Deuteronomy 14.3–21: An Early Exemplar of Rewritten Scripture?

Juni Hoppe

1. Already Driver noted that there existed ‘no verbal parallels between Dt. and P; much that is of central significance in the system of P is ignored in Dt., while in the laws which touch common ground, great, and indeed irreconcilable, discrepancies often display themselves: hence the legislation of P cannot be considered in any degree to have been one of the sources employed by the author of Dt.’ (Samuel Rolles Driver, A Critical and Exegetical Commentary on Deuteronomy [Edinburgh: T&T Clark, ³], pp. xi–xii). In the years since, the perception that the Priestly legislation and the Deuteronomic code were originally independent sources solidified and has now become commonplace.

See, for example, Julius Wellhausen, Die Composition des Hexateuchs und der historischen Bücher des Alten Testaments (Berlin: De Gruyter, ³1899), p. 203; Abraham Kuenen, Historisch-kritische Einleitung in die Bücher des Alten Testaments hinsichtlich ihrer Entstehung und Sammlung I. Entstehung des Hexateuch (Leipzig: Otto Schulze, ), translated into English as The Origin and Composition of the Hexateuch (London: Macmillan, 1886), pp. 278–79.

Yehezkel Kaufmann, The Religion of Israel: From Its Beginnings to the Babylonian Exile (trans. by M. Greenberg; Chicago: Chicago University Press, ).

Excluding the presumed Holiness School texts, believed to have been incorporated in Lev. 17–26, and so called due to their frequently repeated use of the word ‘holy’. The existence of such a Holiness School is still being debated; among those who posit its existence, there is also a discussion as to whether its texts preceded the Priestly code or were composed later, perhaps under the influence of D. See Israel Knohl, The Sanctuary of Silence. The Priestly Torah and the Holiness School (Winona Lake: Eisenbrauns, 2001); Jan Joosten, People and Land in the Holiness Code (Leiden: Brill, 1996); Jacob Milgrom, Leviticus 1–16. A Translation with Introduction and Commentary (AB 3; New York: Doubleday, 1991). See recently Monroe, who dates the Holiness Code to Josiah’s reign: Lauren A.S. Monroe, Josiah’s Reform and the Dynamics of Defilement: Israelite Rites of Violence and the Making of a Biblical Text (New York/Oxford: Oxford University Press, ), esp. pp. 18–20 and her chapter 5.

In this paper we plan to use the terms ‘redactor/redaction/redactional’ and ‘editor/editing/editorial’ indistinguishably, though we grant that their activities can be distinguished. As one might understand it, a redactor introduces himself into a text to change its meaning while an editor intervenes to clarify and improve what is meant without necessarily intending to change the meaning. The distinguishing line is too fine for us to discriminate in the use of these terms when dealing with activities so far removed in time from us. Furthermore, redactors/editors can but need not be scribes. We reserve the terms ‘scribe’ and ‘copyist’ simply for individuals who copy material without necessarily being motivated to do anything other than transmit texts.

Andrew Teeter, ‘The Hebrew Bible and/as Second Temple Literature: Methodological Reflections,’ Dead Sea Discoveries 20 (), pp. 349–377, here: p. 351. He also states: ‘Even those works demonstrably produced in subsequent generations often represent the nearest accessible external witness to the concerns, techniques, and literary methods of those responsible for the production of the Hebrew Bible on its final levels […] Second Temple Jewish literature such as represented in the Dead Sea Scrolls can thus be seen as offering external validation of internal literary processes long detected by scholars within the biblical texts’ (p. 353).

For example, the redactor of Exod. 20 in the Samaritan Pentateuch was ‘[…] flipping back and forth between the Masoretic Exodus and Deuteronomy, adding or dropping a phrase or detail here and there in an attempt to merge and reconcile the conflicting accounts’ (Jeffrey H. Tigay, ‘Conflation as a Redactional Technique’, in idem [ed.], Empirical Models for Biblical Criticism [Philadelphia: University of Pennsylvania Press, ], pp. 53–95 [69], emphasis added). Another example is the Samaritan Exod. 18, where ‘the hand of the redactor is visible in the change from first and second person, which befits the insert’s Deuteronomic home, to third person where necessary, as suits the narrative context of its new home in the Samaritan Exodus’ (Tigay, ‘Conflation’, p. 67, emphasis added). Tigay concludes that ‘the redaction of the Jethro and Sinai pericopes in the Samaritan Pentateuch is as fine an example as one could wish of scissors-and-paste composition, a “patchwork.” But the patchwork is not “crazy.” The main task of the redactor in these pericopes was to reconcile dissimilar accounts of the same events. By interweaving their details in sequence, he facilitated their harmonious coexistence’ (p. 76, emphasis added). This redactional activity extended to omissions, too. One example is Deut. 5.22, where the Samaritan redactor omits what is ‘either substantially covered in parallel material which he preserves, or dispensable on other grounds. This procedure comports with a tendency which has been observed in the redaction of the Pentateuch’ (p. 77).

‘Improvement’ is admittedly a vague term. However, see how it is used in the following sections of the article.

Rather than ‘compositional technique’. Molly M. Zahn defines compositional technique as ‘a specific way of manipulating or altering the base text, such as addition of new text, rearrangement, or paraphrase’ (Rethinking Rewritten Scripture. Composition and Exegesis in the 4Q Reworked Pentateuch Manuscripts [Leiden: Brill, ], p. 12). We do not disagree with her description of what we see happening with the pre-Samaritan texts, but we have tried to stay away from the term ‘composition’, as it denotes an activity that is associated with the creation of a text. Instead, we have assumed that these texts have already been composed and have attained a completed, ‘final’ literary form. They now undergo (further) editing.

Carr calls these additions more precisely ‘innercompositional coordinating revisions’. ‘The innercompositional character of these revisions gains significance because it points to scribal efforts to bridge nonoverlapping compositions that have been combined into a new literary whole […] The secondary scribal accommodations of tetrateuchal materials to Deuteronomy […] are examples of the sort of gradual scribal revision that can occur when divergent literary works are brought together’ (David M. Carr, ‘Data to Inform Debates about the Formation of the Pentateuch: From Documented Cases of Transmission History to Survey of Rabbinic Exegesis’, in Jan-Christian Gertz, Bernard N. Levinson, et al. [eds.], The Formation of the Pentateuch. Bridging the Academic Cultures of Europe, Israel, and North America [Tübingen: Mohr Siebeck, ], pp. 87–106, here: p. 92, emphasis added).

Emanuel Tov, ‘Rewritten Bible Compositions and Biblical Manuscripts, with Special Attention to the Samaritan Pentateuch’, DSD 5 (1998), pp. 334–54. Revised version in idem, Hebrew Bible, Greek Bible, and Qumran (Tübingen: Mohr Siebeck, ), pp. 57–70.

In secondary literature, both terms, ‘pre-Samaritan’ and ‘proto-Samaritan’, appear to be used interchangeably, although ‘because of the lack of a better name’ (Tov, Textual Criticism, 2nd. ed., p. 80). Either of the terms ‘pre-/proto-Samaritan’ can be misleading because these texts were non-sectarian and actually only one of them was the basis for the Samaritan Pentateuch. It seems hard to believe that scholars still do not distinguish between the use of ‘pre-’ versus ‘proto-Samaritan’, given that the terms ‘pre-/proto-Masoretic texts’ have indeed been differentiated and the subject of discussion (Emanuel Tov, ‘“Proto-Masoretic,” “Pre-Masoretic,” “Semi-Masoretic,” and “Masoretic”: A Study in Terminology and Textual Theory’, in James W. Barker, Anthony Le Donne, and Joel N. Lohr [eds.], Found in Translation, Essays on Jewish Biblical Translation in Honor of Leonard J. Greenspoon [West Lafayette: Purdue University Press, 2018], pp. 31–52). Barthélemy, for instance, used ‘pre-Masoretic’ in a chronological sense, while ‘proto-Masoretic’ indicated closeness in content. However, both terms ‘proto-’ and ‘pre-’ are used when it comes to ‘proto-/pre-Samaritan’ texts. Among the scholars employing the term ‘proto-Samaritan’ are J.H. Tigay (1985), Z. Ben-Hayyim (1992), N. Jastram (1998), , M.M. Zahn (2004), and a few of E. Tov’s articles (1989, 2010, 2012). Others, such as E. Owen (1997), S.W. Crawford (2011), and the majority of Tov’s publications, use ‘pre-Samaritan’. All in all, both terms ‘pre-’ and ‘proto-Samaritan’ are an attempt to indicate a close relationship between the end product (Samaritan Pentateuch) and one particular text within the group of the ‘pre-Samaritan’ texts that the Samaritan Pentateuch was based on. In other words, these scrolls formed a group of texts circulated before the Samaritan Pentateuch came into being, one of which was used as the Vorlage of the Samaritan Pentateuch.

Unlike the Samaritan Pentateuch, however, these expansionist texts do not contain sectarian additions. Much has recently been written about these types of texts. See, for example, Tigay, Empirical Models; Zahn, Rethinking Rewritten Scripture; Sidnie White Crawford, Rewriting Scripture in Second Temple Times (Grand Rapids: Eerdmans, 2008), pp. 30ff.; Emanuel Tov, Textual Criticism of the Hebrew Bible, Qumran, Septuagint. Collected Essays, Volume 3 (Leiden: Brill, ).

The reference to the ‘final’ edition of the MT of the Pentateuch is made very loosely. It so happens that the consonantal text of the medieval textus receptus, the MT, follows closely a specific line of text types of the Second Temple period. We are referring to this text type. The MT text was still somewhat fluid going into the Christian Era, as rabbinic citations and medieval manuscripts reveal, but the variants were minute and infrequent. Other texts circulated alongside this text type, the Hebrew Vorlage to the Septuagint, the Samaritan text, and various other popular texts. See Emanuel Tov, Textual Criticism of the Hebrew Bible (Minneapolis: Fortress Press,³ 2012), pp. 22–39. ‘Final’ is placed in quotation marks because the texts probably never reached a final form. As Eugene Ulrich, The Dead Sea Scrolls and the Developmental Composition of the Bible (Leiden: Brill, ), put it: ‘There was no natural conclusion of the process [of rewriting]; it did not “achieve completion.” There was no indication that the developmental process should stop. No so-called “standardized text” was produced. In contrast, the natural dynamic, developmental process of textual growth was simply abruptly halted, frozen, by the results of the two Jewish Revolts against Rome, the loss of the Temple, the crisis of the Jesus movement, and the increasing isolation of the Samaritans’ (p. 112). The best we can do is to identify, roughly, the ancestor text common to the three major witnesses.

The compositional stage is here defined as the time when (the presumably earlier) editors finalized the compiling and re-ordering of various hypothetical sources. Once this task was completed, (later) editors introduced meaningful variants, examples of which are the types of innercompositional harmonizations that we discuss here.

As postulated here, Deut. 14 is one such occurrence. It is not clear, however, that many other such examples can be identified. The parallel Num. 21.33–35/Deut. 3.3–5 passages are at first sight a rare exception. Not only do they not deal with a crucial halakhic question, but they also appear to be an exception to our earlier statement that the early Temple scribes ‘exercised their prerogative in a very limited manner, intervening and editing where they thought it was absolutely necessary’. Was this harmonizing addition necessary? We believe it was. The Numbers tradition appears to have not been aware of the tradition that Moses had defeated two, not just one, Amorite kings of the Trans-Jordan (which in turn paved the way for Manasseh’s inheritance). A plain reading of the Numbers version, in which it is seen that the descendants of Manasseh conquer the territory that subsequently they settle in, contradicts the Deut. story. Though only a narrative, the contradiction may have justified an intervention (Cf. note 50 below).

Instead, later interpretative activity would resolve this type of problem. Appropriately, exegetes of the following centuries worked to harmonize these sections. Rabbinic Judaism chose to smooth away many of these problems via exegesis (midrash). By the early first century BCE, midrashic activity was well underway, though it is unclear how much earlier this interpretative activity began. Daube avers that the Rabbinic interpretative activity derived from Hellenistic rhetoric and that it was first adopted in the first century BCE, coinciding with the philological activities of the Alexandrians. David Daube, ‘Rabbinic Methods of Interpretation and Hellenistic Rhetoric’, in Calum M. Carmichael (ed.), Collected Works of David Daube. Volume One. Talmudic Law (Berkeley: Robbins Collection, 1992), pp. 333–56 (p. 334); see also in the same volume David Daube, ‘Alexandrian Methods of Interpretation and the Rabbis’, pp. 357–79. Lieberman agrees that ‘it was the Greeks who systematized, defined, and gave definite form to the shapeless mass of interpretations’ (p. 62). However, he is only willing to concede that the rabbis borrowed from the Greeks’ ‘formulation, terms, categories and sistematization [sic]’ of their rules of interpretation, but not the logic of these norms (p. 78). While he is not convinced the rabbis ‘adopted a certain method from the Greeks, they may nevertheless have learnt from them the application of that method to a particular question’ (Saul Lieberman, Greek in Jewish Palestine/Hellenism in Jewish Palestine [New York/Jerusalem: The Jewish Theological Seminary of America, ], p. 79). Earlier, Lieberman had stated that ‘it will be demonstrated below that interpretation in general is older than the revelation of the Law at Mount Sinai’ (p. 58), but he neither demonstrates this assertion nor, when he engages the matter, does he deal with the specific ‘quasi-logical’ norms that the rabbis applied to the legal portions of the Bible but rather with interpretation in general, such as those used for dreams and oracles. He claims to have seen Daube’s article, above, only after he had finished his own article but that he, Lieberman, ‘found no reason to change anything in this chapter, as will be self evident from the comparison of Dr. Daube’s article with this paper’ (note 64). One may conclude, with Lieberman, that while there are no grounds for proving a direct borrowing from the Alexandrians, it is quite evident that the interpretative/midrashic impulse gained force, applicability, and acceptance among the rabbis only in the second half of the first century BCE.

As Tigay puts it: ‘Apart from the question of precise recessional relationships, the “conservative, often pristine” Masoretic Text reflects a stage anterior to the expansion which produced the Samaritan’ (Tigay, ‘Conflation’, p. 63). I am calling this anterior stage the stage that produces the common ancestor text.

See note 12.

This designation was first coined by Géza Vermes (‘In order to anticipate questions, and to solve problems in advance, the midrashist inserts haggadic development into the biblical narrative’) in his Scripture and Tradition in Judaism: Haggadic Studies (Leiden: Brill, ²1973). While most agree that the aim of Rewritten Scripture was to interpret texts that are now part of the Hebrew Bible, they often reveal a high degree of freedom in their retelling. The term applied only in the early days to ‘post biblical texts which paraphrase the scriptural narrative whether through conflation, harmonization, and/or supplementation’, in the words of Steven D. Fraade, ‘Temple Scroll as Rewritten Bible: When Genres Bend’, in Binyamin Golstein, Michael Segal, and George J. Brooke (eds.), Haish Moshe: Studies in Scriptural Interpretation in the Dead Sea Scrolls and Related Literature in Honor of Moshe J. Bernstein (Leiden: Brill, 2018), pp. 136–54, here: p. 136. Crawford describes Rewritten Scripture as ‘new works that are still recognizably tied to their scriptural base text and claim the same authority as the base text’ (Rewriting Scripture, p. 85). Andrew Teeter points out that ‘every rewritten Bible composition is defined by its own retelling strategy or program’ and that they are ‘full-fledged literary works in their own right, each with its own distinct profile and its own specific compositional aims’ (‘On “Exegetical Function” in Rewritten Scripture: Inner-Biblical Exegesis and the Abram/Ravens Narrative in “Jubilees”,’ The Harvard Theological Review 106.4 (), pp. 373–402, here: p. 374 and p. 375). Tov addresses the challenge of defining what exactly constitutes ‘Rewritten Bible’ (‘they do not form a well-defined group’) and the question ‘whether or to what extent their writers or readers considered them to be authoritative biblical texts’, in Tov, ‘Rewritten Bible Compositions’, p. 59, and idem, Hebrew Bible, p. 68.

Zahn, Rethinking Rewritten Scripture, pp. 8ff.; see also Crawford, Rewriting Scripture, pp. 13ff.

The initial term ‘4QPP – 4QPentateuchalParaphrase’ was changed to ‘4QRP – 4QReworkedPentateuch’ because, according to Tov, ‘“reworking” is more general than “paraphrase” and would allow for long stretches of unaltered text’ (Emanuel Tov, ‘From 4QReworked Pentateuch to 4QPentateuch (?)’, in idem (ed.), Textual Criticism of the Hebrew Bible, Qumran, Septuagint. Collected Essays, Volume 3 [Leiden: Brill, ], pp. 45–59, here: p. 46). Tov explains that ‘many small exegetical elements are added, while other elements are omitted, or, in other cases, their sequence altered. The exegetical character of this composition is especially evident from several exegetical additions comprising half a line, one line, two lines, and even seven or eight lines’ (Tov, ‘From 4QReworked Pentateuch’, p. 47). Tov struggled with the idea that these fragments can properly be considered biblical. In his words, ‘The extensive additions’ in three of the fragments ‘differed so much from the biblical manuscripts I knew that I could not imagine that 4QRP contained a biblical manuscript’ (p. 47). As a result, he first came to designate these fragments as non-biblical. Eventually, persuaded by evidence from the LXX extensive rewrites of Kings III, Daniel, and Esther, accepted as biblical works, Tov relented and considered 4QRP to be ‘a group of Scripture texts’, or ‘Scripture manuscripts’ (pp. 51–52). Even when 4QRP became accepted as biblical text by most scholars including Tov and was referred to as simply ‘4QPentateuch’, Tov preferred to keep the name ‘Reworked’ in ‘4QPR’ for mainly two reasons. First, not everyone would share the view that 4QPR is indeed a group of Scripture texts. Second, changing the name of published manuscripts generally leads to more confusion in scholarship. For our purposes, we have kept the RP texts in the category of Rewritten Scripture even though they appear to be, as Tov has more recently indicated, a group of Scripture texts.

The five manuscripts 4Q158 and 4Q364–367 altogether cover parts of Gen. 2; 21; 25–48; Exod. 8–39; Lev. 11–27; Num. 1–36 and Deut. 1–14; 16; 19. 4Q365 contains fragments of all five Pentateuchal books, while other manuscripts contain parts of only one book (L in 4Q367), two books (Genesis and Exodus in 4Q158), or three/four books (4Q364 and 4Q366). Scholars now believe that the five manuscripts should be regarded as distinct compositions (George Brooke, ‘4Q158: Reworked Pentateuch^a or Reworked Pentateuch A?’, in DSD 8 (2001), pp. 219–41; Zahn, Rethinking Rewritten Scripture). For more discussion of 4Q158 and 4Q364–367, see Tov, ‘From 4QReworked Pentateuch’, especially pp. 52–56. For the edition of the text, see Emanuel Tov and Sidnie White Crawford, ‘4QReworked Pentateuch’, in Harold Attridge et al. (eds.), Qumran Cave 4. VIII: Parabiblical Texts. Part 1 (Oxford: Clarendon, ), pp. 255–318.

See above, note 22.

4QRP ‘appears to supplement rather than supplant, the earlier form of the Pentateuch’ (Ulrich, Dead Sea Scrolls, p. 208).

For a more detailed discussion of these sections, see Crawford, Rewriting Scripture, pp. 90–102.

Tov, ‘Rewritten Bible Compositions’, p. 2. On the Temple Scroll, see also Johann Maier, The Temple Scroll. An Introduction. Translation and Commentary (Sheffield: Sheffield Academic Press, 1987); Yigael Yadin, The Temple Scroll. The Hidden Law of the Dead Sea Sect (London: Weidenfeld and Nicolson, 1985); Dwight D. Swanson, The Temple Scroll and the Bible (Leiden: Brill, 1995); Sidnie White Crawford, The Temple Scroll and Related Texts (Sheffield: Sheffield Academic Press, ).

As discussed by Ulrich, Dead Sea Scroll, p. 209, the Temple Scroll shows the following features: (1) large-scale expansion, e.g., instructions for the Temple (not found in the Pentateuch); (2) God in the first person as new speaker, e.g., in 45.14; (3) God speaking directly as new implicit claim to divine revelation; (4) new scope, Exod. 34–Deut. 23; (5) rearrangement and harmonization of legal materials; (6) introducing a new theological agenda through instructions for the Temple, exegetical interpretation through ‘conflation, harmonization, and clarification’, expanded festival calendar with new feasts, and a 364-day calendar.

‘Since some of the above analogues (approximate as they are) appear in texts commonly classified as Rewritten Bible (Jubilees and Josephus’s Jewish Antiquities), others are not (the Damascus Document and Philo’s On the Special Laws), we must surmise that this is not a characteristic of legal Rewritten Bible per se, but of legal codification (in a nascent sense) across literary forms and ideologies, finding its most extensive Jewish expression ultimately in the Mishnah’ (Fraade, ‘Temple Scroll as Rewritten Bible’, pp. 143–44).

Where associative rather than topical concepts prevailed. For a discussion of some of the differences between Mesopotamian (Hebrew included, as in the Code of Covenant) and Greek ideas of law, see Jean Bottero, ‘Le “Code” de Hammurabi’, in Annali della Scuola Normale Superiore di Pisa 12 (1982), pp. 409–44, translated into English as ‘“Code” of Hammurabi’, in Mesopotamia: Writing, Reasoning and the Gods (Chicago: University of Chicago Press, 1992), pp. 156–84; and Raymond Westbrook, Studies in Biblical and Cuneiform Law (Paris: Gabalda, ), pp. 2–5.

The English text follows the 1985 Jewish Publication Society (JPS) Tanakh translation. Some alterations are made to be closer to the original Hebrew.

This leads Kuenen to say that ‘the relation of Lev. xi to Deut. xiv.1–21, indicated in the text, is quite unmistakable’ (Kuenen, The Origin and Composition, p. 266).

If we start with two genetically related texts, as these are, one long and one relatively shorter, it is difficult to determine whether the longer text is an expansion of the short one, or vice versa, the shorter text is an abbreviation of the long one. Similar challenges arise, for example, in recognizing pluses or minuses in the biblical text: Is a Septuagint plus, for example, the original reading, and the MT minus a deliberate abbreviation or an accidental textual omission, or is it the other way around, with the Septuagint exhibiting an expansion of the original text? Methodologically, it is easier to argue that pluses are secondary expansions since we have a ‘rule’ that redactors do not generally expunge texts. In our case, we are proposing some rules to help solve this question.

Wellhausen, Die Composition des Hexateuchs, p. 203; Kuenen, The Origin and Composition, pp. 278–92. See also Timo Veijola, Das Königtum in der Beurteilung der deuteronomistischen Historiographie. Eine redaktionsgeschichtliche Untersuchung (Helsinki: Suomalainen Tiedeakat., 1977), pp. 295ff. Moran, too, views Deut. 14 as the earlier text, but not as an abbreviation of Lev. 11—rather, it underwent a later redaction toward L (William L. Moran, ‘The Literary Connection between Lev 11,13-19 and Dt 14,12-18’, CBQ 28 [], pp. 271–77).

August Dillmann, Die Bücher Exodus und Leviticus (KEH 12; Lepizig: Hirzel, ³1897), pp. 525ff.; Bernardus Dirk Eerdmans, Alttestamentliche Studien IV: Das Buch Leviticus (Gießen: Töpelmann, 1912), pp. 61–63; Rolf Rendtorff, Die Gesetze in der Priesterschrift: Eine gattungsgeschichtliche Untersuchung (Göttingen: Vandenhoeck & Ruprecht, 1954), pp. 43–45; Meir Paran, Darkei ha-Signon ha-kohani ba-Torah (Jerusalem 1989), pp. 340–53; Milgrom, Leviticus, pp. 698–704; Eckart Otto, Deuteronomium 12,1–23,15 (HThKAT; Freiburg: Herder, ), p. 1293.

So, for example, Martin Noth regarded L to be first, but thought that D drew upon a pre-text of Lev. 11 (Martin Noth, Das dritte Buch Mose. Leviticus [ATD 6; Göttingen: Vandenhoeck & Ruprecht, 1962], p. 76). Also, Alfred Bertholet, Deuteronomium (Leipzig/Tübingen: Mohr Siebeck, 1899), p. 44; Carl Steuernagel, Das Deuteronomium (Göttingen: Vandenhoeck & Ruprecht, ³1923), p. 105; Gerhard von Rad, Deuteronomy (London: SCM Press, 1966), pp. 101–2; Karl Elliger, Leviticus (HAT I/4; Tübingen: Mohr Siebeck, 1966), pp. 142–44; Andrew D.H. Mayes, Deuteronomy (Grand Rapids: Eerdmans, 1981), p. 237; Walter J. Houston, Purity and Monotheism. Clean and Unclean Animals in Biblical Law (JSOT.S 140; Sheffield: Sheffield Academic Press, 1993), pp. 63–65; Eduard Nielsen, Deuteronomium (Tübingen: Mohr Siebeck 1995), p. 150; Christophe Nihan, From Priestly Torah to Pentateuch. A Study in the Composition of the Book of Leviticus (FAT II/25; Tübingen: Mohr Siebeck, 2007), pp. 283–99; Thomas Hieke, Levitikus 1–15 (HThKAT; Freiburg: Herder, ), pp. 416ff. Nihan argues that the Vorlage ‘[…] was variously adapted by the scribes who composed Lev 11 and Deut 14 respectively. Whereas it was taken up with relatively few modifications in Deut 14, it underwent significant expansion in Lev 11’ (Priestly Torah, p. 415).

Some methodological progress has been made recently in understanding textual dependency. See the article by David McLain Carr, ‘Method in Determination of Direction of Dependence: An Empirical Test of Criteria Applied to Exodus 34,11–26 and its Parallels’, in Matthias Köckert and Erhard Blum (eds.), Gottes Volk am Sinai. Untersuchungen zu Ex 32–34 und Dtn 9–11 (Gütersloh: Chr. Kaiser Gütersloher Verlagshaus, ), pp. 107–40.

Kuenen, The Origin and Composition, p. 266. Regarding oral traditions and familiarity with Priestly procedures as early as the close of the seventh or the early sixth century BCE, see Dalit Rom-Shiloni, ‘How can you say, “I am not defiled…”? (Jeremiah 2:20–25): Allusions to Priestly Legal Traditions in the Poetry of Jeremiah’, Journal of Biblical Literature 133.4 (2014), pp. 757–75. Driver, also a proponent of the priority of Deuteronomy, took what may perhaps be an intermediate position, i.e., the common Vorlage thesis, but clearly acknowledged the Priestly nature of the passages. He states ‘that Deut 14,3–20 is not, as a whole, the composition of D, but borrowed by him (with slight additions, as v. 3–11, and other unessential modifications) from some independent source, cannot be doubted: not only is the general style unlike that of D, but מין‎ kind v. 13.15.18 is a term characteristic of P, and is not likely to have been adopted independently by D’ (A Critical and Exegetical Commentary, p. 163.) In an impressive analysis, Paran lists examples of expressions and stylistic turns appearing in this section that are common in Priestly literature, but rare or non-existent in D: a) למינהו‎/leminehu (=after its kind) appears a number of times in the dietary passage when listing the forbidden birds; b) the particle אך‎/akh (=but) appearing in v. 7 is used in place of the usual רק‎/raq in D (Paran draws an interesting contrast, covering the prohibition to eat blood: in Deut. 12.16, 12.23 and 15.23, the word used is raq. In Gen. 9.4 [a Priestly text], the word used is akh); c) the expression זה אשר‎ zeh/zeh asher (=this, this that), a singular with a collective force, is common in Priestly writings but rare in D. Nonetheless, it appears in Deut. 14.4 and 14.12, most probably under the influence of the Priestly text; d) the term טמא‎ (=impure) is very frequent in the Bible, but the phrase הם/טמא הוא‎ tame hu/hem (=it is unclean/they are unclean) is a Priestly expression, found, e.g., in Lev. 13 (vv. 11, 15, 36, 44, 46) and in Lev. 14.44, 15.2, Num. 19.15 and 19.20; and the parallel version to Deut. 14, in the parts missing from the D version. Contrarily, in D this expression is only found in our passage, 14.3 and 14.21. The same holds true with stylistic turns. Driver stated that ‘[…] to a greater degree than in any other part of the OT, is a preference shown in P for standing formulae and expressions; some of these recur with great frequency […]. Particularly noticeable is an otherwise uncommon mode of expression, producing a particular rhythm, by which a statement is first made in general terms, and then partly repeated for the purpose of receiving closer limitation or definition’ (An Introduction to the Literature of the Old Testament [Gloucester: Peter Smith, ], p. 130). Paran shows that this feature appears frequently in our passage in L but never in D, other than in Deut. 14.4–6 and 9. Even in this passage, stylistic features were not as important to D as substantive matters, leading D to break the rhythm on a number of occasions. Paran also demonstrates a consistent pattern of balance, parallelism and symmetry in the Priestly writings, while only two such examples in D. Examples: Lev. 11.10–11.35, 43. In other instances, we find near parallelism: Lev. 11.2 as against Deut. 14.4, Lev. 11.13 as against Deut. 14.12. In Deuteronomy we find no parallelism in these. Whatever parallelism exists in the D version must have been borrowed from L, which denotes a heavy use of parallelisms. To conclude, linguistic and stylistic analysis is sufficiently persuasive in so far as determining that D drew from a Priestly source. But, as noted earlier, more recent scholars have refused to concede that the Priestly donor/source text in question is Lev. 11; instead, they have postulated a third, common, Vorlage, to solve copying puzzles. It is this problem that we have sought to address in the first part of this paper.

Paran (see note above) is one such exception. He seeks to prove that the text is Priestly solely on linguistic and stylistic grounds. While he concludes that Deut. is dependent on a Priestly text, he ignores the larger question of precedence in composition.

The proponent of a common Vorlage cannot explain why L went out of its way to illogically expand the neat and short formula of the three non-kosher animals, presumably existing in the common Vorlage, nor why Lev. forgot to copy over the bridging sentences, which D succeeded in copying, vv. 11 and 20. Furthermore, even if a common Vorlage is assumed, other questions remain: If the presumed Vorlage had a list of permissible quadrupeds, why did L not copy it? Why do L and D have different sequences if they both copied from one common version? Did the Vorlage use the term sheqetz? If it did, why did D not, in copying it, use the same term?

For instance, the laws prohibiting mutilation of the dead. A cursory look at their differing formulations and terminology confirm that there is no literary dependence between these two pieces of legislation. This is in keeping with the watertight separation commonly postulated between the two early compositions.

^{Lev. 19.28} You shall not make gashes in your flesh for the dead (ושרט לנפש לא תתנו בבשרכם‎) or incise any marks on yourselves: I am the LORD.

^{Deut. 14.1} You are children of the LORD your God. You shall not gash yourselves (לא תתגדדו‎) or shave the front of your heads because of the dead. Disfiguring for the dead was probably another priestly practice that was well known, though not necessarily via literary channels.

The reference to the ‘final’ edition of the MT is made loosely, see note 14 above. It so happens that the consonantal text of the medieval textus receptus, the MT, closely follows a specific line of text types of the Second Temple period. We are referring to this text type. The MT text was still somewhat fluid going into the Christian Era, as rabbinic citations and medieval manuscripts reveal, but the variants were minute and infrequent. Other texts circulated alongside this text type, the Hebrew Vorlage to the Septuagint, the Samaritan text, and various other popular texts. See Tov, Textual Criticism of the Hebrew Bible, pp. 22–39.

Geza Vermes explained that ‘redactor-copyists saw it as their right and duty to correct and improve the work they were propagating.’ (emphasis added). He states that the freedom of textual transmission ‘may be due to variations in local traditions […] to efforts of modernization of spelling and grammar, to literary processes already apparent in the Bible itself, such as stylistic variations and attempts at harmonization, but above all to a phenomenon [… he calls] “scribal creative freedom”’ (The Dead Sea Scrolls Forty Years on. The Fourteenth Sacks Lecture Delivered on 20th May 1987 [Oxford: The Oxford Center for Postgraduate Hebrew Studies, ], p. 14).

An argument has been adduced to support the thesis of D’s dependency on L, a position that we support, based on what appears to be an obvious error appearing in the D text and which is ascribed to sloppy copying. Specifically, it is the prohibition to touch the carcasses of non-kosher animals in Deut. 14.8, a prohibition that speaks of a purity consideration, not a dietary one. Purity considerations are alien to this section, which only deals with dietary rules. (The section opens with ‘you shall not eat anything abhorrent’.) Contrarily, purity matters are an integral part of the L section. Rabbinic interpretation supports this distinction: priests, and only priests, are enjoined from touching a carcass lest they become impure (see Bekhor Shor ad loc.). Israelites, on the other hand, are not prohibited from touching a carcass and becoming impure but are advised not to become impure at the time of the festivals, since they bring offerings to the temple. See Rashi ad loc., see Sifra 4:8–10, and see Encyclopedia Talmudit (Jerusalem: Talmudic Encyclopedia Publishers, 1951–), ‘tumat kohanim’. Since Israelites are not prohibited from becoming impure (though they need to purify themselves if they became impure; see Lev. 11.39–40), it makes no logical sense to copy this clause into this section in D. The argument is that the D redactor copied from L and mindlessly committed an error; ‘sloppy’ copying betrayed his source (see Mark Goodacre, ‘Fatigue in the Synoptics’, New Testament Studies 44.1 [], pp. 45–58): ‘Editorial fatigue is a phenomenon that will inevitably occur when a writer is heavily dependent on another’s work. In telling the same story as his predecessor, a writer makes changes in the early stages which he is unable to sustain throughout’ [p. 45]). The D redactor made changes, creating a more logical classification than Leviticus while saving words, but, presumably, was unable to sustain throughout, when he returned to his Vorlage, and mindlessly added the words ‘nor touch their carcasses’. However, while the ‘sloppy copier’ argument seems reasonable, one can still defend the validity of the sentence ‘nor touch their carcasses’ since touching can also stand for eating. See Joseph ben Isaac Bekhor Shor’s medieval commentary, ad loc., citing Talmudic sources: Bavli Makkot 14b, Yebamot 75a, Zebachim 33b. Synonymous terms are common in biblical poetry, and, as Paran notes, even in Priestly legal material. The text then reads as follows: ‘You shall not eat of their flesh (when you slaughter them ritually) nor eat their carcass (when they die naturally)’. This prohibition would apply to the prohibited animals while v. 21 would apply even to kosher animals. In Hieke’s view, the touching of carcasses in Deut. 14.8 is an argument against D having copied from L Hieke, not believing in the existence of such a ‘sloppy copier’, highlights that if purity laws were not important for D, and therefore purity references and the term sheqetz are omitted when copying from Lev. 11, why did Deut. 14 then not completely eliminate the purity laws? (Hieke, Levitikus, p. 416). Hence, for Hieke this is proof that D was copying from another Vorlage. However, we would counter that even if such a Vorlage existed, the question remains why D copied the reference to touching carcasses at all (regardless of whether it was copied directly from L or from another Vorlage).

Kuenen, Origin and Composition, p. 266. Hoffmann, who argued against Wellhausen’s thesis that L followed D, first noted that both L and D contain self-sufficient kashrut criteria and conceded that there should have been no need to add a specific list of permissible animals. In response, Hoffmann argued that the listing of the ten quadrupeds in D reflected the legislators’ aim to offer a practical list for the non-priestly and uneducated population who would otherwise not be able to determine the signs of kashrut (David Zvi Hoffmann, Rayot Makhriot Neged Wellhausen [Tel Aviv: Netzah], p. 48). Following this explanation, the list of ten quadrupeds is not redundant.

One could of course counter that L intended a cleft hoof, even one that was not necessarily split in two, but the D legislator insisted on a split hoof. In such a case, either of these two texts could stand as the source text; a difference in opinion emerges, with the L legislator taking a laxer, more inclusive position (allowing for a greater number of edible animals). This is possible but not probable, since the Priestly source would want to limit the universe of permissible foods, being the more stringent tradition in matters of purity. More likely, the D redactor clarified the ambiguity left in the L passage.

The same question can be asked of those who postulate a common Vorlage. If the common source is austere in its formulation, why would the L redactor turn it into a verbose formulation? And if, conversely, the common Vorlage is verbose, the L redactor simply copied it and the Deut. redactor abbreviated it, what need is there to resort to such a common-source hypothesis, let the D redactor abbreviate directly from the L passage!

This logic also helps explain an interesting omission in Deut. In referring to the four animals described in vv. 4–7, Lev. 11.8 summarizes the foregoing by adding that ‘they are unclean (temeim)’. This summarizing characterization follows the separate statement ‘it is unclean’ affixed to each of the four animals. The second half of Deut. 14.8 follows verbatim Lev. 11.8 with, however, one exception: it omits the statement ‘they are unclean’. Why? We submit the reason is simple: D listed three animals with common features, grouped them together, and stated that ‘they are unclean’ (end of v. 7). The fourth animal, the hazir, is described in v. 8, and is characterized as ‘it is unclean’. It would be counterproductive to subsequently subsume them all as ‘they are unclean’ since they belong to two different classifications. L, on the other hand, flattened the differences in signs between the four animals, treating them all equally as unclean. Thus, the summarizing statement can characterize them all as ‘they are unclean’.

For clarification, the ‘drafting logic’ we wish to apply here does not relate to text-critical rules such as ‘lectio brevior potior’, which would regard a shorter text to be the older tradition, as redactors would tend to add and embellish a text rather than shorten it. We argue that the (D) redactor produced an improved text, a more concise, precise, and unambiguous version of L. At times the improved text could be shorter (as in this case, where D spends less words to convey the same idea as in L; in this sense, it is contrary to the ‘lectio brevior potior’ principle), and at times the improved text could be longer (see II.c.i, where Deut. 14.6 specifies ‘cleft in two’).

The reverse, namely, that a secondary hand will be adding to a more austere source only when it has new information and not just to produce a more verbose version, can be seen at Num. 21.33–35 versus Deut. 3.1–3 (briefly touched upon earlier in footnote 16). The passages are remarkably similar, but which one is the original and which is the copy? On the basis of style/form, Alexander Rofé perceptively argues that D is the source. Moreover, following a close reading of Num. 32.33–42, Rofé reaches the conclusion that the Numbers editor did not know of the tradition of the defeat of Og of Bashan. The redactional insertion (Num. 21.33–35) ends up harmonizing the two books and they now share the same tradition. The ‘philological rule’ we proposed allows us to confirm Rofé’s finding: Deut. 3.3 reads ‘So the Lord our God also delivered into our power King Og of Bashan, with all his men, and we dealt him such a blow until no remnant was left him’. Num. 21.35 reads ‘They defeated him and his sons and all his people, until no remnant was left him; and they took possession of his country’. The words and his sons that we marked in italics represent a significant addition. Had Numbers been the original, it would be difficult to justify D’s omission. Ergo, the D passage is original and Numbers is a copy that, along the way, contributed a new piece of information. On the other hand, Rofé argues that this copied passage entered Numbers at the time of composition. He says: ‘The material in Num 21:33–35 thus confirms that the method of “completing” one document with the help of another did not first arise after the final editing of the Pentateuch; it operated in earlier stages as well, the stages of the creation of the accounts themselves’, Introduction to the Composition of the Pentateuch (Sheffield: Sheffield Academic Press, ), p. 119. Rofé may be over-reaching here; one cannot be sure that this took place at the compositional stage (‘stage of creation’). As we argue later, these harmonizing redactions are more likely to have occurred only after the ‘final’ literary forms of the various books had been agreed upon and the books are, so to say, published. It is only then that redactors are able to readily compare compositions and coordinate revisions. This latter approach seeks to bring the harmonizing revisions closer to the Second Temple period during which these techniques were commonplace. In general, it is not disputed that we are on safer grounds speaking about redactional activity than theorizing about composition.

V. 12 serves as a resumptive repetition of v. 10, to allow for a secondary and later interpolation contained in the latter part of v. 10 and in v. 11.

Lev. 11.21–23 appears to be secondary. Note the term אך‎ at the beginning of Lev. 11.21, a common marker of secondary additions. More importantly, Lev. 11.23 appears as a Wiederaufnahme, forming an inclusio with Lev. 11.20.

Rabbi Simeón, in Sifre, interprets the rules in a different manner, implicitly taking issue with our taxonomic interpretation. Working with a ‘final’ version of the Masoretic text, which includes the secondary addition in L, he reconciles D with L. In his account, ‘swarming winged creatures’ in the D version refers to the non-edible types of locusts and grasshoppers, while clean flying creatures refer to the edible types of locusts and grasshoppers.

See David M. Carr, The Formation of the Hebrew Bible: A New Reconstruction (Oxford: Oxford University Press, 2011), and more recently, idem, ‘Data to Inform.’ With regard to the phenomenon of scribes writing down an oral transmission in the way they each recollected it, Carr writes: ‘working in an oral-written environment where they had memorized many texts verbatim, scribes often substituted words and phrases that they considered semantically equivalent to the known text. The resulting phenomenon in the textual evidence, where semantically equivalent words, grammatical expressions, and phrases are exchanged, minor particles are added or subtracted, shifts in order are introduced, etc., is what I have called “memory variants”. It’s also what gives rise to synonymous readings’ (Carr, ‘Data’, p. 88). See also Shemaryahu Talmon, ‘The Old Testament Text’, in Peter R. Ackroyd and Christopher F. Evans (eds.), The Cambridge History of the Bible I (Cambridge: Cambridge University Press, 1970), pp. 159–99. As Tov paraphrases Talmon, ‘parallel readings reflect equally early and original constituents and that none can be preferred to the other’ (Emanuel Tov, ‘The Original Shape of the Hebrew Text’, in John A. Emerton (ed.), Congress Volume Leuven 1989 [Leiden: Brill, 1991], pp. 345–59); see also Moshe Greenberg, ‘The Use of the Ancient Versions for Interpreting the Hebrew Text’, SVT 29 (), pp. 131–48.

Interestingly, the Samaritan version has D reproduce exactly v. 7 of L. If the Samaritan text is more original (rather than a harmonization) then the D redactor’s abbreviation was not faulty, and the difficulty disappears.

See Milgrom, Leviticus, p. 699. See also already Kalisch: ‘The specification renders the rule superfluous, and the rule throws doubt on the completeness of the specification; both together are perplexing’ (Marcus Kalisch, Leviticus [London: Longmans, ], p. 128).

Hoffmann (see also note 45 above) offers an interesting but, ultimately, unsatisfactory suggestion. He argues that D, being addressed to non-priests, hence being a more popular work, lists the permissible quadrupeds because the audience would be unable to follow the specialized and complex criteria. Rather than being redundant, the list offers a practical shortcut. He adds that this was not necessary to do with respect to aquatic animals because their signs are easily recognizable (Hoffmann, Rayot, p. 48). On the postulated assumption that the readership was not competent enough to follow the technical criteria, however, why bother to spell these out?

See above and accompanying note 38, for the suggestion that non-literary channels of communication likely existed between priests and non-priests, where the latter could learn ‘correct’ practices. The practice of abstaining from animals that were abhorrent to eat, like swine, must have been widespread among the common people.

For instance, in reference to idolatry/pagan worship and practices (Deut. 7.25–26; 12.31; 18.9,12; 20.18; 27.15 and more), actions against God (e.g., Deut. 17.4), or the wrong offerings (Deut. 17.1), magical practices (18.12), false measures (26.13–16), transgender conduct (22.5), remarrying a divorced woman after she had remarried another man (24.4). Note that תועבה‎/toevah designating food, apart from our case (Deut. 14.3), is elsewhere only found in the Holiness Code (cf. Lev. 18); therefore, Reinhard Achenbach states that the designation of impermissible food as toevah in Deut. 14.3 ‘stellt demnach einen relativ spät gewachsenen, hybriden Sprachgebrauch dar, nach welchem alles “Unreine” zugleich auch “heidnisch” erscheint und mit dem Verdikt des Widergöttlichen belegt wird’ (reflects a relatively late, hybrid use of language where ‘impure’ is identified with ‘pagan’ practices, inimical to God, emphasis added), in ‘Zur Systematik der Speisegebote in Leviticus 11 und in Deuteronomium 14’, ZAR 17 (), pp. 161–210, here: p. 176. Moran argues for the literary priority of Deut. because if D was copying L, why would it use toevah instead of the Priestly term sheqetz? (‘The Literary Connection’, p. 273). Of course, Moran’s question can be reversed, too: If D preceded L as he posited, why wouldn’t L use toevah? It seems that sheqetz has only one use in the Priestly writings, that of detestable foods (BDB, p. 1054). In D, on the other hand, sheqetz is used for pagan objects (Deut. 7.26). The D redactor chose to use toevah for food, because the latter has a much wider meaning and application. This would be one possibility to answer Moran’s question. However, we go even further, by arguing that D’s first redaction containing 14.3–5, 14.11, and 14.20 was not copied from a Priestly source.

It is not at all clear that we are dealing here with a ‘borrowing’. Still, the polemics implicit in the motive clause make it probable that this was indeed a borrowing, and a rare one at that.

We are not taking a position with respect to whether these Priestly additions are P or belong to the HS (the Holiness School). The existence of and attribution to a HS is speculative and would only serve to make present matters even more complex. Suffice it to attribute these pieces of legislation to Priestly sources. See above, note 4.

Driver, Introduction, pp. 75–76.

This confirms Carr’s observation that a text tends to be later than its parallel when it has ‘an element which appears to be an adaptation of an element in the other text to shifting circumstances’ (Carr, ‘Method in Determination’, p. 126).

See prior discussion in note 41.

The clause introduced by the particle כי‎/ki is so read by most commentators, rather than it stating that, through physical purification, Israelites become holy. Weinfeld follows this reading and draws a fundamental distinction: ‘Holiness, in the Priestly view, is a condition that can be secured only by constant physical purification and sanctification, whereas in Deuteronomy it is the effect of a unique act of God—the divine election of Israel—and thus devolves automatically upon every Israelite, who consequently must not profane it by defilement’ (Moshe Weinfeld, Deuteronomy and the Deuteronomic School [Winona Lake: Eisenbrauns, ], p. 226).

Achenbach, ‘Zur Systematik’, see above note 59.

Tigay, ‘Conflation’, p. 82.

The JPS translation used here assumes this to be the case. R. Samuel b. Meir (RaSHBaM, c. 1085–1158), a classic Jewish medieval commentator, points out that alehem serves to emphasize that both Aaron and Moses were being addressed here, to mark the difference with the previous chapter where God addresses Aaron only.

See R. Shlomo b. Yitzhaki’s (RaSHi, c. 1040–1105) commentary, based on Sifra.

‘[…] any part of the older law could be omitted if it did not suit the goals of the new document. If any part of the source law could be omitted in the new law, it is reasonable to assume that entire laws could also have been omitted in the new law collection’ (Juha Pakkala, God’s Word Omitted. Omissions in the Transmission of the Hebrew Bible [Göttingen: Vandenhoeck&Ruprecht, ], p. 123).

Zahn, Rethinking Rewritten Scripture, p. 147, quoting Michael Segal, ‘The Text of the Hebrew Bible in Light of the Dead Sea Scrolls’, Materia Giudaica 12 (), pp. 5–20.

Tov, ‘Rewritten Bible Compositions’, p. 340.

The Temple Scroll, as discussed, goes one step further and introduces God as the direct speaker, instead of Moses.

See Sifra, Dibbura DNedava, chapter 2, but also in various other rabbinic sources, as indicated by Louis Finkelstein in his notes to the passage, in Volume 4, Sefer Torat Kohanim (New York: JTS, 1990). See also Menahem I. Kahana, Sifrei Bamidbar. Part 1: The Edition (Jerusalem: The Hebrew University Magnes Press, ), pisqa 58, and notes to lines 7–12.

Fraade, ‘Temple Scroll’, p. 140.

Tov, ‘From 4QReworked Pentateuch’, p. 47.

This is implicit with the introduction of a new speaker, as discussed above. In this respect, the ‘new speaker’ and the ‘new theological agenda’ features are not clearly separable categories.

Deut. 14.6–21 bears to Lev. 11 the same relationship as, for example, the Temple Scroll bears to the scriptural books of Leviticus and Deuteronomy. At the time of the proposed redaction of Deut. 14.6–21, L was in all likelihood already considered authoritative Scripture. The rules for making a Sukkah discussed in Neh. 8.14–15 are referred to as being written in the Torah, where the Torah is being cited authoritatively. Hence, Deut. 14.6–21 can be considered a piece of ‘Rewritten Scripture’. In terms of dating, since Deut. 14 appears in all three witnesses, we can assume with some certainty that it formed part of the common ancestor text; therefore, the hypothesized interpolations discussed earlier cannot post-date the mid-third century BCE. Ptolemy II is thought to have ordered the translation of the Pentateuch into Greek in the mid-third century BCE. The terminus post quem is unlikely to date to before the time of Nehemiah in the mid-fifth century BCE. As an example, Lev. 23.42–43 is similar to what Neh. 8.14–15 reads, but there are significant discrepancies that indicate L had by the time of Nehemiah still not reached its final compositional form.

More specifically, a text that lies along a continuum of texts that share characteristics of the texts called Reworked Pentateuch and the new literary works based on Scripture, such as Jubilees and Temple Scroll, as discussed earlier.