samekh. It differs clearly from the undoubted mem which is the second letter in this line. Presumably it is a samekh. This is confirmed by another open form in our material.5 Line 17 This line is written by a fifth hand. The first stroke is a waw, zayin or yodh, the second a waw or yodh. This is followed by the right-hand stroke and the middle one of a shıˆn. Of the next letter a small trace is preserved: it might belong to a daleth, ayin or resh. The black line after that seems to be a shadow. The next two signs are in Hebrew script. We came across such a mixture before – in the signatures of the BMD.6 After that nothing can be made out for the space of two or three letters. Then comes an – angularized – oval sign. As there is no tail the letter is, presumably, not a mem but a samekh. It is followed by two dots which reveal themselves to be a single stroke – the middle having flaked off – because there is clearly the continuation under the hole in the shape of an oblique stroke. Thus the letter is either a final nuˆn or a final pe. If the black lines after the heˆ are not a shadow they would make a daleth or resh. Line 18 This line is written by a sixth hand. The first letter (under the beth of line 17) is a blurred heˆ. It is followed by a waw, zayin or yodh, and then by a daleth or resh. The letter after the beth is Hebrew again, and so is the heˆ between the two waws / yodhs which follow, as well as the letter before the last one of the line. But the heˆ between the daleth and the shıˆn is not.

5 6

NS 20 {p. 304 in this volume}. BMDB, pp. 28–31 {pp. 266–269 in this volume}.

000345

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Conveyance

The Kephar Bebhayu Conveyance

345

000346

Birnbaum 01.04.11 18:17

346

[ 52 ]

Temporary Transliteration The foregoing examination of the individual signs and traces yields the following transliteration of our document. The number of different possibilities for many of the traces is so great that it would not have been feasible to arrange them vertically here, as is usually done where there is a choice between two or, at most, three letters only. The letters in question are therefore arranged horizontally and separated by oblique strokes. The result of so large a number of uncertainties is that coherent reading at this stage is out of the question, the more so as we have followed the arrangement of the original, where there is no separation of words. It is hardly necessary to point out that, in spite of this, we cannot say that the writer was intending to write scriptio continua. The fact that final forms of the letters kaph, nuˆn and pe are employed shows that he was fully conscious that words are units, and that as such they are separately written. Although he wrote carelessly he must have been aware that in more careful MSS . there is a space between each word and the next – he must at some time have seen letters like the (practically contemporaneous) BKL, BMD and literary hands like the Murabb aˆt Exodus Fragment.7 This transliteration will form the basis for the discussion in Section II .

7

HS 89.

000347

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Conveyance

347

000348

Birnbaum 01.04.11 18:17

348

[ 52 ]

II. The Words Line 1 There is no difficulty in isolating the first word which is clearly b srh. Our MS . is obviously not a column from a book but just a single sheet. As such it would be a letter, a legal document, or the like. A glance at the end tells us that there are a number of what seem to be signatures appended. This suggests a legal document. If that is correct, srh is the numeral and designates the date, i. e., the day of the month. The next word should reveal the month – either the name of the month, as in the EP, or the word month (yrh, later hwds˙ ) followed by its ˙ lamedh ˙ (see Section I) for name, as in the medieval deeds. The reading the initial letter is suitable because the preposition l is required. The next trace of a letter – part of an aleph or beth (see Section I) – must belong to an aleph, since no name of a month begins with a beth (nor do the words yrh or hwds˙ ). Hence we have a choice between b, ˙ in Section I eliminate the first and last of dr, yr and lwl. Our ˙readings these, leaving us with a choice between dr and yr. The traces appear rather to favour dr. In the rest of the line there is no difficulty in isolating the words. We would expect the year to be followed by an indication as to the era. The next two words (lhrt ys˙r l) are obviously identical with the legend on ˙ the second year of the Second Revolt.8 That gives some of the coins from us the era of our document – it was written during the Bar Kokhba War (M). We need hardly add that this date is confirmed by the evidence of the BKL and BMD which were found together with our MS . The era should be followed by the name of the place where the document was penned. The next sequence of letters fits in with this requirement, kpr being a geographical designation. The word after that should be the name itself. This seems to be bbyw, recalling the sequence b b w/y w/y in the KBK, which was found together with our MS . Line 2 After the place-name we would expect the name of the party making out the document. It can easily be isolated. The daleth / resh ambiguity, unfortunately, prevents us from establishing it with certainty.9 8 9

s˙ [nt] b lhr(w)t ys˙r l. ˙ Hdr (M). We adopt ˙

000349

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Conveyance

349

The following readings (see Section I) of the puzzling sign after yhwdh would make no sense: q, kn, nn, sn, pn. That leaves us with the choice between bn (i. e., beˆn) and mn (i. e., min). The next group of letters – kprbbyw – is easy to isolate since it is identical with the end of line 1, confirming that reading as well as the one in the KBK (line 4). The ending w shows the name to have an Arabic form. That, in its archaeological and historical context, would mean Nabatean. The geographical designation disposes of the reading beˆn which could hardly connect a personal with a place name. Mn is the obvious choice and the derivation of the sign then becomes evident: the tick which forms the right-hand part is just a very open stage of a mem – which letter is anyhow very frequently open; the tail has been merged with the final nuˆn, the top of the letter being dragged down by the mem tail so that this can run into the nuˆn. Next we would expect the name of the person on whose behalf the document has been made out. The word mr tells us that he is being addressed by Hdr. ˙ of the line is baffling. As it looks like a nominal form The final word and follows the word Eleazar, s˙ty might qualify that name. But what was its meaning? We can hardly˙ assume that the man was addressed as ‘the fool’. Line 3 The first sign, though slightly damaged, is clearly identical with what we found to be mn = min in line 2. The same meaning fits in perfectly well here, too. There we had Hdr s address, mn kpr bbyw, and here Eleazar’s, mn tmn. It should be˙ noted that in tmn the sequence mn is written out in full. In other words, the ligature mn is used only for the preposition min – it is, so to speak, a kind of ideogram. After the date and the parties, we get the subject matter of the document. It consists of a declaration by Hdr and is in direct speech. It begins with two adverbial phrases. The˙ first – mn r wty – tells us that the speaker is acting of his own free will. This clause does not occur in the EP but it is frequently to be met with in later Jewish deeds, and is reported (M) to be present in a Hebrew deed found at Murabba aˆt. The object of zbnt is, of course, bth; lk must accordingly be the dative. Hence zbnt cannot mean ‘to buy’ but must be the pa el: ‘to sell’.

000350

350

Birnbaum 01.04.11 18:17

[ 52 ]

Line 4 It seems unlikely that spn is directly governed by ptyh; the two words were ˙ probably linked by a˙preposition. There is room where the hole is, for a lamedh. That the house is ‘open’ to the courtyard presumably means that the house is entered via the courtyard. This idiom is not used in the EP. Eleazar acquires the right to ‘open’ the newly bought house into his old house, i. e., he evidently gains direct access from his old house to his new one. From this it would appear that these two houses were adjacent, otherwise Hdr would have said tpthnh lgh drtk or the like. ˙ ˙ Line 5 Lk lk is an obvious error. Seeing that the statement at the end of the preceding line and at the beginning of this one is a negative one and refers to the 2nd pers. (lk = Eleazar), it is evident that the waw / yodh after m is the suffix of the 1st pers., i. e., that it refers to Hdr. ˙ Line 6 The selection of the correct letters from the possibles (see Section I) presents no difficulties. The text divides easily into words but they are incoherent: ‘zuz which are eight two sela’. The writer obviously forgot ‘eight’ and added it after he had already written ‘which are’. What he intended was zwzyn tmnyh dy hmwn sl yn trtn (AG). Line 7 If we take the trace just before the hole to be the top right-hand corner of a taw, giving the reading s˙ryt (M); the final aleph of this word, followed by a waw, would just fill the gap. Our reading of the next word as hgdy / hgry is compatible with the translation ‘bricks’ (AG, reading gry ) because alternation between the glottal stop and fricative is not rare. Such alternation occurs in another Murabba aˆt document, the BKL, where sl ‘by, with’ is written hsl.10 ˙ ˙ Line 8 The letters suggested by us in Section I for the beginning of the line add up to form prh ‘dust’. The expression prh dqrq is, of course, a 10

See BKA, p. 29 {= pp. 247 f. in the present volume}, where, in addition, examples from the Qumraˆn MSS . are adduced.

000351

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Conveyance

351

parallel to the formulae in the EP and the Babylonian Talmud: mn r w d l 11 and mthwm w d rwm rqy .12 Our document is closer to the Talmudic form in that it adds a qualification to ‘ground’ (although there is a great difference between a matter-of-fact expression like ‘dust of the ground’ and the more imaginative and theoretical wording of the Talmud: ‘abyss of the earth’). But where, in this deed, is the sky part, corresponding to w d l or w d rwm rqy ? Has it been just accidentally omitted? There seems to be rather a lot left out for that to have happened, besides which there would have been omissions in two different parts of the same sentence (a bit of the beginning: mn, as well as the whole second half: w d rwm drgy or w d l ). Or did the writer, by any chance, accidentally reverse the formula in his mind, intending to write mn rwm drqy w d prh dqrq ? If so, he would have omitted only the first half. After the general description of the house, its location is dealt with: the boundaries in all directions are defined. The gap in the line, caused by the big hole on the left, would accommodate seven to eight letters, i. e., two to three words. These no doubt referred to the previous bth, while being at the same time connected with zbnh which follows. The link between the house and the transaction is, of course, the seller or buyer. Accordingly, we would expect the missing words to contain the name of either Hdr or Eleazar. ˙ the name of The trace of a lamedh top is clearly visible giving us Eleazar, i. e., of the buyer. That would mean that Hdr, who is the ˙ speaker in this deed, and who has, until now, employed the pronoun (‘thou’, ‘thy’) in addressing Eleazar, is here using his name. Zbnh, referring to Eleazar, must therefore be in the pe al (‘to buy’). Zbn is, accordingly, either the 3rd pers. pf. sg. or the participle. Everywhere in the body of the deed Hdr addresses Eleazar. This corresponds ˙ grammatically to the 1st / 2nd pers. relationship. Hence zbnh would not be the 3rd pers. but the participle. Accordingly, l zr would not be the subject of the passage. That subject must have been the pers. pron., 2nd pers. sg., nt ( l zr being in apposition to it). nt helps to fill the available space: the word l zr would not fill the gap – the longest in the MS . Even nt l zr would hardly be long enough. ( The form nt is almost certain in view of line 10). As the passage is a relative clause qualifying bth, it would be introduced by the nota relationis. And even this dy, as tests show, might 11 12

EPC no. 5; but the other EP do not have it. B. Babha Bathra 63 b.

000352

352

Birnbaum 01.04.11 18:17

[ 52 ]

still not quite fill the space. Hence the daleth after bth would presumably not be the first letter of dy but of another word. As a matter of fact, yodh was not one of the letters in our series of possible readings (see Section I). In lines 7, 11 and 13, bth is followed by dk. We have therefore probably the same word here – the surviving trace after daleth being compatible with a final kaph. Line 9 After the heading – mdnh – at the end of the preceding line, the new ˙ for this side (east). The text obviously goes line starts with the boundary up to the next heading (spnh), but what it contains is very uncertain. ˙ nine, if there was one in the hole) and that There are eight letters (or would make two to three words. According to our readings in Section I, the first word might be whw, why, whn or yhw . If it is yhw , the next five letters can hardly form more than one word as it is necessarily a noun. However, five letters would only be possible if there was an ending, or if the noun was in the tp wl (or tp yl) form. An ending is excluded because the fourth letter is an ayin, i. e., would belong to the root. The tp wl (or tp yl) form would be feasible. The roots with s˙ as the second and third radicals are: ps˙ , rs˙ and ts˙ . None of these makes sense here. In addition, the use of hw would be very unidiomatic: ‘in the east will be (this or that thing)’ – apart from the fact that hw is not used in describing the other boundaries here; nor does it appear in the EP. To start the line with ‘and’ (whw, why, whn) is also difficult but perhaps not impossible. Whn ‘and if’ would make no sense here. Whw might introduce the noun giving the boundary. The next word would begin with t. The only possible root is tr (since th is a verb), and ‘place’ would be a suitable word here. For the following word we would have a choice between the roots sˇw , sˇr and s´r . Sˇuˆa means ‘to plaster’ and our word might then be sˇy , the pass. ptc. It would qualify the preceding t[r]: ‘plastered’ or perhaps ‘smooth’. Sˇr means ‘to slope, glide’, and there is a Talmudic noun sˇr t ‘slope, sloping ground’. We would then have t[r] sˇr ‘sloping place’ – a possible description here, and perhaps a more likely one than ‘plastered’ or ‘smooth’.13 S´r conveys the idea of ‘joining, adding’, perhaps it 13

Arabic sˇr has among its verbal meanings that of ‘communicating with the road’, and among its nominal ones ‘situate upon a road that is a thoroughfare’, ‘main road’, ‘open track or way’. That would seem suitable but it is rendered

000353

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Conveyance

353

means ‘adjoining’ here. However, t[r] s´r ‘an adjoining place’ makes no sense as a boundary. In all three cases we would perhaps rather expect the stat. determ., trh sˇy h / sˇr h / s´r h. But there is no room for two letters (rh or r ) between the taw and the shıˆn, and there is no heˆ or aleph after the ayin. However, the stat. abs. is not impossible. If all these suggestions should be beside the mark, the next letter, waw / yodh, might perhaps form a unit with the preceding letters, and so give us the ending uˆ or ˆı. Both of these are, however, generally abstract (especially the words in uˆ ). Or else, the waw would go with the succeeding letters, i. e., it would be the word ‘and’. The gap after m rb presumably contained w ‘and’. The western and southern boundaries of the house are formed by bwh / bzh/byh / blh.14 Line 10 The initial clause is a repetition of the one in the middle of line 5. That answers the question (see Section I) as to whether there was a letter between mem and beth. The reading must be my. In line 5 we took it for granted that the general clause wrs˙h l yyty lk my bgw drth dk included the denial of the right of way to the buyer. Here in line 10 the same clause is followed by an explicit injunction to that effect: wl m l [w]l[ ] mpq ly. Perhaps this – the general clause together with the special one – was the usual formula, and, having erroneously omitted the second part, Hdr may now have inserted the complete clause here. On the other hand, ˙it is also possible that Hdr introduced the special clause as an additional safeguard. ˙ of the pronoun at the end of the line must be a slip – it The 2nd pers. seems very unlikely that the 2nd pers. should have preserved the final vowel.15 The 1st pers. is therefore probably the correct reading. This linguistic argument is confirmed by the contents: (1) The reading of nth as the 2nd pers. would mean a reversal of the roˆle between speaker and addressee in the deed. And that seems most unlikely.16

14 15 16

unlikely by the fact that we would expect sr (with samekh) in Aramaic. Our form (s˙r ) would correspond to Arabic saru a ‘to be quick’ and that is clearly useless to us. If our s˙r were taken not as a cognate of this Ar. sˇr but as an Ar. loanword in the local Aramaic, the reading tr would become very doubtful. AG suggest the Latin via. It had already disappeared in the EP, six centuries earlier. In EPK no. 5 there appears to be such a reversal but there we have what are

000354

354

Birnbaum 01.04.11 18:17

[ 52 ]

(2) The pronoun and the name Hdr conclude the whole twofold ˙ clause; this end must therefore correspond to the end of each of the halves, m[y] and ly, i. e., the pronoun must also be in the 1st pers. Even if it were taken to belong to the second half only, the same would apply. Apart from that, it follows immediately on ly, but is at a great distance from lk.17 Does the fact that Eleazar has no right of way in the courtyard perhaps imply that his old house bordered on it – or was, at least, close to it and so situated that the way through the courtyard might be a short cut? This would fit in with our inference, in the discussion of line 4, that his old house was adjacent to his new one. Accordingly, the old house would have been to the east or west of the courtyard. Line 11 At the beginning of the line, the upper part of the traces might be the first stroke of the ligature min – and min is the word to introduce the phrase ymh dnh wl lm which follows. The difficult sequence of letters in the middle of the line can be defined as a word, because we have clear words before and after it: nh and rb, respectively. One uncertainty however remains: does the last of those letters belong to the word in the middle or to rb? If to the latter, that would give us either y rb or w rb. Y rb is unlikely, because it is in the 3rd pers. and this does not fit in with the 1st and 2nd pers. throughout our deed. If rb is taken as a verb, the same holds good for w rb (unless rb is interpreted as an imp. – which we do not get in deeds). Hence rb is most probably a noun. If the preceding letter is w ‘and’, then the word in the middle would also be a noun. But neither hw nor hy can be read as nouns. Thus the waw / yodh must belong to the required word, and we have a choice between hww, hwy, hyw and hyy. The second is obviously correct: ‘I shall be’ – hwy rb ‘I shall guarantee’.

17

really two separate deeds: one is a manumission by a certain Meshullam, the other a paramone by a certain Tpmt. – In a few cases a grammatical change occurs: nh is replaced in the course of the document by the name of the speaker (EPC nos. 5 and 15, EPK nos. 2 and 7). A change from the 1st to the 3rd pers. does not of course reverse the role of speaker and addressee. Cf. to nth also AG.

000355

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Conveyance

355

Line 12 The words hdr/ . . dnh after the tear must obviously have been preceded by ‘the wife˙ of’ (M). The form would be that with nuˆn, as suggested by KBK and two out of the three occurrences on ossuaries.18 The form of the renunciation corresponds to that in the Mishna (R). The expression for ‘law-suit, claim’ (mlyn) corresponds to the second half of dyn wdbrym in the Mishna, while the EP have dyn wdbb. Line 13 Something appears to have gone wrong in this line. If the clause was constructed like that of her husband in line 11, the time should be mentioned at the end of the clause. We may therefore assume that the first l lm is an error – a double error: wrong position in the sentence and the omission of mn ymh dnh – and when the writer reached the word dk, he realised it but was already so confused that he again omitted mn ymd dnh. Since bth occurs so often in this deed that the ]th on the left of the big hole might be a remnant of the same word. For the preceding word I have no suggestion. Could nqnh – which is presumably a pa el: ‘to bring into possession, to sign over’ – be connected with the next line, indicating that the transfer of the property was done by means of this deed? Or might it indicate a special form of transfer, as e. g. the qnyn swdr? Line 14 If the passage with nqnh is a clause concerning the method of transfer, the first half of line 14 – i. e., the end of the body of the document – might be a section of its own, dealing with the form of the document. Our deed is of the ‘plain’ or ‘open’ type: the signatures of the witnesses are not on the outside of the closed sheet but ‘inside’, i. e., at the end of the text.19 The words that follow ps˙yt are therefore simply a ˙ duplication. Shalom’s signature follows immediately on the conclusion of the text. We may nonetheless assume that she was not the first to sign but that her husband had already done so at the beginning of the next line. Perhaps she did not like to leave a blank space at the end of the text. Or 18 19

ntt l zr, s˙pyr ntt s˙m wn (but lys˙b tt trpwn). Cf. R, M and KBK 2.

000356

356

Birnbaum 01.04.11 18:17

[ 52 ]

perhaps she just put the signature where you would not quite expect her to put it. There must have been a heˆ at the end of the line, as the lady would have used the fem. form of the verb. Line 15 The restitution of nuˆn between lamedh and pe is obviously needed (M). The expression l nps˙h is known to us from the BMD.20 At first glance Hdr bar Judah’s writing looks like that of the deed. But we cannot be sure ˙as the heˆ differs considerably from the many specimens in the text. On the other hand, the writing of the document is pretty irregular – clearly not the hand of a professional scribe. So perhaps he is after all the writer of the deed. He would have copied the text from another deed. After the name l zr a letter is missing. Among the possibilities for the next one there was resh (see Section I) so that what we have here might be the word br. The next word starts with ms˙ but the name Moses is out of the question: the letter after the shıˆn cannot be a heˆ.21 It stands to reason that the Eleazar of this signature is the Eleazar b. Eleazar of line 2, the buyer of the house. Accordingly, if the second word is br, the next one, starting with a mem, cannot be a name. We would then have to assume that Eleazar accidentally omitted his father’s name. The letters following his name – or the word br – presumably contain some kind of designation. Looking back to line 2 we see the words Eleazar b. Eleazar followed by the sequence s˙ty . Here, in line 15, we ˙ / yodh between the teth have almost the same, only that there is no waw and the aleph, and that the preceding mem might belong to this word. Ms˙t would then be a nominal derivation from the root s˙t , and thus a ˙ parallel to the nominal form s˙ty of line 2. The end of our˙ word here is certain, because aleph has the˙ final form.22

20 21

22

BMDB, pp. 31 f. {pp. 269 f. in this volume}. Apart from that, the name Moses was not current among the Jews of this period. The earliest occurrences I know of are in the 9th century: there is a gaon of that name at Sura, 825–836, and in the second half of the century there is Moses b. Aaron, the Masorete. (Could this be a case of Moslem influence? Musa was one of the Biblical names adopted by the Moslems from the very beginning. The earliest known bearer of that name seems to be Musa ibn Nusair, the conqueror of Spain [born 640].) Cf. NS.

The Kephar Bebhayu Conveyance

357

The next letters appear to constitute the word memar, i. e., a nominal form of the root mr. Was this word in some way a parallel to Hdr s and ˙ Shalom’s l nps˙h? The heˆ might be the suffix of the 3rd pers. masc. The meaning might be something like ‘matter; affair’. However, we would expect l before it. Line 16 As in the BMD, and on the ossuaries, Jehoseph, the long form of the name is used. In this line the only Hebrew word of the MS . occurs – as a technical term – d. However, the next two witnesses use the Aramaic term. Line 17 The first name consists of four or five letters, with a shıˆn in the middle. We have not a big choice: Jeshua and Manasseh occur in the BKL, BMD and KBK; the name Elisha was in use at the time: lys˙ bn bwyh and lys˙ b l knpym. The traces before the shıˆn can hardly be squared with the letters of the name Manasseh or Elisha, even assuming a spelling ls˙ . There is a better chance for Jeshua, spelt yys˙ ,23 or perhaps Joshua, spelt yws˙ . (In either case there was almost certainly no room for a waw before the ayin.) The two surviving letters of the father’s name leave no doubt that this name was Jehoseph – incidentally, one of the most frequent names in the four Murabba aˆt documents so far published.24 Line 18 Two of the six persons named in this document have the same name as their father. There is one such case among the seven in the BMD. Three out of thirteen is certainly an unusually high percentage. If the custom of naming only a posthumous son after his father was already established at that time, this high percentage might mean that the three fathers in question had been killed in the war of 68–70. In the year of our deed, 134, their sons would be 64– 66 years of age. 23 24

The BKL has ys˙ , the BMD: ys˙w . Here is a list of the names occurring in them, together with the number of occurrences: Eleazar (5), Judah (4), Jehoseph (3– 6), Simeon (3), Ariston, Hdr, ˙ Jacob, Manasseh, Shalom (fem.). (Eliashib is probably not a personal name.)

000358

Birnbaum 01.04.11 18:17

358

[ 52 ]

Reconstruction To simplify matters we have, in general, omitted the dots above the letters used in the Temporary Transliteration (pp. 346–347 above).25 We have separated the words but have not supplied any punctuation at this stage. That will be done at the next step in our interpretation – in the Tentative Translation. uibb rpkb larwi trxl tlt tnw rdëal hrsyb

1

aijw rzyla rb rzylal rma uibb rpk ñm hduhi rb rëdëx

2

il id htbl hnd hmi Ól tnbz hnd amui ituyr ñm hna ñmt ñm

3

Ótb hgl hnxtpt id itrd hgl ñupc [l] xitp id

4

úskb Ól tnbz Ód htrd ugb imy {Ól} Ól itii al hwru

5

iwr mlyl ñirmg ñimd ñtrt ñiyls hinmt ñumh id ñizuz

6

hb id luk airgh [u a]tëirwu ainba Ód htb ñbzb rzyla

7

axndm hnbz rz [y] l [a tna id Óë] d htb amxt ayqrqd hrpy

8

?

al hwru h-b aëmrd [u] abrym htrd hnpcu yr±iw [r] ta uëhu

9

rëdëx h{t}na ily qpm [a] l [u] lym alu il id htrd ugb iëmy Ól itia

10

hnd hmi ñm Ód htb ñb [zb] bry iuha hnau mlylu hnd hmi [ñm]

11

itial ñilm hnd rëdëx [ttn] a ñuymw trb mlw hnau mlylu

12

htë [b (±)±]±t± mlylu »hnd hmi ñm« Ód htb ñbzb {ml [y] lu} il lë (±)(±)±l hnqn idu

13

[h] btk hwpn ly ñuymw t [rb] mlw hugb mtxu jiwp hnd hb [t] k

14

hrmiëm

25

jwm

±(±) rzyla hwp»n« ly hd [u] hi rb rëdëx

dy úsuhi rb ñuymw

16

dhw ús [uhi ] rb [y] wii

17

dhw hduhi rb hduh [i ]

18

15

Key: { } superfluous letter (or word) erroneously added by the writer. 〈 〉 letter (or word) accidentally omitted by the writer.

000359

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Conveyance

359

Tentative Translation 1. On the tenth of Adar of the year three of the Freedom of Israel, in the Village of Bebhayu. 2. Hdr, son of Judah, of the Village of Bebhayu, says to Eleazar, son ˙ Eleazar, s˙ty , of 3. of the same ˙place: Of my own free will {to-day} I sell you to-day the house I own 4. which opens on the north into my courtyard so that you will have entry through it into your house 5. but you shall have no right with me in this courtyard. I am selling for the amount of 6. eight zuz, i. e., two sela, full price, for ever. Entitled is 7. Eleazar through the purchase of this house to the stones and the beams (?) [and] the bricks (?) and all that is in it 8. 〈from the height of the firmament to〉 the dust of the ground. The boundaries of the house which you, Eleazar, are buying, are: On the east side 9. it is the sloping ground (?), on the north side the court yard, on the west and south sides . . . . . . . . . . . . And no right 10. have you (in common) with me in the courtyard which I own, nor right of way on me, Hdr, ˙ forever. And I guarantee the purchase of 11. [from] this day onwards this house from this day onwards 12. and forever. And I, Shalom, the daughter of Simeon, the wife of the said Hdr, have no claim ˙ in connection with the purchase of this house 〈from to-day〉 13. {forever} and forever . . . . . . . . . . [hous]e and which we sign over to (you) . . . 14. This document is open and signed within. Shalom, [daught]er of Simeon wrote on her own behalf. 15. Hdr, son of Judah, on his own behalf. Eleazar ms˙t , in his affairs (?). ˙ ˙ 16. Simeon, son of Jehoseph, witness. 17. Jesh[ua], son of [ Jeho]seph, witness. 18. [ J]udah, son of Judah, witness.

000360

Birnbaum 01.04.11 18:17

360

[ 52 ]

Appendix Spelling The spelling is very fluid: a word is sometimes written in two quite different spellings in the same line, e. g., ywm and ymh (line 3) – a change between scriptio plena and defectiva for the first vowel, alternation between aleph and heˆ for the second vowel. The etymological principle is sometimes neglected: lgh occurs side by side with bgw. Scriptio Defectiva Scriptio defectiva may be employed for all long vowels: bth, dk, hrt, ˙ zbn, htm, ymh, thm , trtn. It is not used in final position. ˙ ˙ Scriptio Plena In medial position scriptio plena is of about the same frequency. However, its use in final position and the spelling of the ending yn give the text, on the whole, an appearance of being in scriptio plena. Aleph designates aˆ or eˆ in final position: bny , hgry , ywm , l , mdnh , m rb , qrq , thm . ˙ˆ designates aˆ, eˆ˙, oˆ or uˆ in final position: nh, bth, gh, dnh, drth, He ymh, k[t]bh, nqnh, srh, prh, s˙pnh, rs˙h, tmnyh. Waw designates o, oˆ, uˆ and u (?) in either position: bbyw, gw, hw (?), hmwn, zwzyn, kwl, r wty. Yodh designates ˆı in either position and eˆ in final: hwy, yty, gmryn, dy, dmyn, drty, zwzyn, ly, mlyn, sl yn, my, ps˙yt, ptyh, r wty, rs˙y. ˙ ˙ In reverse tabulation we have: aˆ is rendered – at the end only – by heˆ: nh, bth, dnh, drth, ymh, k[t]bh, srh, prh, spnh, tmnyh. But aleph is not much less frequent: bny , hgry , ywm ,˙ l , mdnh , m rb , qrq . eˆ is rendered by yodh: ˙hwy, rs˙y; by aleph: thm ; by heˆ: nqnh. ˙ zwzyn, ly, mlyn, sl yn, ˆı is rendered by yodh: yty,26 gmryn, dy, dmyn, my, ps˙yt, ptyh, r wty. o and˙oˆ are˙represented by waw: gw, hmwn, kwl, in one word by heˆ: gh. uˆ and u (?) are rendered by waw: bbyw, hw (?), zwzyn, r wty, in one word by heˆ: rs˙h.

26

Yyty (line 5) is not a matter of spelling. The first yodh reflects the glide which replaced the glottal stop between the aˆ and ˆı of laˆ ˆıtai. (Cf. Kutscher apud AG, and the BKL, where the glottal stop also disappeared but without necessitating a glide: alai et > alaiet.)

The Kephar Bebhayu Conveyance

361

Abbreviations AG:

BKA: BKL: BMD: BMDB: EP: EPC: EPK: HS: KBK: KBK 1:

KBK 2: M: NS: R:

S. Abramson and H. L. Ginsberg, “On the Aramaic Deed of Sale of the Third Year of the Second Jewish Revolt”, in: Bulletin of the American Schools of Oriental Research 136 (1954), pp. 17–19. S. A. B., “Bar Kosba (Kokhba) and Akiba”, in: Palestine Exploration Quarterly 86 (1954), pp. 23–32 {= Chapter 46 in the present volume}. Bar Kokhba Letter. See BKA. Beth Mashku Document. See BMDB. S. A. B., “The Beth Mashku Document”, in: Palestine Exploration Quarterly 87 (1955), pp. 21–33 {= Chapter 48 in the present volume.} Elephantine Papyri. A. Cowley, Aramaic Papyri of the Fifth Century B. C., Oxford 1923. E. G. Kraeling, The Brooklyn Museum Aramaic Papyri. New Haven 1953. S. A. B., The Hebrew Scripts. 2 vols. London 1954–1957 / Leiden 1971. Kephar Bebhayu Marriage Deed. S. A. B., “The Kephar Bebhayu Marriage Deed”, in: Journal of the American Oriental Sociaty 78 (1958), pp. 12–18 {= Chapter 53 in the present volume}. S. A. B., The Bar Menasheh Marriage Deed. Its Relation with other Jewish Marriage Deeds. Istanbul 1962. J. T. Milik, “Un contrat juif de l’an 134 apre`s J.-C.”, in: Revue Biblique 61 (1954), pp. 182–190). S. A. B., “The Negeb Script”, in: Vetus Testamentum 6 (1956), pp. 337– 371 {= Chapter 51 in the present volume}. J. J. Rabinowitz, “Some Notes on an Aramaic Contract from the Dead Sea Region”, in: Bulletin of the American Schools of Oriental Research 136 (1954), pp. 15–16.

000362

Birnbaum 01.04.11 18:17

000363

53.

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Marriage Deed*

So small and fragmentary was the papyrus deciphered in my paper “A Fragment in an Unknown Script”1 that some signs had to remain in doubt. The Kephar Bebhayu Conveyance (KBC ), which was published two years later,2 contained a considerable amount of material in the new script and made it possible to clear up most of these uncertainties. Other difficulties have since then yielded to further study, resulting in the present article. The new document also made it possible to broaden the palaeographical discussion of my “An Unknown Aramaic Cursive”3 although its essentials – the establishing of a new branch of the Aramaic script and of the approximate date of the Marriage Deed – have been confirmed by the Conveyance.4 In the present article we shall first of all discuss the script of the Kephar Bebhayu ktwbh (= KBK ) line by line until a transliteration has been arrived at. This will form the basis for a discussion establishing the words and meanings.5

* First published in: Journal of the American Oriental Society 78 (New Haven, Jan – March 1958), pp. 12–18. {A few slight corrections were made by the author in 1988.} 1 PEQ 84 (1952), pp. 118–120 {= Chapter 49 in the present volume}. 2 J. T. Milik, “Un contrat juif de l’an 134 apre ` s J.-C.”, in: Revue Biblique 61 (1954), pp. 182–190. Here on p. 189, n. 3, the author also refers to a few readings in the Kephar Bebhayu Marriage Deed. He does not indicate whether he, too, had deciphered this fragment prior to his study of the Conveyance, and whether he utilised my articles in PEQ 1952 and 1953 {= Chapters 49 and 50 in the present volume}, when he worked out his readings. In his kind reference to the general correctness of my result he does not refer to the fact that I had had no assistance from the additional material contained in the Conveyance. Cf. now my article “The Kephar Bebhayu Conveyance” in: PEQ 89 (1957), pp. 108–132 {= Chapter 52}. 3 PEQ 85 (1953), pp. 23– 41 {= Chapter 50}. 4 Cf. “ The Negeb Script”, in: Vetus Testamentum 6 (1956), pp. 337–371 {= Chapter 51}. 5 See the facsimile of the document p. 374. A photo may also be found in: S. A. Birnbaum, The Hebrew Scripts. (HS) 2 vols., London 1954–1957 / Leiden 1971, plate 121(4). A comparison of our document with Elephantine deeds and later ones is given in S. A. Birnbaum: The Bar Menasheh Marriage Deed. Its Relation with other Jewish Marriage Deeds. Istanbul 1962.

000364

Birnbaum 01.04.11 18:17

364

[ 53 ]

I. The Letters Line 1: The seventh letter might at first glance be taken as waw or yodh but it differs very much from all the specimens of these two letters in the KBK since they are curved and generally smaller in size. It looks rather like the fourth letter from the left edge in line 3. The next group of signs is puzzling. The first sign is a s(h)ıˆn, the third downstroke constituting its left-hand stroke. This letter is followed, without an intervening space, by a taw; in addition, a dot or little stroke sits on top of both of them, and so looks like the top of the right stroke of the taw. Line 3, however, provides the clue. Here, exactly under that sequence, we get part of the same configuration, undisturbed by the presence of a right-hand part. It is a nuˆn followed by a taw that was started so close to the nuˆn that its right-hand stroke could not avoid contact. The difficulty in line 1 arises from the fact that all three letters (s˙nt ) touch each other in one place. The three letters of the preceding word also touch each other, but only two at a time. Line 2: The third letter, s(h)ıˆn, runs into the second, nuˆn, but, like the examples in line 1, this is not a ligature because the two signs retain their normal form. The sign following the heˆ is, however, a ligature – a sign in which two letters have been fused to form a new unit: the circle of mem has practically become a tick while the tail of the mem has been swallowed up by the top of the final nuˆn; we have here a special sign for mn, i. e., the preposition min. Line 3: The first stroke (at the edge) might have been part of an aleph, zayin, heth or taw. ˙ from the left-hand edge is too pointed at the line bottom The fifth letter to be a nuˆn. The only other possibility seems to be a kaph, and that is what we actually have here – our letter tallies with the specimens in the KBC . The next letter might be a daleth (straightened like the resh in line 1), waw, zayin (of which there is no other specimen in this deed), yodh, resh (like the one in line 1). The same possibilities apply to the subsequent letter as well. The traces of the last sign appear to indicate a mem. Line 4: The letter after the mn is a kaph, like those of the KBC . Next there seems to be a hole in the papyrus – the outline does not suggest a letter or group of letters. Line 5: The little strokes at the end of the line seem to be a group of digits, ending up with a long stroke, like the numerals in the Beth Phage graffiti6 or in the Elephantine Papyri.7 The graffiti, however, have the 6

Cf. Facsimile 121(1) in HS.

000365

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Marriage Deed

365

last stroke slanting in the opposite direction. Half the specimens in the Elephantine Papyri have the last stroke parallel with the preceding ones; twice out of every three times the strokes are all of equal length; also they are all clearly and separately written, not running into each other as here in the KBK .8 There seem to be four little strokes so that the numeral would be 5. Beth Phage, however, as well as Nabatean, have a special sign for 5, so we should perhaps assume that there are only three little strokes here.9 If our sign is a figure then the sign before it might be a higher decimal unit, a 10 or a 100. It does not look like the 10 or 20 of Beth Phage or Nabatean, and a derivation from Nabatean 100 would present difficulties. If the beth in front of it were meant to function as a numeral in a multiplication – say, 2 × 100 – it would be an exceptional combination of a figure and an alphabetical numeral. But perhaps this problem need not arise: might not the mysterious sign be a letter, after all? Identified letters with descenders, occurring in this deed, are final kaph, final nuˆn and qoph, so that our sign would have to be a final pe or a final sadhe. The top angle on the left excludes sadhe. The top stroke on the ˙right, however, does not exclude a pe ˙ because this right stroke might not belong to our sign but be a letter by itself, running into the pe. As such it might be a misshapen waw, yodh or nuˆn. Our word would then be bwp, byp or bnp. Another difficulty militating against the reading pe is the top of the big downstroke. In the specimens known to us from these centuries we do not come across such an angular break as here. Line 6: The letter with the descender is vitiated by a hole and / or effacement of the top part. It cannot be a final kaph, there being no room for the horizontal. Final nuˆn is excluded because of its entirely different shape. Final sadhe is out of the question, there being no trace ˙ nor room for it. Thus only qoph is left, and its of the right-hand stroke top part would fit in here. The sign at the left-hand edge resembles the one immediately above, i. e., the right-hand half of a digit numeral, but as it is preceded by a letter there is perhaps more likelihood of it being the beginning of a s(h)ıˆn. 7

8 9

Cf. Facsimile 122 in HS. Cf. also the Pahlavi papyri from Egypt, 619– 629 (Olaf Hansen, Die mittelpersischen Papyri der Papyrussammlung der Staatlichen Museen zu Berlin. Berlin 1938 (Abhandlungen der Preussischen Akademie der Wissenschaften, Phil.-Hist. Kl., 1937.9). Also in these Pahlavi papyri. The above-mentioned Pahlavi papyri have the digits for 5.

000366

366

Birnbaum 01.04.11 18:17

[ 53 ]

Line 7: What looks like the third sign is only an accidental combination, waw or yodh, with the bottom part of the second sign. This bottom part only happens to be severed from its top stroke: thus the sequence is bw or by. The second difficulty in this line occurs in the sign after the mem. It looks unlike any of the other letters in this hand. The greatest similarity would be with a non-final aleph, the top stroke of which is touched and passed by a little downstroke – either a waw or yodh, or just an accidental stroke. It seems, however, to be too tall for a non-final aleph – of the other specimens even the tallest (line 6) is decidedly shorter. The only other possibility would be to read it as a heˆ that had got misshapen in avoidance of the preceding mem. The right-hand stroke would have stopped short to avoid the slanting tail of the mem. This abnormal shortness would, in its turn, have caused the displacement of the left stroke, which now started from the left end of the top stroke instead of from the right. Line 8: The first letter is incomplete. What we see are the top and left-hand strokes of a heˆ. Next there is a waw (damaged at the bottom) which touches the top of the heˆ and so misleads the eye into regarding this group of strokes as a single letter, i. e., it looks as if it were a samekh, mutilated at the bottom. Just under the misshapen heˆ of line 7 there is a stroke running into the preceding letter (aleph). As this stroke is oblique and straight it is neither a waw nor a yodh, but the right-hand part of a mem – the bottom curve that should link it with the left-hand part has been effaced. Line 9: The first sign must be a beth. It differs from the other specimens but each of its parts is present in one or the other of them: the straight, longish top stroke occurs in lines 2 and 4 (the second specimen); the sharply pointed right top corner we see in line 2; the base beginning at the end of the downstroke is to be found in lines 4 (second specimen), 5 (second specimen), 14, 15, 15, 16. At the end of the line the base of the beth at first glance looks as if it corresponded to the second specimen in line 5, but careful inspection shows that there is a left-hand part to the base, as in the first specimen of line 5 – the wrong impression being created by the right-hand stroke of the subsequent taw running into the base. Line 10: In spite of the break in the papyrus – especially in the lefthand half – all the letters are clear. Line 11: The top of the second letter is almost effaced but is still recognisable as part of a daleth or resh. The two final letters correspond exactly to those of line 9. The base of the beth is clearly written and corroborates our interpretation of the form in line 9.

000367

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Marriage Deed

367

Line 12: The initial letter at first glance suggests a final kaph, because it differs clearly from the qoph just below it in line 13. However, in the kaph of line 7 the top stroke reaches the downstroke somewhat below the top, in contrast with our specimen. This indicates that it is, after all, a qoph, with its top region not as well developed as in line 13. It seems that the indistinct specimen in line 6 is of the same form. The second letter runs into the qoph and into the subsequent letter and thus creates difficulties. However, what we have here is a nuˆn followed by a final aleph. At the end of the line, the black top part is evidently the shadow cast onto the background by the upper edge of a horizontal break in the papyrus. If the two parallel downstrokes belong to one letter, it could only be taw: if to two letters, the first stroke might be daleth, waw, yodh, ayin or resh: the second stroke might be one of these or aleph. Line 13: There is no trace of writing to the right of the first sign so that it seems that this is the beginning of the line. The fifth sign is difficult. That it represents two letters is clearly out of the question. Could it be a heˆ? This is rendered unlikely by the presence of what appears to be a little stroke in the middle, as well as the absence of a horizontal top stroke – which is fully preserved in the distorted heˆ of line 7 and which it would be hardly possible to identify as the oblique stroke here. Upon re-examination we find that we have here not two downstrokes and a little line in between but a downstroke on the right, while the left consists of two parts: (1) an angular hook, starting at the top and failing to reach the bottom, (2) an oblique stroke that issues from the left-hand point of the hook. In other words it is a mem. It is not written in one continuous oval movement – the right and the left are drawn separately, the left-hand stroke being accidentally shortened. Whether the other mems in this document were constructed in the same way it is hard to tell. Our final difficulty concerns the sign(s) after the ayin. This group looks like the ligature mn (the absent downstroke would then be accounted for by the gap) but perhaps a mem with an unusual left-hand stroke is also possible. Line 14: The last sign might be the lower half of a daleth, resh or taw. Line 16: The first letter is followed by what seems to be a heth, although it is not identical in every particular with the specimen in˙line 1. Otherwise we would have to read a waw or yodh, run together with an aleph.

000368

Birnbaum 01.04.11 18:17

368

[ 53 ]

First Transliteration This document shares with the KBC the feature that the separation of words by means of blank spaces is sporadic. In the following text no attempt is made to decide where words were intended to be separated in the original: we are giving the whole in scriptio continua.10 The underlined - stands for the final form of aleph which is used in this script, the underlined m for the single form of mem in both final and non-final ¯¯ position. Alternative possible readings of damaged or indistinct letters or fragments of letters are separated by oblique strokes. K and N stand for the final forms of kaph and nuˆn. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

10

] b hl d/rd/rs˙nthd / rh [ ˙ ] mns˙hmNbnw/y lw/ys˙w / yb [ ¯¯ [ ] /z/h/tthw/y-lw/yl nthkd/w/z /y/rNm ˙ ] -ltmNk - - - bbw / yw/yd/rh/k/t [ ¯¯ ˙ ˙ ] ¯bt t bw/y b - '''' [ ˙ ˙ ] lhw/y-qw/ym m ts˙ [ ¯¯ ¯¯ ] lbw / yt lmhthK [ ¯¯ ] hw/yd/r hw/y mbnNlb¯ [ ¯¯ ] blw / yN w/yhN nhlbt [ ] w/ymtw / ynhw/ymbsw/y- [ ] d/rmlw / yd/rw/yl-mNbt [ ¯¯ ] qn- hd/r w/yNw / yt (or: d/w/y/ /r or: /d/w/y/ /r ) [ ˙ ] w / yqw / yd/rmw / yd/rtw/y m (or: mN )kw/yl [ ¯¯ ¯¯ ¯¯ ] lw/ybs˙td/r-d¯/r¯ [ ˙ bs˙b h [ bhd / rd/rw/y [ ˙ mNbnw/y lw/y [ ¯¯

It need hardly be said that the writer of the Deed did not intend scriptio continua.

000369

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Marriage Deed

369

II. Words and Meanings Line 1: This line divides easily into words. The words s˙nt and hdh strike the eye, suggesting that the line contains a date. We expect ˙a numeral and the name of a month before s˙nt: the first letters are obviously the end of s˙b h ‘seven’. Thus we have isolated four letters for the month: lamedh is the connection with the numeral, the first two letters of the month are d or r. It is very likely that the ayin at the end of the line was followed – as suggested by Milik – by samekh and resh. To which era did this “year eleven” belong? There is every likelihood that the KBK , found together with the Bar Kokhba Letters, the Beth Mashku Document and the KBC , belongs to about the same time as they do. If the number denotes a regnal year it would, in that period and region, presumably refer to a Nabatean king. This would then have to be the last – Rabbel II – and the year would correspond to 81 C. E. Another possibility is the Era of Provincia Arabia,11 and the year would then be 117 C. E. The fact that the KBK comes from the same place as the KBC which bears a date corresponding to 134, inclines one to consider 117 as the more likely of the two. Line 3: The final aleph marks out the first word boundary, and the word is, beyond doubt, thw . It could be either the 2nd or 3rd pers. If the preceding letter is taw, it might be the end of the personal pronoun and thus decide in favor of the 2nd person. The next division, between ly and l nth, provides no difficulties. Line 4: The letters towards the end of the line were read in my first paper as bbw or bby. At that time the KBC had not yet been published so that those letters appeared inexplicable. Now, however, there can be no doubt, that the sequence of letters we have here – b b w/y w/y – constitutes the same word bbyw as there. This suggests that the preceding word was kpr but there are difficulties: at least something of the extremities of pe and resh should be visible on the edges of what seems to be a hole, and why is there a blank between this and bbyw? The same question would apply if what we thought was a hole were simply a smudge (although, to judge by the forms, a smudge seems unlikely). Line 5: tby might be ‘coined’. bwp, byp or bnp makes no sense ˙ (unless it were some kind of abbreviation). The explanation of the sign 11

This era might have been called here that of the eparchy (hprky ) as in Jaussen / Savignac, Mission Arche´ologique en Arabie. Vol. 1, Paris 1909, no. 159, or it might have been omitted, as ibid., no. 17.

000370

370

Birnbaum 01.04.11 18:17

[ 53 ]

after the beth as a numeral therefore seems to be preferable, especially as it is followed by what can hardly be anything but a numeral. Line 6: Since the first four letters make up a word (lhw – a reading supported by line 3) the mutilated sign with the descender must be regarded as the beginning of a new one, and would therefore not be a final letter. This confirms our palaeographical findings in Section I, and enables us to isolate the second word as qym, presumably qayyaˆm ‘existing’ or, possibly, qim ‘established, certain’. The blank space after the second mem is too wide to be accidental, and the third word therefore is m. This can hardly signify ‘if’ – although Targum Onkelos has it – seeing that the hn of line 9 is obviously ‘if’. It is probably identical with the m of Elephantine12 meaning ‘in addition, moreover’.13 Line 7: The blank space after the ayin is only apparent because the lamedh does not come down between the ayin and the mem. The space between the subsequent heˆ and the following taw seems to be intentional. The division into words is obvious: lbyt lmh thk. ‘The house of one’s eternity’ is, of course, one’s eternal home, i. e., the grave. The expression is familiar to us from Phoenician,14 the Bible,15 Targum,16 Talmud,17 Palmyrene.18 ‘To go to one’s eternal home’ for ‘dying’ is used in the Eccl. passage.19 Whether thk is the 3rd or 2nd pers. cannot be decided on the evidence of this or the former lines. On that decision depends whether the h of lmh is to be read as the pronominal suffix or the ending of the stat. determ. 12

13 14 15 16 17 18 19

A. Cowley, Aramaic Papyri {of the Fifth Century B. C., Oxford 1923}, no. 13, line 11; no. 34, line 6; E. G. Kraeling, The Brooklyn Museum Aramaic Papyri. {New Haven 1953}, no. 3, line 16, 19; no. 9, line 21; no. 10, line 11, 14. As suggested by Jacob Rabinowitz apud Kraeling, ibid. hdr bt lm (Corpus Inscriptionum Semiticarum {CIS} I, no. 124, line 1). ˙hwlk . . . l byt lmw (Eccl 12: 5). s˙kbyn . . . gbr bbyt lmyh: rendering of s˙kbw ys˙ bbytw (Is 14: 18, Rashi ad loc.: nqbryn) – while the byt lmw of Eccl 12: 5 is translated byt qbwrtyh. bs˙ t ptyrtw s˙l dm lbyt wlmw (Taan. 11a). ˙ (CIS II , no. 4116, 4121, 4123, 4168). – The earliest examples of the byt lm meaning ‘cemetery’ seem to be in Lev R. 12 : 1 and Lam R. 1 : 32. In an epitaph of 1718 at Aden (cf. plate 219 in HS) the idiom is n sph lbyt wlmh h s˙h etc. In the above-mentioned Talmudic passage the root ptr is ˙ employed. In the liturgy and in epitaphs right up to the present day the idiom is ‘to go to one’s eternity’, i. e., the word byt is omitted. Perhaps it was dropped to avoid ambiguity when the concrete meaning of byt wlm – ‘cemetery’ – had become the only one.

000371

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Marriage Deed

371

Line 8: The sequence hw in the middle of the line indicates a probable word boundary – h would be an ending, w would be ‘and’. The letters before it, hwd h or hwr h, cannot be the Hebrew words of the same spelling. The reading yhwd h suggests itself. The form is either masc. determ. or fem. absol. Towards the end of the line the word bnn leaps to the eye. It invites the vocalisation be˘naˆn and thus the translation ‘daughters’, but in view of the fluidity of the spelling, be˘nıˆn ‘sons, children’ is also quite possible. These two readings have isolated w m. That is evidently the same word as in line 6, preceded by ‘and’. Line 9: lwn might be the preposition l with the suffix of the 3rd pers. plur., the heˆ having dropped out, as it frequently did in Palestinian Aramaic. But as the preceding letter is almost certainly a beth, it seems preferable to assume that there was a qoph before it, resulting in the root qbl, pa el: ‘to receive’.20 If this verb refers to bnn of the previous line that noun would have to be read benıˆn, because for benaˆn the verb would have to be yqbln. The word nh being obvious, the letters between it and blwn – with the final nuˆn showing a word boundary – can easily be divided into w and hn. The bt at the end of the line is possibly a defective spelling of byt, although in line 7 the scriptio plena is used, but divergencies in one and the same MS . are not uncommon in this material: in the KBC we find ywm and ymh in the same line (line 3), and in that document ‘house’ is always spelt bt. If this suggestion is correct, lbt might be a parallel to the lbyt in line 7, and it would be possible on the basis of the preceding word ( nh) to complete the phrase with lmh hk. The stress on ‘I’, as expressed by the use of the separate pronoun, fits well into this interpretation. Line 10: The sequence hw in the middle of the line suggests that there is a word boundary between them – the h being an ending of the first word and the w the link with the second. The first – from the root mtn – might be derived from the verb me˘than ‘to cause delay, to let wait’, being either (a) the 3rd pers. sg. masc. with a pronominal suffix: yamtineh or yamtinah ‘he shall wait for him / her / it’, or (b) the partc. pass. fem., preceded by ‘and’: uˆme˘thıˆnaˆ ‘and waiting, and detained’; or it might be (c) the fem. of the adj. mathyaˆn ‘slow, careful’. The second root is bsy, which in the Talmud means ‘to trample upon, despise; drive, instigate’. In the Aramaic languages the particle min 20

The meaning of the qal – ‘to complain, lodge a complaint’ – need not be considered as it would have to be followed by l and the name of the person against whom the complaint is directed.

000372

Birnbaum 01.04.11 18:17

372

[ 53 ]

seems to be generally written as a separate word. It would therefore be unlikely that wmbsy is to be read uˆmibbisyaˆ / uˆmibbusyaˆ < uˆmin bisyaˆ / busyaˆ ‘out of wilful negligence, indifference’ (for which the Talmud has bbsy ). The m would, therefore, be the nominal prefix, either of a noun, or perhaps more likely, of a partc. pa el. That the first word is a verb is unlikely, seeing that the second word cannot be a verb and that it is therefore evidently linked to it by ‘and’. From among the other meanings we shall have to choose one that will either be parallel to the second word or its antithesis – a pattern common in legal language. From this point of view ‘careful and negligent’ would be a suitable pair. Line 11: Most of the line divides easily into words. At the beginning of it we have however to make a choice in regard to the second letter. dm occurs instead of dm in the Targum21 and the Talmud22 but it is not very likely that we should meet with the word ‘blood’ here. If we read resh, and do not take the subsequent two letters as another word (lw or ly) we get rmlw ‘widowhood’ – a perfectly suitable word in a marriage deed. Line 12: The final aleph provides us with the first word boundary in the line, the final nuˆn with the last one. That leaves six letters to decide about. The choice between daleth and resh is not difficult: hd ‘to hold, ˙ because seize’ would be followed by yn ‘yes’ – yn ‘if’ being excluded of the form hn in line 9; wn does not exist in Aramaic – and this combination makes no sense. hr yn, on the other hand, is a natural ˙ word to find in a deed. n is evidently the pronominal suffix of the 1st pers. plur. As the subsequent word hr yn would not start a new sentence, we would ˙ it, i. e., to be the suffix of a noun in the nominative. expect n to relate to Line 13: This looks like a difficult line: four times we have a choice between waw and yodh, twice between daleth and resh, there is a thoroughly misshapen letter (mem), and something seems to be missing at the bottom in the last third of the line. A number of feasible combinations suggest themselves. One could, e. g., try to read yqyr, or think that yqyd and yrty go well with each other semantically, and finish up the line with a plain mn kwl. However, if our first choice is daleth, the mem, which would be inexplicable after a resh, falls into place: wqwdm. If we read the sign after the ayin as a mem we have m kwl (not less probable than mn kwl) and have isolated the word yrty ‘my heirs’ – an obvious word in a marriage deed. 21 22

I Chron 22 : 7. J. Maas. Sh. 5 : 56 d; B. Git. 47a.

000373

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Marriage Deed

373

Line 14: In view of what we said about the beginning of line 13 (see Section I), the lamedh may have been the first letter of the line. We would then have to read ly. But if it was not the first, it could have been preceded by not more than one or two letters. There might have been an ayin: ly before bs˙tr would be the right word to express obligation. ˙ The possible daleth at the end of the line might be the first letter of the demonstrative pronoun dnh, referring to the preceding noun. After line 14 there is room for one or two more lines but no trace of effaced writing is visible here. Line 15: The identity with the number in line 1 suggests that what we have in line 15 is the date again and that the wording at the beginning of the document is repeated here. Accordingly we may perhaps supply the text of line 1 after s˙b h. Line 16: Reading b w/y d/r d/r does not seem to yield any result. B h ˙ d/r d/r, on the other hand, recalls h d/r d/r, the name of the seller of the ˙ house in the KBC . Is it by any chance the same name? If our deed is dated in the Era of Arabia then this name might conceivably refer to the same person, 17 years earlier. He might then have been a young man – the bridegroom – and his bride would have been Shalom. Line 17: We get the same words here as in the middle of line 2, and so, on the basis of what we said in regard to line 15, line 16 may be presumed to have ended with the word mns˙h.

000374

374

Birnbaum 01.04.11 18:17

[ 53 ]

The Kephar Bebhayu Marriage Deed

000375

Birnbaum 01.04.11 18:17

The Kephar Bebhayu Marriage Deed

375

The Text 1 bs˙ ] b h l dr s˙nt hdh [sr ˙ 2 ] mns˙h mN bny lys˙yb [ ¯¯ [ 3 ] /z/h/t thw- ly l nth kd/w/z/y/rd/w/z/y/rN m ˙ ¯¯ 4 ] -lt mN k - - - - bbyw d/rh/k/t [ ˙ 5 ] ¯b t tby b -'''' [ ¯ ˙ ˙ 6 ] lhw- qym m ts˙ ¯¯ ¯¯ 7 ] lbyt lmh thK [ ¯¯ 8 y ] hwd h w m bnN lb¯ [ ¯¯ 9 yq ] blwN w hN nh lbt [ 10 ] wmtynh wmbsy- [ ¯¯ ¯¯ 11 ] rmlw dy l- mN bt [ ¯¯ 12 ] qn- hr yN w/yt (or: d/w/y/ /r or: /d/w/y/ /r ) [ ˙ 13 ] wqwdm yrty m kwl [ ¯¯ ¯¯ 14 ] ly (or: ]ly) bs˙tr- d[nh ˙ 15 bs˙b h [l dr s˙nt hdh sr ˙ 16 bhd/rd/rw/y [ ˙ 17 mN bny ly [ s˙yb ¯¯ Tentative Translation 1 On the se]venth of Adar of the year el[even (of the Eparchy?) 2 ] Manasseh of the Sons of Eliashib [ 3 ] – you / she shall be to me a wife . . .[ 4 ] . . . of Bebhayu (Village?) . . . [ 5 ] good (?) coined 205 (?) [ 6 ] to be valid / established. Furthermore . . . [ 7 if ] to the eternal home you should go [ 8 ] Jewish. And furthermore: Daughters / Sons . . . [ 9 they will rec]eive. Or if I myself to the house of [eternity should go 10 ] and careful or negligent [ 11 ] widowhood that not from . . . [ 12 ] . . . are sureties . . . [ 13 ] and before my heirs with all [ 14 upon] myself G by th[is] deed [ ] to me On the seventh [of Adar of the year eleven (of the Eparchy?) 15 16 . . . . . . [ . . . Manasseh 17 of the Sons of Elia[shib

000376

Birnbaum 01.04.11 18:17

000377

54.

Birnbaum 01.04.11 18:17

A Sheet of an Eighth Century Synagogue Scroll*

The sheet in question is in the library of Jews College, London.1 How and when it got there is not known but it only recently came to light during the removal of the College. Miss R. P. Lehmann, the Librarian, drew my attention to the MS . and asked my opinion about it. The present article is the result.

Fig. 1. Part of a sheet of an eighth century Synagogue Scroll

* First published in: Vetus Testamentum 9 (Leiden 1959), pp. 122–129. 1 {Now renamed London School of Jewish Studies. – E. B.}

000378

Birnbaum 01.04.11 18:17

378

[ 54 ]

I. Description A) External Features The material is a light brown leather. The sheet is about 155 cms in length, not counting the folded edges on the right and left where it had been sewn to the preceding and following sheets. The present height of the sheet ranges from 48.1 cms to 53.2 cms. The top and bottom edges are not the original ones – except perhaps a small part of the right-hand bottom corner – the sheet as we have it now being cut in a somewhat irregular fashion. However, the original margin was hardly much wider. The height of the written surface is 40.8– 42 cms (including the ascender of lamedh). The width of the columns – of which there are seven – is generally 7.9 cms; one, however, is 7.75 cm, and column VII is only 7.3 cms wide. In most cases this coincides with the actual length of the written line but occasionally the final letter projects beyond the vertical ruling; it now and then happens that two letters do so. There are, on the average, 22 letters to the line. The distance between the horizontal rulings is 1 cm, the height of the letters (excluding ascenders and descenders) about 3.5 mms. There are 42 lines per column. The intercolumns are 3.2 – 4.9 cms wide. A big tear across the lower half of column II has left the text undamaged. In column VII , however, some letters are deleted by a hole crossing four of the lines. There is a patch at the back here but the same hole goes through this too. A few Arabic letters are visible on the patch but they are too mutilated to make any sense. There are a number of stains (grease?). The three biggest are in columns II – V. Here the leather is considerably darker and has lost much of its shine and characteristic surface grain. The areas covered by the biggest stains are slightly crumpled but the lettering is not affected, the ink remains black; as a matter of fact, it is darker there than in the rest – particularly the left half of the fragment.

B) The Text The sheet contains Exodus 9: 18 – 13: 2. The text is identical with the Masoretic Text. The divisions into sections tallies with the traditional one, except that the section following the word m rsw (11: 10) is a ˙ ptwhh, while we have here a stwmh. ˙

000379

Birnbaum 01.04.11 18:17

A Sheet of an Eighth Century Synagogue Scroll

379

There is one very striking feature: At the end of many of the lines we find a dot – occasionally two dots horizontally arranged. Sometimes the dot is placed before the final word, and occasionally before as well as after it; two dots before and one after also occur. Their purpose is obviously to fill in every blank, even a small one, lest it should be mistaken for a ptwhh division. As far as I know there is no mention in our literary sources˙ of this particular device.

II. Date and Country of Origin Needless to say, this sheet once formed part of a synagogue scroll. When and where was it written? There is no internal evidence to help us answer that question. The above-mentioned difference in one of the sectional divisions provides no clue. As to external evidence: that the material is leather and not parchment points, on the whole, to a comparatively early date. But there are recent examples too, and we do not know until what period the use of leather was fairly widespread, otherwise this might have given us at least a possible, very approximate, terminus ad quem. Thus the only kind of external evidence available in the present instance is the writing.

A) General Features The height of the letters (3.5 mms) is small compared with that in the middle, late and post-medieval scrolls: as early as in the ninth century the height ranges from 6 to 9 mms. But we cannot draw a chronological conclusion from this because, from that time onwards right up to the present day, small, medium and large lettering is used. The average of 22 letters per line falls short of the figures prescribed for a synagogue scroll which is 27. The number of lines per column tallies with the prescribed minimum, i. e., 42. Our MS . has no tgyn – the ‘crownlets’ on certain letters of the alphabet. Their absence does not, however, provide a chronological clue. The tgyn are first mentioned in the fourth century,2 which, of course, does not tell us when they began to be used. There are none in 2

By Raba, 299–352 C. E., cf. Men. 29b. Earlier mention refers to parts of letters, not to what was later called tgyn (as I hope to show on some future occasion).

000380

Birnbaum 01.04.11 18:17

380

[ 54 ]

our oldest fragments – that from Murabba aˆt, dating from the early second century B. C. E.3 and the considerably earlier ones from Qumraˆn;4 these, however, were not necessarily parts of synagogue scrolls. But many medieval and later synagogue scrolls have no tgyn either, and this problem needs thorough investigation. There are a few dilated letters in our MS . This feature is of some chronological value to us. The device of dilation is not employed in the MSS from Qumraˆn and Murabba aˆt, nor5 even in the much later Antinoopolis Fragment no. 46 which emanates from sixth century Egypt. We first encounter the regular use of dilation in our sixth to seventh century Palestinian material. Here, as in the present MS ., it is employed only sparingly; similarly, the actual degree of the dilation is very moderate, even where a small increase in the length of the top bar would have fulfilled more adequately the aesthetic purpose of the device by bringing the word right up to the left edge of the column. Moreover, the final letter, or last two letters, of a line are not infrequently written beyond the vertical ruling (a permitted procedure). This, together with the severe economy exercised in the use of dilation, results in the left edge of the column being almost as irregular as the corresponding edge of a carefully type-written page would be at the present day. Although the left edge of our MS . is more regular than in the Qumraˆn and Murabba aˆt material, it bears no comparison with the late- and post-medieval scrolls where the left and right edges are of practically equal straightness.

B) Comparison No detailed examination is needed to see that the writing of our MS . is very old. On the other hand, the script is nothing like as early as that of the Cave MSS . – including the latest of them, those from Murabba aˆt.7 It is also clearly later than our material from the next few centuries after that.8 When we come to the ninth to eleventh centuries Bible codices9 3 4 5

6 7 8 9

The Hebrew Scripts (HS), by the present writer, no. 89. D. Barthe´lemy / J. T. Milik, Qumran Cave I. Oxford 1955, Nos. 1, 2, 4, 5. In the fourth century epitaph from Jaffa, shown in HS 90, the first mem is wider than the second but I am inclined to think that this is not our earliest example of dilation as an accepted method but that it was a device used by this particular mason. Cf. HS 183. Cf. HS 89. Cf., e. g., HS 91*. Cf., e. g., HS 92, 93*.

000381

Birnbaum 01.04.11 18:17

A Sheet of an Eighth Century Synagogue Scroll

381

however, we notice that the forms there are obviously far beyond the stage of development reached by the letters in our MS . Our field of investigation has thus been restricted to the period from about the sixth to the ninth centuries. We shall now see whether a detailed examination of the script will yield a more precise date. For that purpose we need material for comparison, and such material should obviously be as nearly as possible of the same type as the MS . in question, and it should also, if possible, bear a date or be dateable. Material of the same type means synagogue scrolls of the Pentateuch. But such scrolls cannot bear a date for the simple reason that they must not contain a single letter or sign in addition to the Biblical consonantal text. And even undated scrolls or fragments of scrolls which might, with some likelihood, be ascribed to the sixth to ninth centuries do not appear to be known. We shall therefore have to content ourselves with basing our comparison on other MSS . of that period, on the assumption – not, after all, an unreasonable one – that their forms and those of contemporaneous synagogue scrolls were either identical or not sufficiently different to vitiate our result.10 It is not possible here to compare our MS . with a large number of others. I have done this preparatory work and have selected from my collection of photos two representative specimens of the Palestinian branch, one from the early and one from the late part of our period (i. e., sixth to ninth centuries).11 The first, T (for Targum MS .), a Geniza fragment in the Taylor-Schechter Collection of the Cambridge University Library ( T.-S. 20155) {see HS 91A}, is a fragment from a scroll containing the Palestinian Pentateuch Targum, with Palestinian vocalisation.12 It can be dated by the convergence of internal and palaeographical evidence: (a) The text is provided with vocalisation and was thus probably not written before the sixth century. As its system of vocalisation is the Palestinian, the likelihood of a date after the ninth century is probably small. (b) As to the script, it is sufficient here to point out a striking feature: the final mem is open – a form which is characteristic of the fourth to seventh centuries. The dates of (a) and (b) overlap in the sixth to seventh centuries so that this is clearly the approximate time when T was penned. 10

11 12

Such an assumption would not, of course, be justified for recent centuries, when books had ceased to be written by hand and the type-face cutter had lost all connection with the letter forms of the scribal tradition. {Palaeographically, the script of our scroll would indicate Palestinian Square origin. – E. B.} And Tiberian accents, added later.

000382

Birnbaum 01.04.11 18:17

382

[ 54 ]

The second representative specimen selected here for comparison is the first Biblical MS . bearing a date – the famous codex of the Prophets written in 895 at Tiberias by the Masorete Moses ben Asher, and for many centuries in the possession of a Cairo synagogue. We shall designate it P (Prophets) {= HS 92A}. In comparing the three MSS . we shall refer to that from Jews College as E (for Exodus) {= HS 91B}. a) General Features We shall first of all deal with some general features. E has 42 lines to the column, T has 34, P has 22. In this respect E is closer to T than to P. In E the line averages 22 letters, in T the figure is 32, in P it is 12–13. So E stands half-way between T and P. The average height of a letter (excluding ascenders and descenders) in E is 3.5 mms, in T it is 2–3 mms, in P it is 7–8 mms. E is accordingly much closer in this respect to T than to P. An important criterion in comparing the three scripts is that of style. In P there is a strong contrast between the horizontals and the downstrokes (except in aleph), the former being three to four times thicker than the latter, while in T and E there is either very little or no such contrast. Thus in the majority of these particulars there is a closer relationship between E and T than with P. Translated into terms of relative chronology this would mean that E stands in between T and P while being nearer in time to T than to P. In absolute chronology that would suggest a date between roughly 600 and 895 but closer to the upper point – say very approximately around 700. b) The Letter Forms13 We shall now examine the individual letter forms. Aleph: In P the left stroke is split up: the part above the main stroke’s point of issue has become an independent little stroke; in contrast with the slant of all the 13

The sigla T, E, P will be used here to denote not only the MSS . but also the individual letters in them. The first line of this page is T, the second E and the third P. {Unfortunately in HS vol. 1, cols. 163–174, the printer confused the three alphabets reproduced correctly in the original article and above: The Alphabet reproduced in HS as “91A” (= T) at the bottom of col. 163–164 is actually 91B (= E). The alphabets reproduced on cols. 167–168 as “91B” and on cols. 173–174 as “92A” (= P) respectively are identical, being actually 92A. Consequently, alphabet 91A is missing in HS. – Eds.}

000383

Birnbaum 01.04.11 18:17

A Sheet of an Eighth Century Synagogue Scroll

383

Fig. 2. Comparison of alphabets of the Targum MS. T, the Codex of the Prophets MS. P and the Exodus MS. E. Drawing by the author

other strokes of the letter it is vertical. This whole feature is absent in E and T. In T the left stroke has a ‘foot’ turned rightwards, in E the left stroke has a short base, in P we have a ‘clump foot’, turned leftwards. Hence E stands between T and P. Beth: In T and E the right end of the base ends bluntly, while in P it comes to a point with an upward movement. Gimel: In T the left stroke is very long and it starts high up. In E it is shorter and its point of issue from the right stroke is lower down. In P it is short and starts at the line bottom. E therefore stands in between T and P. Heˆ: In T the left stroke is severed from the top bar, in E and P it is not – the writing of the two MSS . is in a more formal style than that of T. Waw: In T the top is an imperfectly developed head. This is also one of the E forms but there an even less developed type having only a small oblique top stroke is frequent. In P the head is well developed. Hence E is closer to T. Heth has the general form – not the specific one employed in synagogue˙ scrolls, where the top bar is split into two and the gap bridged by a circumflex. The earliest literary

000384

Birnbaum 01.04.11 18:17

384

[ 54 ]

reference to this type is in the Talmud (R. Ashi, dd. 427 C. E.). However not all synagogue scrolls have this heth. I have photographs from ˙ absence of this specific form many scrolls with the general form.14 The here provides no chronological clue, in the present state of our knowledge, or rather ignorance. Mem: In T the left stroke and the base are parallel, in E and P the two strokes are at an acute angle. Ayin: In T and E the top of the right stroke is just a bend, in P it is a fully developed head. Pe: In T the top stroke is horizontal, in E and P oblique, so that the outline of the upper part of the letter is triangular. Final pe: Here the same holds good. Sadhe: In T the right stroke is straight, in E bent, in P there is a head˙ obviously developed from a bend. Sˇˆın: In T and E the left stroke is a plain one, in P it has a fully developed head. In T the middle stroke is straight, in E it is bent, in P it has a fully developed head. Looking through the results of the foregoing comparison we find that six features (in aleph, beth, waw, ayin, sˇˆın) link our MS . with T, while four features (in heˆ, mem, pe, final pe) relate it to P. Six more features (in aleph, beth, gimel, sadhe, sˇˆın) indicate that E stands in between T ˙ and P. To sum up the relation between the three scripts we might say that E appears to be at a stage of development which has gone beyond T but has not yet quite arrived half-way towards the forms of P. Transposed into relative chronology this means that on the time graph between T and P, our MS . would be placed somewhat before the middle point between the two. In absolute chronology this would yield a date between c. 600 and the half-way point to 895, i. e., c. 750, but somewhat closer to the latter than to 600 – say, very approximately, round about 700.

C) Conclusion This date coincides with the one we arrived at in section (a), the two results corroborating each other, giving the period around 700 as the time when our MS . was penned. Our fragment is accordingly the oldest complete sheet from a synagogue scroll of the Pentateuch, separated by centuries from what, until now, were the earliest known to exist. 14

I do not remember having noticed any reference to this fact. It is as yet impossible to say to which regions and periods the scrolls with the non-specific type belong. This could only be done in the course of a general palaeographical investigation of extant synagogue scrolls – a vast and difficult undertaking, perhaps a life work.

000385

55.

Birnbaum 01.04.11 18:17

Die Vielfalt der hebräischen Schrift im Mittelalter*

Schon zu Beginn des Mittelalters weist die hebräische Schrift1 kein einheitliches Bild auf. Die weite Verbreitung der Diaspora und die Schwierigkeiten des Verkehrs führten zur Entstehung einer Anzahl von Judenheiten, die sich naturgemäß jede in besonderer Art entwickelten. Das fand selbstverständlich auch in der Schrift seinen Ausdruck. Die paläohebräische Schrift, die im fünften vorchristlichen Jahrhundert von den Juden aufgegeben wurde, erhielt sich bei den Samaritanern ins Mittelalter (und bis zum heutigen Tag). Die Formen der Buchstaben entfernten sich so weit von den paläohebräischen, daß die Kenntnis der einen nicht zum Lesen der andern befähigt. Am ursprünglichen Ausgangspunkt der hebräischen (d. h. jüdischen) Schrift – die sich seit ungefähr dem fünften vorchristlichen Jahrhundert aus der aramäischen entwickelt und bald die volle Form der sogenannten Quadratschrift erreicht hatte – haben wir den palästinischen Typus. Angrenzend im Osten liegt in Mesopotamien der babylonische Typus, im Süden der ägyptische. Diese drei sind eng verwandt. Etwas weiter östlich, in den iranischen Regionen, finden wir den sehr verschiedenen parssischen Typus. Eine ebenfalls sehr individuelle Form stellt das Temanische der jemenitischen Juden dar. Ein fünfter Haupttypus ist das Maarawische (Maghrebinische) Nordwestafrikas. Typen von geringerer Wichtigkeit entstanden in Kurdistan und Georgia, die verwandt zu sein scheinen, sowie die in Indien und China, die Schrift des letzteren dem Parssischen entstammend. Von besonderer Wichtigkeit sind die hebräischen Schriftgruppen Europas: das Jewanische und das Krimtschakische im byzantinischen Raum, das Italkische der Apenninhalbinsel, das Zarfatische Nordfrankreichs, das sich im Aschkenasischen der deutschsprachigen und spä* Zuerst veröffentlicht in: Hebräische Schrift, von der Steinschrift zum Poster. [Ausstellungs-]Wegleitung. Zürich, Kunstgewerbemuseum, 1976 ( Wegleitung Nr. 308). Neuauflage 1987, S. 22, 27. 1 Der Name ‘hebräisch’ für unsere Schrift ist nicht glücklich. Zur Zeit der Hebräer existierte sie ja noch gar nicht. Ihr Alphabet war das ‘phönizische’, das nordwestsemitische. Die richtige Bezeichnung wäre: ‘die jüdische Schrift’, so daß ‘hebräisch’ für das Alphabet der Hebräer frei wäre. Aber da diese Umbenennung offensichtlich nicht durchführbar wäre, ist es besser, der Schrift der Hebräer – die bisher keinen eigenen semitistischen Namen hatte – einen solchen zu geben: ‘paläohebräisch’.

000386

386

Birnbaum 01.04.11 18:17

[ 55 ]

ter der osteuropäischen Gebiete fortsetzt. Den Beschluß macht der sefardische Typus in der iberischen Halbinsel, von wo er nach der Vertreibung in die Länder um das Mittelmeer getragen wurde; er war auch die Schrift Südfrankreichs. Da er mit dem Maarawischen eng verwandt ist, stellt er eine geographische Verbindung zwischen den europäischen und den orientalischen Typen dar.2 Schließlich sind noch die Schriften der Karäer zu erwähnen, die natürlich recht kleine Gruppen repräsentieren: das Südkaräische Palästinas und Ägyptens, das Jewanokaräische des byzantinischen Raumes, das Nordkaräische Polens und das Parssokaräische Irans. Das geographische Klassifikationsprinzip ist mit einem andern zu kombinieren, das die Schrift unter dem Gesichtspunkt des Anwendungsgebietes oder Stils betrachtet. Torarollen und andere Ritualien sind in einer mehr oder weniger besonderen Quadratschrift geschrieben, deren Formen sich nur äußerst langsam ändern. Ein davon ein wenig abweichender Quadratstil erscheint in andern biblischen Handschriften sowie in liturgischen und talmudischen und entwickelt sich recht langsam. Briefe, Notizen, rechtliche Dokumente, Listen etc. gehören in den Bereich der Kursive. Diese ändert sich natürlich sehr schnell. Sie wurde aber auch für das Schreiben von Büchern verwendet und dann selbstverständlich mit Sorgfalt behandelt. Im Laufe der Zeit erwuchs daraus ein besonderer Stil: Maschait.3 Zu bemerken wäre, daß nicht alle Gruppen einen voll entwickelten Maschaitstil haben. Die frühen Drucke unterscheiden sich nach ihrer Herkunft – aschkenasisch, italkisch, sefardisch. Durch historischen Zufall hat sich aber nur das Sefardische bis heute als allgemeine Buchschrift, in Quadrat- und Maschaitstil,4 erhalten, während das Aschkenasische bloß in spezieller Verwendung (Maschait und Kursive) fortlebte oder wieder aufgenommen wurde (Quadratschrift) und das Italkische gänzlich verschwand. Bei 25 Typen und vier Stilen ergibt sich eine Ziffer von fast hundert Schriftgruppen. Abgesehen vom Paläohebräischen hat eine systematische Untersuchung des Gesamtgebietes erst vor kurzem mit einer allgemeinen Übersicht begonnen. Dem hebräischen Paläographen steht also ein ungeheuer reiches Forschungsgebiet offen. 2

3 4

Die Vielfältigkeit der nicht-aschkenasischen Schrifttypen und die großen Verschiedenheiten unter ihnen machen es unmöglich, sie alle – oder auch nur die orientalischen Gruppen – unter dem Namen sefardisch zu vereinigen. Die Bezeichnung ‘rabbinisch’ für diesen Stil ist sonderbar und nicht zu rechtfertigen. Die sogenannte Raschischrift. Der Name kommt natürlich daher, daß Maschait die allgemeine Buchschrift war und so jedermann als die Schrift der von Raschi verfaßten Kommentare bekannt und vertraut war.

000387

56.

Birnbaum 01.04.11 18:17

Die Schrift des Dukus Horant *

Das Manuskript [Universitätsbibliothek Cambridge, T.-S.10.K.22], zu dem unser Fragment gehört, trägt an zwei Stellen ein Datum. Am Ende von fol. 19v steht: g. kslu. q.m.g. lpr. ‘3. Kislew 143 nach der (kleinen) Z(ei)t(rechnung)’, bei der die Tausender als selbstverständlich weggelassen werden. Es handelt sich also um das jüdische Jahr 5143 ‘nach der großen Zeitrechnung’, das dem letzten Drittel des christlichen Jahres 1382 und den ersten zwei Dritteln von 1383 entspricht. (Der Tausender kann nur eine 5 sein, denn im Jahre 4143 = 382 / 383 gab es noch kein Jiddisch.) Der 3. Kislew fiel auf den 9. November 1382. Auf fol. 20v steht nach einer Liste der ‘Wochenabschnitte’ (die an den Sabbaten aus der Pentateuchrolle rezitiert werden) sliq q.m.g.l. (d. h.) ˘

˘

‘Finis, 143’ (d. h. 5143 = 1382/1383). Auf den ersten Blick liegt kein Grund vor, diesem Datum mit Mißtrauen zu begegnen. Das Datum auf fol. 19 ist offensichtlich das Schreiberkolophon, und solche werden nicht von einem späteren Kopisten gedankenlos nachgeschrieben. Bei Kolophondaten, die aus einer Vorlage übernommen werden, handelt es sich um das Datum des betreffenden Werkes, nicht der betreffenden Handschrift. Auch eine Fälschung ist ausgeschlossen. Ein anspruchsloses Bändchen wie unser Manuskript konnte niemanden dazu reizen, ein falsches Datum einzufügen. Wertvoll ist es erst jetzt, und eine moderne Fälschung kommt natürlich nicht in Frage. Die Wiederholung des Datums auf fol. 20 ist vielleicht mit dem Ordnungssinn des Schreibers zu erklären: als er nach dem ursprünglichen Kolophon noch einen Text (die Liste der ‘Wochenabschnitte’) gab, fühlte er vielleicht die Notwendigkeit, hier das formale Ende der Handschrift anzumerken. (Allerdings tat er es nicht mehr, als er darunter noch die Edelsteinliste setzte; vielleicht weil sie auf derselben Seite steht.) Die Wiederholung des Datums macht es auch klar, daß die Jahreszahl auf fol. 19, obwohl sie unmittelbar auf das Ende des Gedichtes folgt, doch ebensowenig dessen Abfassungstag bezeichnen kann, wie die Jahreszahl auf fol. 20 das Abfassungsdatum der Liste ist. * Zuerst veröffentlicht in: P. F. Ganz, F. Norman, W. Schwarz (Hrgg.): Dukus Horant. Tübingen 1964. Exkurs, S. 7–14. {Siehe auch Kapitel 35, Abschnitt 38, in diesem Band.}

000388

388

Birnbaum 01.04.11 18:17

[ 56 ]

Abb. 1: Codex T.-S. 10.K.22 der Universitätsbibliothek Cambridge, fol. 21 r (Anfang des Dukus Horant)

000389

Birnbaum 01.04.11 18:17

Die Schrift des Dukus Horant

389

Zusammenfassend können wir also sagen, daß 1382 bedenkenlos als das wahre Datum der Handschrift aufgefaßt werden darf. Bei ihrer Wichtigkeit für die Erforschung der altjiddischen Literatur mag es aber vielleicht doch angemessen sein, noch ein Übriges zu tun und das Alter des Manuskriptes unabhängig von den obigen Ausführungen festzulegen. Das kann auf Grund einer paläographischen Untersuchung geschehen. Schriftformen – wenn richtig analysiert – sind von untrüglicher Charakteristik und erlauben gewöhnlich eine recht genaue Datierung. In unserem Falle ist das Vergleichsmaterial – Handschriften, die ein Datum tragen – so reich, daß ein Zweifel an der Richtigkeit des Ergebnisses ausgeschlossen ist. Unsere erste Aufgabe ist es, die Grenzen abzustecken, innerhalb derer wir die Vergleichung der Formen vornehmen wollen. Die Schrift unseres Kodex [vgl. Abb. 1] – sie stammt in allen seinen Teilen nur von einer einzigen Hand – gehört zur aschkenasischen Gruppe der eine beträchtliche Zahl von Gruppen umfassenden hebräischen Schrift. Die Formen, klar und deutlich geschrieben, sind die der Kursive und weisen keinerlei Besonderheiten auf. Für dieses Gebiet haben wir datiertes Material seit 1237, aber selbstverständlich brauchen wir nicht alle Jahrhunderte vom dreizehnten bis zum zwanzigsten heranzuziehen. Eine provisorische Durchmusterung zeigt, daß die Schrift unseres Kodex im allgemeinen Eindruck den Schriften des dreizehnten bis sechzehnten Jahrhunderts entspricht.1 Für diesen Zeitraum haben wir nun eine in Einzelheiten gehende Vergleichung der Buchstabenformen durchzuführen, um ihre chronologische Stellung – und damit die des Manuskripts – ausfindig zu machen. Hier können wir einer außerpaläographischen Erwägung Raum gewähren, die den zu untersuchenden Zeitraum nicht unwesentlich verkürzt und uns damit unnötige Arbeit erspart. Die Handschrift trägt das Datum 1382. Sie kann also jedenfalls nicht vor diesem Datum geschrieben worden sein – denn niemand wird doch ein Zukunftsdatum unter sein Werk setzen. Folglich brauchen wir unsere Unternehmung erst mit dem Ausgang des vierzehnten Jahrhunderts zu beginnen. Was das andere Ende der Vergleichsperiode anlangt, so erfährt unsere provisorische Annahme des sechzehnten Jahrhunderts eine außerpaläographische Verstärkung durch die Tatsache, daß die letzten direkten Ausläufer dieser Art von Literatur noch bis ins siebzehnte Jahrhundert reichen. 1

Man vergleiche die Faksimiles 349–365 in meinem Buche The Hebrew Scripts. {Bd. 2} London 1954–1957.

000390

390

Birnbaum 01.04.11 18:17

[ 56 ]

Die Entwicklung der Schriftformen vom Ende des vierzehnten bis zum Ende des sechzehnten Jahrhunderts ist, wie deutlich aus der Tabelle [Abb. 2] zu ersehen ist, nicht unbeträchtlich. Sie ist hier durch drei Alphabete illustriert, die ich nach den folgenden, je um rund ein Jahrhundert voneinander getrennten Manuskripten gezeichnet habe:

Abb. 2: {Alphabete aus Handschriften des 14. bis 16. Jhs. zum Vergleich mit der Cambridger Hs. (C), gezeichnet von S. A. Birnbaum}

000391

Birnbaum 01.04.11 18:17

Die Schrift des Dukus Horant

K: W: H:

391

Das Kölner Archivdokument No. 4 von 1396, also ein Beispiel einer mit der unsern gleichzeitigen Handschrift. Ein Dokument von 1484 aus dem Wiener Neustädter Stadtarchiv. Der Hamburger Cod. hebr. 250 von 1574.

Auch diese Manuskripte sind, wie unser Büchlein, recht sorgfältig geschrieben, so daß je eine Form für die Darstellung eines jeden Buchstabens hinreicht. Zu den obigen Siglen kommt noch C

für unsere Cambridger Handschrift, deren Formen in der ersten Zeile der Tabelle erscheinen.

Wir gehen nunmehr zur Vergleichung der Formen im einzelnen über: Alef:

In C und K biegt der untere Strich der rechten Buchstabenhälfte nur ein wenig nach unten, während er in W und H bis an den Zeilenboden reicht. B: In C und K springt die Basis ein wenig nach rechts vor, im Gegensatz zu W und H. Sie ist so lang wie der Oberbalken, in K ist sie nicht viel länger. In W und H ist der Unterschied bedeutend größer. G: Der linke Strich ist in C und K ziemlich gerade, während er in W gewölbt ist und in H die Tendenz dazu hat. In C und K ist er kürzer als der Hauptstrich, in W und H so lang wie er oder etwas länger. D: In C und K beginnt der Abstrich in spitzem Winkel, in W und H sind die beiden Striche zu einer Kurve vereinigt. H: In C und K ist der Oberbalken gerade und schließt mit dem Abstrich einen Winkel ein, während in W und H die beiden Striche in einer Kurve ineinander übergehen. Z besteht in C und K aus zwei Teilen, Kopf und Abstrich. Dieser setzt in der Mitte des Kopfes an. In W bilden Kopf und Anfang des Abstrichs einen Winkel, der dann in H zu einer Kurve wird, wie es auch dem Rest des Abstriches ergeht. In C, K und W trifft der rechte Strich den linken etwas über dem T: ˙ Fußende, in H kommen die beiden in einer Spitze zusammen. K finale: Siehe D. M: Der mittlere Strich geht bei C und K von links unten nach rechts oben. Bei W setzt er in der Mitte des linken Striches an und ist horizontal. In H ist der Ansatzpunkt noch höher, ans Kopfende, hinaufgerückt, und der Strich läuft nun – umgekehrt wie bei C und K – von links oben nach rechts unten. M finale ist in C und K ein Viereck, in W und H ein Zweieck: zwei zu einer schrägen Achse ungefähr symmetrische Kurven treffen

000392

392

Birnbaum 01.04.11 18:17

[ 56 ]

einander links oben und rechts unten. Die obere Kurve entspricht der oberen und rechten Seite des früheren Vierecks. Die linke Hälfte der unteren Kurve dagegen entspricht nicht der linken Seite des Vierecks, sondern ist eine sekundäre Bildung, entstanden dadurch, daß die Feder nicht vom Papier abgehoben wurde, als sie das linke Ende der Basis erreicht hatte, sondern nach oben gezogen wurde, um dort den linken Abstrich des Vierecks anzufangen. Das Zweieck wurde nun als der eigentliche Buchstabe gefühlt und der linke Abstrich wurde eine Art Anhängsel. S: Die obere und rechte Seite entwickeln sich wie bei D. In C und K wird die linke Seite durch eine Kurve gebildet, in W und H durch eine Gerade. P: In C geht der linke Strich vom linken Ende des Oberteiles aus, in K ist er losgelöst, in W ist er zum Punkt geworden, steht aber außerdem etwas weiter rechts, d. h. im Innern des Buchstabens; in H ist er wieder ein Strich und verbindet den Oberbalken mit der Basis, beginnt aber nicht am linken Ende des Oberbalkens, sondern etwas weiter rechts, entsprechend der Stellung des Punktes in W. P finale: In C und H ist der linke Teil vom rechten losgelöst, im Gegensatz zu K und W. In C, K und W ist der linke Teil schräge, in H vertikal. S und S finale: In C und K ist der rechte Strich in der oberen Hälfte der ˙ ˙ Zeilenhöhe, in K steckt er ein wenig nach oben hinaus, in W und H ragt er fast eine ganze Zeilenhöhe aufwärts. Q: Was den Oberteil in C, K und W betrifft, siehe D. In H ist der Oberteil verkümmert. Er geht unmittelbar in den Abstrich über. Dieser hat am Ende eine große rechtsläufige Kurve entwickelt. Sˇ: In C und K ist der Mittelstrich gerade und freischwebend. In W und H hat er sich auf dem Weg über eine Kurve zu einer fast geschlossenen, ungefähr elliptischen Figur entwickelt, die an der Spitze des linken Striches haftet. Diese Vergleichung zeigt klar, daß C in allen Dingen – mit Ausnahme des P finale – vollständig mit K übereinstimmt. Wenn C kein Datum trüge, müßten wir es also in die Zeit des Kölner Schriftstücks setzen, ans Ende des vierzehnten Jahrhunderts. Mit andern Worten, das von unserer Handschrift selbst gegebene Datum, 1382, ist als echt erwiesen. Wie überall in jüdischen Manuskripten ist das erste Wort eines Abschnittes – hier: eines Gedichtes – mit großen Buchstaben geschrieben.

000393

Birnbaum 01.04.11 18:17

Die Schrift des Dukus Horant

393

Die Titel sind in keiner Weise hervorgehoben, sie sind meist sogar in derselben Zeile untergebracht wie das Ende des vorhergehenden Gedichtes. Verse sind durch einen Punkt voneinander getrennt, Strophen durch einen Doppelpunkt. Der Gebrauch des Punktes als Trenner kleinerer, und des Doppelpunkts als Trenner größerer Abschnitte ist in jüdischen Handschriften allgemein üblich. (Der Doppelpunkt ist häufig horizontal gelagert. Auch Dreipunkt kommt oft vor. Da der Bereich der jüdischen (‘hebräischen’) Manuskripte sowohl räumlich als auch zeitlich sehr ausgedehnt ist, kann von einem einheitlichen System natürlich keine Rede sein.) Wie allgemein in jüdischen Manuskripten läuft der Text am Ende von Abschnitten – hier: von Gedichten – keilförmig zusammen, indem jede Zeile etwas kürzer als die vorhergehende ist, und alle nach der Mittelachse des Schreibspiegels geordnet sind.

000394

Birnbaum 01.04.11 18:17

000395

57.

Birnbaum 01.04.11 18:17

Manuscripts in Old Yiddish*

The great libraries of Europe and America contain a considerable number of Old Yiddish manuscripts. These are valuable sources for research into the linguistics and phonology of Old Yiddish, literature in Old Yiddish, Ashkenazic biography, the history of Ashkenazic settlements, history of the halakha and paleography of the Ashkenazic type of the Hebrew script. Regrettably, only few have been published so that most of the material remains a hidden treasure. In many cases a manuscript is of value to research only if it is known when and where it was penned; it is also sometimes necessary to know as much as possible about the scribe himself. In some manuscripts there is a colophon, a formal ending which contains this information, or at least part of it. But this is rare: either the scribe provided no colophon, or the end of the manuscript is missing. Hence the editor or the cataloguer has to determine the date and provenance of the manuscript. He can draw conclusions from a number of facts: the kind of writing material, the kind of paper used, watermarks (if any), the writing instrument, the ink, internal evidence. Conclusions based on these facts will only in rare cases lead to sufficiently precise dates. That, however, can be achieved by the paleographical method. In order to make use of it, the editor or cataloguer of a manuscript written in the Hebrew alphabet must know not only the language very well but the script, too. Manuscripts written in the Square style will not be difficult to read, but cursive or mashait writing presents in general various degrees of difficulty – and these are the styles in which very many, probably the great majority, of the manuscripts are written. Mashait is an old term which I have revived, in order to replace the nonsensical term ‘Rabbinic writing’. I have, however, narrowed its meaning to make it more precise. It is a formal book hand developed from Cursive.1 The old sense was ‘Non-Square’, i. e., it included both cursive and mashait. However, in the case of the Ashkenazic script there is, to the practiced eye, at no period any great difficulty in reading cursive or mashait. * First published in: Dovid Katz (ed.), Origins of the Yiddish Language. Oxford 1987, pp. 7–11. 1 See my book The Hebrew Scripts, Part I: The Text. Leiden 1971, cols. 189 f., 311 f.

000396

396

Birnbaum 01.04.11 18:17

[ 57 ]

The Ashkenazic type is only one among many kinds of very varied form. Some differ so much from Ashkenazic as to be at first unrecognizable as being Hebrew scripts. The difference is not simply between Ashkenazic and Sephardic, as one might imagine on the basis of the present ignorant and ridiculous usage by which all non-Ashkenazim, including Baghdadim, Bukharim, Temanim, etc., etc., are included in a single entity termed Sephardim. This misconception has made it possible to encounter a Jew whose ancestors, far from having seen Spain, never left Asia, yet who thinks he is a Sephardi! As far as determining dates is concerned, Yiddish has a certain advantage when compared to Hebrew. Hebrew manuscripts are mostly written in Square or mashait style – rarely in cursive. Yiddish manuscripts, on the other hand, are mostly cursive, occasionally mashait, but never Square. Now, Square script changes very, very slowly over time, so that the determination of even a very approximate date is usually extremely difficult. But the pace of cursive development is fairly quick, so that one can often arrive at rather precise results. Let me give an illustration of how a manuscript was actually dated by the paleographical method. About forty years ago I ordered photostats of folios 129v – 130r of the Paris MS . He´b. 589. These pages contain a little epic poem (incipit Iüdis´er s´tam), part of which I intended to publish.2 The catalogue estimated the date as “XVI sie`cle”. One should realize that the authors of catalogues, or scholars who had worked on manuscripts in their research, had gained a certain amount of experience in matters paleographical, but had not had systematic paleographical training and were not paleographers. In fact, no catalogue of Hebrew manuscripts has had the assistance of a paleographer. No disrespect is intended, but the expertise of these cataloguers and scholars was in the contents of the manuscripts rather than the forms of their script. In other contexts it might well be satisfactory to date a manuscript within a century, but in this case more precision was needed, and an exact examination of the script was necessary. The natural method is to compare the form of each letter with its form as it appears in dated manuscripts. Those whose forms are closest to the forms in our manuscript provide us with its date. I had at that time a considerable collection of photostats and photos of specimens of dated manuscripts (and inscriptions), as well as alphabet 2

See now my Yiddish. A Survey and a Grammar. Toronto 1979, p. 157 f.

000397

Birnbaum 01.04.11 18:17

Manuscripts in Old Yiddish

397

tables, which I had myself drawn from manuscripts and photostats. Having sufficient experience, I did not have to examine every letter of the alphabet but was able to simplify the process by examining only certain characteristic letters (see fig. 1: a page of the manuscript, and fig. 2: my sketch of relevant letters and the documented lifespans of their forms). On looking through my dated material, I noted down for each form of letter close to the Paris manuscript, the first and last years in which it had appeared. I was able to use this method because the forms of the various letters do not change all at the same time, but each letter goes its own individual way. Thus each letter has its own particular span of time and there are considerable differences in the length of these periods. It is therefore reasonable to state that such a manuscript as ours was penned during the time span which all the letters have in common. This was the period between the latest year when a certain form came into existence, and the earliest year when a certain form had disappeared. In our case, examining the forms of final pe and ayin, I found these years to be 1551 and 1596. This is a span of less than two generations. It would, therefore, not be too risky to think that the middle of this period would not be far from the time our manuscript was penned. The middle year, 1574, we may for the sake of convenience call the paleographical date. Many years later I learned that there is a date in the manuscript; the cataloguer obviously did not notice it. And that year is 5339 i. e. 1579. The paleographical date and the real one were identical. This degree of precision is of course fortuitous, but to have arrived at the correct quarter of a century is not. Another illustration. Earlier, in 1934, I made a dating experiment. Sitting in the British Museum Library with a manuscript in Old Yiddish (Add. 19972) written in the mashait style, I did not know its date but did know that it was, in fact, dated. I should, therefore, be able to test the paleographical method. I examined each letter of the alphabet and arrived at a dating between near the end of the fourteenth century and the first quarter of the fifteenth. This resulted in a paleographical date of 1410. I then looked up the colophon: the year was 5153, i. e. 1393. The difference between the paleographical and the real date was, accordingly, only seventeen years – a perfectly satisfactory result. Later in the same year I made a further test and determined the paleographical date of British Library manuscript Add. 26919, written

000398

398

Birnbaum 01.04.11 18:17

[ 57 ]

in Ashkenazic cursive, as 1553. On then referring to the colophon, I found the date to be 5309, i. e. 1549.3 The paleographical method is both natural and reliable if the eyes of the researchers – and preferably also their hands – are familiar with form, for paleography is concerned only with form: the form of writing, letters or other symbols. Anything else connected with writing, such as writing materials, their form (scrolls, codices), type of books, binding, etc., is beyond the purview of paleography. It is most regrettable that today no one seems to specialize in research in Hebrew script during the three millennia of its existence. My own paleographical work has been only a beginning.

3

Details of both experiments are to be found in my book The Hebrew Scripts, Part I (see above, n. 1), cols. 343–346.

000399

Birnbaum 01.04.11 18:17

Manuscripts in Old Yiddish

Fig. 1. A page of Paris MS. He´b. 589

399

000400

400

Birnbaum 01.04.11 18:17

[ 57 ]

Fig. 2. Select Yiddish letter forms and their documented lifespans

000401

58.

Birnbaum 01.04.11 18:17

Design for an Ornamental Ashkenazic Printing Type

Design for a Hebrew ornamental printing type, based on the style of Ashkenazic manuscripts of the 15th century. Designed by S. A. Birnbaum in 1931. First published as no. 348 in his The Hebrew Scripts (vol. 2, London 1957)

000402

Birnbaum 01.04.11 18:17

000403

59.

Birnbaum 01.04.11 18:17

Das Datum des Codex Zimt Sand (Maase Bukh)*

{Introductory Note by the Author’s son Eleazar Birnbaum: The Maase Bukh, a collection of more than two hundred stories, was popular reading among Jews of both Western and Eastern Europe from the sixteenth through the nineteenth centuries. After circulating in manuscript, it was first printed in 1602. The book was the subject in 1956 of an M. A. thesis at Columbia University written by Ilse Zimt Sand. A bookseller acquired a manuscript of the Maase Bukh and asked Dr. Birnbaum to examine it and provide a date. Immediately afterwards Ms. Zimt Sand bought this same manuscript from the bookseller and in 1958 she published a 12-page brochure about it titled: “An extract from an unpublished manuscript of the ‘Mayse-Bukh’ of 1541 (?). Prepared for the Conference on Yiddish Studies, New York, April 7–10, 1958. By Ilse Z. Sand.” In it she wrote: “The date is ambiguous: it appears to read aw = 1541 C. E., although another interpretation would make it correspond to ñw = 1590 or unw = 1596” (p. 10). Presumably it was in order to correct this statement in her pamphlet that Dr. Birnbaum added the final two paragraphs of his article. To place this brief article in context and to give an idea of Dr. Birnbaum’s working methods, I have appended: (1) A photocopy of a maase [no. 48] from this manuscript (reproduced from Sand’s brochure, p. 7). (2) A copy of his working sketches for dating this manuscript, in which he had drawn various letters in this manuscript and compared them with the same letters in photostats of Ashkenazic manuscripts dating from the 15th through 17th centuries. From these he concluded, on purely paleographic grounds, that the Zimt Sand manuscript must date from the period 1585–1610. The authenticity of the colophon date 5356 A. M. = 1596 is thus confirmed. It is likely that Dr. Birnbaum had intended, for the purposes of this chapter (59), to explain in more detail how he arrived at his paleographical conclusions. See also fig. 2 of Chapter 57 of this volume, which shows selected Yiddish letter forms and their documented lifespans from 1421 through 1700, which the Author also used for the purposes of comparison.}

Der Herausgeber Dr. Franz J. Beranek erwähnte auf den Seiten 89 und 150 des ersten Bandes dieser Zeitschrift {Mitteilungen aus dem Arbeitskreis für Jiddistik, Januar 1960} den Codex Zimt Sand. Diese Handschrift wurde mir vor gerade drei Jahren von einem Buchhändler zur Begutachtung vorgelegt und daraufhin an Frau Ilse Zimt Sand in New York verkauft. Da das Datum der Handschrift inzwischen als zweideutig * Zuerst veröffentlicht in: Mitteilungen aus dem Arbeitskreis für Jiddistik 2.11 (1960), S. 9–10. [Hrsg. von Franz J. Beranek.]

000404

404

Birnbaum 01.04.11 18:17

[ 59 ]

(genauer gesagt, dreideutig) bezeichnet wurde, möchte ich jetzt den Wortlaut meines damaligen Gutachtens veröffentlichen. (Die Wörter, die dort in hebräischen Schriftzeichen gegeben wurden, sind hier in Lateinschrift umgesetzt.)1 “School of Oriental and African Studies University of London London W. C. 1 November 8, 1956 I have made a careful study of the script of the M"SH BWk {Óub hwym} ¯ manuscript. My examination of it was completely independent of the date in the colophon. The palaeographical result is the period 1585–1610. This confirms the genuineness of the MS . because the date in its colophon is 1596. The MS . was written by Samuel ben Zelikman Bak in Ispruk (i. e. Innsbruck), ‘under the auspices’ of Mordecai Shmaye Segal. It was intended as a present for the copyist’s aunt, LJM T {jmil}, the wife of Moses Segal. The copyist hailed from RWBRJ {ir¯ bur}. This name can be hardly read other than Rovere. The place in question was perhaps Rovere di Caorso near Piacenza. S. A. Birnbaum” Aber auch ohne paläographischen Beweis steht das Datum außer Zweifel. Die Schrift Samuels ist die eines Berufsschreibers, deutlich und schön, das Datum ist ganz klar: S NW {unw}, also 356 [= 5356], d. h. ¯ 1596. Das N hat eine besondere Zierform, wie sie auf anderen Seiten der Hs. in den mit großen Buchstaben geschriebenen Initialwörtern vorkommt. Die Lesung S {aw}, also 301, d. h. 1541, ist unmöglich. ¯ Das gilt auch für die Lesung Sn {ñw}, also 350, d. h. 1590. Der Name der Tante beginnt¯ mit einem klaren und unbezweifelbaren L {l} (nicht mit einem Z {c}).2

1 2

{Die hebräischen Zeichen wurden der besseren Verständlichkeit halber aus der Handschrift hinzugefügt. – E. T.} {Ilse Z. Sand hat 1965 (“A linguistic comparison of five versions of the Mayse-bukh”, in: The Field of Yiddish II , pp. 24– 48, n. 5) S. A. Birnbaums Richtigstellungen voll akzeptiert. Ein Faksimile des Kolophons ist heute zu sehen in: Chava Turniansky / Erika Timm, Yiddish in Italia. Manuscripts and Printed Books. Mailand 2003, p. 128. Die Handschrift befindet sich seit 1970 in der Jewish National and University Library of Jerusalem und hat dort die Signatur Heb. 80 5245. – E. T.}

000405

Birnbaum 01.04.11 18:17

Das Datum des Codex Zimt Sand (Maase Bukh)

Abb. 1. {Maase Nr. 48 aus dem Codex Zimt Sand.}

405

000406

406

Birnbaum 01.04.11 18:17

[ 59 ]

Abb. 2b. {Arbeitspapier von S. A. Birnbaum: Vergleich von Schriftzeichen aus verschiedenen Manuskripten zwecks Datierung des Codex Zimt Sand.}

000407

Birnbaum 01.04.11 18:17

Das Datum des Codex Zimt Sand (Maase Bukh)

Abb. 2a. {Arbeitspapier von S. A. Birnbaum: Vergleich von Schriftzeichen aus verschiedenen Manuskripten zwecks Datierung des Codex Zimt Sand.}

407

000408

Birnbaum 01.04.11 18:17

000409

60.

Birnbaum 01.04.11 18:17

Palaeographical Examination of the ‘Klein Manuscript’*

As will be seen from the attached photograph,1 this document in Hebrew is addressed to the Head of the Exile hlugh war (line 2), and bears what purports to be the signature of Hillel son of Judah the Prince: ina aiwnh hduhi unbru unnuda ñb llh (lines 13–15). In the preceding lines (12–13) it is dated as having been written in the month of Ziv of the year 261 of the aera destructionis ñbrxl dxau Õiwwu Õitam tnwb uiz wdxb, which corresponds to the year 329 or 331 of the Christian era. Although we do not have a great amount of material from that period, there is sufficient for a comparison with our manuscript, which I have set out in some detail below. I have designated the forms of the palaeographical material from the fourth (and occasionally the fifth) centuries C. E. with the letter “F” (for fourth / fifth),2 those of our document are designated with the letter “K” (the Klein Manuscript).3

* Previously unpublished. {This chapter was written in 1972 in the form of a letter in response to an enquiry from the owner of the manuscript, Mr. Max Y. Klein, Director of Le Petit Musee´ Ltd., and dealer in antiquities in Montreal. The drawings below are simplified quick sketches prepared by Dr. Birnbaum for the purposes of the letter to Mr. Klein. – E. B.} 1 See fig. 1 on the following page. 2 {For examples of Palestinian Square script from the fourth century C. E., see S. A. Birnbaum, The Hebrew Scripts, plates 90, 90*. – E. B., D. J. B.} 3 {See also p. 411, the author’s simplified working sketches of the Hebrew alphabet in the ‘Klein Manuscript’.}

000410

410

Birnbaum 01.04.11 18:17

[ 60 ]

Fig. 1. The ‘Klein Manuscript’

000411

Birnbaum 01.04.11 18:17

Palaeographical Examination of the ‘Klein Manuscript’

Fig. 2. The author’s simplified working sketches of the Hebrew alphabet in the ‘Klein Manuscript’

411

000412

Birnbaum 01.04.11 18:17

412

[ 60 ]

Comparison of F and K {Below is a comparison between the forms of letters appearing in F (palaeographical material of the fourth / fifth century C. E.) and K (the Klein Manuscript). The Hebrew letters in the left marginare, as explained above, simplified quick sketches by Dr. Birnbaum. – E. B., D. J. B.} F: K:

Aleph: In F, the right and left strokes end straight while in K they have developed a ‘flag’ and ‘foot’, respectively.

F: K:

Beth: In F the top bar has a tick on the left, being the remains of an original triangle. In K the top bar starts straight.

F: K:

Daleth: In F the downstroke generally starts higher up than the bar, in K it has moved leftwards, starting from the bar.

F: K:

Zayin: In F the letter is just a straight stroke or starts with a little piece sloping diagonally down leftwards – the first indication of what later developed as a ‘head’, sloping diagonally down towards the right, as shown in K.

K: F: K:

Teth: In F the right side consists of three strokes forming two angles. In K the two lower right-hand strokes now form a right angle, while the top one has become a head. In F the left stroke has no head, unlike K.

F: K: F/K:

Lamedh: In F the horizontal bar is short, in K it is long, in one specimen very long – an example of dilatation. A similar difference exists in regard to the downstroke.

F:

Mem: In F the letter has not yet reached the stage when its outline is almost rectangular, which is the case in K. The ‘inner’ [left-side] stroke and the bottom stroke are parallel in F, while in K the lower part of the ‘inner’ stroke approaches verticality. In F the left top stroke is comparatively long, in K short.

K:

F: K:

Final mem: In F the letter is more often than not open at the left-hand top corner, but never in K, where it is much narrower than in F.

F:

Nuˆn: In F the letter starts with a little bend from the right – really the beginning of a flat S-line which sometimes changes into a diagonal downstroke and an almost horizontal base.

000413

Birnbaum 01.04.11 18:17

Palaeographical Examination of the ‘Klein Manuscript’

413

K:

In K we have a head on the left, a vertical downstroke and a much smaller horizontal base.

F:

Final nuˆn: In F, the letter is a wavy line or a straight one. In K it is also straight but has a left-looking head, developed from the wavy type as in nuˆn.

K: F: K: F: K:

Ayin: In F the main stroke slants to the left, after beginning with a little vertical bend. The left stroke has a slight slant. In K the main stroke is broken into a vertical downstroke and a horizontal base, and both downstrokes have a left-looking head. Pe: In F the top is the point of an angle while in K we have a short or medium-sized horizontal bar. In F the right-hand stroke is slanting, in K it is in a vertical position. In F the bottom is long and slanting, in K it is almost horizontal and not longer than the top part. Final pe: Regarding the top part, see the description of pe.

F: K:

Sadhe: In K there are two heads, each looking outwards. In F the tops of the two strokes have no heads.

F: K:

Qoph: In F the downstroke usually starts from the top bar, in K it is always severed from it (as sometimes does occur in F).

F: K:

Shıˆn: In F the outline is completely, or almost, triangular; in K it is quadrangular: in K the right stroke is broken up into an almost vertical upper part and an almost horizontal lower part, which often approximates to a base. In F the inner stroke slants to the middle of the left stroke; in K to its end. In F the tops of the strokes, especially of the two right ones, sometimes start with a little bend; in K all three have a head.

F:

Taw: In F the left stroke very often starts higher than the top bar, passing and touching it at its left end. Sometimes the top piece is missing, occasionally the left stroke is slightly further in to the right, but if so there is a remnant of the loop from which this position developed. In K we have no indication of this.

F: F: K:

Finally I shall comment on some general features. In K the letters are rather narrow which is a sign of lateness. The forms of F are those of a

000414

Birnbaum 01.04.11 18:17

414

[ 60 ]

more square Square script. The difference in thickness between horizontals and verticals is considerable in K, in contrast with F. Such a difference in thickness does not occur earlier than the ninth or tenth century.

Conclusion In this comparison we find no features at all which could connect our document with the date given within it. The differences between the forms of K and F are so great that the time between the script of this document and that of fourth century documents must be very considerable. How great this may be is actually of no consequence for our purpose. However, for the sake of completeness I shall add a few words on the possible date of the script. The hand of our document is individually very bad. It is also very untypical – I have not seen any similar handwriting in libraries, in published facsimiles or in my photographic collection. This makes me think that the writer penned his letter purposely in as unusual way, for whatever reason. It would, therefore, be hopeless to try and date the manuscript from its script. However I think I am safe in saying that a medieval date is absolutely out of the question and a very recent one is almost certain. This is also shown by the fact that this document, a non-Biblical text, is written in the Square script style, whereas (pre-medieval) scribes would have written such texts in a careful cursive, which later developed into special book hands called Mashait in my scholarly terminology. The writer of your manuscript had probably never seen a real medieval manuscript. What could have been his purpose? His intention could hardly be other than to deceive. It would be almost impossible to explain the document as just a copy in a halachic context – in which case it would probably be part of a book. Such a book would most likely not be written in Square script, and not in the form of an original document, and of course it would not bear a stamp, as this one does. In Arabic manuscripts stamps were used at least since the Middle Ages (I emphasise: stamps, not seals). The language of the document is peculiar. A man of such eminence as Hillel II or his secretary could not have written in such a way. They could not be credited with having made mistakes of gender such as (line 8) ñkwi hbha uniihb hihi untuqt.

000415

Birnbaum 01.04.11 18:17

Palaeographical Examination of the ‘Klein Manuscript’

415

Finally: The document appears to be written with a quill. As far as I know, the quill came into use in the seventh century, and in the Orient the reed remained the writing implement for many centuries thereafter. That would exclude a fourth century date for the document. For many reasons, therefore, this document cannot possibly belong to the fourth century, as is claimed in its text. It is a forgery.

000416

Birnbaum 01.04.11 18:17

000417

61.

Birnbaum 01.04.11 18:17

A Scrap from a Hebrew Manuscript of a Kita¯b al-Mala¯him / {Kita¯b fı¯ ilm al-jafr}* ˙

No. 4100 of the Adler collection of Hebrew manuscripts – they now belong to the Library of the Jewish Theological Seminary of America in New York – contains miscellaneous pieces, one of which is described in Adler’s catalogue as a “letter dated 1043, brti tnidmb”.1 Such a document would be of great interest for Hebrew palaeographical research because we have no examples of Hebrew script from those regions of Arabia. I was, however, somewhat sceptical about the correctness of the description. So I ordered a photocopy and my suspicions were confirmed. In order to enable the reader to follow the details of my short note I shall start by reproducing a photocopy of the original (fig. 1, p. 419), together with two transliterations into printed type, one in Hebrew Square script (fig. 2), the other in Naskh (fig. 3).2 The Arabic script transliteration forms the basis of Dr. Khu¯lus¯ı’s article in which he deals ˙ fragment.3 with the literary and historical aspects of the Unless a transliteration makes use of special or new signs, it cannot be perfect. In the case of a document which is in the Arabic language but expressed in the Hebrew script, the Hebrew letter g (gimel) may be used to represent two quite different sounds which are normally indicated in the Arabic script by either  ( jı¯m) or A (ghayn). The Hebrew letter t (taw) may be used to express either (ta’) or (tha’). I have * Previously unpublished. {Parts of this chapter have been expanded from the Author’s brief incomplete draft by the Author’s son, Eleazar Birnbaum, in order to increase its clarity. Significant modifications are indicated generally by { }.} 1 {Catalogue of Hebrew Manuscripts in the Collection of Elkan Nathan Adler. Cambridge 1921, p. 152, no. 4100 (penultimate item). Now at the Jewish Theological Seminary, MS ENA 4100, 20c, verso. – E. B.} 2 {Fig. 1 was found among the Author’s papers, and only a rough draft of 3 was located. The present figs. 2 and 3 were prepared by me. – E. B.} 3 {Dr. Khulu ¯ s¯ı’s article was not found among the Author’s papers, nor could it be traced in˙ bibliographical sources. Perhaps Dr. Khulu¯s¯ı showed Dr. Birn˙ may never have baum a draft of an article in Arabic by Dr. Khulu¯s¯ı which ˙ been published. Dr. Safa¯ Khulu¯s¯ı, an Iraqi Arabic poet, was an instructor in Arabic at the School ˙of Oriental ˙and African Studies, University of London in the late 1940s and early 1950s, and thus a colleague of Dr. Birnbaum when he drafted his article. – E. B.}

000418

418

Birnbaum 01.04.11 18:17

[ 61 ]

therefore transliterated t into Arabic script as  or  without any diacritics, when it may represent either ta¯’ or tha¯’. It would be wrong to transpose the text {of this fragment which is in Hebrew script but in the Arabic language} into the orthography of standard literary Arabic by supplying in the Arabic script the diacritical marks which, though normal in Arabic, are absent in the Hebrew script original. To do so would be expressing a view as to the way the copyist of the Hebrew script pronounced the text, when in fact we do not know anything about his phonetics. Hence, we are not entitled to infer, for instance, from the form transliterated as 3d
that it was pronounced ~3d
{aydan as in Classical Arabic}, or that the writer’s t in the word tlt ˙ indicated that he pronounced it th rather than t. This fragment of the text is not part of a “letter”, as Adler’s catalogue erroneously describes it. In fact, it contains the final words of a literary work, with two colophons appended. Furthermore, the wording employed shows clearly that this work is not Jewish but Islamic. “Yathrib” {(line 13) the pre-Islamic name of the town that later became Medina}, is not the name of the place where this fragment was written, but where the book of which it is a remnant was supposedly composed. {In line 9 the work is referred to as ‘the book of the science of jafr, that is, the book of mala¯him’.} The work is attributed (lines ˙ was the son-in-law of the founder 13–14) to Alı¯ son of Abu¯ Ta¯lib {who of Islam, Muhammad}, and˙ the time of composition stated as being the ˙ twelfth year after the prophet’s death {which occurred in 632 C. E.}. The date of copying given as 1043 C. E. in the Catalogue was erroneously computed from the number 354 mentioned in the last two lines of the fragment, 354 being taken as an abbreviation for 1354, and assumed to be a year of the Seleucid Era. But in fact the number refers expressly to another era and that, a particular one, ‘354 years after the death of them both’ (wafa¯tihima¯ ). {As Muhammad died in 632 C. E. ˙ use of the dual {‘both’} and Alı¯ ibn Abı¯ Ta¯lib died in 661 C. E.}, the ˙ makes the date ambiguous. But it appears probable that the writer was actually reckoning from the death of the latter. Thus the date of the second colophon should probably be computed as 1015 C. E. Moreover, that is not the date of our present fragment but of the Arabic manuscript from which it was transposed by a Jewish copyist into the Hebrew script – either visually or from the dictation of a Muslim. For the second colophon cannot refer to a Jewish scribe. It is hardly imaginable that a Jew would have used such an individual, Islamic, method of dating. He would have given the Jewish date, or else, – and even such a thing is of rare occurrence, – used an ordinary

000419

Birnbaum 01.04.11 18:17

A Scrap from a Hebrew Manuscript of a Kita¯b al-Mala¯him ˙

419

Fig. 1. Cairo Genizah MS : Jewish Theological Seminary (JTS ), New York, MS . ENA 4100, 20c verso.

000420

420

Birnbaum 01.04.11 18:17

[ 61 ]

Fig. 2. Transliteration into Hebrew Square printed characters: JTS , ENA 4100, 20c verso.

000421

Birnbaum 01.04.11 18:17

A Scrap from a Hebrew Manuscript of a Kita¯b al-Mala¯him ˙

Fig. 3. Transliteration into Arabic Naskh printed characters: JTS , ENA 4100, 20c verso.

421

000422

Birnbaum 01.04.11 18:17

422

[ 61 ]

non-Jewish era. It is even possible that the original Jewish copyist omitted the final portion of the second colophon intentionally and that another Jewish writer decided to reinsert it, for it is in a different hand. Palaeographically, the fragment belongs to about the eleventh century. This manuscript is therefore probably not much younger than its prototype. Thus by pure coincidence the date in the Adler Catalogue is approximately correct. {Additional note by E. B. The Name and Subject of this Work The author of the colophon refers to the work by two designations in tandem (line 9): hada¯ kita¯b ilm al-jafr, hu¯ kita¯b al-mala¯him. These terms may be under˙ or formal book titles: jafr and stood equally as either descriptions of the contents mala¯him (plural of malhamah) are technical terms, originally referring to two ˙ ˙ different occult arts of divination, and foretelling the future,4 but for many centuries these words have been used widely as synonyms for divination, often interchangeably or as doublets. Some parts of this document are difficult to read and understand because it is damaged, incomplete or partly illegible. The authorship is not easy to determine. The colophon states (lines 9–15) that ‘this book of the science of jafr, that is the book of mala¯him derives from the ˙ gives information prophet Daniel (Da¯niya¯l al-Siddı¯q, ‘Daniel the Truthful’),’ and ˙ on ‘what has happened [. . .] and what will happen ultimately ( fı¯ a¯khir shay) to kings and states (al-mulu¯k wa-al-duwal); and he made that clear (bayyana) in the year 12 after the death of [Muhammad] the gentile [or illiterate] Arab prophet (al-nabı¯ al- Arabı¯ al-ummı¯ ) [cf.˙ Qur’a¯n, sura 3: 157–158] in the city of Yatrib / Yathrib. Alı¯ ibn Abı¯ Ta¯lib transcribed it (nasakhahu). It was written [or: he wrote ˙ hundred and fifty-four after the death of both of them.’ (it) ktb] in the year three Many manuscripts of works of jafr and malhama in Arabic (or Persian or ˙ Turkish) are to be found in libraries. Only a minority has been catalogued, and very few of those catalogued include in their descriptions the ‘explicit’ (final words of the body of the text). Of the explicits cited in the manuscript catalogues which I have consulted, not a single one corresponds to that of the Adler MS . Printed texts of jafr works which I have seen also do not contain a corresponding conclusion. As for works expressly described as malahim, the only printed one that I have seen, a Malhamat Da¯niya¯l from Najaf5 has˙ an ending quite different from that of the Adler ˙fragment. 4

5

For further information, see e. g. Encyclopaedia of Islam. New (2nd) edition, Leiden 1960 ff., art. Djafr and references there; T. Fahd, La divination arabe. Leiden 1966. Published by al-Matba ah al-Haydarı¯yah in Najaf, Iraq, no date, but about the middle of the 20th ˙century. It˙ has been reproduced in facsimile by A. Fodor (together with his study and English translation), “Malhamat Da¯niya¯l” in: The ˙ ed. G. Ka´ldy-Nagy. Muslim East. Studies in Honour of Julius Germanus,

000423

Birnbaum 01.04.11 18:17

A Scrap from a Hebrew Manuscript of a Kita¯b al-Mala¯him ˙

423

Curiosity about the future is a basic human characteristic, so divination texts meet a perceived need and are found in many different civilizations. As to the Adler fragment, one may well wonder why this particular work, which contains wording so explicitly Islamic in expression, was transcribed into Hebrew script, obviously for a Jewish reader not well acquainted with Arabic script. I suggest that the answer may be found in the well-attested fact that religious boundaries tend to be porous where the occult is concerned.6 In this particular case, some one thousand years ago, an Arabic-speaking Jew, whose traditional Jewish education gave him fluency in reading the Hebrew script but not the Arabic script, had an interest in divination and foretelling the future. He had the financial means to commission a Jewish copyist with a good hand to transpose this text into the Hebrew script. The later insertion, by a different hand, of the word al-nabı¯ (‘the prophet’), above line 13 may indicate a reluctance of the original Jewish scribe to include that particular word from his Vorlage (the manuscript from which he copied), as it would seem that, by doing so, this Jewish scribe would apparently be assenting to the terming of Muhammad as a genuine prophet. The later (different) Jewish hand ˙ the omitted al-nabı¯ above the line. most likely restored Yet another Jewish hand is responsible for inserting the word sanah (line 12, penultimate word), doubtless carelessly omitted by the original scribe because its appearance is similar to the preceding word minh(u), and for completing most of the remainder of the colophon (last word of line 13, all of line 14 except for the first two words, and all of line 15). This would probably indicate that the original Vorlage lacked this portion. It was most likely torn off, except for a fragment of the first two words of line 14. The missing words were evidently copied later from another manuscript of the same work.}

6

Budapest 1974, pp. 85–133 + facsimiles. Fodor shows clearly that this kind of divination existed also in ancient Assyrian and Babylonian texts. Jewish magical / mystical texts or formulae were also used in Islamic works, particularly in the form of Hebrew divine names or attributes which were rendered from Hebrew into Arabic script. Below I note two examples from Islamic works in Turkish: (I) Kita¯b fı¯ riya¯z˙at il-khayl, a 14th century anonymous work in the Mamluk-Kipchak Turkic language. See E. Birnbaum, “The mysterious incantation to cure sick horses: a formula in a Mamluk-Kipchak treatise”, in: the Journal of Turkish Studies 21/I (1997), pp. 96–101; and (II ) A ceb ül- ucca¯b by Manya¯sog˘lu (mid-16th century in Ottoman Turkish). See E. Birnbaum, “Turkish Manuscript victims of the 1992 war in Bosnia: Section 1”, in: Journal of Turkish Studies, 24/III (2000), pp. 47– 49.

000424

Birnbaum 01.04.11 18:17

000425

62.

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait *

Table of Contents Part I: The Material and its Sources A. B. C. D. E. F.

The Forms of the Term Mashait, with Primary References Phonological Examination Lexical Examination The Origin of the Term Conclusion Secondary References

Part II : Sources and Biographical Notes.

The term mashait, in the Hebrew sources, denotes a non-Square style of writing, i. e., it is an undifferentiated designation for both plain cursive and the elaborate book hand developed by professional scribes from cursive. This writer has revived the term, but applies it only to the book hand.

Part I: The Material and its Sources A. The Forms of the Term Mashait, with Primary References I shall first give the forms1 of the word which I have collected, arranging them in the chronological order of the authors in whose works they appear. The manuscripts and early printed books from which most of the present examples have been culled are not easily accessible, and each specimen is therefore given in its context in Part II . These quotations are referred to here by bold-face numbers in square brackets thus: [1], [2], etc., corresponding with the numbers that they * {This article appears to have been drafted originally in the early 1960s. The author made some revisions, probably in the late 1970s, with some further minor revisions in the later 1980s, but did not prepare a ready-for-publication final draft. It is therefore likely that he would have made further refinements to his manuscript. The final editing was completed in 2003 by his sons Eleazar Birnbaum and David J. Birnbaum to the best of their ability. – E. B., D. J. B.} 1 See the writer’s work The Hebrew Scripts, vol. 1, col. 189, and the Sephardic, Zarphatic and Ashkenazic specimens in vol. 2, nos. 255–257, 324–326, 331– 334, 366, 367, 369–381.

000426

426

Birnbaum 01.04.11 18:17

[ 62 ]

have in Part II of this study. (For the convenience of the reader each text is preceded by some biographical data concerning its author.) The first author on our list belongs to the twelfth century: Jacob b. Meı¨r (Rabbenu Tam) [1], the grandson of Rashi. The term does not occur in those of his writings which have come down to us nor in a quotation from his writings, but it is mentioned by two later authors in references to him. We have therefore no means of ascertaining whether he himself used it. Those authors are: Jeroham b. Meshullam [2] where ˙ b. Jacob hakKohen who we find it in the form iqwm [3], and Aaron writes qwm [4]. Thus if Rabbenu Tam did use the word we do not know in which form it was. There are two quotations from Jeroham, one [5] ˙ by Joseph b. Ephraim Caro [6] and one [7] by Moses Isserles [8]. The next author is almost contemporaneous with Rabbenu Tam but again we do not encounter the word in the source itself. It is mentioned in a reference to a responsum by Joel b. Isaac halLewi [9]. Joseph (Yoslin) b. Moses [10] mentions this responsum in his rwui jql from hearsay, so that there is no certainty whether Joel used it himself. Its form here is juwym [11]. As a matter of fact, another reference made by Isaac b. Moses [12] – and thus more than a century before Yoslin – to Joel’s decision does not contain the word at all (at least not in the printed edition) [13]. Not much later than Joel we have Simha b. Samuel [14]. Again we ˙ His pupil Isaac b. Moses have only indirect evidence and a quotation. [12] uses the word in a summary of an opinion of Simha’s. The form is jijwm [15]. The quotation occurs in tuinumõm tuhgh by˙ Meı¨r hakKohen [16]. The form here is jwym [17] and jiwm [18]. The first time we come across the term in an author’s own work and not in a reference to an earlier writer is in the thirteenth century, when it occurs in the yurz rua by the above-mentioned Isaac b. Moses [12]. There it is used a few lines before his reference to Simha b. Samuel [14] ˙ by Isserlin (see above). Here too the form is jijwm [19]. Isaac is quoted b. Petahia [20] in his ñwdh tmurt but in a completely different text and in ˙ jiwm [21]; the printed editions have the form ajiwm [22] and the form hjiwm [23]. Isserlin’s reference to Isaac b. Moses is re-quoted by Joseph Caro [6] in his úsui tib, with the spelling jõwm [24]. Later in the same passage, however, in a second reference to Isaac, he writes jõqwm [25]. Next in chronological order is Moses b. Nahman (Nahmanides) [26]. Simeon Duran b. Semah [27], in a reference to˙ him, uses˙ the term qwam ˙ passage ˙ [28], but the relevant in Moses b. Nahman’s works is without it ˙ Caro’s reference to the (in the printed edition) [29], nor does Joseph same passage [30] contain the term.

000427

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait

427

Isaac b. Moses’ [12] son, Hayyim [31], in his yurz rua inmis, the ˙ work, writes jijwm [32], the same as compendium of his father’s great his father. In one manuscript the letters are provided with vowel signs: jijÈwmÅ [33]. This indicates – on the basis of Yiddish phonology – the pronunciation mastet or masˇtet, perhaps with [e:]. Another manuscript ˙ corruptions of jijwm by the copyist, has jimwm and jjw˙im, obviously who does not seem to have known the word [34]. Abraham b. Moses of Sinsheim [35] in a work on the writing of ñilipt – edited and annotated by his pupil’s pupil, Samson b. Eliezer Barukh She-Amar [36] – has, according to the manuscripts, jiwym or jiwm [37]. In Mordecai b. Hillel’s [38] great work I found the term in only one of the manuscripts accessible to me (in the ñjq dyum passage), and there, unfortunately, the passage is corrupt [39]. Solomon b. Abraham Ibn Adret [40] in his Responsa writes, according to the manuscripts, qwÂaÅm [41], qwam [42] or qwm [43]; the printed editions have 'qwm [44] and aqwm [45]. Solomon Ibn Adret is quoted by Joseph b. Ephraim Caro [6] (the latter himself quoting Joseph Colon b. Solomon [46]) but with the spelling jõqwm [47]. Solomon Ibn Adret is quoted also directly by Joseph Caro, but with the spelling qawm {presumably a misprint transposing the w and the a}, and in a reference immediately following it, in the form jõqwm [48]. Alexander Ziislin hakKohen [49] in his hduga has jiwym [50]. Jeroham b. Meshullam [2] and Aaron b. Jacob hakKohen [4] use the word in˙ reference to Jacob b. Meı¨r [1]. Aaron is referred to by Joseph Caro as having, in his Õõx tuxra, interpreted bdbdm as qwm [51]. I have not found this passage in the edition. The code ub-lk, which is considered by some scholars to be a work of Aaron b. Jacob, is referred to by Joseph Caro in the form jõqwm [52], but again I have not been able to find such a passage in the printed edition – and nothing is said there about the style of the script [53]. In Samson b. Eliezer’s [36] rmaw Óurb we have jiwym in one manuscript [54], and in another manuscript of the same text: jõcam [55]. The rmaw Óurb printed editions give jiwmg [56] and jiwim [57]. The next author is Samuel b. Aaron of Schlettstadt / Se´lestat [58], in his ikdrm ruciq [59]. The manuscripts give jõcm [60], ajõcm [61], jiwm [62] and unõwm [63]. In this last case the copyist had misread the j as un, and another hand had written jiwym as a correction above it. Simeon Duran b. Semah [27] has qwam both in quoting the enquiries addressed to him and˙ his ˙answers, as well as when recapitulating [64]. Isserlin b. Petahia [20] is only an indirect source (see text above, p. 426). ˙

000428

428

Birnbaum 01.04.11 18:17

[ 62 ]

In a book list on a flyleaf of a manuscript [65] our word occurs five times in the form jõwm and once as jigwm [66]. Joseph Colon b. Solomon [46] uses the form jõqwm (see above [47]). Joseph (Yoslin) b. Moses [10] in his autograph collectanea rwui jql employs jiwym [67]. The editor of the first edition of Solomon b. Abraham Ibn Adret’s [40] Responsa gives, in the index, the form jqwym whilst the text itself has 'qwm (with the abbreviation sign). Whether Meı¨r b. Isaac [68] who quotes it from Simha b. Samuel [14] ˙ used the term himself we do not know. Joseph b. Ephraim Caro [6] has, in his úsui tib, a passage in which he discusses non-Square writing: the terms mentioned are: (A) jõwm (B) jõqwm, (C) iqwm, (D) qalyum, (E) lacniburp, (F) arjil aidim, (G) hbitk h³lt [69]. Elia Levita (Elijah b. Asher halLewi) [70] deals with the name of the script in a reply to his contemporary Sebastian Muenster (an important German Christian Hebraist) [109], who had asked him why the cursive was called dwm by the Jews. Elijah replied that jiwm was w mÅ (masˇket )2 in use the Ashkenazic development of the word jiqÈÏ amongst the Zarphatic, Sephardic, Italkian and Yevanic Jews. He went on to explain that it was derived from the Arabic, where its meaning was ‘thin, meagre’ [71]. Later, in his ibwt, he withdrew the Arabic derivation [72]; he now vowelled the Ashkenazic form, giving the wÈ ymÃ.3 pronunciation jiÏ Joshua Boaz b. Simeon Barukh [73] in his compendium of the ikdrm writes jiqwm [74]. Asher Antshel b. Joseph [75], in speaking of the Ashkenazic type face in which the Yiddish columns of his uisndruqnuql hnwmh tbkrm are printed, uses jiwymg [76]. Moses Isserles [8] in Responsum no. 34 writes 'jiwm [77]. In his ikrd hwm we find hjiwm [78] but in his hpm we find ajiwm [79]. In Õiwn tucm, a Yiddish manuscript, we have jiwym [80]. 2

3

As this word is vowelled, we can tell from our knowledge of his Yiddish phonology how he pronounced it. Since he speaks of the Ashkenazim swallowing the q we may infer that he pronounced jiwm as masˇet. Muenster’s spelling indicates the same pronunciation – the final d was, of course, unvoiced in his German phonology. The stress must be on the second syllable because it could not otherwise contain a full and even long vowel. That the w was a /ʃ/ follows from his non-Ashkenazic form, where the letter is dotted in both specimens. – Levita was a source for Dukes, Grünbaum and Ben-Yehuda (see section F below).

000429

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait

429

Joel Serkes b. Samuel halLewi Yafe [81] in his wdx tib writes jiwym (three times) [82]. David b. Samuel halLewi [83] in bhz iruj has jiwm [84]. Abraham Abbale Gombiner b. Hayyim halLewi [85] refers to Joel ˙ is thus quoted [88] by Isaiah b. Serkes but writes ajiwm [86, 87], and Abraham [89]. The latter uses the same form in his bjih rab [90]. In a witness deposition made to the Jewish court of Piko´w (Poland) a certain thing is likened to ‘the letter z of our ajiwm script’ [91]. Moses Sofer / Schreiber [92] in his rpus Õtx employs ajiwm [93]. Israel b. Gedaliah Lipshitz [94] in his larwi trapt has the same form [95]. I shall now make an alphabetical list of all these various forms:4 ,jqwym ,jiwÈymà ,jiwym ,jwym ,juwym ,jiwim ,jjwim ,qwÂamÅ ,qwam ,jõcam ,hjiwm ,ajiwm ,'jiwm ,jiwm ,jijÈwmÅ ,jijwm ,jigwm ,qawm ,ajõcm ,jõcm .jõqwm ,jiqÈÏ w mÅ ,jiqwm ,iqwm ,aqwm ,'qwm ,qwm ,jimwm ,unõwm ,jõwm

That there are so many is perhaps not surprising, seeing that they have been collected from a period covering many centuries and from a number of different regions, which will certainly not make our task of finding the etymology of the word any easier. How are we to establish which of the above forms is the earliest or the original one? Could we perhaps discover a criterion for placing them in chronological order? The chronology of the authors, according to which we have listed them above, does not aid us in this respect because we cannot simply assume that the form in which the term occurs in any of the authors is actually that which he himself employed. On the contrary, it is quite clear that the immediate responsibility for this very often rested with a later copyist or the printer. Could we then, perhaps, use the chronology of the manuscripts containing the respective forms? No, unfortunately not, because the manuscripts in question were written within too short a space of time for significant linguistic changes to have taken place. Having no help from any chronological evidence we will now see what we can discover by examining our material phonologically and lexically, and by trying to work back to find what could be the archetype from which all our forms developed.

4

{Note: The use in two sources of the forms jiwmg and jiwymg is merely the Germanic prefix ‘ge-’ to the forms jiwm and jiwym. – E. B.}

000430

Birnbaum 01.04.11 18:17

430

[ 62 ]

B. Phonological Examination Upon classification we find that our material falls into the following nine groups. The forms shown in parentheses are clearly errors made by copyists who were unfamiliar with the word: .hjiwm ,ajiwm ,'jiwm ,jiwm (1) .jiwim ,jwym ,(juwym) ,jiwÈymà ,jiwym (2) .(unõwm) ,jõwm (3) .(jimwm ,jjwim) ,jijÈwmÅ ,jijwm (4) .ajõcm ,jõcam ,jõcm (5) .(qawm) ,qwÂamÅ ,qwam (6) .iqwm ,aqwm ,'qwm ,qwm (7) .jõqwm ,jiqÈw Ï mÅ ,jiqwm ,jqwym (8) .jigwm (9)

The vowelled specimens show us how to pronounce most of the unvowelled ones. But even without these, that should present no difficulties. On the basis of what we know of the spelling systems of the Jewish languages we should be able to determine what our forms were like: õ stands for Zarphatic (‘Judaeo-French’) /e/; i stands for /i/ or /e/; absence of vowel sign indicates /a/ or absence of vowel; y stands for Ashkenazic /ε /; c in certain languages stands for /s /;5 the letter w without a diacritic stands for either /s / or /ʃ/.6 We can now transcribe the above list:7

7

8

e

6

e

5

mas¯et, mas¯et(–?), mas¯et( ?), mas¯et( ?). mε s¯et, mε s¯et, [mε s¯et],8 mε s¯ [e]t, mes¯et. mas¯et, [mas¯et]. mas¯tet, mas¯tet, [mas¯tet, mas¯tet]. maset, maset, maset( ?). mas¯k, mas¯k, [mas¯k]. mas¯k, mas¯k-, mas¯k(a?), mas¯ke / mas¯ki. mε s¯k[e]t, mas¯ket, masˇket, mas¯ket. mas¯get. e

(1) (2) (3) (4) (5) (6) (7) (8) (9)

{e. g. in North African Arvic (‘Judaeo-Arabic’) the initial consonant of the word sultan is written in Hebrew script sometimes with a c, and sometimes with a w, even for the same word on the same page of a book. – E. B.} This double possibility is here indicated by the sign s¯. {Words in square brackets represent Solomon A. Birnbaum’s restoration to the original form (Vorlage) which the copyists must have had in front of them, but which they mistranscribed because they were unfamiliar with this technical term. – E. B.} u clearly is an error for i.

000431

Birnbaum 01.04.11 18:17

431

The Etymology of the Term Mashait

The differences in the spelling of the originals have now disappeared and the list has shrunk to the following items: mas¯et, mas¯et( ?), mas¯et-, mε s¯et, mas¯get, mas¯k, mas¯k(a?), mas¯ke / mas¯ki, mas¯ket, masˇket, mε s¯ket, mas¯tet. These forms clearly fall into two main divisions according to whether the third consonant of the word is a t or k / g: (1) mas¯et, mas¯et-, mas¯et( ?), mas¯tet. (2) mas¯ke / mas¯ki, mas¯ket, masˇket,9 mε s¯ket. Thus our task is narrowed down to establishing which of the two is the original. To derive t and k from some other phoneme, as their common ancestor, seems phonetically hardly feasible. We have therefore to ask ourselves whether the t-type is derived from the k-type or vice versa. Looking through our list we notice that the t / k difference goes hand in hand with another: s¯ and t are separated by vowels, while s¯ and k are not. There is one exception where s¯ and t are not separated: mas¯tet. This exception is significant – it constitutes a bridge between the two groups. The link between the k-type and mas¯tet is the absence of a vowel after the s¯, and the link with the t-type is the fact that it has the same ending: -et. Mas¯tet is thus a parallel to mas¯ket, the form which combines the k-type with the t-type. We now seem to have advanced considerably in our quest: Those two appear to indicate the stage of development at which the parting of the ways took place. Which then was earlier, the t or the k? Was it mas¯tet > mas¯ket or vice versa? There is no obvious, phonetical solution. Changes between k and t take place in either direction. If we assume the development s¯t > s¯k, i. e., mas¯tet > mas¯ket, then mas¯k would have come into being through the loss of -et. An apocope of a vowel and a consonant would be very difficult to explain. A vowel disappears from an unstressed syllable. But the e here is in a stressed syllable. This follows from the fact that we have here, in a number of cases, a ±È (sere), in others õ. In Ashkenazic phonology a full vowel cannot survive˙ as such in an unstressed position. Hence we had better assume the opposite development – s¯k > s¯t, i. e., mas¯ket > mas¯tet. There would then be no such difficulty: we could derive mas¯ket from an original mas¯k plus an accretion -et. So it seems probable that the original form was mas¯k. We shall turn now to the forms mas¯et etc. and mas¯get, which we left out of account while pursuing our main line of argument. Having decided that the development went from mas¯ket to mas¯tet, we would e

e

9

{Includes mas¯get: see footnote 10 below. – E. B.}

000432

Birnbaum 01.04.11 18:17

432

[ 62 ]

have no difficulty in explaining mas¯et as a later stage, resulting from dissimilation. Mas¯get could have developed from mas¯ket and might, in turn, be the source of mas¯et, as follows: mas¯ket > mas¯get > mas¯yet > mas¯et. However, mas¯get occurs only once in our material [66], and then at a late period. Moreover, on the very same page the writer uses the form mas¯et five times. From that it seems improbable that mas¯get belonged to the main stream of the development.10 The real number of different forms is thus reduced to only four: mas¯k w mÅ, jiqwm, jõqwm, is represented by qwÂamÅ and qwam; mas¯ket by jiqÈÏ jigwm; mas¯tet by jijÈwmÅ, jijwm; maset / mas¯et / mε set by jõcam, jõcm, jõwm, jiwm, jiwym, jiwÈymÃ. The other forms: qawm, aqwm*, qwm*, 'qwm*,11 juwym, jwym, ajiwm, hjiwm, 'jiwm,12 may be attributed to unfamiliarity with this technical term, especially in later centuries when the term was no longer in current circulation. We may summarise thus: mas¯et < mas¯tet < mas¯ket < mas¯k.

C. Lexical Examination Turning now to the lexical aspect and for the moment ignoring the results of our phonological examination, we shall take a look at those linguistic areas which might be the source of the name mashait. All the above-mentioned specimens are Ashkenazic or Sephardic origin. But Elijah Levita tells us that the Italian, French, Greek and Arabic Jews were also using this term [72]. From this we could conclude that it must go back to very early times, and that Aramaic or perhaps even Hebrew might be its source, although it could have originated in either of these at a later period also. However, the Hebrew dictionary does not afford us any help in this respect. Those few words in it which bear some resemblance to any form of our term, are sew mÅ ‘beverage’; qÏ wÄ mÅ ‘jostling, rushing’; mantically impossible: [96] hqÃÏ 10

11 12

To explain this form here might not be difficult. It could hardly be a slip of the pen since the writer used it side by side with mas¯et, and was not copying a literary work but was writing on a flyleaf, making a list of books. It was, presumably, his intention to differentiate. I would suggest that he used the form mas¯et for his own, Ashkenazic type of mashait, but mas¯get (his South German pronunciation of mas¯ket ) to designate a non-Ashkenazic type of this style of script. That would go well with the fact that this list contains five books in mas¯et and one only in mas¯get. {Perhaps the three forms which I have asterisked should have been included earlier in this sentence, as representing forms of mas¯k. – E. B.}. {And the others included in the list at the beginning of “B. Phonological Examination” (p. 430). – E. B., D. J. B.}

000433

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait

433

qÏ wà mà ‘household, administration’; yqÄÏ w mÇ ‘sediment, clear water’; hTÃÏ w mÇ

‘feast’. After the Hebrew period in Jewish linguistic history, there are the Aramaic, Persian, Greek, Latin, Romance and German sources of Jewish languages to be considered. In none of them does there appear to be a stem resembling any of the forms of our word. That leaves us with Arabic. Here there are a number of words with a similar sound but of these none have a suitable meaning: [97] masata ˙ ‘to press, squeeze, wring out, thrash’; mas¯it ‘a purgative plant, salt˙ water’; masıˆt ‘mud, slush, dirty water’; masˇˆıta the same, and ‘muddy ˙ scent’; masˇata ‘to comb’; ma ˙ˆ sˇit ‘hairdresser’; maˆsˇita river, a certain ˙ ˇt ‘comb’; masˇita˙ ‘to be chapped, hurt, ˙ ‘lady’s maid’; misˇt / musˇt / mas ˙ ˙ ˙ ˙ torn’; musˇitt ‘one who exceeds, goes too far, is unjust to, overcharges’; musˇˆıt ‘who˙˙/ what singes, scorches, burns slightly’. Nor is the root sqt a ˙ source: masqit ‘one who causes to fall, lets fall, drops’ and ˙ likely ˙ masqat / masqit ‘a place where something falls’ can hardly describe the ˙ result ˙of writing. Last in alphabetical order comes the root msˇq. Here we have masˇaqa ‘to comb (hair), card (wool), stretch (linen), tear, shred, lash, strike with a lash, strap, beat with a strap’; imtasˇaqa ‘to snatch, draw (sword)’; masˇq ‘ruddle, red ochre’; misˇqa ‘refuse of hemp, oakum’. These do not suggest a likely etymology. We also find misˇq, masˇˆıq and mamsˇuˆq ‘slender, slim’. Could there be a connection between this meaning and the Hebrew expression hqd hbitk ‘thin script’ = ‘cursive/mashait’? If our word were a translation into Arabic of Hebrew hqd, its form would be either misˇq or masˇˆıqa (feminine, because of hbitk). But it is neither. Moreover, no adjectival meaning from the sphere of writing is listed for them in the Arabic dictionaries. However, the root qwm (msˇq) actually is indeed used in connection with writing. The classical Arabic dictionaries quote two idioms containing it: (a) The verb masˇaqa with a noun indicates that the action in question is performed quickly. ‘Writing’ is one of the examples given. A semantic development ‘to write quickly’ – ‘to write cursively’ would be natural. (b) Masˇaqa with the object ‘writing, script’ means ‘to write elongated letters’. This presumably refers to the elongated letters of Arabic calligraphy. That, however, is just the opposite of cursiveness. The dictionaries do not speak of cursiveness in either of these idioms nor do they list the next stage of a semantic development which would have resulted from writing cursively, i. e., there is no noun denoting ‘cursive script’. Only such a noun could have become the Jewish word.

000434

434

Birnbaum 01.04.11 18:17

[ 62 ]

Finally, let us turn to the modern Arabic dictionaries, in spite of the considerable age gap. They might contain meanings which did not find their way into the classical dictionaries. Here we come upon two additional meanings from the sphere of writing [98]. One is that of the noun: masˇq ‘the model from which the pupil copies out his writing exercise’. The other is a verbal meaning: masˇaqa ‘to draw letters or ruling in order to exercise one’s hand or learn writing’.13 That the meanings ‘writing practice’ and ‘copybook’ are old – although they are not mentioned in the classical dictionaries – seems proved by the fact that they exist both in Persian [99] and Turkish [100]. Still, however closely these meanings may be linked with writing, they do not seem to be semantically connected with the Jewish word for ‘cursive’.14 But perhaps the gap can be bridged in this way: Writing exercises belong to school life. The writing that was, and is, taught in school is not the Square script. Square belonged – and belongs – to the sphere of the specialist, the scribe employed in the making of scrolls and certain codices. Children learnt – and learn – plain cursive. Here, as in English, the action of writing gives its name to that which results from it: writing (action) > writing (script). Thus ‘script’ and ‘cursive’ were identical ideas. 13

14

This meaning was unknown to some educated Arabs whom I consulted. They only knew the noun meaning ‘copybook’. However, the verbal meaning must have existed, otherwise the noun would not have come into being. Perhaps the development was something on these lines: Masˇaqa is used for describing actions where the arm makes a straight movement: ‘combing hair, carding wool, stretching linen, striking with a lash, beating with a strap, drawing a sword’. Thus it would be natural to use it for ruling – cf. English ‘drawing a line’. Ruling plays a special role in learning to write, and here we might have the transition to ‘practising writing’ – which practice is preceded by ruling. ( This connection seems to appear in qwim ‘the ruling material’, i. e. ‘ruddle’, because the ruling was done with ruddle.) Finally, ‘doing one’s exercise’ was identical with ‘copying from a model’, and so the noun masˇq ‘model for writing practice’ emerged. The intimate relationship masˇaqa has assumed with writing is paralleled by a similar development in English: the specific meaning of ‘drawing’ as ‘the making of pictures by outlines’. We might mention here the form qwam. It could be read as the Arabic participle maˆsˇiq. But ‘one who writes quickly’ or ‘one who draws letters’ or ‘is ruling in order to exercise his hand’ or ‘is learning’ cannot lead to a meaning describing the result of the action indicated – ‘the quick script’ or ‘the long-lettered script’ or ‘the practice script’. Even if masˇaqa had the meaning ‘to write in cursive’, our qwam could not be the participle of this verb because it would mean ‘one who writes cursive’. Our qwam is simply a spelling with a vowel letter (a, standing for a), and the shewa of the specimen qwÂamÅ confirms this.

000435

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait

435

This derivation seems attractive. However, our word does not by any means always appear in the form mas¯q – or even in a similar one. On the contrary, in most of our material the third consonant is a t. How can one account for that?

D. The Origin of the Term In looking through our sources we find that the forms with j (t ) are used by the Ashkenazim, those with q (k) by the Sephardim. If our Arabic etymology is correct, the use of the k-form may be easily explained: the Sephardim could have acquired the word during the period when they were under Muslim rule. To the Ashkenazim it would then have come via the Sephardim. How would that have happened? We do not know of any Sephardic movement northwards to the Rhine in times when some of the Sephardim spoke Arabic or were the children of Arabic-speakers – or, for that matter, at any time. Should we then conclude that the Ashkenazim adopted the word as a technical term from written Sephardic sources? On the face of it this seems improbable. The Ashkenazic and Sephardic forms are rather too different. Such phonetic divergence is more likely to occur when a word is adopted orally and undergoes the normal linguistic process of changes. Let us ask ourselves in what essential way the Ashkenazic and Sephardic forms differ. If we leave out of account the doubtful final a, h and i, the Ashkenazic form is disyllabic, ending in ji- (-et ), while the Sephardic one is monosyllabic, ending in qw (s¯k). If the derivation from Arabic is correct, then the second syllable of the Ashkenazic form must be an accretion. Where would it have come from? Not from the German, of course. In view of the fact that the Ashkenazim are of Zarphatic (‘Judaeo-French’) origin, and that they remained for a considerable time in close contact with their parent group, we might recognise in this second syllable the French suffix -et.15 The diminutive meaning of this ending would seem to be quite compatible with the sense of the word: cursive is a small script in comparison with Square, which is generally written very large indeed in scrolls and Biblical codices. 15

The spelling with j does not speak against a non-Semitic derivation. In Zarphatic, as in other non-Semitic Jewish languages, j is used for the unvoiced dental stop. (Inversely, the q of qwm is no proof of the Semitic origin of the word, since q is the Jewish spelling for the unvoiced palatal stop in non-Semitic languages.)

436

[ 62 ]

A Zarphatic origin for certain Ashkenazic words would not be in any way surprising. A few words from that source actually have survived in Yiddish to this very day. There are two additional points that seem to corroborate a Zarphatic etymology for the term mashait. One of our authors, Abraham b. Moses of Sinsheim [35] (or possibly his editor, Samson b. Eliezer Barukh She-Amar [36]) accompanies the word jiwym with the formula z`ylb, i. e., ‘in Romance’.16 This indicates that he was still aware that the word belonged to the Zarphatic heritage of the Ashkenazim. The second corroboration appears to lie in some of the spellings: the vowel of the second syllable is written with õ: jõwm, jõcm, jõcam, jõqwm. Now this spelling for /e/ is specifically Zarphatic.17 Another spelling, while being Zarphatic in origin, does not, however, necessarily point to that source: the final a of the form ajiwm etc. There are two reasons: First, this a is not restricted to that word – in early Yiddish it was the regular sign employed for final / /. Second, the form ajiwm appears only late in our sources – our earliest example in a manuscript showing a final a is from the beginning of the fifteenth century (cf. [61]), and all the others belong to the period after the introduction of Hebrew printing. It had long before then become purely a spelling symbol without any phonic reality. Here is a possible explanation as to the origin of the forms hjiwm / ajiwm. The little stroke at the end of the word 'jwm denoted that the word was not a Hebrew one (corresponding to the use nowadays of italicisation for foreign words). Such a stroke, however, could also be used as a sign that the end of the word had been abbreviated. A later copyist, not knowing the word jiwm, mistook the stroke for the mark of e

16

17

In a printed edition the z`ylb has been replaced by a`lb (i. e. znkwa ñuwlb ‘in the Ashkenazic tongue’, i. e., Yiddish). This, if not simply a slip, might reveal that the writer regarded the word mashait as Yiddish, which would, in fact, have been correct, since it formed part of his traditional vocabulary; it would not mean that the word was of German origin. In other Jewries õ designates non-syllabic /i/ as in post-Biblical Hebrew, cf. my articles `Õ³gh lkk ñiùaùÄ õÈùc`, in: bqyi tib 8, no. 71 (Lodz 1931); ñup` `hduhi rbdm ryd ñia lõh ryd zib ÕzirymwjõÅd, in: Óarpw ywidõ 13 (1953), pp. 109–120 (reprinted in: Óub-yp Æ ai .aÅ aÅdui [Yuda A. Yofe-bukh], New York 1958, pp. 195–206. We would, accordingly, have to read mas¯yat, mas¯kyat or mas¯ayt, mas¯kayt. This, however, is impossible because the /e/ of the second syllable is certain in a number of specimens, it being expressed there by a vowel sign.

000437

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait

437

an abbreviation.18 A subsequent copyist, wishing to write out the word in full, replaced the mark by a, which is frequently the final letter of zyl and Yiddish words; other copyists used a h for the purpose,19 presumably influenced by the fact that, in Hebrew, h was the most common victim of abbreviation. To sum up: It is very probable that the Ashkenazic form was derived from Zarphatic but we have still to discover the course it would have followed between the Zarphatic and the earlier Arabic stages. There was no immigration wave of Sephardim into northern France but the connection might have come about through the Jews of Provence. From 1113 to 1245 this country belonged to the House of Count Berengar of Barcelona, which might explain why this Jewry formed culturally a part of the Jewry living immediately south of the Pyrenees. A clear indication of this is that their script was Sephardic, not Zarphatic. In 1246 the lands of Provence passed to the House of Anjou. Contact between Provenc¸al and Zarphatic Jews might have carried Provenc¸al influences northwards, including those of a linguistic nature. Here, presumably, the word acquired the ending -et, although it could already have done so in Provence.20 Our results might be summed up diagrammatically: Arabic masˇq ‘writing exercises’ > Arabic Jewish masˇq > ‘writing lesson’ > ‘writing’ > ‘script’ > ‘cursive’ > Sephardic masˇk ‘cursive / mashait’ > Provenc¸al Jewish masˇk > Zarphatic masˇk > masˇket > Ashkenazic masˇket > masˇtet > masˇet.21

18 19 20 21

In this article, our earliest example of this abbreviation mark is from the year 1640 (cf. [77]). Our earliest example here is from the year 1546 (cf. [23]). In the Provenc¸al sources no nouns with -et seem to be preserved but there are adjectives with -et and nouns with -at and -it. The only difficulty regarding this derivation is [70] Elijah Levita’s statement that masˇket is common to all non-Ashkenazim (cf. notes [71, 72]). How would Jews speaking Italkian (‘Judaeo-Italian’), Yevanic (‘Judaeo-Greek’) and Arvic (‘Judaeo-Arabic’) come to use a Zarphatic form? The answer might be that Levita was writing from hearsay and that he might have been misinformed. Or perhaps it was that he or his informants equated masˇk and masˇket because both shared a common difference from Ashkenazic masˇet. That would still leave us to explain how the Arabic word masˇq came to be employed by those speaking Italkian and Yevanic. The explanation, as regards Italy, might be that it was introduced by Zarphatic refugees after 1394 or / and by Sephardic refugees after 1492; the latter might have been instrumental in conveying it to the Yevanim, i. e., those long resident in the Balkans.

000438

Birnbaum 01.04.11 18:17

438

[ 62 ]

E. Conclusion We have treated the phonological and lexical examinations in two watertight compartments. If we now place the two results side by side, both roads of approach are seen to have led to the same result: the multiplicity of the forms is best explained by assuming that they are derived from mas¯k, and the most likely original lexical source is Arabic masˇq. The fact that the internal and external evidence, so to speak, converge to one point seems to justify the hope that we are on the right track. As regards the chronology of our term, we know that by the twelfth century, at the latest, it was already employed in Zarphatic and Ashkenazic. In the latter it had developed into masˇtet not later than by the thirteenth century, and during the fourteenth century it had become masˇet. The e had been growing long, presumably since the late Middle Ages, and was, in due course, diphthongised. The most widespread wÈ mÅ or jõwaÅm pronunciation of ±È (sere) today is /ai/. My spelling jõÏ ˙ mirrors the final Ashkenazic form of our word. The advent of printing put an end to the style of writing designated by the term mashait and so to the common use of the word, but it survived in the sphere of halakhic study, discussion and codification.

F. Secondary References Paulus Fagius [101], in his Compendiaria Isagoge {Constantiae 1543} writes jiwm and describes it as a cursive script. Elias Schadaeus (Schade) [102] in his Ein gewisser Bericht etc., giving the name in transcription – Masket – only, also describes it as a cursive [103]. Thomas Erpenius [104] in his Grammatica Arabica compares the relationship between Diwani and Nischi {= Naskhi} to that between Masket and Assyrian (i. e., Hebrew Square) script [105]. J. J. Schudt [106] in his Juedische Merckwuerdigkeiten writes jiqÈsÂmÅ (with shewa under the s),22 transcribes Masket and translates ‘falling’, deriving the word from Arabic [107]. The Handlexicon speaks of “gemaschete Buchstaben” [108]. We do not know whether this is taken from the above-mentioned hnwmh tbkrm [76]. These (as was Sebastian Muenster – see above, and [109]) were all Christian scholars. The first modern Jewish scholar to mention the term was I. B. Levinson [110]. He based himself on Caro [111]. L. Zunz [112], the father of 22

As he is the only one to use s, this is presumably a retranscription from Schade or Erpenius.

000439

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait

439

the “Wissenschaft des Judentums”, writes Meschit, thereby revealing that the word was not known to him from oral tradition [113]. L. Löw [114] in his Graphische Requisiten lists qwm, jiwm, jiqwm, aqwm, qwam, jywmig, jiwym, jiwim, ajiwm, jwõm and compares the Syriac akwm (msk ) ‘nachlässig’ – cursive, careless writing [115]. M. Grünbaum [116] in his Jüdischdeutsche Chrestomathie transcribes the word in the passage as Maschkit [117], disregarding Elijah Levita’s vowelling. Grünbaum is quoted by J. Perles [118, 119]. M. Peritz [120] suggests derivations from Arabic msˇq and msˇt [121]. ˙ in his Vorlesungen [123] cites Zunz, Löw, M. Steinschneider [122] Dukes [124] and Perles; he compares qwm ‘schnell sein’ and ‘ausziehen’ = Ówm. N. Prilutzki [125] employs the name jyqwaÅm (maskε t or maskε it ) for the Ashkenazic cursive type face used for Yiddish books until the early decades of the nineteenth century. It is unlikely that he knew the word from living tradition. Perhaps he took it from Elijah Levita. E. BenYehuda (Ben-Jehuda) [126] in his Millon [127] lists it in three places: w mÅ (with shewa under the Ï w) [130]; (a) dwÃmà [128], jwÃmà [129], jiwÇmÂ, jiqÈÏ wà mà umk ± ajÄiÏ wÇ m ,jiÏ wÇ m [131]; (c) qÏ wà mà [132] deriving our term from (b) dÏ Arabic msˇq ‘to be long and thin’.23 A. Harkavy [133] in his Yiddish-English-Hebrew Dictionary lists ajiwÇm (i. e. m sˇˆıt ) with the translations ‘cursive Hebrew writing’ and irby ,Ñr-btk [134]. He gives no etymology. The present writer introduced the term mashait to replace the unsuitable designation ‘Rabbinic’24 [135]. M. Weinreich in his book id ÓylyjniP ycraÅùw mentions jiqwmÅ (i. e. masˇkit ) and jiwym (i. e. mε sˇit ), of unknown etymology, as the name of the cursive [136]. A. M. Habermann in his irbyh rpsh tudlut gives (without sources) as names of the cursive: jiqwm, jiwym, hjiwm, juqwm, hjuqwm, hjiqwm, deriving them from the Talmudical juqw ‘stubby, abnormally short (neck); abruptly receding (forehead)’ [137]. I. Sonne, publishing the book list of ca. 1425,25 renders our term as ‘semi-cursive’, and considers jigwm as a possible transition form from jiqwm to jiwm [138]. S. D. Sassoon in the introduction to his edition of Maimonidis Commentarius in Mischnam, quoting from Simeon b. Semah, translates qwam as ‘clear, bold writing’ ˙ form is iqwm [139]. and mentions that the Baghdadi e e

23 24

25

Ben-Yehuda, accordingly, thinks of cursive. The two catalogues of manuscripts in the Jewish National and University Library, Jerusalem, do not employ this term but use the words Cursive and Rabbinic. G. Scholem in his hlbqb di-ibtk ( Jerusalem 1930) writes hbisruq and tinbr, and I. B. Joel in his Õirbyh di-ibtk tmiwr ( Jerusalem 1934) has bizruq (North German voicing of the s) and tinbr. Cf. notes [65, 66].

000440

Birnbaum 01.04.11 18:17

440

[ 62 ]

Part II : Sources and Biographical Notes 1 Jacob b. Meı¨r (Rabbenu Tam). 1100 – 1171. Ramerupt, Troyes. 2 Jeroham b. Meshullam. Late 13th century and first half of 14th century. Spain,

˙ Provence.

3 ³nw alb hqd hbitk aihw iqwm tbitkb dyumh luxb butkl t`r ritmw ñuilgb itacm .(huxr Õda tudlut) .hsg hbitk qr inisb 'ntin al irhw ³nw hmcy hbitkhw

Constantinople 1516, fol. 31r. 4 Aaron b. Jacob hakKohen. End of 13th century and first half of 14th century.

Provence. .(Õõx tuxra) .iuniw Õuw alb qwm tbitk butkl ritm t`rhu

MS . Sulzberger, fol. 185v. Script: Maaravic mashait, ca. 16th century. MS . Adler 1770, fol. 116v, of 1554. Both in the Jewish Theological Seminary of America, New York. Ed. Florence 1750, fol. 91r, col. 2, bottom. 5 tbitkb dyumh luxb butkl t`r ritmw ñuilgb itacm hzh ñuwlk 'd'n úusb Õda rpsb butk .(h`mqt 'is ,Õõx xrua ,Óury ñxluw) .iqwm

1st ed., Venice 1550, fol. 371v. Venice 1589, fol. 372v. Later edd. have / msqh /: Berlin 1702, fol. 157r, col. a. Dyhernfurth 1790, fol. 139r, 2, near top. 6 Joseph b. Ephraim Caro. 1488–1575. Toledo, Lisbon 1492, Constantinople 1496, Adrianople, Nikopolis 1522, Safed 1525. His úsui tib was written 1522–1542; 1st

ed.: Venice 1550. 7 irhw ³niw alb 'jiwm 'bitkb h`xb butkl t`r ritmw ñuilgb itacm l`zu 'k h`x d`n u`atbu .('g 'is ,h`mqt x`a ,hwm ikrd) .l`ky hsg hbitk qr inisb hntin al

Berlin 1702, fol. 257r. 8 Moses Isserles. 1525/30–1572. Cracow. 9 Joel b. Isaac halLewi. Died ca. 1200. Bonn, Ratisbon / Regensburg, Mayence /

Mainz, Würzburg, Cologne. 10 Joseph (Yoslin) b. Moses. Ca. 1423–1490. Höchstadt a. d. Donau, Augsburg,

Erfurt, Ratisbon / Regensburg, Landau or Bamberg, Padua, Mestre, Wiener Neustadt, Rhineland ca. 1449–1454. His work: 1463–1475. 11 tuxlg tbwb 'tk Õa bõx 'iaw l`z laui unibr 'buwtm dxa ruxb qityhw 'biwih ly itymwu .(rwui jql) .hligmd q`p ikdrmb ñõyu juwym 'ipa 'itk raw ua

MS . Hebr. 405 (fol. 44r, top) of Bayerische Staatsbibliothek, Munich. Autograph. J. Freimann, in his edition {Leket joscher des Joseph b. Mose. Collectaneen seines Lehrers Israel Isserlein. 2 Bde., Berlin 1903/1904}, p. 65, middle, has erroneously replaced the word ñõyu by itacm ñku. The edition has not been accurately copied. In a note at the end of our quotation Freimann refers to no. 786. No doubt the Venice 1552 edition is meant, but the subject dealt with there is the writing of a Pentateuch in Roman characters and has nothing to do with mashait or cursive. 12 Isaac b. Moses. 1180/1200 – 1260/70. Bohemia, Ratisbon / Regensburg, 1214–1217

Paris, Coucy, Worms, Speyer, Würzburg, Ratisbon / Regensburg, Saxony, Bohemia, Vienna.

000441

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait 13

441

.u`y 'is ,tbw tuklh ,yurz rua

Zhitomir 1862, vol. 2, p. 32. 14 Simha b. Samuel. Lived in the decades around 1200. Speyer.

˙

15 alb tujuidh btkb 'iqusp butkl rtih ughnw 'ikmus hz lyd l`cz hxmw unibr irum 'ipu irhw btk buwx unia tujuidh irjw ñibtuk unaw j`ijwm ñiruqw hm hqd hbitkd Õuwm jujrw .(u`jwt 'is ñijig tuklh .yurz rua) .ñguhb ñqutm unia

Zhitomir 1862, vol. 1, fol. 101v, below. 16 Meı¨r hakKohen. Second half of 13th century. Rothenburg ob der Tauber. We

know when he lived from the fact that he was a brother-in-law of Mordecai b. Hillel (see [38]) and is said to have been a pupil of Meı¨r b. Barukh of Rothenburg ob der Tauber. 17 unia ik ub ñiwwux ñia jwym ñiruqw unlw hqd hbitkw btk duyu rumybk btk hxmw unibru ,tuinumõm tuhgh) .l`ky hrumg hbitk ala [ . . . ] ñuiz rsx tuah Õa ulipa ahd btk buwx .(úus ,'z 'is .hrut rps tuklh

1st ed., Constantinople 1509, fol. 29r, top. 18

.(úus ,'g hhgh ,z`i 'is ,Õw)

Venice, edd. 1524, 1550 (Giustiniani). .[ . . . ] ñiwwux ñia j`iwm unõhd unlw hqd hbitkw duy btku rbxmh unibrk btk hxmw unibru .(z`j 'is ,Õw)

Venice 1550 (Bragadin). 19 hbitk btknw turix jgb hih hwymu lusp Óklh ñk ³wy unia j`ijwm ñiruqw hm hqd hbitkbu .(u`jwt 'is ,ñijig tuklh ,yurz rua) .unitubr uhulspu j`ijwm ñiruqw hm hqd

Zhitomir 1862, fol. 101v, bottom. 20 Isserlin b. Petahia. ‘1390’ – 1460. Ratisbon. Austrian ancestry.

˙

21 tmurt) .këy bkym dui lw ucuq ulipa jiwm alu hsg bëitk butkl Óirc lëzu ñijg ñinyl tëk zëaëbu .(ñwdh

Jewish Theological Seminary of America, New York: MS . Adler 1412, fol. 117r; palaeographical date: ca. middle of 15th century, and MS . Bamberger, fol. 149v; palaeographical date: decades around 1500. 22

.(ñwdh tmurt) .a`jiwm alu hsg 'bitk butkl Óircu l`zu ludgh zëaëb btk irhw

Venice 1519, and Fürth 1778, no. 230. 23 .(ñwdh tmurt) .h`jiwm alu hsg hbitk butkl Óircu unuwl hzu ludgh yurz ruab btk irhw

Venice 1546, no. 230. 24 ñwdh tmurtbw Õiqsuph ñib qulix wi iruwa btk tluz unlw tubitk rawb jgh butkl ñinylu .(u`kq 'is ,rzyh ñba ,úsui tib) .jõwm alu hsg hbitk butkl Óircu z`a Õwb btk l`r ñmis

Sabbioneta 1553, fol. 152v. 25

.(u`kq 'is ,rzyh ñba ,úsui tib) .j`õqwm butkl alw urma alu ikhb udum ublku z`a

26 Moses b. Nahman (Nahmanides). 1198–1270. Gerona.

˙

˙

27 Simeon Duran b. Semah. 1361–1444. Palma, Teruel, Algiers 1391.

˙

˙

28 hz ñiaw ipl qwam btkb turgab jujrw alb Õiqusp butkl l`z ñ`bmrh rith imyj iahmu .(Ñ`bwrh tubuwt) .btk

London, British Library, Add. MS . 27 117, fol. 20v, bottom (no. 8).

000442

Birnbaum 01.04.11 18:17

442

[ 62 ]

29 Õiaibnu t`s ñugk ala urma alw jujrw alb butkl rtum tiruwa hniaw hbitkw l`n lba ibtk lkb urwkuh alu llk hbitk hniaw trtum iadu uz hbitk lba tiruwa Õtbitkw Õibutku .(l`z ñ`bmrhl ñijig iwudix) .rbd amy ñku 'ujrw alb uz hbitkb btkiml Óims ana ad lyu wdqh

Sulzbach 1762, fol. 56r, col. 1, bottom. 30 hbitk lba jujrw iyb 'ibutku Õiaibn hrut hb Õibtukw tiruwa tbitk aqud ñ`bmrh btk .(d`pr 'is ,hyd hrui ,úsui tib) .jujrw hkirc hnia Õiwuripu Õibtk hb Õibtukw

Venice 1551. .(1565 ,u`kq 'is y`ha ,úsui tib) .bdbdm ibryb arqnh rxa btkb butkl gihnh [ . . . ] ñ`bmrh

31 Hayyim b. Isaac. Approximately 1230–1305. Wiener Neustadt, Cologne.

˙

32 turix jgb hih hwymu [ . . . ] jgb bkym d`ui lw ucuq 'ipau .jijwm alu .hsg 'bitk 'utkl 'ircu .(yurz rua inmis) .[ . . . ] 'wam hl hl 'irmgd 'uwm 'itubr ulspu .jijwm btknw

Oxford, Bodleian Library: MS . Opp. 78, fol. 194r (no. 42, end); palaeographically roughly contemporaneous with the later times of the author. London, British Library: Add. MS . 23 974, fol. 213r, col. 1, top; palaeographical date: approximately first half of 15th century. 33 London, British Library: Add. MS . 27 297, fol. 120v (no. 43). Palaeographical

date: approximately second half of 14th century. 34 Oxford, Bodleian Library: MS . Mich. 524, fol. 141v – 142r. Palaeographical date:

roughly the decades around 1400. 35 Abraham b. Moses of Sinsheim. Approximately 1245–1320. Pupil of Meı¨r b.

Barukh of Rothenburg. Alsace, France, Sinsheim / Elsenz. 36 Samson b. Eliezer Barukh She-Amar. 14th century. Saxony, Prague, Palestine,

Germany, Austria, Poland. 37 hbitk ik hlyml rcq alu hjmlu hlyml Õ³w Õinp hl hwyiu huwb warh [ . . . ] Ók butki .uiu 'iruzxmb itbtukw umk alu 'ujriw 'ùi ñuwlbu z`ylb jiwym umk tiacxu hrubw alu 'iht hmt .(rmaw Óurb) .hjml hlymlm Óluhu jymtm ala

Hamburg, Staats- und Universitätsbibliothek: MS . Hebr. 291, fol. 59v – 60r. Palaeographical date: ca. end of 15th century. In the British Library Add. MS . 17 338, fol. 14r, middle (palaeographical date: decades around 1500) the form is jiwm. The MS . has Õibtukw instead of the slip in the above MS . There is no mention of ñuwl 'uui but there is in the Hamburg MS . Hebr. 145, fol. 9v (palaeographical date: roughly the middle of the 17th century), written out: ñui. The edd. have ajiwm: Dubno 1796, fol. 13v (here the etymological passage is missing), Shklov 1804, fol. 9r, col. 1, bottom (here z`ylb is replaced by a`lb). 38 Mordecai b. Hillel. Ca. 1240–1298 (Nuremberg pogrom). 39 .rusal xk ñiau hbitk ala dumlth rithw yicm alm úau tuquspu tukutx tuitua 'utkl ughn .(ikdrm)

New York, Jewish Theological Seminary of America: MS . Sulzberger, fol. 289v. Palaeographical date: ca. 16th century. 40 Solomon b. Abraham Ibn Adret. 1235–1310. Barcelona. 41 alym jg arbia .hbuwt .ual Õa rwk auh Õa qwÂamÅ ñiruq unaw btkb btknh hwa jg tlaw duy ahiw armiml ual btku bitkd gyau rwk ñuwl lkb ala iruwa btk jgb ñniyb ald auh rwku .(a`bwrh tubuwt) .úskb twrgtm hniaw jyml ala ab alu ymwmb btk lk la iruwa btk Óirc

000443

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait

443

London, Jews College: MS . 103, fol. 265r, no. 386. Palaeographical date: ca. 14th century. 42

.(a`bwrh tubuwt) .rwk q`wam ñiruq unaw hbitkb butkw jg

London, British Library: Add. MS . 22 090, fol. 219v, no. 6. Palaeographical date: ca. 15th century. 43 dblm ñijw b'i Óirc auh Õau qwm tbitkb rwk auh Õau al ua jujrw Óirc hwa jg tlaw duy aidhbu hlitkl úau ñijigb rwk ñuwl lku btk lk qwm tbitk ñinylu [ . . . ] 'uwt .Õidyh tmitx .(a`bwrh tubuwt) .[ . . . ] aqud ual tinuiu rwk tirby uidyu tinui uidyu tirby ñibtukw ñnint

London, British Library: Add. MS . 26 983, fol. 6r, no. 1. Palaeographical date: ca. 15th century. 44 jg arba .hbuwt .rwk auh Õa 'qwm ñiruq unaw btkh ñugk iruwa 'iaw tkb 'tknh jg .hlaw .(a`bwrh t`uw) .[ . . . ] rwku auh ailym

1st ed.: Venice 1519, no. 122. 45 atad auh 'mlyb Õirbd rupsl auhh rpsh btkd Õuwmu [ . . . ] rwk aq`wm butk auhw jgb duy biwh .(j`ywt 'is ,a`x ,a`bwrh t`uw) .rwk aq`wm tbitkb ulipad Óirc unia tuitual ñinuiz Õgu .(a`cq'a 'is ,a`x ,a`bwrh t`uw) .q`wm ñiruq unaw hbitkm butkw jg

Bologna 1539; Venice 1546; Hanau 1610. 46 Joseph Colon b. Solomon. Ca. 1420–1480. Savoy, Seville, Pieve di Sacco, Mestre,

Bologna, Mantua, Pavia. 47 l`acniburp aih a`bwrh rikzhw jõqwmw [ . . . ] a`bwrh tbuwt irbd a`y wruwb btk qirhmu .(1565 ,u`kq ,úsui tib) .j`õqwm utua ñiruq umuqmbw ala 48 auh a`bwrh rikzhw j`õqwmw l`nu [ . . . ] q`awm tbitk ñinylu l`zu hbuwtb btk a`bwrhu .(u`kq 'is ,rzyh ñba ,úsui tib) .j`õqwm itua ñiruq umuqmbw ala lac`niburp Sabbioneta 1553, fol. 152v. Berlin 1702. But in the edition of 1565: jõwm {instead of qawm}. 49 Alexander Ziislin hakKohen. Died 1349. Erfurt, Worms, Cologne, Frankfort/M., Erfurt. 50 tbitku hnwh lk ligr uniaw ybca ly sumluqh suptlu ñimib ua lamwb butkl buj ik itacm .('duga) .hbitk irq al jiwym

Paris, Bibliothe`que Nationale: MS . He´b. 409, end of chapter II of Mo ed Qatan: ˙ marginal note. Palaeographical date: roughly first half of 15th century. .('g qrp war ,ñjq dyum ,hdugah rps) .hbitk irq ald j`iwym tbitku

1st ed.: Cracow 1571, fol. 169, no. 38 (erroneous, should be no. 27). 51

.(u`kq 'is ,rzyh ñba ,úsui tib) .q`wm bdbdm wripw x`ab itacm ñku

Sabbioneta 1553, fol. 152v. 52 ikhb udum ublku z`a [ . . . ] l`ky hmt 'bitk 'iybd hsg hbitkm hihiw Óircu .'tk ublkb Õgu btkb lba .iruwal hmudw i`qwm ñiruq unaw btkb ala jõqwm butkl alw urma alu .(u`kq 'is ,rzyh ñba ,úsui tib) .ñibtukw hdum hmud uniaw lacn`iburp

Sabbioneta 1553, fol. 152v. 53 Chapter 60, 1st ed., Naples 1490 (?). 54 rawb w`ku Ó`irj`wua 'nidmbw 'ixmumh 'ipa hzh rudbw ñilipt ñibtuk bur iqupal [ . . . ] .(rmaw Óurb) .[ . . . ] hzk tiacx 'bitku ji`wy`m 'mk ñhlw ñ`iùh ñibtukw tunidm

Hamburg, Staats- und Universitätsbibliothek: MS . Hebr. 291, fol. 60v. Palaeographical date: late 15th century.

000444

Birnbaum 01.04.11 18:17

444

[ 62 ]

55 burm iqupalu [ . . . ] j`õcam umk hrubw alu tiacx hbitkb butkl alu hb rhzil wi ñkl ñibtukw tunidm rawb w`ku ñilu`pu Óir`jwia tnidmbw ñixmumh 'ipa hzh rudb ñilipt ibtuk .(rmaw Óurb) .[ . . . ] hzk hrubw ua tiacx hbitk ua j`õcam umk ñhlw ñ`iùh lk

London, Jews’ College: MS . 150. Palaeographical date: ca. first half of 15th century. 56 buj Õw rtk rpsb btk rwa azimrgm rzyila 'uhm brh lw udimlt l`z Õxnm 'r irbd Ól ahu hrubw hbitk llk ubtukl ñiaw 'uh ly Õkxh hz gilph hmk Óiniy xqp hnhu [ . . . ] l`zu :ulw ñixmumh 'ipa hzh rudb rwa ñilipt ibtuk burm iqupalu [ . . . ] hzk ruzxmb butkw umk ñiacxu [ . . . ] hzk 'muqy ua jiwmg umk ñhlw 'uh ñibtukw tunidm rawb w`ku ñilupu Órjsõa tnidmbw .('ipth tbitk inidm rps [rmaw Óurb]) .l`ky ñibi [ . . . ] 'dqdmh lku

Dubno 1796, fol. 14r. 57 buj Õw rtk rpsb u`³h ly btk rwa 'zimrgm rzyla unibr lw udimlt Õxnm r`xh l`zu h`gh umk ñiacxu hrubw llk hbtukl ñiaw ñ`iùh ly 'kxh hz gilph hmk haru Óiniy xqp hnhu [ . . . ] ulw tnidmbw ñixmumh 'ipa hzh rudbw ñilipt ibtuk burm iqupalu k`y [ . . . ] 'iruzxmb Õtua ñibtukw lku ñiacx hbitka jiwim umk Õhlw ñ`iùh lk ñibtukw tunidm rawb w`ku ñilupu Óõrjwua .(rmaw Óurb rps) .h`kyu l`ky ñibi ñibmhu hkrbh tazb 'h hkrbi Õnidk tuituab qdqdmh

Shklov 1804, fol. 9r, col. 1. 58 Samuel b. Aaron of Schlettstadt / Se´lestat. Died probably 1380. Strasbourg, Bagh-

dad, Jerusalem, Ratisbon / Regensburg, Strasbourg. 59 Finished end of 1376. 60 lugiyb butkl rusau biwxd uhnh ala butkl rusaw ñakm .arzy rpsb txa 'ua ñihigm ñiau .(g`p úus ,dyumh lux ,ikdrm ruciq) .ñijig wirb 'ig 'ip ñku 'tk unia unlw j`õcm iluau

Hamburg, Staats- und Universitätsbibliothek: MS . Hebr. 247, fol. 61v, top. Palaeographical date: ca. 1400. 61 In the margin of the passage in the preceding note, there is this remark, written by

the same hand: rhm hntinw hbitkm ñia tuituah ik h`xb ³niw tuwyl a`jõcm ñiruqw hqd 'bitkb ñia 'ut ./ut / inis

62 As 53, with these variants: .ñijig wirb 'ip ñku btk ñia [ . . . ] jiwm [ . . . ] hnwmb biwxd [ . . . ] uhnh k`aa [ . . . ] hrzy rpsb .(g`p úus ,dyumh lux ,ikdrm ruciq)

Paris, Bibliothe`que Nationale: MS . He´b. 409. Palaeographical date: ca. first half of 15th century. 63 butkl rusau biwhd uhnh k`aa 'ubtkl rusaw ñakm hrzy rpsb 'ipa txa 'ua ñixigm ñiau .(Õw) .'i'b'g ñijig wirb 'urip ñku 'tk 'ia unlw unõwm iluau lugyb

Paris, Bibliothe`que Nationale: MS . He´b. 408, fol. 124r, end of top sections. Written 1454, at Ahrweiler, Germany, by rkwi tb tumuru, a woman copyist. The form unõwm might be due to a slip, anticipating the following unlw, but as the copyist made so many mistakes in the Hebrew, it is more likely that she did not know the word and misread the j as n+u, an error which is not rare in these centuries in Ashkenazic cursive and mashait. 64 Óirc ñiau ñõzl ñikirc ñiad l`z a`bwrh Õwb btku [ . . . ] btk Õõx tuxra lyb Õg [ . . . ] h hlaw unibr brh ubr Õwb 'buwtb btk l`z Õ`bmrh ik 'tk dy rwk q`wam 'bitkb ulipau jujrw utua ñmznu Õidla btkm hzb wmtwn Óaih rmau tiruwa 'bitkb jgh butkl rusaw l`z úsuhi

000445

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait

445

inwil irt ymwm hzm 'biw rmau dam ³lt auhw inpm Õidrpsh unww hbitkb Õg 'usau ñuizbl [ . . . ] hbuwt .l`ky bdbdmla ñiruq unaw btkh hz 'umkb butki lba utua lsup hzu ajgb btkb jgh butkl ñiaw l`z Õ`bmrh tbuwt Õwm tbtkw hmu [ . . . ] ñijgl 'rwk q`wam tbitku hbitkb jgh butkl ughn al hz Õyjmw il harnu [ . . . ] l`z ñtnw Õyjhu atlm il ahimt iruwa .('h 'is ,Ñbwth 's) .uldgiw dy hturql ñiydui unia 'uqunithw ipl q`wam h³rqh

Amsterdam 1738, fol. 9v – 10r. As above, with variants as follows: 'rh ± úsuhi unibr brh ,duy ± dy ,tbitkb ± 'bitkb ,c`au ± Óirc ñiau ,x`a ± Õõx tuxra ,abiw ± 'biw ,rsau ± rusau ,ñmdznu ± ñmznu ,Õiqla ± Õidla ,tbitkb ± 'bitkb ,³lh úsui ± unia ,tuqunithw ± 'uqunithw ,l`nu ± il harnu ,tumkb ± 'umkb ,ajigb ± ajgb ,ymwt ± ymwm .('h 'is ,Ñ`bwt 's) .Õnia

London, British Library: Add. MS . 27 117, fol. 18v, middle. Palaeographical date: 18th/19th century. Órcuh alu Õyj uz hlawb ñia hbuwt bdbdm arqnh qw`am btkb hzuzm butkl uhm tlaw duy alb Õiqusp butkl l`z ñ`bmrh rith amyj iahmu llk btk unia qw`am btk irhw btkil hz .(Õw) .hmt hbitk 'iyb hzuzmbu btk hz ñiaw ipl qwam btkb turgab jujrw

Ibid., fol. 20v, bottom. 65 Cincinnati, Hebrew Union College: MS . 675. Ca. 1425. 66 hruw) jigwm 'ugrt Õy ;(d`i hruw) jõwm 'zxm icx ;('a hruw) jõwm dxa 'bitk 'b 'g'm's 'g 'd Õkutbu 'unjq 'ulipt 'z ;(z`k hruw) jõwm hlipt ;(d`k hruw) jõwm 'ludg hlipt ;(z`i .jõwm 67 'ibtukw jiwym lw 'bitkh hzd 'utb 'irmad 'uwm jujriw alb tubit 'gm rtui 'ibtuk unaw hm .hbitk unia Õibtkl

Munich: Bayerische Staatsbibliothek: MS . Hebr. 404, fol. 50v. Ashkenazic cursive, 1474. 68 Meı¨r b. Isaac. Ca. 1480–1565. Probably Katzenellenbogen (Hesse), Prague, Padua. 69 Õiritmh úau btk irqm ald Õuwm laicn`iburp btkb butkh jg ly Õimgmgm wiw itymwu iq`wm Õidrpsh lca arqnh auhw jõwm btkb ala urith al iruwa uniaw btkb butkl btk ³rq unia umcy auhu iruwa btkl imd ald lacn`iburp btk lba iruwa btkl hmud auhw ahd iniyb rùxm uniau lusp jgh ubtk Õau ub jgh ñibtuk ñia y`kl Õblm tujuidh udbw ala [ . . . ] Õ³gh btkm yrg al iadu Õirpsu turjwu Õibtk lacn`iburp btkb butkl ughnw ñùik ³lh i`r irbdl lxtkl ub ñibtukd ajiwp h³lt hbitk uniaw lacn`iburp hbitk ñk Õau arj`il a`idim arqnh btkb jgh btk Õa ulipau [ . . . ] ub butkl gihnhw bd`bdm btk uhzw rwpau bkiyw umku ajigb inwil irt ymtwil ald ikih ik Õhb ñibtuk ñia hlxtlw ala rwk qalyum ua il harnu ajigb inwil irt ald ikih ik 26qal`ym tarqnh aihu h³lt hbitkb butklm ³lh i`rh ina duyu .j`õqwm utua ñiruq umuqmbw ala lacn`iburp auh ab`wrh rikzhw j`õqwmw rmulu igulpl akiad akpia ul hmud uniawl iruwal hmudh btk ñib qlxl unab Õaw rmua .(u`kq 'is ,úsui tib) .iq`wm ñiruq unaw btk ñugk iruwal hmudw 'tkb jgh butkl ñiaw 70 Elia Levita (Elijah b. Asher HalLewi). 1469–1549. Ipsheim (Franconia), Padua,

Venice, Rome, Isny, Venice. 71 uydi al Õiznkwah Õiduhih ik ydt yudi .dwm hqdh hbitkh ñiaruq unaw uhm tlaw duy 26

{The term qalym mu allaq is Arabic, and means ‘hanging’. Its Hebrew translation is '³lt hbitk, ‘hanging script’. R. Joseph Caro rightly recognizes that they are the same script. – E. B.}

000446

Birnbaum 01.04.11 18:17

446

[ 62 ]

Õitprc Õiduhih lk htua ñiaruq ñku jiqÈwÂmÅ hbitk aihh hbutkh Õw ik ñguhk umw ajbl aihw drpsl hnuwar abu hbitkh taz hxqln Õhm ik ibry ñuwlm auhu Õinuiu Õizyul Õidrps Õnuwlb auhu ibry ñuwl auh jiqwm ik llkhu [ . . . ] :Õiduhih lkb hjwptn Ók rxau Õhl burq ñiyilbm unaw turxa tulm ñk ik jiwm rmanu úuqh ñiyilbm Õiznkwah unxnau .tuzru tuqd ñuwl .Óirahl lkua alu hbith ycmab txa tua

From a letter printed in Sebastian Muenster’s Commentarium Rabi David Kimhi in Amos prophetam. Basel 1531, sheet E, fol. 1ar. The letter is entitled “Epistola Eliae Levitae ad Sebastian Munsterum, eruditione non vacans”. Reprinted by M. Peritz in Monatsschrift für Geschichte und Wissenschaft des Judentums 38 (1894), pp. 256 f. 72 il dghÁw Õibr Õimi hzu jiqÈÏ w mÅ tyburm hbitk hniaw rmul hcur hqd hbitkl ñiruq jiqÈÏ w mÅ ÕiduhilÅ itlawu llk ibry ñu´wl uniaw il ydun Ók rxau hwuxku hzr wurip ibry ñuwl auhw Æ pÄrcu Õizyul Õibr auh hm urtpl uydi alu ñk hl ñiruq Õlku ÕiiÇbÇryu Õinuiu Õidrpsu Õit [ . . . ] auh hm unydi alu jiwÈymà Õhl ñiruq ÕiznÄkwah unxnau Elia Levita, ibwt. Isny 1541. 73 Joshua Boaz b. Simeon Barukh. 16th century. Sabbioneta, Sarigliano, Italy. 74 ji`qwm iluau :rtih Ók lk ñia ik úau tuyujqu tuquspu tuktuxm tuitua butkl ughn lba .(ikdrm inid ruqx) :btk ñia unlw In the Alfasi edition, Sabbioneta 1554, fol. 30r, no. 3 ‘to Alfasi, no. 859’; and ruciq uinmisu ikdrm, Cremona 1557, fol. 16, col. 1 ‘to Alfasi, no. 859’. Written in 1521. 75 Asher Antshel b. Joseph. First half of 16th century. Poland. 76

.jiwymg 'õbryd wj³d zd 'nua hb Æitk irb Æy ñia ñibirwig

‘[ The Hebrew words] written [i. e., printed] in Square script and the German [i. e., the Yiddish translation] in mashait script’. Cracow, ca. 1534, fol. 1r. In the second edition, named liwna ibr lw rps, Cracow 1584, that statement is reprinted but does not tally with the facts: the Ashkenazic script has here been replaced by the ‘Rashi’ type face.27 77 rmul wi Õirpsb btkn rwa iruwa btk uniaw 'j`iwm 'tkb 'ibutkh unlw 'irpsb Õau irps rqiyu [ . . . ] iruwa btkb 'ibtknh ñiliptb rimxhl wi ñklu .k`k arimx al uhõtwudqd .(umcyb ñuagh di btkm qtyuh) .[ . . . ] iruwa uniaw btkb 'ispdn ua 'ibutk unlw

Responsa, Cracow 1640, fol. 82v; Hamburg 1710. In the edition Hanau 1710, fol. 46v, col. 1, however, there is no abbreviation mark. 78

.³niw ub ukirch h`pa hjiwm 'bitk ghun hih Õinurxah imib irhd n`ku Õõx xrua, no. 545. Ed. Berlin 1702, fol. 257r; Dyhernfurt 1790; in the edition Fürth 1760, the h of hjiwm is replaced by an abbreviation mark.

79

.aj`iwm hbitk aihw unlw hbitkb ulipa rimxhl ughnu Õõx xrua, no. 545,5. Cracow, 2nd ed. 1580.

80 Parma, Biblioteca Palatina, MS . 2511 (de Rossi Jud.-Germ. 2). 1575. 81 Joel Serkes b. Samuel halLewi Yafe. 1561–1640. Lublin, Pruzˇany, Lublin, Cracow,

Brest Litovsk. 27

{The above line is quoted from the second edition (1584) whereas the first edition (around 1534) has jiwym (see both editions in the Yiddish Books on Microfiche, selected by Chone Shmeruk. Zug / Leiden 1976 ff.). – E. T.}

000447

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait

447

82 tuzuzmu ñilipt Õhb ñibtukw tiruwa btkb ala dyumb hbitk ursa ald juwp rbdh d`ypld ulipa butkl urith h`mu ñmua hwym unia unlw j`iwym hbitk lk lba auh ñmua hwymd t`su unia ulipa j`iwymd unlw hbitkb Õibr Órucl butkl irwd w`kmu uitunubwxu Õulw tliaw ³niw alb dyumb btk umcy auhw hard h`tb aibhw ludgh tydku ³niw albu dyumh Órucl al ulipa ji`wym btuk hih ludgh iadud harnu dyumh Óruc hih alw ip ly úa Õibr Órucl .(h`mqt 'is ,Õõx xrua) .uitunubwxu Õimulw trga ñugk dixi Óruc ala Õibr Óruc hih

Cracow 1631, fol. 351a. 83 David b. Samuel halLewi. Ca. 1586–1667. [Vladimir Volyn´sk, Lemberg], Cracow,

Potylicz, Poznan´ / Posen, Ostro´g, Lublin, Germany, Moravia, Lemberg. 84 auh rtihhu yburm hbitk ñiruq unaw t`s lw btkb 'ipa rtum i`bhld ymwm k`au [ . . . ] (b) z`y a`mr w`m k`yu turga Órucl ala ñniaw ñuik ñnuqitk tuituah ñtua butkl ñiqdqdm ñiad Õighun unaw rmaq a`jiwm btk ly ahd il ñbum unia ñmua hwym uniaw unlw btkb lqhl ughnu rmaq Õibtk rawa ala yburm hbitk Õhd y`wh jqnd Õirps thgh ly iaq al k`au ub hbitk ly iau Õulw tliawb ub rimxhl ughnw rma 'h 'iysb ahd unia hzu lqhl ub ughnd hzbu ñiliptu hrut rps tbitkb umk Ók lk ñqutm uniaw unlw btkb ub ñiliqmd rmaq yburm l`cu h`xb Õhb highl ñiliptu hrut rps btkb llk lqhl ughn al iadud lqhl ughn Óõw ñia dyumh Órucl uniau ub acuiku Ñ`w rxa xulwl lqhl Óruch btk ñinyl h`tb rkzumh ly iaqd btku hsg hbitk qr inisb hntin al irhw a`jiwm btkb ulipa rurb rtih ñia hzb Õg tmabu hzl úrucmu [ . . . ] (h) .hzb lqhl ghnmh dcm a`mr btk k`y iniyb rurb hz ñiau z`y úsui tib qr rtih ñia Õ`bmrh Õyjld i`b btku [ . . . ] (u) .llk ñmua hwym unia unlw a`jiwm tbitkw lqhl Óumsl wi j`iwm auhw ñdid hbitkbu b`sb d`mqt 'isb w`mk dyumh Órucl auh Õa hbitkw hzl úrucmu ñnbrd ala hnia h`x tkalm l`iqd ñuik d`yln [ . . . ] (j) .a`bwrh Õyjk ughn ñkw z`y diyhw jrpb i`b qsp ly Óumsl wi ripw llk ñmua hwym unia unlw a`jiwm (.h`mqt 'is ,Õõx xrua) .Óumsl im ly ul wi liqmhu umyjl rimxmhu 85 Abraham Abbale Gombiner b. Hayyim halLewi. 1635–1683. Ga¸bin, Kalisz.

˙

86

:h`sy

x`bh k`ku ñmua hwym unia y`klw a`jiwm unõh [:ñmua 'wym uniaw btkb »:a`mr«] Õhrba ñgm to Õõx xrua, no. 545,1 (Ñra ingm, Dyhernfurt 1692, fol. 278r).

87

.('h h`mqt 'is ,Õhrba ñgm) .llk btk irqm al a`jiwmd btk x`bu

88 bjih rab to Õõx xrua, no. 545,1. 89 Isaiah b. Abraham. Died 1723. 90

.llk btk irqm al a`jiwmd ubtk Õiqsup buru bjih rab to Õõx xrua, no. 545,5.

91 ly uaru Õinmis ub uwpxu uar gurhh uaibh rwakw udiyh l`nh q`qd d`bhu [ . . . ] buqip [ . . . ] ruxw ñizyùg zia lamwh dih tsipl Óums yurzh ly Õg amug ñiyk 'h ñimi lgrh lw rwbh by .unlw ajiwm btk lw z tua ñiyk hduhib ydun to rzyh ñba, no. 31. Prague 1776. 92 Moses Sofer / Schreiber. 1762–1839. Frankfort / M., Mainz, Boskowitz, Prostic /

Proste´jov / Prossnitz, Dresnitz, Matersˇborg / Mattersdorf, Presborg / Bratislava / Pressburg 1806. 93 spdnw hmb uyj tuxlg btkb 'iruad 'ia akild ñbmr Õwb rmul Õicruph uyiqphw hm ydu adx itrtb uyji ajiwm hbitk m`hxb rithl Õda trut rps Õwb hmqt ñmis úus x`a i`bb Õw m`du i`brh uily gilpu dn 'is huxu Õda tudlut 'sb ala ñ`bmrl Õda trut rpsb unniaw rumg btk tuxlg btk lba hmt hbitk 'iybd ñuik unlw ajiwm btk ala rith al duyu 'g tua .(rpus Õtx) .auh

000448

Birnbaum 01.04.11 18:17

448

[ 62 ]

94 Israel b. Gedaliah Lipshitz. 1782–1860. Kolmar / Chodziez´, Poznan´ / Posen; Danzig /

Gdan´sk 1837. 95 .[u`kqs Õw a`gm] hnuwar hruw Õqyl ñinwmu 'ibtkh lkb rtih ughn unlw ajiwm btkb uhim To ñjq dyum, ch. 3, m. 3, n. 31. Danzig 1844. 96 (a) jiqÇÏ w mÅ ‘one who gives rest, calms, eases, shows quietness, causes quietness, is quiet, is calm’; qÏ wà mà ‘housekeeping, household management, [acquisition, possession]’; qÏ wÄ mÅ ‘running and pressing of a crowd, rushing, running about’; hqÃw ÏÂmÅ ‘one

who causes to drink water, gives to drink, cup-bearer, butler, butlership, beverage, drink’; hqÃw ÏÂmÅ ‘well-irrigated land’; qiw ÏÇmà ‘that which overflows’; ju´w ÏmÄ, ju´w ÏmÇ ‘oar’; jiw ÏÇmÈ ‘one who causes to swim, teaches swimming’; hjÃw ÏÅmÂ, hjÃw ÏÂmÅ ‘one who jests, fool’. 97 Cf. Muhammad ibn al-Hasan Ibn Duraid (873–933, Basra), Jamharat al-lug˙a.

˙ Haidarabad 1925–1933,˙ vol. 3, p. 66. – Isma ¯ıl Ibn Hamma¯d al-Jawharı¯ (died 1007–1010), Ta¯j al-lug˙a. Bulaq 1865, vol. 2, p. 121. ˙– Mahmu¯d ibn Umar alZamahsˇari (1143), Asa¯s al-bala¯g˙a. Cairo 1922/23, vol. 2, p.˙ 387. – Muhammad ˘ ˙ ibn al-Mukarram Ibn Manzu¯r (1232–1311), Lisa¯n al- Arab. Bulaq 1302 {=1884}, ˙ vol. 2, p. 221. – Muhammad ibn Ya qu¯b al-Fı¯ru¯za¯ba¯dı¯ (1329–1414, Persia), al˙ vol. 2, p. 1334; Bulaq 1884, vol. 3, p. 274. – Muhammad Qa¯mu¯s. Calcutta 1817, al-Murtada ibn Muhammad al-Zabı¯dı¯ (1731–1791, India, Egypt), Ta¯j ˙al- aru¯s. ˙ Cairo 1889/90, vol.˙ 7, p. 70. – Butrus al-Busta¯nı¯, Qatr al-muh¯ıt. Beirut 1869, ˙ 1870, vol. 2, p. ˙1980. – ˙Sa ˙ ¯ıd asˇ-Sˇartu¯nı¯, p. 2053, and Muhit al-muh¯ıt. Beirut ˙ ˙ Beirut˙1889, p. 1215. – Abd al-Fattah as-Sa ¯ıdı¯ and Husain ˙ Aqrab al-Mawarid. ˙ ˙ Yu¯suf Mu¯sa¯, al-Ifsa¯h fı¯ fiqh al-lug˙a. Cairo 1929, p. 104. ˙ 98 A. de Biberstein Kazimirski, Dictionnaire arabe-franc¸ais. Paris 1860, vol. 2, p. IV . – A. von Kremer, Beiträge zur arabischen Lexicographie. Vienna 1883, p. 54 (= Sitzungsberichte der phil.-hist. Classe der Kaiserl. Akademie der Wissenschaften 103, p. 480). – H. A. Salmone´, An Arabic-English Dictionary on a New System. London 1890, p. 980. – W. T. Wortabet (and H. Porter), An Arabic-English Dictionary. Beirut 1913, p. 800. – J. G. Hava, Arabic-English Dictionary. Beirut 1915. – S. Spiro, Arabic-English Dictionary of the Modern Arabic of Egypt. Cairo, 2nd ed. 1923, p. 433. – E. Fagnan, Additions aux dictionnaires arabes. Algiers 1923, p. 164. – A. Hindie´, The Student’s Dictionary, EnglishArabic and Arabic-English of the Classical Arabic. Cairo 1927, p. 389. – J. B. Belot, Vocabulaire arabe-franc¸ais. Beirut, 14th ed. 1929. – E. A. Elias, Modern Dictionary Arabic-English. Cairo, 2nd ed. 1930, p. 585. – A. Barthe´lemy, Dictionnaire arabe-franc¸ais, Dialectes de Syrie: Alep, Damas, Liban, Je´rusalem. Paris 1935–1954, p. 789. – H. Wehr, Arabisches Wörterbuch für die Schriftsprache der Gegenwart. Leipzig 1952, p. 811. 99 Muhammad Husain ibn Halaf Burha¯n, Burhan-i qa¯ti (1651). Calcutta 1858, Supple-

˙ p. 77. ˙– Tikcˇand ˘Baha¯r (1768), Mustalaha˙¯¯t i Bahar-i Ajam. Delhi 1853. ment, ˙¯˙n Ra ˙¯ mpu¯rı¯, G˙iya¯s al-lug˙a¯t. Cawnpore ˙ iya¯s ud-Dı¯n ibn Jala¯l-ud-Dı – Muhammad G ˙ ¨ ¨ to English Diction1868. – A. A. K. Muhammad, Beginner’s Pronouncing Persian ary. Calcutta, 2nd ed.˙1924, p. 277. – F. Johnson, A Dictionary, Persian, Arabic and English. London 1852. – A. Berge´, Dictionnaire persan-franc¸ais [etc.]. Leipzig / Paris / London {1868}, col. 409. – F. Steingass, A Comprehensive Persian-English Dictionary. London 1892, p. 1247. – J. J. P. Desmaisons, Dictionnaire persan-franc¸ais. Rome 1908–1914, vol. 3, p. 620. – A. A. Romaskevich, Persidsko-ruskiı˘ slovar’

000449

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait

449

[etc.]. Leningrad 1931, p. 215. – S. C. Paul, The New Royal Persian-English Dictionary. Allahabad 1932. – S. Haı¨m, New Persian-English Dictionary. Teheran 1934–1936, vol. 2, p. 906; cf. also p. 899. 100 Muhammad ibn Mustafa¯ Va¯nqu¯lı¯, Lug˙at-i Va¯nqu¯lı¯ [transl. of al Jawharı¯’s Ta¯j al-

˙ Istanbul 1756,˙˙fol. 118. – Ahmad A ¯¯ sim Ainta¯bı¯, Tibya¯n-i na¯fi [transl. of lug˙a]. ˙ ˙ – Idem,˙ al-Uqı¯ya¯nu¯s al-bası¯t [transl. Burha¯n-i qa¯ti ]. Constantinople 1799, p. 763. of Fı¯ru¯za¯ba¯dı¯’s al-Qa¯mu¯s]. Constantinople 1220 A.H/1805, vol. 3, p. 54. ˙– Shams ed-Dı¯n Sa¯mı¯, Qa¯mu¯s-i türkı¯. Constantinople 1899 ff., p. 1353. – A. Hindoglu, Dictionnaire abre´ge´ turc-franc¸ais. Vienna 1838, p. 442. – H. Tewfik, Türkischdeutsches Wörterbuch. Leipzig, 3rd ed. 1921, p. 321. – A. Seydi, Resimli yeni türkc¸e lug˘at, Istanbul 1929, p. 824. – ˙I. Alaˆettin, Yeni türk luˆg˘ati. Istanbul 1930, p. 723. – Türk Dili Tetkik Cemiyeti, Osmanlıcadan türkc¸eye söz kars¸ılıklar Tarama Dergisi. Istanbul 1934, p. 519. – Türk Dil Kurumu, Türkc¸e Ingilizce sözlük. Istanbul 1945, p. 408. – A. V. Moran, Turkish-English Dictionary. Istanbul 1945, p. 814. – H. K. Kadri, Türk luˆg˘ati. Istanbul 1927–1945, vol. 4, p. 387. – H. C. Hony and F. Iz, Turkish-English Dictionary. Oxford 1947, p. 226. – F. a Mesgnien Meninski, Lexicon arabico-persico-turcicum. 2nd ed. by B. de Jenisch / F. de Klezl, Vienna 1780, vol. 4, p. 562. – T. X. Bianchi / J. D. Kieffer, Dictionnaire turc-franc¸ais. 2nd ed., Paris 1850, p. 911. – Müntehaba¯t-i lug˙a¯t-i osmaˆniyye. Constantinople 1853, vol. 2, p. 271. – J. T. Zenker,˘ Dictionnaire turc-arabepersan. Leipzig 1866–1876, vol. 2, p. 851. – J. W. Redhouse, A Turkish and English Lexicon. Constantinople 1890 (and reprints), p. 869. – L. Bonelli, Lessico turco-italiano. Rome 1939, p. 235.

101 Paulus Fagius. 1504–1549. Rheinzabern, Heidelberg, Strasbourg 1522, Isny 1527,

Strasbourg 1544, Cambridge 1549. 102 Elias Schadaeus (Schade). 1541–1593. Liebenwerda (Saxony), Strasbourg. 103 “Demnach soll man wissen, dass die Juden auch wie wir zweierlei Formen des

Alphabets haben, [. . .] die andere, so sie Masket nennen, wir aber die Kurrent, die sie in ihren Kommentarien, Briefen und anderen gemeinen Sachen in täglichem Gebrauch haben.” Mysterium [. . .] Sampt [. . .] einem gewissen Bericht von der Teutsch-Hebreischen Schrifft, deren sich die Juden gebrauchen. Strasbourg 1592. Reprinted as: Judenmissions-Tractate des M. Elias Schadäus. Leipzig 1892 (Schriften des Institutum Judaicum in Leipzig, Nr. 33), p. 59. 104 Thomas Erpenius. 1584–1624. Gorinchem, Leiden. 105 “Diwani [. . .] habetque se ad Nischi ut Masket Hebraeorum ad scripturam As-

syriacam.” Grammatica Arabica. Leiden 1613, p. 5. Repeated in his Arabicae Linguae Tyrocinium. Leiden 1656, p. 6. 106 J. J. Schudt. 1664–1722. Frankfort / M., Wittenberg, Hamburg, Frankfort / M. 107 “Das zweyte Hebraeische ist das sogenannte Rabbinische. [. . .] So haben doch die

Rabbinen wieder besondere Hebraeische Buchstaben zu dem Rabbinischen / welche kleiner / und etwas rundlich / fast wie im Lateinischen die Cursiv-Schrifft ist / und mit dem Arabischen Nahmen jiqÈsÂmÅ [samekh with shewa] Masket d. i. fallend / genennet werden / sind einem / der das Bibel-Hebraeische versteht / gar leicht zu erlernen und zu verstehen.”

000450

Birnbaum 01.04.11 18:17

450

[ 62 ]

Juedischer Merckwuerdigkeiten Vorstellende Was sich Curieuses und denckwuerdiges [. . .] mit denen [. . .] Juden zugetragen. II . Teil, Frankfort M. / Leipzig 1714, pp. 286 and 288. Our passage is followed by a discussion about “Das dritte Hebraeische”, i. e., the cursive. Did Schudt, in connecting jiqÈsÂmÅ [samekh with shewa] with Arabic / sqt / ˙ ‘to fall down, drop’ perhaps conceive of masket as a parallel to ta lı¯q ‘the hanging suspended script’? His source was obviously Elijah Levita and there the designation qalyum actually occurred. 108 {Nützliches Handlexicon der jüdischen Sprache.} Prague 1776, p. 17, sv. tua; I

have not seen the first edition, that of 1773. 109 Sebastian Muenster. 1489–1552. Ingelheim, Heidelberg, Basel. 110 I. B. Levinson. 1788–1860. Kremynyc / Kremeniec (Volhynia), Radziwil 1806,

Brod / Brody (Galicia), Kremynyc / Kremeniec. 111 ajiwm btk Õgu i`wr lw btk Õuih unilca arqnh btkh ta uaicmh Õhh Õimib drps imkh hbrh Õw aibmw (u`kq ñmis) rzyh ñba ruj ly úsui tibb ñõy .lacniburp btk ua ,jõqwm ua Õyjh (g`pr ñmis) d`i ruj ly úsui tib ñiyu ,Õõrbyh Õiwmtwmw Õtumwu Õibtk inim wjõj btk ,ajiwmu i`wrm dbl Õuih unta btk duy wiu ,hlah Õibtkh Õidrpsh uaicmhw unl wiw Õibtkh tunumt lkw ari ñibm lku ,znkwa ñuwlb Õirpsh lk ub ñisipdmw ,tiduhi hnuwar hpqwhb hmdiw dy ñmzh burb untwn lba ,iruwah btkm Õruqm Õluk hp Õirkznh .Õirxa Õibtk hmhw hduhi tib. Vilna 1838, ch. 129. 112 L. Zunz. 1794–1886. Detmold, Wolfenbüttel, Berlin, Halle, Berlin 1815. 113 “in [. . .] dem sogenannten Raschi-Cursiv, Meschit genannt, das auch französische, provenzalische oder spanische Schriftart heißt [. . .]” [with the forms jiwm, wwõm, ajiwm, jiwim, jiwym, and their sources]. “[. . .] entstanden aus dem bei den spanischen Juden üblichen qwm, aqwm, jiqwm [with sources], vielleicht dem ara-

bischen Neschi verwandt.” Zur Geschichte und Literatur. Berlin 1845, pp. 206–207. 114 L. Löw. 1811–1875. Crnahora (Moravia), Nagy Kanizsa 1840, Papa 1846, Szeged

1850 (all in Hungary). 115 “qwm und jiwm sind die zwei kürzesten Formen eines Schriftnamens, denen noch andere, erweiterte Formen zur Seite stehen. Ersterer Form stehen zur Seite: jiqwm, aqwm, qwam; letzterer: jywmig, jiwym, jiwim, ajiwm, jwõm. Schon Zunz bemerkt,

daß die erste Gruppe bei den spanischen, die zweite bei den französischen und deutschen Juden üblich ist. Nach seiner Meinung ging die Benennung von jenem zu diesem über, und ist sie vielleicht dem arabischen Neshi verwandt. Diese Verwandtschaft dürfte jedoch nur eine scheinbare sein. Das vorliegende Wort bezeichnet ganz entschieden eine Cursiv-Schrift, während dem arabischen Neshi das Diwani als Cursiv-Schrift entgegengestellt wird. Dann ist das arabische xsn ‘ein Buch abschreiben’ buchstäblich in die rabbinische Literatur übergegangen, [. . .] der Ausdruck findet sich zuerst bei orientalischen oder arabisch redenden spanischen Schriftstellern. [. . .] Den Schriften der Gaonen entnahmen ihn die Frankogermanen. Den selben Weg hat vielleicht das dunkle qwm genommen. akwm heißt syrisch: ‘nachlässig’. Die Cursiv-Schrift, die sich gehen ließ und die für die Quadratschrift bestehenden Formen vernachlässigte, hieß ‘nachlässige’ Schrift.

000451

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait

451

Der Name, dessen Ursprung in Vergessenheit gerieth, wurde in verschiedenen Gegenden auf verschiedene Weise ausgesprochen.” Graphische Requisiten und Erzeugnisse bei den Juden. Vol. 2, Leipzig 1871, p. 74. 116 M. Grünbaum. 1817–1898. Seligenstadt (Hesse), New York 1858, Munich 1870. 117 Jüdischdeutsche Chrestomathie. Leipzig 1882, p. 494. 118 J. Perles. 1835–1894. Baja (Hungary), Breslau, Poznan´ / Posen 1862, Munich 1871. 119 Review of Grünbaum’s Jüdisch-deutsche Chrestomathie, in: Monatsschrift für Ge-

schichte und Wissenschaft des Judentums 31 (1882), p. 134. ( This is presumably the place mentioned by Steinschneider in his Vorlesungen etc., p. 30, below [123]: “Perles, Mtschr.”) 120 M. Peritz. I have not found any information about this scholar, except for three

items in Moı¨se Schwab’s Repertoire des articles relatifs a` l’histoire et a` la litte´rature juives parus dans les pe´riodiques de 1783 a` 1898. Paris 1899–1905. 121 “Uns aber scheint es näher zu liegen an den arabischen Stamm qwm zu denken, man sagt nämlich h¨ batkla ip qwm extendit literas in scriptura (siehe Freytag, Lexicon arabico-latinum, Tom. IV , S. 182 b.). qwm aber heisst auch pexuit comam, und ebenso wird das arabische jwm übersetzt. Es ist vielleicht daher zu erklären, dass die Schreibart auch jwm genannt wurde.”

M. Peritz, “Ein Brief Elijah Levita’s an Sebastian Münster, nach der von letzterem 1531 besorgten Ausgabe desselben” [etc.], in: Monatsschrift für Geschichte und Wissenschaft des Judentums 38 (1894), p. 261. 122 M. Steinschneider. 1816–1907. Prostic / Proste´jov / Prossnitz, Leipzig 1838, Berlin

1839, Prague 1842, Berlin 1845. 123 Vorlesungen über die Kunde hebräischer Handschriften, deren Sammlungen und

Verzeichnisse. Leipzig 1897, p. 30. 124 L. Dukes. 1810–1891. Pres´borg / Pressburg / Bratislava / Pozsonyi, Vienna. 125 N. Prilutzki. 1882 – 194? (killed by the Nazis). Bardic´yv / Berdic´ev, Kremynyc /

Kremeniec, Couzmer / Sandomierz, Warsaw 1898. 126 E. Ben-Yehuda (Ben-Jehuda). 1858–1922. Lusky (Lithuania), Polotzk, Glubokoye,

Dwinsk / Denynborg / Dünaburg, Paris 1878, Algiers 1881, Jerusalem, America 1914, Jerusalem 1918. 127 tirbyh ñuwlh ñulm, Thesaurus totius Hebraitatis. Vol. 7, Berlin 1926. 128 This is, of course, Muenster’s form (cf. [71]) and it is strange that Ben-Yehuda

lists it as the basic one. 129 jwm does not occur in Elijah Levita (nor anywhere else as far as I know). BenYehuda evidently postulated it as a transition form between dwm and jiwm. 130 Õidrps Õhb uwmtwi tuligrh tuqdh tuituah trucl ³nk lkh ± ,jiqÈÏ w mÅ ,jiÏ wÇ m ,jÏ wà mà ,dÏ wà mà -úlah tuitua trucm jym jym ñmzh tubrb upytsnw ,hzb acuiku Õibtkm tbitkb Õiznkwau .ybrÁmh btkh lw tib

P. 3357.

000452

Birnbaum 01.04.11 18:17

452

[ 62 ]

:hbitkb

131

guhnh unlw ligrh btkh dÏ wà mà umk ± ,ajÄiÏ wÇ m ,jiÏ wÇ mÂ

He continues with quotations from Moses Isserles, Piko´w deposition, Moses Sofer / Schreiber; p. 3374. 132

,hbitkb Õiduhih lca guhnh qdh btkh ÕÏ wÈ ± qÏ wà mÃ

– with quotations from Solomon b. Abraham, Aaron b. Jacob. As regards the etymology, Ben-Yehuda quotes Elijah Levita, adding: wmwm 'bryb qÏ wm ik ,tibryh ñuwlb auh uz hlm ruqm ik ul urmaw im uqdc ik aih tmah Óa qwm wurp hbitkbu :l`r ahpurx dm hbatkla ipu :hbitkh ñinyb jrpbu ,Óirah 'mwmb Õg ,utÄb ,tiqwm auh urqyu wbw Á m auh jiqwm Õwh ik rbdl Õilgru .tuqdu Óra ´ 'mwmb Õg wmwm .hqdu hkura hbitk aih hnukhu (P. 3413.) Apart from jiqÈwÂmÅ (the w with shewa), all the vocalisations in his Millon

are Ben-Yehuda’s own inventions. He did not consult any manuscripts. If he had used Elijah Levita in the original – instead of the form in larwi izng (Paris) – he would have found that the vowel of the second syllable was an e. He omits the specimens with y, no doubt because he saw in that letter a Yiddish spelling. – In spite of the fact that not one form is written with a t in the sources, he took it for granted that the j was a corruption. In this way he got an adjective, meaning ‘long and thin’ (scil. ‘writing’). Arabic masˇaq + Hebrew / feminine suffix -ıˆt. 133 A. Harkavy. 1863–1939. Novarydek / Nowogro´dek, Vilna, Bialystok, New York

1882, Paris 1885, New York. 134 2nd ed., New York 1928, p. 316. 135 Soon after 1935. First mentioned in the Calendar of the School of Oriental and

African Studies, University of London, 1947/48, p. 80. He had never used the designation ‘Rabbinic’ in his palaeographical research (which he began in 1928). At first he called the script in question ‘Ornamental Cursive’ (‘Zierkursive’), first mentioned in his note “Aschkenasische Handschriften” in: Zeitschrift für die Geschichte der Juden in Deutschland 3, Heft 4 (1931), p. 277. 136 zia ñymaÄn aÅ ÓaÄn .`hb ÆiÇtK ynõlq` jnõm saÄd saÄù ,hnjÅq hb ÆiÇtK ñp Æ uryg ñym jaÄh jp Æ irwkig id .jin ñym sõù Óiz jmyn ñymaÄn ryd ñynaÅù ñup ;jiwym rydaÄ jiqwmÅ :ñyùyg

Vilna 1939, p. 177. 137 ,`jiqwm` ,`hj (i) u´qwm hbitk` ,`juqwm` (`bqyi ñiy`b) lcniburp Õg arqn ibisruqh btkh (.b ,us Õirdn ,'iy ,tuqdu tunjq tuituab btkl hnùkh) .`hjiwm`u `jiwym` Õiznkwah ipbu

Jerusalem 1945, p. 17. 138

`.jõwm 'au 'bitk 'b [ . . . ]`

(Speaking of books:) “[. . .] two in [square?] writing, one in semi-cursive. hbitk is in all probability an elliptic form of tyburm hbitk or hsg or hmt, used to indicate square characters. As to jõwm, cf. Löw, Graphische Requisiten . . . (Leipzig 1870–1871), II , 74 and 156, note 342. The form jõwm is mentioned neither by Löw nor by L. Zunz, Zur Geschichte und Literatur (Berlin 1845), p. 206. There still is no certainty about its meaning, but all indications point to a certain ductus between square and cursive. Hence our rendering: ‘semi-cursive’.” (P. 66.) “It may be noted that neither Zunz nor Löw quoted above [. . .] have the form jigwm. If Zunz is right that jõwm is a German transformation of the Arabic qwam, jiqwm then we may consider jigwm as a transitional form.” (P. 70.) I. Sonne, “Book Lists Through Three Centuries”, in: Studies in Bibliography and Booklore 1 (Cincinnati 1953), pp. 55–76.

000453

Birnbaum 01.04.11 18:17

The Etymology of the Term Mashait

453

139 “This cursive is still recognised and used in the East, and in Baghdad we find the same name given to it under the variant form of iqwm for qwam”.

S. D. Sassoon, Maimonidis Commentarius in Mischnam. Vol. 1, Copenhagen 1956, p. 19.

Index of Persons (in Chapters 28– 62) A page number in italics leads to a complete bibliographic information. Persons named in letters or deeds are marked with ° Biblical and Talmudic names are not included.

Aaron b. Jacob hakKohen 426, 427, 440, 452 Abraham b. Moses of Sinsheim 427, 436, 442 Abramson, S. 361 Abu l Fath 176 ˙ Nathan 417, 418, 422, Adler, Elkan 423 ¯¯ sim Ainta¯bı¯ 449 Ahmad A ˙ ˙ ˙ 248, 250, Akiba°, Rabbi 241, 251/252, 253 Alaˆettin, I˙. 449 Albeck, Ch. 242 Albright, William F. 9, 17, 151, 152, 156, 175, 176, 180, 181, 203, 213, 217, 257, 265 Alexander Ziislin [Suslin] hakKohen 427, 443 Alı¯ ibn Abı¯ Ta¯lib 418, 422 Alt, Albrecht ˙ 42 Asher Antshel (Rabbi) 428, 446 Avigad, N. 47

Baha¯r, Tikcˇand 448 Bar Kosba° (Kokhba) 241, 246–252, 253–255 Barthe´lemy, A. 448 Barthe´lemy, D. 300, 380 Bauer, Hans 39, 172 Bea, A. 39 Belot, J. B. 448 Ben Aflul° 248, 251/252 Ben-Yehuda, E. 428, 439, 451, 452 Beranek, Franz J. 403 Berge´, A. 448 Bianchi, T. X. 449 Biberstein Kazimirski, A. de 448 Birnbaum, Eleazar 94, 141 Boer, P. A. H. de 68 Bonelli, L. 449

Brockelmann, Carl 260 Brønno, Einar 179 Burha¯n, Muhammad Husain ibn Halaf ˙ ˙ ˘ 448 Burrows, Millar 175, 186 Busta¯nı¯, Butrus al- 448 ˙

Cantineau, Jean 28, 279, 289, 309 Caro, s. Joseph Charles, Robert H. 147 Chwolson, Daniel 197, 207 Clermont-Ganneau, Charles 31, 168, 175, 177, 182, 198, 203, 208, 213, 314, 332 Colon, s. Joseph Cooke, George A. 137 Cowley, A. E. 18, 28, 160, 182, 242, 282, 361, 370 Cross, F. M. 40 Crowfoot, John W. 17 Cumont, Franz 162, 189 Dalman, Gustaf 310 David b. Samuel halLewi 429, 447 Delcor, Mathias 243 Desmaisons, J. J. P. 448 Diringer, David 7, 10, 17, 18, 174 Driver, G. R. 150, 162, 242 Dukes, Leopold 428, 439, 451 Dupont-Sommer, Andre´ 39, 180 Duran, s. Simeon Dussaud, Rene´ 177, 182, 203, 213, 282, 324, Eissfeldt, Otto 179 Eleazar° b. Eleazar 349–351, 354, 356, 358, 359 Eleazar° b. Jehoseph 271 Elia (Elijah) Levita 428, 432, 437, 439, 445, 446, 450, 451, 452

456

Index of Persons

Elias, E. A. 448 Epstein, J. N. 259, 260 Erpenius, Thomas 438, 449 Euting, Julius 327

Fagius, Paulus 438, 449 Fagnan, E. 448 Fahd, T. 422 Fı¯ru¯za¯ba¯dı¯, Muhammad ibn Ya qu¯b al˙ 448 Fodor, A. 422 Freimann, J. 440 Freytag, Georg Wilhelm 451

Ganz, Peter F. 387 Ginsberg, H. L. 148, 150, 241, 243, 244, 257, 262, 264, 361 Gombiner, Abraham Abbale 429, 447 Gottheil, Richard 172 Grimme, Hubert 332 Grünbaum, Max 428, 439, 451 Gulak, Asher 169, 198, 208 Haı¨m, S. 449 Hansen, Olaf 274, 365 Harding, G. Lankester 53, 57, 61, 69, 151, 153, 299 Harkavy, Alexander 439, 452 Hava, J. G. 448 Hayyim b. Isaac 427, 442 ˙ dr° b. Judah 348–359, (373, H ˙ 374/375) Hillel° b. Judah 409, 410, 414 Hindie´, A. 448 Hindoglu, A. 449 Hoffmann, D. 241 Hony, H. C. 449 Hooke, S. H. 27, 242 Horsfield, G. 324 Ibn Adret, s. Solomon Ibn Duraid, Muhammad ibn al-Hasan ˙ ˙ 448 Ibn Manzu¯r, Muhammad ibn al-Mukar˙ ˙ ram 448 Isaac b. Moses 426, 440, 441 Isaiah b. Abraham 429, 447 Isma ¯ıl Ibn Hamma¯d al-Jawharı¯ 448 ˙ Israel b. Gedaliah Lipshitz 429, 448

Isserles, Moses 426, 428, 440, 446, 452 Isserlin (Israel) b. Petahia 426, 427, ˙ 441

Jacob° b. Jehoseph 271 Jacob° b. Judah 261, 264, 271 Jacob b. Meı¨r (Rabbenu Tam) 426, 440 Jaussen, A. 292, 332, 369 Jawharı¯, s. Isma ¯ıl Jehoseph° b. Ariston 261, 264, 271 Jehoseph° b. Jehoseph 271 Jenisch, B. de 449 Jeroham b. Meshullam 426, 427, 440 ˙ (Hieronymus) 176 Jerome Jeshua° b. Eleazar 271 Jeshua° b. Galgola 242, 246, 248, 251/252, 263, 264, 271, 299 Jesh[ua]° b. [ Jeho]seph 358, 359 Joel b. Isaac halLewi 426, 440 Joel Serkes (Sirkes) 429, 446, 447 Joel, I. B. 439 Johnson, F. 448 Joseph b. Ephraim Caro 426, 427, 428, 440,442, 443, 445 Joseph b. Moses 426, 440, 445 Joseph Colon b. Solomon 427, 428, 443 Joshua Boaz b. Simeon Barukh 428, 446 Judah b. Eliezer 100 [ J]udah° b. Judah 358, 359 Kadri, H. K. 449 Kahle, Paul 179, 180, 179–190 Ka´ldy-Nagy, G. 422 Katz, Dovid 395 Khulu¯s¯ı, Safa¯ 417 ˙ J.˙ D. 449 Kieffer, Klein, Max Y. 409 Klein, Samuel 168, 169, 198, 208 Klezl, F. de 449 Kraeling, E. G. 361, 370 Krauss, Samuel 169, 198, 208 Kremer, A. von 448 Kutscher, Eduard 360

Lacheman, E. R. 162, 179, 187 Lagrange, M. J. 332

Index of Persons Lehmann, O. H. 257, 269 Lehmann, R. P. 377 Lethbridge, T. C. 152 Leveen, J. 152 Levinson, I. B. 438, 450 Levita s. Elia Lewis, Alkin 53 Lidzbarski, Mark 18, 177, 182, 203, 213, 303, 332 Limt° (wife of Moses Segal) 404 Lipshitz, s. Israel Littmann, Enno 182, 279, 284, 288, 302, 332 Longperier, A. de 198, 208 Löw, Leopold 439, 450, 451, 452 Luzzatto, Moshe Chaim 100

457

Neubauer, Adolf 172 Norman, Frederick 387 Origenes 176 Paul, S. C. 449 Peritz, M. 439, 446, 451 Perles, Joseph 439, 451 Pilcher, E. J. 33 Plenderleith, Dr. [Harold] 163, 164 Ploeg, J. van der 186 Pognon, Henri 37 Porter, H. 448 Prilutzki, Noah 439, 451 Pritchard, James B. 3

Rabinowitz, Jacob J. 257, 264, 265, Macalister, Robert A. S. 169, 197, 198, 208, 302 Macler, F. 324 Mahmu¯d ibn Umar al-Zamahsˇari 448 ˙ Maimonides 95, 99,159 ˘ Malamat, A. 39 Manasseh° (of the Sons of Eliashib) 274, 374/375 Meı¨r b. Isaac 428, 445 Meı¨r hakKohen 426, 441 Meissner, Bruno 39, 172 Meninski, F. a Mesgnien 449 Milik, J. T. 241, 244, 246, 249, 268, 300, 303, 361, 363, 369, 380 Montgomery, James A. 175 Moran, A. V. 449 Mordecai b. Hillel 427, 441, 442 Mordecai° Shmaye Segal 404 Moscati, Sabatino 17 Moses b. Asher 382 Moses b. Nahman (Nahmanides) 426, ˙ 441, 442 ˙ Moses Sofer 429, 447 Muenster (Münster), Sebastian 428, 438, 446, 450, 451 Muhammad, A. A. K. 448 Muh˙ ammad al Murtada ibn Muham˙ al-Zabı¯dı¯ 448˙ ˙ mad ˙ Muhammad Giya¯s ud-Dı¯n ibn Jala¯l-udDı˙¯n Ra¯mpu¯rı¯ ¨448 Muhammad ibn Mustafa¯ Va¯nqu¯lı¯ 449 ˙¯ , Husain Yu¯suf˙˙ 448 Mu¯sa ˙

361, 370 Ra¯mpu¯rı¯, s. Muhammad Rashi 96, 100,˙ 159, 370 Redhouse, J. W. 449 Reisner, Georg A. 18, 174 Romaskevich, A. A. 448 Rostovtzeff, M. I. 168, 197, 207

Sa ¯ıdı¯, Abd al-Fattah as- 448

Sachau, Eduard 18,˙ 160, 172, 180, 197, 207 Salmone´, H. A. 448 Sa¯mı¯, Shams ed-Dı¯n 449 Samson b. Eliezer 427, 436, 442, 443, 444 Samuel b. Aaron of Schlettstadt 427, 444 Samuel° b. Zelikman Bak 404 Sˇartu¯nı¯, Sa ¯ıd asˇ- 448 ˙ Sassoon, S. D. 439, 453 Saul° b. Eleazar 271 Savignac, R. 292, 310, 324, 332, 369 Sayce, A. H. 18, 41, 160 Schadaeus, Elias 438, 449 Schechter, Solomon 147, 165 Scholem, Gershom 439 Schubart, W. 190 Schudt, J. J. 438, 449, 450 Schwab, Moı¨se 451 Schwarz, Werner 387 Sellers, O. 151 Serkes, s. Joel

458

Index of Persons

Seydi, A. 449 Shalom° (daughter of Simeon) 355, 357–359, (373) Shmeruk, Chone 446 Simeon Duran b. Semah 426, 427, ˙ ˙ 441, 444f. Simeon° b. Jehoseph 358, 359 Simha b. Samuel 426, 428, 441 ˙ Solomon b. Abraham Ibn Adret 427, 428, 442f. Sonne, I. 452 Spiegelberg, Wilhelm 9 Spiro, Socrates 448 Staples, W. E. 17 Starcky, J. 311 Starkey, J. L. 53 Steingass, Francis J. 448 Steinschneider, Moritz 451 Stern, S. M. 257, 269 Sukenik, Eleazar L. 17, 28, 29, 32, 38, 40, 41, 147, 148, 175, 182, 186, 219, 231, 234, 332

Teicher, J. L. 165 Tewfik, Hacki 449 Theodor, J. 242 Thureau-Dangin, Franc¸ois 34 Timm, Erika 404 Torczyner, Harry 18, 53, 54–77, 174

Trever, John C. 81, 83, 153, 163, 180, 181, 217 Turniansky, Chava 404

Va¯nqu¯lı¯, s. Muhammad Vaux, R. de (O.˙P.) 79, 151, 152, 173, 186, 188, 190, 257, 259–265, 269, 297 Vincent, H. 332 Vincent, L. H. (O. P.) 151, 182, 221 Wartski, I. 241, 261, 264, 265 Wechsler, Tuviah 151 Wehr, Hans 448 Wenham, John W. 217 Wickes, William 172 Wortabet, W. T. 448 Yalon, [Henoch] 259 Yeivin, Shemuel 17, 19 Zabı¯dı¯, s. Muhammad Zamahsˇari, s. ˙Mahmu¯d ˘ Solomon ˙ 148, 149, 158, 165, Zeitlin, 187 Zenker, J. T. 449 Zimt Sand, Ilse 403, 404 Zuntz, G. 190 Zunz, Leopold 438, 439, 450, 452