Seeking the Autographs of Jeremiah

Abilene Christian University

Seeking the autographs of Jeremiah

A Research Paper submitted to

Dr. Mark hamilton

Graduate School of theology

by

benjamin covington

abilene, texas

November 30, 2009

1. Introduction

Rather than resolve the tension between LXXJer and MTJer, the fragments found in the caves of Qumran, especially 4QJerb, have made the inquiry in many ways a Teumessian fox, seemingly destined never to be caught. Pre-Dead Sea Scrolls scholarship has often discounted the Old Greek version as an abridgement of the Proto-Masoretic tradition, deleting double readings and translating away information considered superfluous. On account of support from 4QJerb foran early Vorlage reflecting the Old Greek version, later scholarship has favored LXXJer due in large part to “expansion theory.” When scholars began searching for new defenses for the priority (even validity) of the Masoretic tradition to balance this trend, it became apparent that text-criticism had a limited ability to move behind either tradition to a single autograph. Accordingly, I will give here an overview of critical interpretation of the text and then give special attention to mediating theories emphasizing the independence of the two traditions. Since most recent scholarship has dealt with test cases in the text from which to extrapolate broader trends, I will also draw upon case studies in light of general theory.

2. Features and Witnesses of the Text Forms

The most marked forms of discrepancy between the two traditions are:

1. The text of LXXJer is approximately 14 percent shorter than MTJer, according to the calculation of Y-J. Min.[1]The difference between the two versions of Jeremiah is greater than that of any other Old Testament book.[2]While most of the pluses of the MT are minor, significant pluses appear in numerous prose sections. The biographical account includes 39:4-13, the largest plus in the section. The MT also includes several repetitions (8:10b-12; 17:3-4; 30:10-11; 48:40b.41b) and prophecy headings (2:1-2; 7:12; 16:1; 27:1; 46:1; 47:1).[3]
2. Many sections of the two versions appear in significantly different orders. Many of the bisectional differences are touched on by Tov and Thackeray, and T. Michael has recently produced a more thorough study.[4] The most significant difference is in the placement of the oracles against the nations. Where the MT appends them to the end of the book (46-51), the LXX places them just after the oracle against Judah (25:14-31:44).[5] In the overall interpretation of Jeremiah’s origin many consider this discrepancy indicative of a late addition, though many have reacted strongly against that claim.[6]

Between the two traditions, majority representation falls to the MT. The Qumran witnesses which favor it include 2QJer and 4QJera,c,d. 4QJera bears the closest resemblance to the proto-Masoretic text, with only insignificant differences.[7] While the differences in 4QJerc are more numerous, they consist primarily of the addition of wawand yod.[8]2QJer, though representative of the same text form, differs more regularly.[9]In favor the LXX version, 4QJerb stands alone in the Qumran collection. Note that in Janzen’s discussion elaborated below ‘4QJerb’ is an umbrella term denoting 4QJerb, 4QJerd and 4Jere. As the differences between these texts became clearer (dimensions of the material in reconstruction and style of handwriting), the three came to be considered separately.[10] Nonetheless,Emanuel Tov does admit that certain tendencies of 4QJerd also parallel LLXJer, from details of language to major recensional trends. Therefore he suggests that they might have existed in the same scroll as part of a single collection, though written by different scribes.[11]The LXX version is also witnessed in the Vetus Latina, some of the implications of which have been handled by Pierre-Maurice Bogaert.[12] Tov has produced an excellent overview of the text evidence for either version, and Ziegler’s critical edition provides more information regarding ancient versions.[13]

3. The Effects of Qumran

J. Gerald Janzen showed LXXJer to reflect an early revision of the text. Through conflate readings he characterized the two text forms in response to the generally accepted position of Otto Eissfeldt.[14] In many places he also defended the superiority of the Greek text and overall evinced expansion theory’s explanation of the textual traditions.[15] As Watts explains, “[he] concluded that the Hebrew Vorlage of the LXX was very conservative and separated by only a few text generations from the original.”[16] Much of the scholarly opinion on the subject has followed suit on the subject. Yohanan Goldman initially contested Janzen’s view strongly. He later changed his mind based on case studies of key passages throughout Jeremiah, arguing against the assertions that LXX minuses should be explained through scribal error or interpretive translations.[17] Sven Soderlund, in response to Janzen’s work, argued through a limited case study on Jeremiah 29 that a discontinuity exists in the focus of the Septuagint text, and it must therefore be a secondary inclusion.[18] Janzen later turned the argument around, showing that it was in fact the MT which bears the discontinuity, further strengthening his original claim in favor of the LXX.

Sigmund Mowinckel proposed a reconstruction of the text which has long been a centerpiece of the dialogue, with many variations stemming from it. He divided the book into four primary sections including the oracles of Jeremiah, the biography, the dueteronomistic material and the positive oracles. This is not only important because it emphasizes the different sources contributing to the composition of Jeremiah (labeled by Mowinckel sources A through D respectively); but it also labels the OAN a secondary addition to the text. Mowinckel suggested that it circulated separately for a time before becoming part of prophetic material attributed to Jeremiah.[19] Janzen later advocated this position and explained the inclusion of the OAN in a section of his overall reconstruction for LXX priority.[20]

E. Tov and P-M. Bogaert furthered Janzen’s claims by positing theories about howthe conservative Vorlage of the LXX could undergo the patterns of redaction they collated to become the expansionistic MT. Both evinced two progressive stages of development from a mutual Hebrew precursor. Though not perfectly represented in every instance, the LXX represents the earlier development.[21] While Thackeray believed the work of many translators helped develop the Greek version,[22] Tov argued after Qumran that the differences were the work of a single translator and an editor working from the early Vorlage; this would later be developed again in MTJer. Watts further notes an important emphasis of Tov: “[he] argued that the differences between the two are not textual but redactional. It is therefore a mistake to "correct" one on the basis of the other.”[23] In respect to the idea of translation, Michael argues that the LXX translator’s inconsistency with recurring Hebrew phrases across the two sections limits the weight of evidence which can be placed on redactional studies emphasizing Hebrew-Greek consistency.[24]

Bogaert produced one of the most detailed assessments of the redactional process behind the MT tradition. After his extensive assessment, he attributes the majority of differences to the work of a single editor/translator. Moreover, he argues also for a date for the latter of the two stages. Based on his assessment of the texts, he estimated the origins of MTJer around the third century BCE.[25]

Discoveries among the Dead Sea Scrolls from the “Babylonian” and “Egyptian” textual traditions have shown that the differences between the texts are not the poor or prejudicial work of translators.[26] Becking argues fervently, “Even in the explicit differences in content no obvious interpretative or theological tendency can be detected.”[27] Therefore whatever differences exist, they are either informative or indicative of independent content.

4. Case Studies

4.1 In Favor of LXXJer

Johan Lust gives a compelling argument for the superiority of the LXX in a study of Jer 33:14-26, the longest plus in the MT. After surveying several perspectives of the textual data from general theory, he assesses the case with all of them. He attributes many anomalies of language in the passage to a later hand, including hapaxes, unusual word forms, an uncommon literary device and a rarely evidenced substitution. Changes in content and referents also suggest a redactional purpose behind the plus. For example, he shows that wnqdxhwhy has been changed from the original personal name referring to Zedekiah for the purpose of refocusing the passage on the revival of Jerusalem, to which the phrase refers in MTJer. He concludes that the Hebrew of the MT in this section is not composed by Jeremiah. Furthermore, the expansions change the intended purpose of the section. For these reasons Lust prefers the Greek version as nearer to the original. He also conjectures a possible time for the inclusion of this plus by comparing it to a similar textual focus in Zech 12:10-14 and similar poetry in Sirach 45:23-26. He suggests that the plus may have been composed around the same time as Zech 12 and of Ben Sirach’sgrandfather.[28]

Anneli Aejmelaeus also defends the priority of the LXX when considering the dueteronomistic tendencies and “ideological novelties” of 25:1-14. His basic premise is that dueteronomistic language is easy to imitate in later redactions, and since features of ideologies from 2nd century BCE Palestine appear often in the section, scholars should be wary of asserting the homogeneity of MTJer.[29] Among these features he notes the regular clarification of “the enemy from the north,” the change of the first person referent in 25:3a-7 and the identity of those receiving punishment. Aejmelaeus concludes from the nature of expansions and the sensitivity to contemporary theological concerns in emendations, that the shorter of the two is closer to the autograph. He further proposes how LXXJer might have remained unaffected by the further redacted form of MTJer: “The existence of two so different editions of the text side by side may be explained by the fact that the Vorlage of the Septuagint (i.e., the Hebrew source text of the translation), having been brought to Alexandria, remained there in geographical isolation and was largely untouched by the final Palestinian edition.”[30] He claims that this disconnection would be due in large part to the fact that a work of such strongly Palestinian influence would not have found popularity in Egypt over the simpler version.

4.2 In Defense of MTJer

From here we will look at a few of the case studies which have been done in defense of the MT tradition. In light of criticism favoring the LXX, Duane Christensen defends the textual integrity of the MT in his study of chapter 25. His main point is that the internal cohesion of the chapter is defensible through prosodic analysis. Using the LXX in a reconstruction of Jeremiah’s poetry here, Christensen shows that the metrical integrity of 25:15-38 (which is not present in the LXX form) suggests that the MT pluses cannot be written of as glosses. Nonetheless, he admits that the parallels between chapter 25 and chapters 46-51 allow for strong argument in favor of the LXX.[31] That is to say that Christensen does not hope to establish MT priority, simply integrity as an independent tradition. This trend of emphasizing the dynamic nature of the text over the relative priority of either version has become more prevalent in later scholarship; it is important for our final assessment of the issue.

Among the passages assessed by Yohanan Goldman, Jer 27:5-15 (LXX 34:5-15) has faced able criticism from Arie van der Kooij. Goldman claimed that the internal cohesion of the passage was better preserved in LXX; this position is primarily based on linguistic structures that change the contextual function and semantic range of the Hebrew when translated to Greek. Yet Goldman admits that the structure in the MT is acceptable, simply less smooth. Concerning the task of weighing the value of each form, Kooij says, “Goldman (and others) favours the assumption that LXX being the smoother text is the better and earlier one. However, one can also argue that MT being the fuller text has been shortened in order to get a smoother and easier text.”[32] This problem of two acceptable explanations for the same problem characterizes much of the text-critical and redactional-critical research being done in Jeremiah. Kooij addresses the issue by contesting many of the assumptions behind Goldman’s study, most of them dealing with homoioteleuton in the LXX translation and translation of Hebrew idioms (a position which Aejmelaeus likewise considers relevant, though to a limited extent[33]). Finding many of Goldman’s supports weak, Kooij concludes that MTJer reflects an earlier text in 27:5-15. These assumptions with which Kooij wrestles are common with many scholars preferring LXXJer, and Kooij’s challenge brings their value into question. Having viewed a number of case studies in both directions, it is important to consider the contribution of mediating theories.[34]

5. Mediating Theories

Generally speaking, early attempts to produce mediating theories for the relationship of the two forms were not well received, but their value in tempering the dialogue is worthy of note. These theories originally tended to argue that the dynamic textual traditions made the relative priority of either text indiscernible.[35] Some have furthered that principle by proposing alternative theories of textual origin. Here we will address the value of some of these theories in establishing alternative questions for redaction criticism.

A summary of some of the assumptions which preclude mediating theories appears in the Jewish Encyclopedia by Victor Ryssel. Ryssel admits that many of the pluses of MTJer may be later additions, but he shows how this does not make the MT dependent on the LXX version. His qualifying claim is that though the Greek may reflect an older version, the pluses of MTJer might have been in development even during the process of translation. He clearly argues that the translation of the Septuagint occurred before work on the Hebrew version had come to an end, simply interrupting or representing a specific moment in the redactional process of the intended final copy. He says concerning the nature of their side by side progression, “the correspondence between the Septuagint and the Hebrew is too great, and their relationship too close, for one to be able to speak of two redactions. They are rather two editions of the same redaction.”[36]It is important to remember that mediating theories do not necessarily attempt to dismantle theories of when the text forms developed in relation to one another; they simply address the nature of the relation as non-derivative and therefore non-preferential.

Bob Becking deals thoroughly with the relationship between the two forms of the Book of Consolation. He addresses the many issues present in this central section, including zero-variants, corruptions, linguistic exegesis, content variants and arrangement of units. While he concedes that many arguments made for intentional interpretive translation and redaction are possible, he reminds us that they are not necessary. He shows that the majority of the discrepancies between the two versions can be attributed to linguistic exegesis and/or minor differences in content, and he emphasizes that these changes do not change the theological purpose of the passage. He says, “the comparison of the two texts of the Book of Consolation did not suggest that any support for the hypothesis that the Old Greek version is a text which was intentionally abridged from its Vorlage or that the MT contains an expanded version of a superior text: the Hebrew text underlying the Old Greek.”[37] Becking admits that his study does not prove different original versions of the text; nonetheless, if we accept his premise it does prohibit preference being given to either text form. Though he does not consider it certain, he does promote the theory of two originals as one of two primary explanations for the difference in order in the Book of Consolation:

It is possible to assume that two diverging collections of material related to the prophet Jeremiah once circulated independently. The question of the superiority of one text over the other, consequently, is of a literary-critical and redaction-critical nature. An answer to this question relates to one’s overall view regarding the genesis of the book of Jeremiah.[38]

He cites Christoph Levin’s theory that each developed with an independent Fortschreibung and in a different location from the other.[39] While Carolyn Sharp clearly shows in her study of the oracles against Egypt and Babylon that entirely independent Vorlagen cannot be evinced beyond reasonable doubt, she does say that “it should no longer be assumed that secondary editorial activity, whether reflected in expansion in individual readings or in the larger literary ordering of the OAN, precludes the possibility of original variants existing as well.”[40]

While the scope of this study is limited, we see from what has come before that the dynamic nature of the textual traditions is in many places so uncertain that establishing a clear relative priority is impossible. Moreover it is not the case, as Bogaert suggests, that valid defenses of the MT are few in number. The application of many text-critical tools provides valuable arguments for either side in a great number of passages, some of which are very significant pluses or minuses (most notably the OAN and the Book of Consolation). As Christensen states in the summation of his study: