K.A. Kitchen, “The Aramaic of Daniel,” D. J. Wiseman, ed., Notes on Some Problems in the Book of Daniel. London: The Tyndale Press, 1965. pp. 31-79. The Aramaic of Daniel K. A. Kitchen* [p.31] INTRODUCTORY This subject has been closely studied by two or three generations of modern scholars1—S. R. Driver,2 R. D. Wilson;3 G. R. Driver,4 W. Baumgartner,5 H. H. Rowley;6 J. A. Montgomery,7 H. H. Schaeder,8 F. Rosenthal,9 and various others. Nevertheless, there is today ample scope for reassessment. The inscriptional material for Old and Imperial10 Aramaic and later phases of the language is constantly growing. One need only mention the Brooklyn and * My thanks go to Mr. David Clines, Department of Biblical History, University of Sheffield, for kindly reading the manuscript of this paper in the light of his own investigations, and for comments and corrections; also to Mr. Alan Millard, Librarian at Tyndale House, Cambridge, for some additional references. [p.32] Borchardt-Driver documents published in 1953 and 1954 or the Aramaic documents from Qumran and other cave-sites of Graeco-Roman Palestine. Furthermore, some earlier views require revision in the light of facts hitherto unknown or neglected. In dealing with the book of Daniel, theological presuppositions are apt to colour even the treatment and dating of its Aramaic.11 The only fair way to proceed is to leave open the whole period c. 540-160 BC until the end of any inquest on the Aramaic, as far as date is concerned. 1 See F. Rosenthal, The Aramaistische Forschung seit Th. Nöldeke’s Veröffentlichungen (1939, repr. 1964), pp. 60-71. Throughout this paper no attempt is made to give the luxuriant bibliography of Aramaic either inside or outside of Daniel. Besides Rosenthal down to 1938-1939, cf. (e.g.) J. J. Koopmans, Aramäische Chrestomathie, I-II (1962). 2 LOT9, pp. 502-504, 508, and xxxiv-xxxviii; S. R. Driver, Daniel (Cambridge Bible, 1900), pp. lix-lx. 3 In Biblical and Theological Studies by the Members of the Faculty of Princeton Theological Seminary... (1912), pp. 261-306. 4 JBL, XLV, 1926, pp. 110-119, 323-325. 5 ZAW, XLV, 1927, pp. 81-133. 6 AOT and indirectly HSD. In English, at least, Rowley’s book of 1929 is a classic statement from the point of view of a second century date for the Aramaic of Daniel, with a wealth of data, and was the last substantial work on the subject to appear in English. Therefore his book has been taken as a convenient starting-point at certain junctures in this study. While Rowley’s failure adequately to recognize the distinction between orthography and phonetics (and to take any note of relevant ancient literary methods) raises grave doubts of his results, it is a pleasure to acknowledge his convenient and careful marshalling of so much of the basic material at issue. 7 J. A. Montgomery, The Book of Daniel (International Critical Commentary, 1927), pp. 15-20. 8 Iranische Beiträge I, 1930, being Heft 5, 6. Jahr of the Schriften der Königsberger Gelehrten Gesellschaft (Geisteswissenschaftliche Klasse). Pages are quoted both by Heft and as book. 9 See work cited n.1, above. 10 I use ‘Imperial Aramaic’ merely as one possible English equivalent among Several others for the usual German term Reichsaramäische. 11 Thus, on ‘inhaltlichen Kriterien’, ‘content’ (Rosenthal, Aramaistishe Forschung, p. 71 apud J. Lindner, Zeitschrift für Katholische Theologie, LIX, 1935, pp. 503-545; H. L. Ginsberg, JAOS, LXII, 1942, p. 231 end, apud Schaeder), i.e. prejudging the book a priori as pseudepigraphic. Wilson (n. 3, above) and Rowley (n. 6, above) represent opposite viewpoints. K.A. Kitchen, “The Aramaic of Daniel,” D. J. Wiseman, ed., Notes on Some Problems in the Book of Daniel. London: The Tyndale Press, 1965. pp. 31-79. In this study, the Aramaic of Daniel is examined compactly in relation to (a) vocabulary, (b) orthography and phonetics, and (c) general morphology and syntax. This enquiry is necessarily limited to the main points at issue, without digressing into secondary literature or investigating the details of all the older discussions. It is not a total presentation of work done or of material available; the entire analysed word-list of Biblical Aramaic is omitted, and also detailed references to the Nabataean and Palmyrene data (sparse in any case) and Targums.12 A. VOCABULARY 1. Basic West Semitic. For the purpose of this study, the entire word-stock of Biblical Aramaic, in particular of that in Daniel, has been listed and analysed against the comparative background of objectively-dated inscriptions and papyri in Old and Imperial Aramaic, and cognate West Semitic and Akkadian (Babylonian and Assyrian).13 The result is that nine- tenths of the vocabulary is attested in texts of the fifth century BC14 or earlier. The slender one-tenth remaining consists of words so far found only in sources [p.33] later than the fifth Century BC (e.g. Nabataean, Palmyrene or later Aramaic dialects), or so far not attested externally at all. The meaning of these facts in either case is clear. Among the nine-tenths, words found in Old and Imperial Aramaic documents in the ninth to fifth centuries BC would in themselves allow of any date for the Aramaic of Daniel from the sixth century BC onwards. Words found in other early West Semitic texts15 as well as in the Aramaic of Daniel are to be taken as being not merely early Aramaic but as common, early West Semitic, not even peculiar to Aramaic. Words found in Akkadian may be relevant in one of three or four main ways: they may be taken as Common Semitic; they may indicate old West Semitic terms borrowed by Akkadian (e.g., as in the Mari texts of the eighteenth century BC);16 they may represent Aramaic words so far only attested as loan-words in later Akkadian (Neo-Assyrian and Neo-Babylonian, eighth-fifth centuries BC), or else may be Akkadian vocables borrowed as loan-words by Old or Imperial Aramaic and so turn up in our text. Whichever the individual case, the result ultimately is the same—all such words could obviously occur at any time from the sixth century BC onwards. 12 A sufficiency of this later material is included in AOT, passim. 13 Apart from the Ararnaic papyri (which also contain Persian words), what follows principally concerns the main Semitic word-stock of the Aramaic of Daniel. It is regretted that the detailed analysis of the vocabulary of Biblical Aramaic cannot be included with this paper—even set out compactly, it would require up to forty pages in small quarto. 14 The accidents of preservation and discovery have so far produced very many more Aramaic papyri for the fifth century BC than for the sixth, the upper time-limit for Daniel. However, attestation of a Semitic word in a fifth- century document is ordinarily a sufficient presumption for its existence and use in the late sixth century BC. The onus of proof would lie on anyone who might prefer to believe the contrary. The sixth-century P. Meissner differs in no essentials from fifth-century documents. 15 Including: Ugaritic (fourteenth-thirteenth centuries BC), Canaanite glosses in the Amarna tablets (fourteenth century BC), Phoenician, Hebrew and Moabite inscriptions (tenth century onwards). 16 Good examples, cf. E. A. Speiser, JBL, LXXIX, 1960, pp. 157-163; M. Noth, Die Ursprünge des alten Israel im Lichte neuer Quellen (1961), pp. 14-22, 34-40 (on which see D. O. Edzard, ZA, LVI, NF XXII, 1964, pp. 142-149); A. Malamat, JAOS, LXXXII, 1962, pp. 143-150. K.A. Kitchen, “The Aramaic of Daniel,” D. J. Wiseman, ed., Notes on Some Problems in the Book of Daniel. London: The Tyndale Press, 1965. pp. 31-79. Turning to the one-tenth of unattested and so far late-attested words, two interpretations are theoretically possible for the latter class. One may assume that they are genuinely ‘late’ words, and therefore bring down the date of the Aramaic of Daniel—or of their introduction into that Aramaic—to a suitably late date. One may with equal justification assume that in fact this is the familiar vicious fallacy of negative evidence17—that these words may yet turn up in ‘early’ sources, so disproving their apparently late date.18 Otherwise-unattested words, of course, cannot serve as [p.34] dating evidence at all. Where nine-tenths of the vocabulary is clearly old-established (fifth century BC and earlier), it is a fair assumption that the lack of attestation of the odd tenth represents nothing more than the gaps in our present knowledge—gaps liable to be filled by new material in the course of time.19 Hence, as far as the main (i.e. Semitic) vocabulary is concerned, we have no warrant whatever to draw any conclusion about the date of Daniel from its Aramaic except to say that any date from the sixth century BC onward is possible.20 2. Hebrew and Akkadian Loan-words. However, separate consideration must be given to foreign loan-words in Biblical Aramaic. These are attributable to Hebrew, Akkadian, Persian and Greek sources. Hebraisms in Biblical Aramaic require no consideration—its writers were Hebrews, one expects them to betray this, and such Hebraisms have no bearing on date. Akkadian words in Aramaic are also of very limited value in this connection, for Aramaean penetration of Mesopotamia from the Middle Euphrates down to southern Babylonia persisted from the twelfth century BC (Tiglath-pileser I)21 for six or seven hundred years before the period of the late sixth-mid-second centuries BC that is our concern here. During that long period there was ample time for Aramaic words to enter Akkadian, and for Akkadian words to penetrate Aramaic. From the eighth century BC, Aramaic enjoyed official recognition in 17 See my Ancient Orient and Old Testament, section I.B.5 (iib). It is no unusual occurrence in Egyptian, for example, for words once known only from texts of the Graeco-Roman age suddenly to turn up a millennium or two earlier in Old, Middle or Late Egyptian texts. For Semitic, cf. (e.g.) the next notes. 18 A random example or two from Semitic must suffice here. Sixty years ago, hðmr ‘wine’, in Biblical Aramaic and Hebrew could be dismissed as ‘late’ in the latter (cf. BDB, p. 1093a, s.v.). But the word is now attested in Ugaritic (fourteenth-thirteenth century BC; UM, III, no. 713) and also in Mari Akkadian (M. L. Burke, Archives Royales de Mari, XI (1963), p. 133, §11). Or take šÖpr, ‘be fair, acceptable’, in the Aramaic of Daniel, also in Hebrew, labelled as ‘rare and mostly late’ (BDB, p. 1117, s.v.). This word occurs at least twice in the Ahiqar papyrus of the fifth century BC (lines 92, 108 and perhaps 159), once in the mid eighth century BC (Sfiré stela III, line 29), and as a proper name (like Shiprah of Ex. 1:15) back in the eighteenth century BC in the slave-list of Papyrus Brooklyn 35.1446 (W. C. Hayes, A Papyrus of the Late Middle Kingdom (1955), p. 96; W. F. Albright, JAOS, LXXIV, 1954, p. 229). Rare it may be; late, it is not. 19 Cf. previous note; a good example is the supposedly late yt of Biblical Aramaic turning up in an Aramaic papyrus published in 1953, cf. below, p. 69. 20 When such vocabulary is common to both ‘early’ sources (Old Aramaic; papyri) and to ‘late’ ones (e.g. Targums), this does not a priori date Daniel any closer to (say) the Targums than to the Old Aramaic texts or the papyri; it could be at either end, or anywhere in between, or even slightly before or after (seen strictly from this vocabulary point-of-view). 21 See E. Forrer in E. Ebeling and B. Meissner (eds.), Reallexikon den Assyriologie, I, 1928, pp. 131-139: Aramu. K.A. Kitchen, “The Aramaic of Daniel,” D. J. Wiseman, ed., Notes on Some Problems in the Book of Daniel. London: The Tyndale Press, 1965. pp. 31-79. Assyria.22 Thus, if one combines the words of possible Akkadian origin given by Rowley in AOT, pp. 134-135, with those given by Rosenthal, GBA, §§188, 190 (pp. 57-58, 59), one obtains a total of some 37 words and phrases, [p.35] including bel-t»’m (but not t»’m) and ebr-nhr as phrases. One of these 37 is probably not Akkadian (’usŒsŒarna, in Ezra only) but Old Persian (with Rosenthal, GBA, §189). Of the remaining 36 words, 15 occur only in Ezra, leaving 21 in Daniel, 16 of them in Daniel alone. The vocabulary analysis already mentioned would indicate that practically all of these 21 words are either attested in Imperial Aramaic well beyond Daniel, or have a long history prejudicial to their entering Aramaic only in the sixth century BC (or later) or occur already in Old Aramaic or West Semitic, or in Hebrew outside Daniel (ruling out special usage there). In other words, the Akkadian loan-words are probably simply part of the multicoloured fabric of Aramaic, and have no real bearing on the date of the language of Daniel within the sixth to the second centuries BC. Only ’aŠsŒepa, ‘enchanter’ (Akkadian asŒipu), is restricted so far to the Hebrew (1:20; 2:2) and Aramaic (2:10, 27; 4:4; 5:7, 11, 15) of Daniel, outside of Akkadian itself, and is not attested (so Rowley) in the later Aramaic of, for example, the Targums. This might speak for a direct loan from Akkadian into the Hebrew and Aramaic usage of a Hebrew in Babylon in the mid-sixth century BC—but it is so isolated that as a single word it cannot constitute proof. A single occurrence of ’sŒp in some future discovery of a West Semitic text or Aramaic document would soon dispel any such assumption. Hence, for dating purposes, the Akkadian words in the Aramaic of Daniel must be accorded the same status of non liquet as the rest of the Semitic vocabulary. 3. Persian Loan-words. The Persian words in the Aramaic of Daniel are some 19 in number. Rowley gave 20 such words.23 From his list, zmn and srbl should probably be omitted,24 and to it pt»ysŒ (a garment) be added.25 With these changes, 13 out of the 20 words are attested in later literature, particularly the Targums,26 while [p.36] 22 See R. A. Bowman, JNES, VII, 1948, pp. 75-76; in 729 BC, tribute for Tiglath-pileser III is shown as recorded by an Aramaean scribe with pen and parchment alongside his cuneiform colleague; note also for 701 BC, 2 Ki. 18:13-37, Is. 36:1-22. 23 AOT, p. 138. 24 Zmn, ‘time’, is probably derived from Akkadian simānu (refs. in KB, p. 1972a), rather than from Old Persian zurvan (m = rv seems far-fetched). Srbl, ‘trousers’ (?), is of obscure origin, but known to the fifth-century papyri (AP, no. 42:9). It can hardly be native to Old Persian, as the latter does not use l except in foreign words that already contain l, and only in those not assimilated (OP, p. 8, §6, and especially p. 33, §107; cf. W. Brandenstein and M. Mayrhofer, Handbuch des Altpersischen (1964), pp. 32-33, 35, §§28, 32). 25 With F. Rosenthal, GBA, p. 59, §189; KB, p. 1112b, after Nyberg. 26 Rowley’s twelve words are reduced to ten by omission of zmn and srbl; but these losses are more than made up by (i) pt»ysŒ which also occurs in Talmudic literature (BDB, p. 1108), (ii) gdbr which is merely a variant of gzbr and should not count separately (see below, pp. 61-62), and (iii) dt, ‘law’, which is now attested in the Qumran Targum to Job (see J. van den Ploeg, Le Targum de Job de la Grotte II de Qumran (1962), p. 7—where ptgm also occurs). K.A. Kitchen, “The Aramaic of Daniel,” D. J. Wiseman, ed., Notes on Some Problems in the Book of Daniel. London: The Tyndale Press, 1965. pp. 31-79. 6 of them so far are not.27 (In the Aramak of Ezra, Rowley listed 9 words, which would become 8 with the omission of zmn as in Daniel. But to this 8 may be added a further 5 words (with Rosenthal, §§189-190 et al.): ’sprn’, ’prs(t)k, ’šrn’, ’štdr, šršw. Of these 13 words, 5 occur in later (e.g. Targumic) literature as survivals, that is, Rowley’s five, minus zmn, but plus dt instead.) In the Aramaic papyri collected together by Cowley,28 Rowley with him would find (p. 139) some 26 Persian words of which only 2 occur in the Targums. He further noted that only 2 of these words (t(y)pt, zn) recurred in Daniel; and 3 (’zd’, gnz, nštwn) in Ezra, probably 4 (adding ’šrn’),29 as preferred here. All this led him to observe that (p. 139) ‘a very large proportion of the Persian words found in Daniel is known to have persisted in Aramaic until Targurnic times’ [i.e. until the first century BC and later], ‘while a very small proportion of those found in Egyptian Aramaic so persisted’. So far so good, on the facts so far adduced. Rowley then interpreted these facts as follows (p. 139): ‘In this matter [survival of words in Daniel], therefore, Biblical Aramaic... stands very much nearer to the Targums than do the Papyri.’ And later in the same paragraph, ‘It thus appears that in the matter of Persian loan- words Biblical Aramaic is also very much nearer to the Targums than it is to the Papyri.’ These inferences are open to question on various grounds. (1) As generally admitted, the Persian vocabulary in the Aramaic of Daniel amounts to barely a score of words. When the material at hand consists of words by the thousand or even in hundreds, there is some hope of sound results, but a mere score or so of words is altogether too fragile a basis for statistical argument.30 This will become more evident from what follows. (2) If one compares the Persian vocabulary of Biblical Aramaic with what the Targums contain—as Rowley does in the interests of a second-century date for Daniel—one should also compare that vocabulary equally with what is found in Imperial Aramaic docu- [p.37] ments of the sixth-fifth centuries BC.31 This omission must now be rectified. Of the 19 words here accepted as Persian in the Aramaic of Daniel, 8 or g occur in Imperial Aramaic and contemporary sources. In Imperial Aramaic, we find ptgm, ’zd’, zn, t(y)pt, gzbr,32 and [rz]33 (all in the Elephantine papyri from Egypt); and via Elamite,34 ’h©šdrpn, dt, 27 The proportion is thus now 12:7 instead of 12:8 as in Rowley’s time. More Qumran material could easily increase the body of such survivals; the new Job Targum contains some other Persian words, e.g. dh©št, ‘desert’, known neither to Biblical Aramaic nor to the papyri as yet (van der Ploeg, loc. cit.). 28 AP. 29 See for this word C. G. Tuland, JNES, XVII, 1958, pp. 275-274, and as Old Persian, H. S. Nyberg, Le Monde Oriental, XXIV, 1930, pp. 538-139. But cf. also E. G. Kraeling, BMAP, p. 101, n.6. 30 A fact realized at least in part by Rowley (p. 136: ‘While in literature so scanty as our texts, all arguments on Vocabulary are liable to be precarious...’), but not sufficiently. It is clearly stated by F. Rosenthal, Aramaistische Forschung, p. 63, in another connection. 31 In practice, mainly of the fifth century BC, due simply to the accidents of preservation and discovery. 32 Of which gdbr is merely a variant; see below, pp. 61-62. 33 Ahiqar, 141, restored—but what else of two letters would fit? If this word be omitted, then we have eight, not nine words here. K.A. Kitchen, “The Aramaic of Daniel,” D. J. Wiseman, ed., Notes on Some Problems in the Book of Daniel. London: The Tyndale Press, 1965. pp. 31-79. Dt, ‘law, decree’, also occurs in Akkadian cuneiform as early as the second, sixteenth and thirty-fifth years of Darius I, and databara in a later group of documents.35 In other words, nearly half of the Persian words in the Aramaic of Daniel are attested (mainly in Aramaic itself) in the sixth-fifth centuries BC. Or, nearly half of the Persian words would speak just as much for a sixth-fifth century date as would the 13 words found in the Targums for a second- century date. Now, of the 8 or 9 words here listed as occurring early, 4 or 5 (ptgm, gzbr, dt, zn, [rz]) recur in the Targums and 4 (’zd’, t(y)pt, ’h©šdrpn, dtbr) do not. In other words, on Rowley’s kind of reasoning, half of these 8 or 9 words would stand nearer to the Targums and half to the sixth-fifth centuries BC―but as those recurring in the Targums also occur in documents of the sixth-fifth centuries BC, could these documents, too, stand ‘nearer’ to the Targums than to other documents of their own date? Clearly they could not; the plain fact is that all of these 8 or 9 words came into Aramaic (and Akkadian and Elamite) in the sixth-fifth centuries BC, and some of them happen to be retained four centuries later (cf. (4) below). The occurrence of 4 or 5 of them in both Persian imperial documents and Targums merely leaves the date of Daniel’s Aramaic where it was before: in the sixth to second centuries BC. (Similarly with Ezra’s Aramaic. Of the 13 words here, 9 occur in Imperial Aramaic or contemporary sources (’ptm, perhaps in Akkadian), only 4 so far do not (’prs(t)k,36 pršgn, ’drzd’, ’štdr). That [p.38] is, three-quarters of the Persian words in the Aramaic of Ezra are attested in the sixth-fifth century BC documents available, and speak for an early date. Of these 9 words, 4 recur in the Targums (ptgm, gnz, gnzbr, dt) but 5 do not (’sprn’, ’šrn’, nštwn, šršw, ’ptm).37 In other words, again, half of the words are ‘early’ only, and half both early and late, with the same result as in Daniel, i.e. sixth-fifth to second centuries BC for scope in dating.) (3) As for the Aramaic of Daniel having only 2 Persian words (and that of Ezra, 4)38 in common with the Aramaic of the fifth-century papyri from Egypt, much new material has become available since 1929, and permits of some interesting comparisons. In the eighty-seven documents collected by Cowley,39 some twenty-seven words were isolated as probably Persian. Since then, two further (but much smaller) collections of Aramajc documents have been published, by Kraeling in 195340 and Driver in 1954-1957,41 besides a trickle of lesser items.42 34 Both under Darius I; for the view that treasury-orders at Persepolis were turned into Aramaic and then Elamite (hence Old Persian words in these texts would come via Aramaic), see PTT, chapter 2, especially pp. 27 ff., nuanced by G. G. Cameron, JNES, XVII, 1958, p. 163. 35 CAD, 3/D, pp. 122-123. These, too, are likely to be indirect evidence for dt, dtbr in Aramaic in so far as the latter language was used as intermediary between Persian officials and cuneiform scribes in Babylonia. Dāta is, of course, directly attested from Darius I onwards in actual Old Persian texts (OP, p. 189). 36 This word is probably also attested within the Persian period in Akkadian as iprasakku (CAD, 7/I-J, pp. 165- 166). This would give Ezra fourteen words attested early and only three not (so far). 37 One word (pršgn) so far only in Ezra and Targums (apart from biblical Hebrew, left out of account in this study, except where cited). 38 AOT, p. 139, but including ’šrn’; in fact, the proportions are quite different: Daniel, five or six words in the papyri and three more in parallel material; and Ezra, eight in the papyri and one probably in the parallel material—cf. above, pp. 37f. 39 I.e. AP, nos. 1-83, plus Ahiqar, Behistun-text, and two documents (A; B + C) in Appendix, pp. 317-319. 40 BMAP. K.A. Kitchen, “The Aramaic of Daniel,” D. J. Wiseman, ed., Notes on Some Problems in the Book of Daniel. London: The Tyndale Press, 1965. pp. 31-79. The Kraeling series comprises seventeen documents (four very fragmentary), practically all legal texts. As might be expected, they contain numerically fewer Persian words than the Cowley corpus―11, of which (’bygdn, hnbg’, hngyt, ’šrn’) occur in Cowley, 1 in Driver and Cowley (the ubiquitous krš), 1 in the Talmud,43 and 5 (hpth©pt, hnpn’, ’drng, drmy,44 ’zt)45 nowhere else. In even so few documents, half the Persian words are otherwise unattested in contemporary records! The Driver series consists of thirteen letters and various lesser fragments. The bulk of these letters came from Babylon (and perhaps Susa) to Egypt, sent by the Persian Arsames, satrap of Egypt (while absent from Egypt) and three of his adjutants. This small [p.39] group of documents contains no less than 26 Persian words46—three-quarters as many as the entire Cowley collection of eighty-seven documents! Nor is this all. Of those 26 words, 19 are entirely flew to Imperial Aramaic, 2 are shared with Cowley and Kraeling (krš, pt), and only another 3 words with the Cowley series (gnz, hmrkry’, hndyz), and 2 more in other sources (’sprn, ptgm). One of the new words (srwšyt’, ‘punishment’) recurs in a slightly different spelling (šršw, probably to be taken as šrśw originally, initial śin assimilated to shin?) in Ezra (7:26, ‘banishment’).47 Five others of the new words recur in the Talmud (bg, ‘domain’; dšn’, ‘gift’; ptkr, ‘image’; gnz, ‘treasury’; ptgm, ‘decree’48). In other words, the Driver documents show as many affinities in their Persian vocabulary with the Talmud (3 words; plus gnz and ptgm attested early and late) as with other records of the fifth century BC (five items in Cowley and Kraeling)—but no-one would use such statistics to prove that the Driver documents should be placed half-way chronologically between the fifth and second centuries BC (c. 250?) on such a basis!49 Ezra is in a far more convincing position with his 9 Persian words in sixth-fifth century sources out of only 13 (and 1 word unattested so far except in Biblical Hebrew and the Targums), than are the Driver documents with only 5 or 7 contemporaneously attested Persian words out of 26 (and 3 words unattested elsewhere until the Talmud) Likewise, Daniel with 8 or 9 Persian words in sixth-fifth century sources (5 or 6 actually in the papyri) out of 21 words (and 7 words unattested until the Targums) compares perfectly well with the 5 or 7 words in the sixth-fifth centuries out of 26 (3, only in Talmud) of the Driver documents—these latter have less in the late sources only, but less in the early documents also! The whole of this section (3) simply throws into relief the following facts. (i) With only a score or so Persian words in each writing or group of documents, statistics are virtually worthless. (ii) The supposedly few Persian words common to the Aramaic of Daniel and Ezra and that of the eighty-seven papyri in Cowley prove only that our knowledge of the total 41 G. R. Driver, Aramaic Documents of the Fifth Century BC (1954) and abridged and revised version 1957). 42 And others yet to be published, e.g. the fifth-century Aramaic papyri from Tuna el-Gebel (Hermopolis West) in Middle Egypt (cf. M. Kamil, Revue de l’Histoire Juive en Égypte, I, 1947, pp. 1-3). 43 Mgwš, ‘magian’, in AP, p. 254: ‘Behistun text’, line 60, cf. M. Jastrow, Dictionary of the Targumim, Talmud… II 1926, p. 727a/b. 44 Unless this word is Greek? Cf. below, pp. 46, 47. 45 This word, J. De Menasce, BO, XI, 1954, p. 161; LAP, p. 38. 46 Excluding, as throughout this discussion, personal and place names. 47 N.B. šrōši in Qrē—from śrwšy(t)? 48 M. Jastrow, Dictionary of the Targumim, Talmud… I, pp. 134b, 326b, II, p. 1254b. 49 And it is as striking that the Kraeling legal documents have barely half their Persian vocabulary in common with the other Elephantine documents collected in Cowley. K.A. Kitchen, “The Aramaic of Daniel,” D. J. Wiseman, ed., Notes on Some Problems in the Book of Daniel. London: The Tyndale Press, 1965. pp. 31-79. impact of Old Persian upon Imperial Aramaic (and its continuations) is grossly inadequate, when one small group of closely similar documents yields 50 per cent new [p.40] Persian loan-words, and another and separate group (from the East) 19 out of 26 words new (c. 80 per cent). (iii) It should be noted that in fact several more words in the Aramaic of both Daniel and Ezra are common to them and the Aramaic papyri (besides other documents) than was allowed by Rowley in 1929. (iv) When words are attested outside of Daniel (or Ezra) both in the sixth-fifth century texts and in the late (Targumic/Talniudic) sources, this proves only that these words had along life in Aramaic, and in themselves leave open the whole period of the sixth-second centuries BC for Daniel. Words attested only in Daniel (or Ezra) and, say, the Targums can be balanced by other words occurring , only in Daniel (or Ezra) and the sixth-fifth century documents—each represents merely negative evidence for the periods unattested, and hence is useless for specific dating purposes; they cancel each other out. This matter of survivals must now be further considered. (4a). From the preceding, it should be plainly evident that Persian words in Daniel that survive in Targums or Talmud prove only that the words in question could have been used as easily by a writer of Aramaic in the second century BC as by one in the sixth or fifth century BC. It is unjustifiable to infer therefrom (with Rowley, p. 139) that Biblical Aramaic is ‘nearer’, i.e. chronologically, to the Targums than to the papyri. The numerical ‘prepon- derance’ of 13 such words in Daniel as opposed to 5 in Ezra or 7 in Cowley, Kraeling and Driver has thus no necessary bearing on the date of Daniel at all. Words must be weighed, not merely counted.50 As already stated, the impact of Old Persian upon Imperial Aramaic was considerable. The Persian kings appointed Persian and Median officials to govern their empire, and Aramaic was the means of communication between these and the polyglot nations so ruled. In the administrative sphere, the impact was intense—note the list of about 100 Old Persian words in the eighty-four Elamite ‘money-order’ tablets from the Persepolis Treasury published by Cameron,51 to which still more may be added,52 not least when 2000 more Persepolis tablets (fortifications archive) are eventually published.53 The effect on Aramaic must have been the greater, as its use was infinitely wider than Elamite. In the two centuries between c. 540 BC and c. 330 BC many such words had [p.41] ample time not merely to enter Aramaic but to become a regular part of it, assimilated in fact. Thus, when Alexander and his Macedonians supplanted the Persian rulers, Aramaic-speaking peoples would continue to use those Persian words that had lodged securely in Aramaic usage. A fair number of words would drop away—those for institutions and practices that ceased with Persian rule, or received new, Greek names under the Macedonian kings, for example. 50 Cf. F. Rosenthal, Aramaistische Forschung, p. 63. 51 PTT, Texts 1-84 (no. 85 is an Akkadian ‘stray’), and list of Old Persian words on pp. 42-43. 52 E.g. nidani, pitgam, karnuvaka, ras©akara, garda (± pati), fratama, pansukaš (G. G. Cameron, JNES, XVII, 1958, pp. 161-162, 165, n.8, 9). 53 By R. T. Hallock; cf. JNES, XVII, 1958, pp. 256-257. K.A. Kitchen, “The Aramaic of Daniel,” D. J. Wiseman, ed., Notes on Some Problems in the Book of Daniel. London: The Tyndale Press, 1965. pp. 31-79. The scope of Persian words borrowed during the Persian Empire must also have been very broad—from specialized technical terms and titles from administration, law and military, through names of specific ‘cultural elements’ (clothes, materials, etc.) down to apparently ordinary sorts of words like zn, ‘kind’, ptgm, ‘message’, etc.54 Furthermore, the almost unconscious assumption55 that Persian words would take some time to penetrate into Aramaic (i.e. well after 539 BC) is erroneous.56 The Persepolis Treasury tablets run from the 30th year of Darius I to the 7th of Artaxerxes I, while those yet to be published from the fortifications archive go back to the 11th to 28th years of Darius 1.57 (Akkadian was much less receptive to Old Persian loan-words (so far as is known), but as already seen, dāta, ‘law, decree’, occurs in years 2, 16 and 35 of Darius 1.58) Had we similar Aramaic and Elamite documents for Cyrus and Cambyses, the same result might be anticipated. In other words, if a putative Daniel in Babylon under the Persians (and who had briefly served them) were to write a book some time after the third year of Cyrus (Dn. 10: 1), then a series of Persian words is no surprise. Such a person in the position of close [p.42] contact with Persian administration that is accorded to him in the book would have to acquire—and use in his Aramaic—many terms and words from his new Persian colleagues (just like the Elamite scribes of Persepolis), from the conquest by Cyrus onwards. (4b). It is necessary, also, to note what words are involved in practice, and not merely how many. Given the two centuries of unhindered Persian penetration of the Aramaic language (c. 540-330 BC), one cannot be surprised that such ‘ordinary’ words as zn, ‘kind’; rz, ‘secret’; nbršt, ‘lamp’; ndn, ‘sheath, body’; nbzb’, ‘gift’, should be assimilated and survive later. The same applies to items of apparel (hnmk, ‘necklace’; pt£yš, ‘shirt’(?)), and words of such wide application as ptgm, ‘word’; dt, ‘law, decree’. Hdm, ‘limb’, came from legal usage59 (and gnz, ‘treasure’, in Ezra, from administration), as in Daniel, but is an ordinary word in the Targums. The survival of these 10 words (plus one in Ezra) can of itself prove nothing about Daniel’s date. Three words remain over (the first in early sources as well as the Targums): gzbr/gdbr, kr(w)z, srk: ‘treasurer’, ‘herald’, ‘chief minister’. These are administrative titles—but for obvious and basic functions. Any Ancient Near Eastern state had treasurers and chief ministers, and courts their heralds also.60 The non-occurrence so far of srk and krz60a in early 54 Elamite in the Persepolis Treasury texts similarly borrowed quite ‘ordinary’ words as well as technical and other terms (PTT, p. 19, n.125). It should be noted that these so-called ‘ordinary’ words in most cases probably came in at first as technical or cultural words, but became everyday terms in Aramaic. Thus, ndn, ‘sheath’, may come initially from military parlance; rz, ‘secret’, from magicians’ usage, and so on. I owe comment on this point to Mr. David Clines of Sheffield University. 55 Cf. LOT9, p. 508: ‘The Persian words presuppose a period after the Persian Empire had been well established.’ 56 Note also G. G. Cameron’s remarks (PTT, p. 19) on the penetration of Elamite by so many Old Persian words from as early as Darius I. 57 PTT, pp. 32-33, and R. T. Hallock, JNES, XIX, 1960, p. 91 (year seven for five). 58 CAD, 3/D, p. 123a. Other Old Persian words in Akkadian include also dātabara (ibid., p. 122a), ahãšadrapannu (ibid., vol. 1:1/A:1, p. 195a), hãamarakara (ibid., vol. 6/H, pp. 59-60), gardu, gardupatu (ibid., Vol. 5/G, p. 50), etc. See W. Eilers, Iranische Beamtennamen in der keilschrfftlichen Überlieferung, I (1940). Also the lists of loan-words in reviews of successive volumes of the CAD by D. O. Edzard in ZA, LIII, NF XIX, 1959, LIV, NF XX, 1961, and LVI, NF XXII, 1964. 59 E.g. GBA, §189. 60 E.g. the wh©m-nsw, ‘Royal Herald’, in Egypt (A. H. Gardiner, Ancient Egyptian Onomastica, I, 1947, p. 22*, no. 80). K.A. Kitchen, “The Aramaic of Daniel,” D. J. Wiseman, ed., Notes on Some Problems in the Book of Daniel. London: The Tyndale Press, 1965. pp. 31-79. sources, beyond Daniel and the Targums (likewise pršgn in Ezra), is merely negative evidence (cf. also (5), below). It should also be remembered that survival of words from Daniel or Ezra in the Targums is to be expected a priori—after all, they belong to one literary tradition, Jewish, biblical and commentary/interpretation therefor! On the other hand, 2 words are so far unique to the Aramaic of Daniel: ’drgzr and hdbr, both high titles (‘counsellor’, ‘companion’); and 4 words occur so far only in the Aramaic of Daniel and early (i.e. sixth-fifth century) documents: ’h©šdrpn (‘satrap’), dtbr (‘judge’), t(y)pt (‘magistrate’); ’zd’ (‘certified’, etc.). Here also, 3 are titles and 1 part of official style (cf. Rosenthal, GBA, §189). There is as yet no evidence that any of these 6 terms survived the Persian period (i.e. after c. 330 BC). This in itself is negative evidence, and therefore is inadequate. But there is limited positive evidence in its support, from the LXX (Old Greek and Theodotion).61 Among the official titles in the Aramaic of Daniel (Dn. 3:2-3, etc.), Persian [p.43] ’h©šdrpn and Semitic sgn and ph©h, and the general phrase ‘all the rulers of the provinces’ are reasonably well rendered. But for ’drgzr, ‘counsellor’; gdbr, ‘treasurer’;62 dtbr, ‘law-officer’; t(y)pt, ‘magistrate, police chief’, the Old Greek (and later) renderings are hopelessly inexact— mere guesswork. If the first important Greek translation of Daniel was made some time within c. 100 BC-AD 100, roughly speaking, and the translator could not (or took no trouble to63) reproduce the proper meanings of these terms, then one conclusion imposes itself: their meaning was already lost and forgotten (or, at the least, drastically changed) long before he set to work.64 Now if Daniel (in particular, the Aramaic chapters 2-7) was wholly a product of c. 165 BC, then just a century or so in a continuous tradition is surely embarrassingly inadequate as a sufficient interval for that loss (or change) of meaning to occur, by Near Eastern standards. Therefore, it is desirable on this ground to seek the original of such verses (and hence of the narratives of which they are an integral part) much earlier than this date, preferably within memory of the Persian rule—i.e. c. 539 (max.) to c. 280 BC (allowing about fifty years’ lapse from the fall of Persia to Macedon). At maximum, this could affect the whole book of Daniel as we have it; at minimum, it could indicate that a second-century writer used in his work some pre-existing Daniel-narrations, but adapted them so little that he did not even eliminate words meaningless to him and his readers, such was his archaeological conscience. On the use of Persian words in reference to the Babylonian kingdom, and their preponderance over Greek terms, see section 4(5) below. (5) One further point should be made here: the Persian words in Daniel are specifically Old Persian words.65 The recognized divisions of Persian language-history within Iranian are: Old down to c. 300 BC, Middle observable during c. 300 BC to c. AD 900, and New from c. AD 60a But see now A. Schaffer, Or., XXXIV (1965), pp. 32-34, for evidence of krz about 1500 BC. 61 See J. A. Montgomery, The Book of Daniel, pp. 199-200, on LXX of Daniel. 62 Presumably not recognized by the LXX translator as identical with gzbr. 63 The LXX of Daniel is known to have been a relatively ‘free’ rendering—but if the translator gave reasonable renderings for some terms (e.g. sgn, ph©h, ’h©šdrpn), one would expect him to do this for the rest, had he known them equally well. 64 This point has also been made by W. St. Clair Tisdall, JTVI, LIII, 1921, p. 206. 65 Even when words in the Aramaic of Daniel cannot be compared with attested Old Persian words, they correspond with reconstructible forms in Old Persian, not later forms. Old Median is here kept with Old Persian.