Nov/947/1-11 1/2/1904 6:34 AM Page 1 THE NATURE OF THE GREEK TEXTS FROM THE JUDEAN DESERT by EMANUEL TOV Jerusalem 1. The evidence1 The Greek texts found in the Judean Desert constitute merely a small part of the texts found in the area, which are best known for the Hebrew and Aramaic texts, especially the texts found at Qumran. However, the Greek texts are by no means negligible, since in several sites their number equals that of the Hebrew/Aramaic texts, and in one site they even constitute the majority. Thus, while for Qumran in general the number of the Greek texts may be negligible, for cave 7 it is not, since all 19 items found in this cave constitute Greek papyri. This cave thus witnesses activity in the Greek language, but only lit- erary activity, since probably all the fragments found in this cave are non-documentary. Turning now to absolute numbers of texts, a word of caution is in order. Obviously we can only refer to the numbers of the texts which have survived, but as we will turn to statistics, it should be recognized that there is no reason why Greek texts should have perished into a larger or smaller degree than the other documents. Comparative sta- tistics of the various texts found should therefore be considered legit- imate. The majority of the texts found in the Judean Desert are Semitic, mainly Hebrew, but also Aramaic. The Qumran corpus consists of 1 A. R. C. Leaney, “Greek Manuscripts from the Judaean Desert,” in: J. K. Elliott (ed.), Studies in New Testament Language and Text. Essays in Honour of George D. Kilpatrick on the Occasion of His Sixty-Fifth Birthday (NTSup 44; Leiden: Brill, 1976) 283-300; L. Greenspoon, “The Dead Sea Scrolls and the Greek Bible,” in: P. W. Flint and J.C. VanderKam (eds.), The Dead Sea Scrolls after Fifty Years—A Comprehensive Assessment (Leiden/Boston/Köln: Brill, 1998) 101-127; E. Ulrich, “The Septuagint Manuscripts from Qumran: A Reappraisal,” in: G. J. Brooke and B. Lindars (eds.), Septuagint, Scrolls and Cognate Writings: Papers Presented to the International Symposium on the Septuagint and Its Relations to the Dead Sea Scrolls and Other Writings (Manchester, 1990) (SCS 33; Atlanta, GA 1992) 49-80. © Koninklijke Brill NV, Leiden, 2001 Novum Testamentum XLIII, 1 Nov/947/1-11 1/2/1904 6:34 AM Page 2 2 emanuel tov remnants of some 900 compositions which were once complete. Of these some 150 are in Aramaic (including 17 Nabatean texts), 27 in Greek, and the remainder are in Hebrew (including texts written in the cryptic scripts and in paleo-Hebrew). The Greek texts in Qumran thus comprise a very small segment of the complete corpus, namely 3%. This small percentage is matched only by the finds in Wadi Daliyeh, beyond the Judean Desert, while Greek texts have been found in much larger quantities at all other sites in the Judean Desert. Because of the fragmentary state of many texts, especially papyri, statistics for these sites can only be approximate:2 Table 1: Greek Texts from the Judean Desert Sites (North to Total number of Greek texts percentage of South) texts (leather, papyrus) total texts Wadi Daliyeh 29 1+ 3% Jericho 30 17+ 56%+ Qumran 900 27 3% Wadi Nar 4 2 50% Wadi Ghweir 2 1 50% Wadi Murabba'at 158 71 45% Wadi Sdeir 4 2 50% Na˙al Óever3 157+ 55+ 35%+ Na˙al Mishmar 3 1 33% Na˙al Íe"elim 6 2 33% Masada 48 11+ 23%+ 2 The precarious nature of statistics may be illustrated by the following: The numer- ous Greek fragments from what is named XÓev/Se and which are grouped on twodif- ferent plates (DJD XXVII, plates XLVIII and XLIX, are numbered XÓev/Se 74-169 for the sake of convenience, and likewise Óev/Se? 1-57 are grouped on plates L-LIII in the same volume. It is hard to know how these collections should be accounted for in a statistical analysis. The author responsible for these texts (H. Cotton) did not want to imply that these items have to be counted as respectively 96 and 57 different compositions, and hence they should probably be counted as six different ones, although both types of accounting are imprecise. Many of the fragments in these collections will have belonged to other documents from Na˙al Óever published in DJD XXVII, while other fragments must have belonged to different texts, not published in the volume. The collections of fragments known as 1Q69 and 1Q70 are treated similarly. On the other hand, the manifold fragments written in the cryptic A script and recorded by S. J. Pfann as 4Q249a-z were represented by the author as no less than 27 different compositions. 3 Including Óever/Seiyal. Nov/947/1-11 1/2/1904 6:34 AM Page 3 greek texts from the judean desert 3 We now turn to some detailed remarks about the Greek leather and papyrus texts found in the Judean Desert, not counting ostraca. First, attention will be directed to sites other than Qumran, with the exclusion of the approximately fifty texts from Óirbet Mird because of the Byzantine date of that site. Greek texts, most of them documentary, have been found in vari- ous places in the Judaean Desert (North to South): Wadi Daliyeh (1+ [undeciphered]), Jericho (17 and several fragments), Wadi Nar (2), Wadi Gweir (1), Wadi Murabba'at (71), Wadi Sdeir (2), Na˙al Óever (32 from cave 5/6; 2 from cave 8; 21, and many unidentified frag- ments from ‘XÓev/Se’ and ‘Óev/Se?’),4 Na˙al Íe"elim (2), Na˙al Mishmar (1), and Masada (remains of probably 11 texts [a few in either Greek or Latin] and several fragments).5 The largest groups of Greek texts thus derive from Murabba'at and Na˙al Óever, originally wrongly denoted as ‘Seiyal,’6 and involving two archives in Greek and Aramaic from Na˙al Óever (the archive of Salome Komaïse daugh- ter of Levi and that of Babatha). The documentary texts found in these sites relate to such matters as marriage contracts (e.g., 5/6Óev 18, 37), receipts (5/6Óev 27; XÓev/Se 12), deeds of gift (5/6Óev 19), registration of land (5/6Óev 16), summons (5/6Óev 23, 25, 35), let- ters (5/6Óev 52), etc. The nature of the documents found in the loca- tions outside Qumran thus shows that Greek was in active use among the persons who left these documents behind. That Greek was in active use beyond Qumran can also be seen from the percentage of the doc- umentary Greek texts among the Greek texts found at the individual sites. In all sites this percentage is relatively high, but not at Qumran. 4 See N. Lewis, The Documents from the Bar Kochba Period in the Cave of Letters—Greek Papyri (Jerusalem: Israel Exploration Society, the Hebrew University, and the Shrine of the Book, 1989). 5 See H. M. Cotton and A. Yardeni, Aramaic, Hebrew, and Greek Documentary Texts from Na˙al Óever and Other Sites, with an Appendix Containing Alleged Qumran Texts (The SeiyâlCollec- tion II) (DJD XXVII; Oxford: Clarendon, 1997) 134-135; H. M. Cotton and J. Geiger, Masada II, The Yigael Yadin Excavations 1963-1965, Final Reports, The Latin and Greek Documents (Jerusalem: Israel Exploration Society and the Hebrew University, 1989); E. Tov with the collaboration of S. J. Pfann, Companion Volume to the The Dead Sea Scrolls Microfiche Edition (Leiden: Brill, 1995). 6 See Cotton and Yardeni, DJD XXVII, 1-6. Nov/947/1-11 1/2/1904 6:34 AM Page 4 4 emanuel tov Table 2: Documentary and Non-documentary Greek Texts found in the Judean Desert Sites (North to Total Doc. Percentage Non- Percentage South) number texts of total doc. of total number texts number Wadi Daliyeh7 1+ — Jericho 17+ 17+ 100% 0 0% Qumran 27 1 3% 26 97% Wadi Nar 2 2 100% 0 0% Wadi Ghweir 1 1 100% 0 0% Wadi Murabba'at 71 66 93% 5 7% Wadi Sdeir 2 2 100% 0 0% Na˙al Óever 55+ 54 98%+ 1 2% Na˙al Mishmar 1 1 100% 0 0% Na˙al Íe"elim 2 2 100% 0 0% Masada 11+ 9+ 82%+ 2 18% Beyond the documentary texts, a few non-documentary, that is, lit- erary Greek texts, sometimes ill-defined, have been found in various sites outside Qumran, included among the statistics in Table 2: five papyri from Wadi Murabba'at, mostly of undetermined nature (DJD II [Oxford: Clarendon, 1961] 108-112), probably two from Masada (Mas 743 [Mas wood Tablet gr] from 73 or 74 ce; Mas 739 [Mas papLiterary Text? gr]),8 and one from Na˙al Óever (8ÓevXII gr), but none from the other localities of Wadi Gweir, Wadi Nar, Wadi Sdeir, Na˙al Íe"elim, and Na˙al Mishmar. The best preserved of these lit- erary texts was found in Na˙al Óever, viz., the Greek Minor Prophets Scroll, 8ÓevXII gr (publication: DJD VIII [Oxford: Clarendon, 1990]). In striking contrast to the texts found outside Qumran, all but one of the 27 Greek texts found in Qumran are literary, although admit- tedly it is hard to express certainty in the case of small papyrus frag- ments, viz., 4Q119-122, 126-127; 7Q1-19 (all the preserved texts of cave 7 are Greek papyri), altogether five texts on leather and three on papyrus from cave 4, and 19 papyri from cave 7. Almost all of these texts contain Greek Scripture texts in the wide sense of the word (in- cluding 7QpapEpJer gr). This characterization includes the literary papyri 7Q4–18, which are too fragmentary for a precise identification 7 These papyrus fragments (PAM 43.962; Rockefeller Inv. 550) have not been deci- phered. They are published as pl. XL. in DJD XXVIII. 8 See Cotton and Geiger, Masada II, 90. Nov/947/1-11 1/2/1904 6:34 AM Page 5 greek texts from the judean desert 5 of their contents. The one item among the Qumran Greek texts which is not literary is the documentary text 4Q350 (4QAccount gr, written on the verso of frg. 9 of a Hebrew text, 4QNarrative Work and Prayer [4Q460]), the nature and date of which cannot be deter- mined easily (DJD XXXVI). The nature of 4QpapUnidentified Fragment gr (4Q361) cannot be defined either (see DJD XXVII, plate LXI, with- out transcription). The picture emerging from an analysis of the Greek texts found in the Judean Desert is that the situation at Qumran differs totally from that of the other sites. In all sites, all the Greek texts (and in Wadi Murabba'at and Masada, the great majority) are documentary, show- ing that Greek was actively used among the persons who deposited the texts. These texts include documents showing that the adminis- tration was conducted in Greek in the Roman provinces of Syria, Arabia, and Judaea, and that letters were written in that language (see, i.a., Greek letters written by Bar Kokhba’s followers, found in the Cave of Letters in Na˙al Óever). On the other hand, there is no proof that Greek was a language in active use by the inhabitants of Qumran. It is possible that at least some of them knew Greek, since fragments of Greek Scripture were deposited in caves 4 and 7. But cave 4 proba- bly served as a depository of some kind (not a library) in which the Qumranites placed all their written texts (mainly Hebrew and Aramaic literary works, but also tefillin and mezuzot). This depository in cave 4 contained eight Greek texts, which may signify that the person(s) who brought these texts to Qumran had used them prior to their arrival which would imply knowledge of Greek. But it is not impossible that these texts came directly from an archive in which case no knowledge of Greek by the Qumranites needs to be assumed. Furthermore, the small number of Greek texts found at Qumran is also in striking con- trast with the other sites in the Judean Desert. The difference is partly chronological (most of the sites in the Judaean Desert are from a later period than Qumran), but more so in content: the Qumran corpus is mainly religious, which at that time would involve only Greek Scripture texts, and not other compositions. The evidence does not suggest that the Greek texts from cave 4 were read or consulted at Qumran or that they were written there. Cave 7 is a different issue. The contents of that cave which was prob- ably used for lodging (thus R. de Vaux, DJD III, 30) or as a work- place, consisted solely of Greek literary papyri, probably all Greek Scripture, and possibly all of these were brought directly to the cave from an archive outside Qumran or from a specific spot within the Nov/947/1-11 1/2/1904 6:34 AM Page 6 6 emanuel tov Qumran compound. No relation between the Greek texts of cave 4 and cave 7 need to be assumed, and there is no reason to believe that any of these texts was penned down in Qumran. Since the documents found in Na˙al Óever show that Greek was used actively by the persons who left the texts behind, including a Scripture scroll, some or much use of that scroll by the persons who deposited the texts in Na˙al Óever may be assumed. Indeed, that Minor Prophets scroll contains a Jewish revision of the Old Greek,9 and as such a version would have suited the freedom fighters of Bar Kochba, it was probably used by them. The situation was completely different for the Scripture finds at Qumran, which attest to an earlier period, up till 70 ce. In the period which is attested by the settlement at Qumran, the kaige-Th revision of the Old Greek, such as is reflected in 8ÓevXII gr, already existed. But neither this revision nor similar ones, found its way to Qumran, probably not because the Qumran covenanters disagreed with the concept behind these revisions, but because they did not turn to the Bible in Greek. For them the Bible existed mainly in the source lan- guages, and among the 220 biblical texts found at Qumran, Greek and Aramaic translations (4QtgLev, 4QtgJob, and 11QtgJob) form a very small minority indeed. In light of the preceding, special attention should be paid to an opisthograph, the recto of which formed fragment 9 of a Hebrew text named 4QNarrative Work and Prayer, while the verso contained a Greek documentary text, 4QAccount gr (4Q350).10 It is hard to char- acterize that Hebrew composition, which was described by its editor, E. Larson, as ‘somewhat akin to the Hodayot.’11 It may be sectarian, 9 See E. Tov with the collaboration of R. A. Kraft: The Greek Minor Prophets Scroll from Na˙al Óever, (8ÓevXIIgr) (The Seiyal Collection I) (Discoveries in the Judaean Desert VIII; Oxford: Clarendon, 1990) 102-158. 10 This text has been published by H. Cotton in: S. J. Pfann, Cryptic Texts; P. Alex- ander and others, in consultation with J. VanderKam and M. Brady, Miscellanea, Part1: Qumran Cave 4.XXVI (DJD XXXVI; Oxford 2000). 11 E. Larson, ibidem, 372: ‘It is difficult to discern the overall character of the work in its present state of preservation. The major part of the extant fragments is given over to prayer, exhortation, and admonition. It is possible, therefore, that 4Q460 is a collection of psalms somewhat akin to the Hodayot. This may be suggested by the para- graphing of material which is clear on frg. 9 and is supported by the fact that the material before the vacat is addressed to God while that occurring after the vacat is addressed to Israel with little or no intervening narrative to explain the change. If this understanding of the nature of the manuscript is correct, then the person speaking in the first singular in frg. 9 i 2 is some unknown psalmist.’ Nov/947/1-11 1/2/1904 6:34 AM Page 7 greek texts from the judean desert 7 and in any event, its orthography and morphology suggest that it was copied (not necessarily authored) by a sectarian scribe,12 while the verso contains a documentary Greek text. Parallels to the Greek Account from Qumran are found in various sites in the Judean Desert: Mur 810A, 89-102, 118-125; 1Mish 2; 34Íe 5. While the evidence implies that Greek was not in active use among the Qumranites, as no doc- umentary Greek texts have been found on the spot,13 the Greek 4Q350 may indicate an exception, and may imply that Greek was neverthe- less in use in Qumran at some stage prior to 70 ce, or that this doc- ument did not derive from Qumran. With regard to the first possibility that Greek was in use at Qumran, and that there once was a small corpus of administrative documen- tary texts in Greek, attention should be directed to the documentary texts 4Q342-360 in Aramaic and Hebrew, the idea being that if doc- umentary texts were written in Qumran in Hebrew and Aramaic, they could have been written in Greek as well. However, serious doubts regarding the Qumranic origin of 4Q342-360 have been raised by A. Yardeni, DJD XXVII, 283-317.14 Some of these texts may have derived from other, later, sites, and may have been sold to scholars as ‘Qumran,’ possibly in order to enhance their price.15 We therefore resort to the assumption that 4Q350 was written on the verso of frg. 9 of the Hebrew text 4Q460 after the occupation of the site by the Qumranites when some of the documents were still laying around, and were re-used due to the scarcity of writing material: (1) Only the verso of frg. 9 of 4Q460 was inscribed, which 12 See the arguments developed in my articles “The Orthography and Language of the Hebrew Scrolls Found at Qumran and the Origin of These Scrolls,” Textus 13 (1986) 31-57; “Hebrew Biblical Manuscripts from the Judaean Desert: Their Contribution to Textual Criticism,” JJS 39 (1988) 1-37; “Further Evidence for the Existence of a Qumran Scribal School,” in: L. H. Schiffman and others (eds.), The Dead Sea Scrolls: Fifty Years After Their Discovery—Proceedings of the Jerusalem Congress, July 20-25, 1997 (Jerusalem 2000) 199-216. 13 The same argument cannot be used for Hebrew and Aramaic. For the Qumran community, Hebrew was the central language, even if they left very few documentary texts in that language (the main text showing use of this language within the com- munity, beyond the many literary texts, is 4QRebukes Reported by the Overseer [4Q477]). No Aramaic community texts have been preserved, although the influence of the Aramaic language on the community scribes is evident in many writings. 14 In some instances Yardeni points to joins between the Qumran texts and texts which definitely derived from Na˙al Óever (note especially XÓev/Se papDeed F ar [= XÓev/Se 32] which forms one document together with 4Q347). Furthermore, carbon-14 examinations point to a late date of some documents. 15 This assumption has been rejected orally by J. Strugnell (February 2000) who stated that the Beduin were questioned very thoroughly regarding the origin of thetexts. Nov/947/1-11 1/2/1904 6:34 AM Page 8 8 emanuel tov necessarily points to a period in which that manuscript had already been torn into pieces or had partially disintegrated. (2) The writing of a documentary text on the back of a literary text is parallelled by many Greek papyri from Hellenistic Egypt (see the analysis by Gallo),16 by Elephantine papyri,17 and by 4QCal. Doc. Cc (4Q324)—a docu- mentary/literary text—which has on the verso 4QAccount C ar or heb (4Q355). Likewise, Mur papLiterary Text (Mur 112) has on its verso Mur pap Proceedings of Lawsuit (Mur 113). (3) As a rule, writ- ing on the flesh side (the verso) of the leather (4Q350 in this case), is subsequent to that on the recto (4Q460). At the same time, it remains difficult to understand the realia of the writing on 4Q350 and 4Q460: if fragment 9 was hidden in cave 4 by the Qumran community and was indeed found in that cave, how could it have been re-used by those who were to occupy the site after the Qumran community? The writing of the Greek text 4Q350 on the verso of the Hebrew text 4Q460, frg. 9 must have been later than the writing of the recto (4Q460), but the Greek writing could in principle have been performed within the period of the occupation of Qumran by the Qumran cove- nanters themselves, which seems to be a possible alternative. However, E. Larson argues that the Qumran sectarians would not have reused a scroll that contained the Tetragrammaton on the recto (4Q460 frg. 9 i 10) for such a profane use as recording a list of cereals in Greek (DJD XXXVI, 369). Larson adds: ‘If not, then this list could become evidence of a later occupation of the Qumran caves in the wake of the destruction of the settlement in 68 ce.’ If this explanation is adapted, it may imply that this text is irrelevant to our analysis of the use of Greek within the Qumran community. Beyond the enigmatic Greek 4Q350, the Qumran corpus bears a clearly religious character with regard to both the Hebrew/Aramaic texts and the Greek documents. Alongside the Hebrew biblical texts, the following Greek literary texts have been found, mainly containing Greek Scripture. One such text was also found in Na˙al Óever. 16 I. Gallo, Greek and Latin Papyrology (Classical Handbook 1; London: Institute of Classical Studies, University of London, 1986) 10 i; M. Manfredi, “Opistografo,” Parola del Passato 38 (1983) 44-54. 17 See B. Porten and A. Yardeni, Textbook of Aramaic Documents from Ancient Egypt, vol. 3 (Jerusalem: Akademon, 1993). Occasionally even a biblical text was re-used, as theGreek P. Leipzig 39 of Psalms (4 ce) has a list on the reverse. Nov/947/1-11 1/2/1904 6:34 AM Page 9 greek texts from the judean desert 9 2. Greek texts in the Judean Desert The discovery of Greek biblical texts in caves 2, 4 and 7 at Qumran as well as in Na˙al Óever probably implies that these texts were owned by the persons who brought them to these sites. We do not know to what extent the scrolls were also used by them, but some compara- tive evidence is available regarding the use of the Greek language in the same archaeological environment. Thus, in Na˙al Óever many Greek documentary texts have been deposited (see DJD XXVII), show- ing that Greek was in active use, and hence the find of 8ÓevXII gr causes no surprise. The nature of the Jewish kaige-Theodotion revi- sion18 contained in this scroll fits what is otherwise known about the persons who deposited texts in Na˙al Óever at the time of the Second Jewish Revolt. On the other hand, active use of the Greek versions of the Pentateuch at Qumran is unlikely, as virtually no Greek docu- mentary texts have been found there. The opisthograph 4QNarrative Work and Prayer (4Q460) in Hebrew, with a documentary Greek text 4QAccount gr (4Q350) on the verso of frg. 9 is unique, but possibly irrelevant as the Greek text may have been written after the period of occupancy of Qumran by the Qumran community. The fact that the Greek Scripture texts found in cave 4 in Qumran are from the Torah only may be relevant to our understanding of the distribution of that text and of the community’s interest. The identity of many of the texts from cave 7 is unclear. Greek was in active use in all sites in the Judean Desert, showing an administration conducted in Greek and letters written in that lan- guage, with the exception of Qumran. The percentages of Greek texts compared with Semitic texts found in these sites is much larger than that of the Greek texts found at Qumran. The text of the Greek Bible If de Lagarde’s theory on the history of the LXX needed any further support, it is provided by the texts from the Judean Desert. The newly found texts share important details with the manuscript tradition of the LXX known so far, so that all the known Greek texts reflect one single translation, rather than different translations, as suggested by 18 See D. Barthélemy, Les devanciers d’Aquila (VTSup 10; Leiden: Brill, 1963) and Tov, DJD VIII. For further studies, see C. Dogniez, Bibliographie de la Septante 1970- 1993 (VTSup 60; Leiden: Brill, 1995). Nov/947/1-11 1/2/1904 6:34 AM Page 10 10 emanuel tov Kahle.19 Two of the Qumran texts probably reflect the Old Greek translation better than the manuscript tradition contained in the later uncial manuscripts (4QLXXLeva, 4QpapLXXLevb; the evidence for 4QLXXNum is less clear). Furthermore, the transliteration of the tetra- grammaton in 4QpapLXXLevb as IAV presumably represents an earlier text as well.20 By implication, these three texts should also share cer- tain features, but the evidence is too limited. The differences between the Greek texts from Qumran and Na˙al Óever are remarkable. Two of the texts from Qumran provide insights into the early history of the LXX as they are probably better repre- sentatives of the OG than the later uncials. On the other hand, 8ÓevXIl gr, an early Jewish revision of the LXX, belonging to the kaige-Th group, represents a translation which is typologically later than the uncials and early papyri of the LXX, even if the particular copy found in Na˙al Óever is earlier than most surviving representatives of the LXX. The differences between the types of Greek text found in the two localities reflects the different nature of the groups who deposited the texts there. The status of the Greek manuscripts from the Judean Desert thus runs parallel to that of the Hebrew manuscripts from the same area. The Hebrew manuscripts from Qumran reflect a variety of textual forms, among them proto-Masoretic texts, while those of the later sites of Na˙al Óever, Wadi Sdeir, Murabba'at, and Na˙al Íe"elim (as well as the earlier site of Masada) exclusively reflect the proto-Masoretic texts (also named proto-rabbinic texts) later to be contained in MT. To be precise, the texts from the sites other than Qumran are closer to the medieval text than the Qumran proto-Masoretic texts.21 Similarly, at least some of the Greek texts from Qumran probably reflect an ear- lier form of Greek Scripture, while 8ÓevXII gr reflects a later Jewish revision deriving from proto-rabbinic Jewish circles. Both the Hebrew 19 The argumentation was used already by Leaney, “Greek Manuscripts,” 293 and Skehan, “The Biblical Scrolls from Qumran and the Text of the Old Testament,” BA 28 (1965) 87-100, esp. 91-92. 20 The full evidence, included in a paper read in 1998 at a conference organized by the Scriptorium in Hereford, England, is to be published by the British Library. 21 The evidence is described in several monographs. For my own summary, see “The Significance of the Texts from the Judean Desert for the History of the Text of the Hebrew Bible—A New Synthesis,” in: F. H. Cryer and Th. L. Thompson (eds.), Qumran between the Old and the New Testament (Copenhagen International Seminar 6; Journal for the Study of the Old Testament, Supplement Series 290; Sheffield 1998) 277-309.