Critical Jurnal Ibrani Frimus.docx

Abstract This article aims to confront and question the theoretical distinction between textual criticism and redaction criticism from a pragmatic perspective. In order to accomplish this goal, we will examine the Hebrew manuscripts of Ben Sira as a test case and a paradigmatic example. The following situations will be examined: cases of irreducible divergences between the Hebrew witnesses, scribal “mistakes,” doublets in MSS A and B, and the so-called Hebrew II. It is well known that the different Hebrew medieval manuscripts1 of Ben Sira present significant divergences and raise many questions about the “Evolution of the Hebrew Text,” to recall the title of the seminal article by Moshe H. Segal.2 This complex textual situation offers a real challenge for producing a critical edition. The different witnesses and their ancient translations have been compared mainly by scholars to reconstruct the “original” Hebrew text written by Ben Sira or at least “the earliest inferable textual state.”3 Typical examples are the eclectic edition produced by Rudolf Smend at the beginning of the twentieth century,4 the Anchor Bible commentary of Patrick Skehan and Alexander Di Lella,5 and the preliminary work of Hans Peter Rüger.6 While such an approach could be useful, little attention has been paid to the textual plurality and fluidity of these medieval witnesses from a linguistic, philological, and hermeneutical point of view. The edition of the Hebrew fragments that Eric Reymond, Jan Joosten, and I are preparing is meant to join this important discussion. We decided to edit each manuscript independently, without trying to propose any reconstruction of the supposed original or archetype to emphasize the dynamic dimension of the text and the transformation it has undergone over time. Our edition aims to reproduce the text preserved in the various manuscripts as accurately as possible. We will provide each text with a critical apparatus,7 but also a translation and a philological commentary for each specific textual situation (main text, marginal readings, word crossed out, etc.). This edition is intended to highlight the pluriformity of the text and to present it as a product in perpetual development, and as a process of writing through the insightful works of scribes. One of the major theoretical problems in textual editions of ancient Jewish literature is the relationship between textual criticism and redaction criticism. Textual criticism generally distinguishes between two stages in textual history: the redaction stage of literary growth and development, and the stage of text transmission. As noticed by Ron Hendel: “A critical edition tracks changes that occur during the textual transmission of books, and not to literary states prior to the ‘finished literary product’ (Tov’s phrase) of a given book or edition.”8 According to Emanuel Tov: At a certain point in time, the literary growth of the biblical books necessarily ended, at least for those who accepted the present canonical form of the books as final. The date of this moment cannot be determined easily, and it differed from book to book, but despite these and other difficulties it is reasonable to assume that at that time the actual textual transmission began.9 The discussions related to this distinction are not new and focus mainly on the possibility of distinguishing clearly between these two stages. In 1975 in his article “The Textual Study of the

Bible: A New Outlook,” Shemaryahu Talmon already concluded that it is difficult to maintain a clear distinction between the redaction and transmission processes.10 Similarly, the work of Eugene Ulrich questioned this assumption.11 For example, in 2006, he noticed that [t]he organic process that characterized the growth of the biblical texts over centuries relegated the concept of an Urtext to a more distant and foggy position or at least into a more blurred series of Urtexte, since it becomes difficult to decide on principle which one from a series of editions should be chosen as the text.12 In the same vein, in 2013, George Brooke rethought this distinction between higher and lower criticism,13 and more recently, Ronald Troxel published successively two articles related to this topic.14 In this article, I would like to question this theoretical distinction from a pragmatic perspective, using examples from the Hebrew manuscripts of Ben Sira. I start from the assumption that the pluriformity of these documents may illustrate one of the major theoretical problems in the edition of texts. More specifically, I suppose that the process of textual transmission and transformation of the Hebrew text of Ben Sira illustrates what happened with the biblical text before it acquired an authoritative status. The book of Ben Sira could be a paradigmatic example of the overlapping of textual development and textual transmission: (1) it is one of the rare books of Jewish antiquity in regard to which we have a good idea of its date of composition; (2) each textual witness is a photograph of the text at various stages of its transmission: from the Hellenistic period with the Masada and Qumran manuscripts, and from the medieval period for the Genizah witnesses that can be supplemented with the Greek and Syriac translations. From the numerous examples, I select the following:    

Irreducible discrepancies between Hebrew witnesses; Scribal mistakes; Doublets in MSS A and B; The so-called Hebrew II (Hebrew verses missing in the Greek and the Syriac).

1 Irreducible Discrepancies between Hebrew Witnesses Irreducible discrepancies between Hebrew witnesses provide good illustrations of scribal interventions and the evolution of texts in the course of the transmission. One example, among many others, is sufficient to illustrate situations where two different Hebrew texts agree with the ancient translations—Greek and Syriac—but disagree with one another: Sir 3:14, Ms A ‫צדקת אב לא תמחה‬

Sir 3:14, Ms C ‫צדקת אב אל תשכח‬

‫ותמור חטאת היא תנתע תנטע‬ ]‫ותחת ענותו תתנצ[ב‬ Kindness to a father will not be blotted out

Sir 3:14, Ms A Sir 3:14, Ms C And instead of sin it will be crushed ⟨mg. it will be Do not forget to show kindness to your father, planted⟩ And instead of sins (?) it will sta[nd] In this verse, manuscripts A and C say, more or less, the same thing: “we must not forget charity/justice/kindness (tsedaqah) to a father and this tsedaqah will be placed, planted, instead of sin.” With the exception of the syntagm .sdrow yna erahs ton od stxet owt eht ,‫צדקת אב‬ The two sentences present easily recognizable scribal mistakes: In MS A, dedneme eb tsum ‫תנתע‬ a sseltbuod si ”ytilimuh sih“ ‫ ענותו‬,C SM ni 51;ebircs eht yb nigram eht ni detcerroc sa ‫ תנטע‬ot ”.snis sih“ ‫ עונתו‬rof sisehtatem Both texts present linguistic divergences. Hans Peter Rüger solved the problem by suggesting that the reading of MS C is older than MS A,16 but the problem seems more complex: 





)evititev ro( evitacidni na stneserp A SM .)C SM( ‫ אל תשכח‬susrev )A SM( ‫לא תמחה‬ C has a prohibitive form. As the reading of A agrees with the reek MSstatement where eht taht timda dluohs ew ,snoisrev )‫ (ܐܠ ܡܬܛܥܝܐ‬cairyS eht dna )[ιατεσήθσηλιπἐ] κὐο( ariS neB ni erehwesle detsetta sa ,teY .C SM fo gnidaer eht dedecerp A SM fo ‫ לא‬evitagen snoitcurtsnoc citcatnys owt eht neewteb noitcnitsid citnames eht ,sllorcS aeS daeD eht dna emaceb snoitcurtsnoc owt eht dna raeppasid ot dednet ‫ אל תקטול‬dna ‫לא תקטול‬ So, although the reading of 17interchangeable in the Hebrew of the Hellenistic period. .dedulcxe eb ot ton dluohs C SM fo taht ,elbissop si A SM The distinction between ylesolc era sbrev owt ehT .detacilpmoc erom si ‫ תשכח‬dna ‫תמחה‬ related semantically and while the first agrees with the reek, the second agrees with the .cairyS ,narmuQ ta ron werbeH lacilbiB ni rehtien detsetta si ,)A SM( 01:4 ni tneserp osla ,‫תמור‬ .elbareferp eb yam C SM fo ‫ תחת‬taht timda dluow I ,esac siht nI .mituyyip ni srucco tub cairyS eht edistuo detsetta ton ,msiamarA etal a ylraelc si C SM fo ‫ תתנצב‬,emit emas eht tA .elbareferp si ,A SM fo nigram eht ni ,‫ תנטע‬taht edulcnoc I ,eroferehT .egaugnal

In conclusion, the manuscripts attest to two quite different Hebrew texts. Both partially agree with ancient Greek and Syriac versions and both seem to have undergone transformations during the transmission. The fact that there are such clear differences between the two texts leads us to think that neither one reflects the “original” text. But where do they come from? We cannot explain all divergences as scribal errors: how can we explain the change of ot ‫ חטאת‬fo ,‫ תמור‬ot ‫ תחת‬fo ro ,‫ תשכח‬ot ‫תמחה‬ citsiugnil ,gnidrower ,gnitirw evitaerc fo kaeps ew naC ?asrev eciv ro ,‫ תתנצב‬ot ‫ תנטע‬fo ,‫עוון‬ adaptation, or retroversion from languages other than Hebrew (for example, in ort.t)? Is Sir t:14 ?mituyyip eht fo emit eht ta gnidrower etal a ro doirep citsinelleH eht ni ‫ תמור‬fo noitatsetta tsrif eht ,rO ?‫ אל תקטול‬noitcurtsnoc evitibihorp eht rof evitanretla citsiugnil a ro evitacidni na ‫ לא תמחה‬sI ?sralohcs emos yb detseggus sa evitacidni eht fo esnes eht ni desu ‫ אל תשכח‬si ,ylesrever

In conclusion, both manuscripts attest to the existence of radically different Hebrew texts. Both agree partially with ancient versions (Greek and Syriac) and both seem to have undergone transformations in their transmission. In this case, it seems impossible to reconstruct the “earliest inferable state of the text.” We seem to be dealing with a scribal practice of stylistic and textual interchangeability of lexemes also attested in biblical variants, as demonstrated by Shemaryahu Talmon in 1961.18 In such cases, it seems necessary to edit both forms of the sentence to illustrate the phenomenon of textual development produced by scribes. 2 Scribal Mistakes I would like to address what may appear as the weakest point of our principle of edition: evident scribal “errors.” The concept of “scribal errors” is, by itself, a complex theoretical issue that implies a number of cognitive factors that are beyond the scope of this study. By “error,” I consider cases where we can reconstruct the etiology of the final form, which textual critics call the principle of explicability (“principe d’ explicabilité”).19 In Sir 4:30 (MS A), we read: lufraef dna desserppo dna esuoh ruoy ni god a ekil eb ton oD‫אל תהי ככלב בביתך ומוזר ומתירא במלאכתך׃‬ .stca ruoy ni MS C gives a slightly different reading: .skrow ruoy ni tnagorra dna esuoh ruoy ni noil a ekil ton eB‫אל תהי כאריה בביתך ומתפחז בעבודתך‬ The LXX provides the following translation: μὴ ἴσθι ὡς λέων ἐν τῷ οἴκῳ σου καὶ φαντασιοκοπῶν ἐν τοῖς οἰκέταις σουDo not be like a lion in your home, even conceiving vain fancies among your domestics. On the other hand, the Syriac exactly follows MS A: lufraef dna elbirret dna esuoh ruoy ni god a eb ton oD‫ܒܥܒܝܕܬܟ‬ ̈ ‫ܐܠ ܬܗܘܐ ܟܠܒ ܒܒܝܬܟ ܘܙܥܝܦ ܘܕܚܝܠ‬ .skrow ruoy ni The transformation process is clear. As recognized by Smend, Lévi, and Segal,20 the Vorlage of the Greek certainly was dna C SM ni ‫ כאריה‬ot ymynonys fo yaw yb demrofsnart ”,noil a sa“ ‫כלביא‬ ni ”god a sa“ ‫ ככלב‬ot ,yhpargottid fo yaw yb ,yllanif dna )cairyS eht ni sa( ”god“ ‫ כלב‬ot deretla ςῖοτ νἐ keer eht fo egalroV( ‫ בעבדיך‬fo noitamrofsnart eht snialpxe ssecorp emas ehT 12.A SM dna C SM ni detsetta ”,skrow ruoy ni“ ,‫ בעבודתך‬ot )”scitsemod ruoy gnoma“ ,υοσ ςιατέκἰο eht taht sesoppus gnidaer sihT .A SM ni ”,stca ruoy ni“ ,‫ במלאכתך‬mynonys eht yb detutitsbus .cairyS eht yb demrifnoc sa ,A SM fo esoht dedecerp C SM fo snoitamrofsnart

In this case, one may reconstruct the earliest inferable text state ( emos htiw )‫אל תהי כלביא בביתך‬ A and C do not represent the earliest stages of MSScertainty. But even if we are convinced that the text, it is these transformed texts that must be published for the following reasons: 1) A pragmatic reason: in many cases we are not able to reconstruct the preceding state of the text, and must be consistent; 2) We want to present the text as it has been preserved, and to the edition highlight the impact of scribes on the text and the transformations it has undergone; and t) The he preservation of A and C make perfect sense and generate new proverbs. T MSSHebrew texts of the new state of the proverb is fundamental for paremiological studies in diachronic, linguistic, 22 .sevitcepsrep larutluc dna 3 Doublets in MSS A and B The Hebrew manuscripts of Ben Sira share a characteristic that has intrigued scholars: the numerous doublets are related to the history of the transmission of the text and to scribal practices. The fact that these doublets are missing in the Greek and Syriac translations prove that they have appeared during the transmission process in later periods. MS B presents doublets absent in MS A. As a consequence, these doublets must be later than the text of MS A and are likely the result of the conflation of different textual traditions. 3.1 Textual Variation or Textual Development: Sir 11:3 MS A fo feihc eht si tiurf sti tub ,eeb eht si sdrib gnoma tnacifingisnI‫אליל בעוף דברה וראש תנובות פריה׃‬ .stcudorp MS B tub ,eeb eht si sdrib gnoma llamS‫קטנה בעוף דבורה וראש תנובות פריה׃אליל בעוף דבורה וראש תנובות ̇פ ̇רי̇ ה׃‬ its fruit is the chief of products.Insignificant among birds is the bee, but its fruit is the chief of .stcudorp LXX μικρὰ ἐν πετεινοῖς μέλισσα καὶ ἀρχὴ γλυκασμάτων ὁ καρπὸς αὐτῆςSmall among flying creatures is a bee, and the origin of sweet things is its produce. Syriac eeb eht serutaerc degniw gnoma esuaceB‫ܥܠܠܬܐ ܐܒܗ‬ ̈ ‫ܡܛܠ ܕܫܝܛܐ ܗܝ ܒܥܘܦܐ ܕܒܘܪܬܐ ܘܪܝܫ ܛܘܒܐ ܕܟܠ‬ .ecudorp lla fo tseb eht si tiurf sti tey,desipsed si The only variant between the two forms of the verse in MS B is the alternation between dna ‫קטנה‬ .)B91:0t riS .fc( ”slodi“ etangised ot larulp eht ni desu ylniam si ‫ אליל‬,werbeH lacilbiB nI .‫אליל‬ is rarely attested (rob —B MSthe verse of A and in the second part of MSfound in —The singular tnelaviuqe cairyS eht .fc( ”niav ,tnacifingisni“ snaem dna ,)71:11 hceZ ;41:41 rer ;01:01 asI ;4:t1

werbeH cinibbaR ni sraeppasid drow ehT .eeb eht fo ”ytliarf ,ssenkaew“ eht sekove dna ,)‫ܐܠܝܠ‬ tation of rob 1t:4 in b. uul. 121a) but reappears frequently in piyyutim in the (except for a quo singular to designate alternatively an “idol” or “nothingness” (see, for example, Yose ben Yose, .)selpmaxe tseilrae eht rof auhsoheY neb nehokaH nanahoY ,rîlaQaH razaelE ,iannaY Finally, both forms of the verse in MS B, eht dna keer eht htiw ylevitcepser eerga ,‫ קטנה‬dna ‫אליל‬ seod ylsuoiruc hcihw ,cairyS eht elihw t2,‫ קטנה‬ot sdnopserroc άρκιμ keer ehT .noitalsnart cairyS .‫ אליל‬sredner ”,sselhtrow ,elbitpmetnoc ,desipsed“ ,‫ ܫܝܛܐ‬tub ,‫ ܐܠܝܠ‬esu ton Concerning the historical process, one of the two forms of the doublet is pristine. One possible scenario is that ni sraeppasid drow eht sA .noitalumrof lanigiro eht stcelfer B dna A SSM ni ‫אליל‬ Rabbinic Hebrew, or may be confused with idol, a scribe may have replaced it with the more ,rehto eht retfa eno ,htob deipoc dna esrev eht fo smrof owt eht wenk ebircs retal A .‫ קטנה‬nommoc 42 .elbissop osla si )yradnoces ‫ אליל‬dna enitsirp saw ‫ (קטנה‬oiranecs etisoppo eht ,revewoH .B SM ni Should we consider this doublet as a textual variant and try to reconstruct the earliest inferable textual state? In that case, we would have to include in the edition only one form of the verse lanoitcader fo mrof a eb ot telbuod siht redisnoc ew dluohs rO .‫ אליל‬dna ‫ קטנה‬neewteb gnisoohc process even if we are convinced that this process is late, probably later than the text preserved in ?A SM This case may be compared to the numerous doublets attested in the Hebrew Bible and carefully studied by Shemaryahu Talmon in 1960,25 as illustrated by one of Talmon’s examples (Exod t0:6): ‫ל־הֵ֣עֵ ֹ֔דת‬ ָ ַ‫ֲשׁר עַל־א ֲֵ֣רֹ ן הָ עֵדֻ֑תלִ פְ נֵ ֵ֣ י הַ כּ ֹ֗ ַֹפּ ֶרת אֲ שֶׁ ר֙ ע‬ ֶ֖ ֶ ‫וְ נָתַ ָ ָ֤תּה אֹ תֹ ו֙ לִ פְ נֵ ֵ֣י הַ פָּרֹֹ֔ כֶת א‬ The LXX, several manuscripts of the MT, and the SP do not attest the second part of the verse. The second part “is distinguished from the other by the omission of the word fo esu eht dna ‫ארן‬ lanif siht ,nomlaT yb detartsnomed sA 62”.srettel fo noisrevni na htiw ‫ פרכת‬fo tnelaviuqe na ,‫כפרת‬ he numerous form is a conflation of two independent textual forms. But in such a case, and in t similar examples of doublets documented by Talmon, should we consider the second part of Exod t0:6 to be part of the redactional process or to be a textual variant? More generally, should we d in Ben Sira to be textual variants or the consider the numerous doublets in the Hebrew Bible an .drawrofthgiarts ton si noitseuq a hcus ot esnopser ehT ?ssecorp lanoitcader a fo tluser 3.2 Sir 4:3–4a: When Textual Variations Imply Stylistic Improvement In Sir 4:3–4a, which is only preserved in MS A, we observe another doublet (vv. 3a and b) missing in the Greek and Syriac translations: desserppo na fo hcamots eht ekam ton oD‫אל תמנע מתן ממסכינך‬4‫אל תחמיר מעי דך וקרב עני אל תכאיב׃‬t o not refuse a gift to the person churn,and the entrails of the poor, do not impose hurt on them.D uoy raen si ohw tnegidni LXX

καρδίαν παρωργισμένην μὴ προσταράξῃςκαὶ μὴ παρελκύσῃς δόσιν προσδεομένουAn angry heart do not trouble,And do not delay giving to one in need Syriac .‫ܡܥܘܗܝ ܕܐܢܫܐ ܡܣܟܢܐ ܐܠ ܬܟܐܒ‬ ̈ roop eht fo sliartne eht ekam ton oD‫ܘܐܠ ܬܟܐܠ ܡܘܗܒܬܐ ܡܢ ܨܪܝܟܐ‬ ydeen eht morf tfig a dlohhtiw ton od dnAreffus nam If one of the two stichs of v. 3 has been added secondarily, perhaps by conflating two alternative readings, we realize that the scribe takes an active part in the poetic construction of the text with a beautiful chiastic parallelism (verb—complement versus complement—verb). Moreover, the combination of synonyms is not without biblical echoes: 02:1 maL ni detcennoc era ‫ קרב‬dna ‫מעה‬ knil eht rof ;41:02 bor dna ,11:61 asI ,)11:2 maL osla ees ;‫ חמר‬dna ‫ מעה‬setaicossa osla hcihw( .12:47 sP ees ,‫ עני‬dna ‫ דך‬neewteb It seems that the doublet, even if it comes from the conflation of two textual traditions, strengthens the poetic value of the images used here, and should be considered a textual development rather that a textual corruption. 3.3 The Doublet in Sir 31:16: When Scribes Retrovert from the Syriac Translation into Hebrew The scholarly discussion about retroversion from Syriac into the Genizah manuscripts has a long history. However, except for Sir 51:13–30,27 the numerous examples provided by Di Lella, among others, are not convincing.28 At the end of the nineteenth century, Israel Lévi and Victor Ryssel suggested that the doublets of MS B of Ben Sira were retroverted from the Syriac translation.29 Unfortunately, the examples proposed by Ryssel are unconvincing as numerous doublets differ radically from the Syriac. But Lévi’s and Ryssel’s intuition may have been partially correct, as shown by Sir 31:16:30 ‫הסב כאיש אשר נבחר‬

(1) Sit down (at a banquet) like a man chosen,

‫אל תעט פן תגעל׃‬

Do not fall upon (food) lest you be rejected.

‫[אכל כאיש נכח‬

[(2) Eat like a honest man,

]‫ואל תעט פן תגלו תגלע׃‬

And do not fall upon (food) lest you be revealed.]

‫דע שרעך כמוך‬

(3) Know that your neighbor is like you

‫ ואכול כאיש דבר ששם לפניך‬And eat like a man whatever that one set before you, ‫ולא תהיה גרגרן פן תמאס׃‬

And do not be a glutton lest you be abhorred

)…(

(…)

17b‫ואל תלע פן תמאס‬

17b (4) Do not gulp down lest you be abhorred.

The textual situation of this verse is particularly complex. The proverbial saying concerning food is repeated four times, while it appears only once in the Greek and the Syriac. The first occurrence (1) is in the body of the text; the second (2) is in the margin;31 the third (3), which is disrupted from the preceding sentence by a misplaced phrase related to one’s neighbor ( = ‫דע שרעך כמוך‬ he left margin. minally, t1:15a), becomes a tristich and for this reason has its last stich written in t the second colon of the sentence is also attested in t1:17b for the fourth time (4). In this series of in addition to the fact that the scribe copied alternative —doublets, I would like to demonstrate that the third form of the sentence (t) is retroverted from the Syriac —forms of the proverb (1, 2, 4) .noisrev The first form (1) has undergone several transformations independently of the Greek and Syriac traditions: rehto eht lla ni detsetta ‫ אכל‬nommoc erom eht fo tnelaviuqe detacitsihpos a si ‫הסב‬ ‫ כאיש‬esarhp eht ;)‫ (ܠܥܣ‬cairyS eht dna )εγάφ( keer eht sa llew sa ,ecnetnes eht fo smrof werbeH cihparg sa txet eht fo noissimsnart eht ni sraeppa ylniatrec ”,nesohc si ohw nam a ekil“ ,‫אשר נבחר‬ cairyS eht dna ιοσ άνεμίεκαραπ ὰτ keer eht ees( ”uoy erofeb si tahw“ ,‫ אשר נכחך‬htiw noisufnoc 2t .)2( ecnetnes siht fo noisrev lanigram eht ni detsetta ‫ נכח‬ot esolc ,)‫ܕܣܝܡ ܩܕܡܝܟ‬ The marginal version of this verse (2) has also undergone various transformations: hcihw ,‫כאיש נכח‬ a meaning surprisingly not so far ttis certainly misleading, could mean: “Eat like an honest man,” detcerroc niaga si brev sihT .txetnoc eht ni erucsbo si ‫ תגלו‬mrof eht ;)1( txet eht fo ydob eht morf ni detsetta ‫ תגעל‬rof sisehtatem a eb dluow dna retteb yllaer ton si hcihw 4t,‫ תגלע‬ot nigram eht ni .ecnetnes eht fo mrof suoenorre na deipoc ylsuolupurcs sah ebircs ehT .txet niam eht The third version (3) of this proverb is certainly the most interesting. Indeed, (a) this version differs clearly from the two previous forms; (b) Both stichs are unusually long. (c) The Hebrew ( ‫ואכול‬ taE“ ,‫ (ܠܥܣ ܐܝܟ ܓܒܪܐ ܡܕܡ ܕܣܝܡ ܩܕܡܝܟ‬cairyS eht fo tnelaviuqe mitabrev a si )‫כאיש דבר ששם לפניך‬ ‫ ܘܐܠ‬:cairyS eht no sdneped ylraelc noloc driht ehT 5t.)”uoy erofeb tup si )reve(tahw nam a sa eht dna xatnys ehT )d( ”.detah eb uoy tsel nottulg a eb ton od dna“ ‫ܬܗܘܐ ܓܓܪܢ ܕܐܠ ܬܣܬܢܐ‬ vocabulary of this stich is uncommon in Classical Hebrew and follows isomorphically the syntax eht ot yltcefrep sdnopserroc hcihw ,‫ לא תהיה‬,evititev werbeH eht :cairyS eht fo yralubacov eht dna eman eht ;‫ אל תהי‬evissuj eht neeb evah dluohs ylbareferp dna nommoc ton si ,‫ ܐܠ ܬܗܘܐ‬cairyS dna werbeH cinibbaR ni detsetta ylno si ,‫ ܓܓܪܢ‬cairyS eht fo noitisopsnart a si hcihw ,‫גרגרן‬ far from what may have been the supposed Babylonian Aramaic. (e) minally, this version is .)‫תמאס‬/‫ (אכל כאיש אשר נכחך ואל תעט פן תגעל‬lanigiro Based on these insights, we conclude that this third version of the sentence was retroverted by a scribe from the Syriac version. This hypothesis is confirmed by the misplaced doublet of 31:15a ruoy fo taht wonK“ ‫ ܕܥ ܕܚܒܪܟ ܐܝܟ ܕܝܠܟ‬,revoeroM 6t.61 .v fo elddim eht ni ‫דע שרעך כמוך‬ ” has been retroverted into Hebrew, word for word, but has a companion as that of your own, fo ssecorp ehT 7t”.uoy ekil si noinapmoc ruoy taht wonK“ ‫ דע שרעך כמוך‬:gninaem tnereffid yllatot 8t .‫ ידע כי‬fo ecalp ni ‫ ידע אשר‬noitcurtsnoc suolamona eht snialpxe osla noisrevorter

In conclusion, this example in Sir t1:16 shows clearly the scribe’s impact on the text. The history of the verse may be reconstructed as follows. Upon copying the first form of the proverb (1) and noticing that the meaning of the colon was not correct (but on which basis?), the scribe decided to retranslate the proverb into Hebrew from the Syriac version (3), but he mistakenly retranslated vv. 15 and 16ab together as a tristich. Finally, one of the scribes responsible for the marginal notes added in the margin a third version of the proverb (2) as attested in its other copy and corrected .‫ תגלע‬ot ‫תגלו‬ This reconstructed process shows that scribes involved in the transmission of this text were not simply copyists but were actively involved in the creation of the text as learned authors. These scribes compared their own copies with ancient translations, retroverted readings from Syriac, inserted synonymous readings, and adduced variants from other copies. The recognition of these scribal practices in antiquity and the Middle Ages complicates our practice of creating a critical edition: how should we present such a process, such a textual mouvance? For example, should we consider the retroversion from Syriac to be a textual irregularity or a legitimate part of the textual development? 4 The So-called Hebrew II (Hebrew Verses Missing in Greek and Syriac) The tension between redaction criticism and textual criticism is particularly tangible in the case of the so-called Hebrew II. In numerous cases, the Hebrew fragments present texts that are missing in the Greek and Syriac translations or in the Greek only, to be illustrated by Sir 15:14: MS B margin MS B MS A ‫ אלהים מבראשית ברא אדם הוא ֯מראש ברא אדם [א]ל[הי]ם מבראשית‬14a ‫וישתיהו ביד חותפו וישיתהו ביד [חו]ת[פ]ו‬ 14b ‫ויתנהו ביד יצרו׃‬ ‫ויתנהו ביד יצרו׃‬ 14c ‫אם תחפץ תשמר מצוה אם תחפץ תשמר מצוה‬ 15 ‫ותבונה לע רצונו׃‬ ‫ואמונה לעשות רצו֯ ן אל‬ ֯ ‫ותבונה לעשות רצונו׃‬ Verse 14b (“And delivered him into the hand of his spoiler”) is missing in reek and Syriac and is probably a later addition. This addition changes radically the interpretation of this passage on free will and, moreover, contradicts the preceding and following verses. This possible addition has been interpreted diversely by scholars: For Lévi, it is a doublet resulting from a scribal mistake;39 Osterley thinks that this stich has been added for doctrinal purposes;40 Ginzberg identifies ‫חותף‬ for Di Lella, 42and then to Essene additions; II Kearns attributes this verse to Hebrew 41with Satan; (see CEir 4:19b after the ninth century this stich was retroverted from the Syriac version of S mituyyip etiaraK ni dna ariS neB ni ereh detsetta ylno si ‫ חותף‬elpicitrap ehT t4.)241 ,dnemS ydaerla entury (roseph Ibn Abitur, Sadid Ibn Bashdad, and Tobiah ben Moshe) with from the eleventh c 44 ”.rebbor ,reppandik“ fo gninaem eht In this example, history of redaction and history of textual transmission overlap. For a critical edition, it raises numerous theoretical questions. If we decide to compose an eclectic text, should we consider this verse as a textual variant to be deleted from the edition (as done by Skehan, Di

Lella,45 and Smend,46 for example) or relegate it to the critical apparatus? Or should we consider this stich to be a legitimate textual development as the numerous cases attested in the Hebrew Bible? We are convinced that we have to present the text as it stands, even if v. 14b is a later addition. Scholars should take this crucial addition into account when considering the history of ideas and the challenging question of evil. 5 Conclusion This article started with pragmatic evaluations examining the following textual situations: (1) irreducible discrepancies between manuscripts that make it impossible to reconstruct an archetype, for which the reconstruction of an archetype would reduce the meaning and poetical potentiality resulting from this textual plurality; (2) doublets, most of which resulted from collation by scribes on the basis of variations between textual recensions; for example, the numerous marginal readings of MS B that present two or three alternatives for one verse as collated by scribes; (3) poetical and theological variations or additions; and, finally, (4) retroversions from the ancient translations. On the basis of these examples, the aim of this paper was to pose several theoretical questions. The examples did not convince me that it was possible to define an objective criterion capable of distinguishing between redaction criticism and textual criticism, not even the criterion of explicability. Texts are perpetually moving and this mouvance, to recall the concept of Paul Zumthor,47 is ontologically inherent to the concept of text. Our model of a critical edition would be likely to highlight this phenomenon. While this model is not totally Bédierist, I would like to quote his own conclusion: Aussi la méthode d’rédition la plus recommandable est-elle peut-être, en dernière analyse, celle qui régit un esprit de défiance de soi, de prudence, d’rextrême ‘conservatisme’, un énergique vouloir, porté jusqu’rau parti pris, d’rouvrir aux scribes le plus large crédit et de ne toucher au texte d’run manuscrit que l’ron imprime qu’ren cas d’rextrême nécessité: toutes les corrections conjecturales devraient être reléguées en quelque appendice. ‘Une telle méthode d’rédition, a écrit dom Quentin, risque d’rêtre bien dommageable à la critique textuelle’. Peut-être; mais c’rest, de toutes les méthodes connues, celle qui risque le moins d’rêtre dommageable aux textes.48 More generally, Ben Sira’s textual model of transmission may work as a paradigmatic example of the process of textual development of the Hebrew Bible and the implication of scribal interventions in the text before its tendency to standardization after the Hellenistic period. Indeed, for each textual phenomenon presented here we can find parallels in the Hebrew Bible, most of them recorded in Talmon’s study.49 If, in certain cases, we can partially reconstruct the historical development of the text, and reconstruct one form of an archetype, the result will be just one stage of the text among many others. This was neither the original nor the final text, but just one stage within a long process of transformation. The question, then, is: is it methodologically pertinent to produce an edition of an ideal form that we do not have and is not the original and that presents just a hypothetical and accidental photograph of one stage of the text among many in a long process?

1. This contribution will be limited to some examples from the Genizah manuscripts and will not consider Qumran and Masada material. I would also like to thank Myles Schoonover for his editing work and for correcting and improving my English. This article has been written with the support of the ANR-DFG and MSH Lorraine Project PLURITEXT and the center of research Écritures (Université de Lorraine). 2. Loshe H. Segal, “The Evolution of the Hebrew Text of Ben Sira,” JQR 25 (1934): 91–149. 3. Ronald Hendel, “The Oxford Hebrew Bible: Its Aims and a Response to Criticisms,” HBAI 2 (2013): 63–99, esp. 68. 4. Rudolf Smend, Die Weisheit des Jesus Sirach: Hebräisch und Deutsch (Berlin: Georg Reimer, 1906). 5. Alexander A. Di Lella and Patrick W. Skehan, The Wisdom of Ben Sira: A New Translation with Notes, AB 39 (New York: Doubleday, 1987). 6. Hans P. Rüger, Text und Textform im hebräischen Sirach: Untersuchungen zur Textgeschichte und Textkritik der hebräischen Sirachfragmente aus der Kairoer Geniza, BZAW 112 (Berlin: De Gruyter, 1970). 7. The critical apparatus will present variants attested in the other Hebrew witnesses. The Greek and Syriac translations will be mentioned only when they highlight Hebrew variants. 8. Hendel, “The Oxford Hebrew Bible,” 69. 9. Emanuel Tov, “Criteria for Evaluating Textual Readings: The Limitations of Textual Rules,” HTR 75 (1982): 429–448, esp. 431. Emanuel Tov changed his position on this point; see, e.g., Emanuel Tov, Textual Criticism of the Hebrew Bible, 3rd ed. (Minneapolis, MN: Fortress, 2012), 283–284. 10. Shemaryahu Talmon, “The Textual Study of the Bible: A New Outlook,” in Text and Canon of the Hebrew Bible: Collected Studies, ed. Shemaryahu Ṭalmon (Winona Lake, IN: Eisenbrauns, 2010), 83–84. 11. Most recently, Eugene Ulrich, The Dead Sea Scrolls and the Developmental Composition of the Bible (Leiden: Brill, 2015). 12. Eugene Ulrich, “The Dead Sea Scroll and the Hebrew Scriptural Texts,” in The Bible and the Dead Sea Scrolls. The Princeton Symposium on the Dead Sea Scrolls, 3 vols., ed. James H. Charlesworth (Waco, TX: Baylor University Press, 2006), 1:77–100. 13. George J. Brooke, “The Qumran Scrolls and the Demise of the Distinction between Higher and Lower Criticism,” in Reading the Dead Sea Scrolls: Essays in Method, ed. George J. Brooke (Atlanta, GA: SBL, 2013), 1–18. 14. Ronald L. Troxel, “What is the ‘Text’ in Textual Criticism?” VT 66 (2016): 603–626; idem, “Writing Commentary on the Life of a Text,” VT 67 (2017): 105–128. 15. The marginal readings of MS A are corrective elements written in the same hand as the main text. They treat systematically what the scribe himself recognized as an erroneous word or sentence; see Jean-Sébastien Rey, “Scribal Practices in the Ben Sira Hebrew Manuscript A and Codicological Remarks,” in Texts and Contexts of the Book of Sirach— Texte und Kontexte des Sirachbuches, ed. Gerhard Karner, Frank Ueberschaer, and Burkard M. Zapff, SBLSCS 66 (Atlanta, GA: SBL, 2017), 99–112. 16. Rüger, Text und Textform im hebräischen Sirach, 27. 17. See Elisha Qimron, The Hebrew of the Dead Sea Scrolls, HSS 29 (Winona Lake, IN: Eisenbrauns, 1986), or400.1t; Ahouva Shulman, “The munction of the ‘russive’ and ‘Indicative’ Imperfect morms in Biblical Hebrew Prose,” ZAH 13 (2000): 168–180; JeanSébastien Rey, “Quelques particularités linguistiques communes à 4QInstruction et à Ben

Sira,” in Conservatism and Innovation in the Hebrew Language of the Hellenistic Period: Proceedings of a Fourth International Symposium on the Hebrew of the Dead Sea Scrolls & Ben Sira, ed. Jan Joosten and Jean-Sébastien Rey, STDJ 73 (Leiden: Brill, 2008), 155– 174. 18. Shemaryahu Talmon, “Synonymous Readings in the Textual Traditions of the Old Testament,” in Studies in the Bible, ed. Chaim Rabin, ScrHier 8 (Jerusalem: Magnes, 1961), 335–383. 19. “ramais on ne supposera une faute sans rechercher quelles conditions l’ront rendue inévitable, probable ou possible. La recherche en question est d’rune extrême importance et forme une obligation absolue pour la critique.” Louis Havet, Manuel de critique verbale (Paris: Librairie Hachette, 1911), ort81, 110, cited in art. “erreur” in mrédéric Duval, Les mots de l’ édition de textes, Magister (Paris: École des Chartes, 2015), 131. 20. Smend, Die Weisheit des Jesus Sirach, 46; Israël Lévi, L’ Ecclésiastique, ou la Sagesse de Jésus, fils de Sira, BEHER 10 1–2 (Paris: Leroux, 1898–1901), 2:23; Moshe S. Segal, Sepher Ben Sira Hashalem (Jerusalem: Bialik Institute, 1958), 29 (Heb.). 21. The same confusion may have occurred in Sir 11:30 (MS A) regarding the crook ( ‫ככלב הוא‬ ot tluciffid si ti elihw ,)”esuoh a gniruoved [?‫ כלביא‬rof ekatsim] god a ekil si eH“ ‫באוכל בית‬ .cairyS dna keer ni devreserp neeb ton sah hcits eht sa evisulcnoc eb 22. This transformative phenomenon is typical for proverbial sayings through culture and language; see, for instance, the Latin: Homo sine pecunia est imago mortis (“A man without money is the image of death”); in Italian: Uomo senza quattrini è un morto che cammina (“A man without money is a dead man walking”); in English: A man without money is a bow without an arrow; and finally in French: Un homme sans argent est un loup sans dents (“A man without money is a wolf without teeth”). See Augusto Arthaber, Dizionario comparato di proverbi e modi proverbiali: Italiani, latini, francesi, spagnoli, tedeschi, inglesi e greci antichi: con relativi indici sistematico-alfabetici: supplemento ai dizionari delle principali lingue moderne ed antiche (Milan: Ulrico Hoepli, 1995), 575; Roumyana Petrova, “Contrastive Study of Proverbs,” in Introduction to Paremiology: A Comprehensive Guide to Proverb Studies, ed. Hrisztalina Hrisztova-Gotthardt and Melita Aleksa Varga (Berlin: De Gruyter, 2015), 243–261. 23. Cf. Papyrus Insiger 25:2: “The little bee brings the honey” and Nuria Calduch-Benages, “Animal Imagery in the Hebrew Text of Ben Sira,” in The Texts and Versions of the Book of Ben Sira: Transmission and Interpretation, ed. Jan Joosten and Jean-Sébastien Rey, JSJSup 150 (Leiden: Brill, 2011), 66–67. 24. This is the solution suggested by Rüger: “HB1 überliefert die ältere, HA B2 die jüngere Textform, vgl. Biblisch-hebräisch dnu txeT ni ”,‫ אליל‬hcsiärbeh-hcsinhcsiM .wzb ‫קטנה‬ ton si ‫ אליל‬.gnicnivnoc ton si noitseggus s’regüR tuB .56 ,hcariS nehcsiärbeh mi mroftxeT nwonk llew si ti ,revoeroM .‫ קטנה‬naht lacilbib ssel on si dna werbeH cianhsiM ni detsetta now that scribes had the strong tendency to biblicize late Hebrew forms; see Yigael Yadin, (rerusalem: Israel hT o3 Rinal Feports–3 The cigael cadin Excalations hT :VV aasada .t61 ,)9991 ,melasurer fo ytisrevinU werbeH eht dna yteicoS noitarolpxE 25. Shemaryahu Talmon, “Double Readings in the Masoretic Text,” Textus 1 (1960): 144–184, republished in Text and Canon of the Hebrew Bible, 217–271. A similar example in CD VII 14–18 has been provided by Moshe Bar Asher, “Mistaken Repetitions or Double Readings?” in Hebrew in the Second Temple Period The Hebrew of the Dead Sea Scrolls

and of Other Contemporary Sources, ed. Steven E. Fassberg, Moshe Bar Asher, and Ruth Clements, STDJ 108 (Leiden: Brill, 2013), 15–28. 26. Talmon, “Double Readings,” 240. 27. See especially the convincing argumentation of G. Bickell, “Der hebräische Sirachtext eine Rückübersetzung,” WZKM 13 (1899): 251–256. 28. Alexander A. Di Lella, The Hebrew Text of Sirach: A Text-Critical and Historical Study, Studies in Classical Literature 1 (The Hague: Mouton, 1966), 106–147; Di Lella’s hypothesis was criticized by John Strugnell in his review in CBQ 30 (1968): 89–91; Rüger, Text und Textform, 1–11; and Wido van Peursen, “The Alleged Retroversions from Syriac in the Hebrew Text of Ben Sira Revisited: Linguistic Perspectives,” Kleine Untersuchungen zur Sprachen des Alten Testaments und seiner Umwelt 2 (2001): 47–95. 29. Israël Lévi, “Les Nouveaux fragments hébreux de l’rEcclésiastique de résus, fils de Sira,” REJ 39 (1899): 1–15; 177–190 (esp. 11–15, 181); Lévi, L’ Ecclésiastique, part 2 (Paris: Ernest Leroux, 1901), XXVII-XXXV; V. Ryssel, “Die Herkunft der Hebräischen mragmente des Buches resus Sirach,” in Verhandlungen des XIII. Internationalen Orientalisten-Kongresses (Leiden: Brill, 1904), 248–253; see also N. Herz, “Dr. Ryssel on the Origin of the Doublets in the Hebrew ‘Ben Sira,’r” ExpTim 19 (1908): 189–190. 30. After I wrote this article, Menahem Kister informed me that he identified the same verse as a retroversion from the Syriac: “ :)9891( t5 unénošěL ”,‫נוספות למאמר ׳בשולי ספר בן־סירא׳‬ .ecnerefer siht rof hcum yrev retsiK mehaneM knaht I .25–15 .pse ,t5–6t 31. Contrary to the marginal notes of MS A, those of MS B are not corrective readings, but alternative readings. Most of the marginal readings of MS B have been added by the same scribe who also wrote the notes in Persian. These annotations are based on at least one other incomplete copy of Ben Sira from the same family as the Hebrew text of Ben Sira MSS A and D. Consequently, the marginal readings of MS B represent another textual tradition and have to be considered in the edition as such. For more details on these marginal notes, see Jean-Sébastien Rey and Marieke Dhont, “Scribal Practices in Ben Sira Manuscript B: Codicological Reconstruction and Material Typology of Marginal Readings,” in Discolering, Deciphering and Dissenting3 Ben Sira’s Hebrew Text, h8T – 2016, ed. Renate Egger-Wenzel, Stefan Reif, and James Aitken, DCLS (Berlin: De Gruyter, 2017); and idem, “The Relationship between Manuscripts A, B, and D and the Marginal Readings of Manuscript B,” in Proceedings of the Eighth Symposium on the Hebrew of the Dead Sea Scrolls and Ben Sira, ed. Stephen Fassberg, STDJ (Leiden: Brill, forthcoming). 32. See Segal, Sepher Ben Sira Hashalem, 195. 33. 9:8 vorP ni sa( evitcejda na deredisnoc eb dluoc ,‫ נכחך‬rof ekatsim a eb dluow hcihw ,‫נכ)ו(ח‬ and 24:26), while such a usage is not common when qualifying a person in Classical and .werbeH cinibbaR 34. From eb ot“ ‫ גלה‬fo mynonys a ro ”gnihctarcs ro gnibbur hguorht denepo eb ot“ lahpin ‫גלע‬ .681 ”,un à sim ertêr’d ruep ed“ setalsnart kispoM ;”delaever eb ot ,nwonk 35. For the unclear equivalence between .06 .n ,71 ”,snoitiddA“ ,retsiK ees ,‫ ܕܣܝܡ‬dna ‫ששם‬ 36. In the Syriac translation, the second colon of 31:15 is missing, creating a tristich where the sentence on the companion directly precedes the sentence on food. 37. I thank Jan Joosten for this suggestion. 38. The construction 7:2 rhCr2 ees ,7:11 doxE( werbeH lacilbib ni detsetta ylerar si ‫ידע אשר‬ eht ot desoppo sa ,)62:02 kezE ;01:6 hoQ ;5:9 bor ;11:4 htsE ;[‫ ידע כי‬02:5 sgKr1 .fc]

noitcurtsnoc eht sa rafosni dettuber eb yam tnemugra siht tuB .)x 0t2( ‫ ידע כי‬noitcurtsnoc 172Q4 ;[7:01 boT =] t 4 002Q4 ;7 VIX aHQ1( stxet lacilbib etal ni detsetta ylno si ‫ידע אשר‬ .)[‫ ידע כי‬t1:t2 hsor .fc] 2 6 97tQ4 ;7–6 t [D] 39. Lévi, L’ ecclésiastique, 110 considers a ”,ymene“ ‫ צר‬rof gnidaer suomynonys a ‫חותפו‬ ehT ni ”,hcariS“ ,yelretseO .E.O.W dna xoB .H. .)c41( ‫ יצרו‬lanigiro eht fo gnidaersim , ed. apocrcpha and lseudepigrapha of the eld Testament in Englisha apocrcpha ., t71: “A later scribe, realizing the ICharles (Oxford: Clarendon Press, 191t), R.H. l things od must have created difficulty which could be urged, that as the Creator of al with its tendency to evil as well as to good, added the gloss that od delivered ceserthe . 19. This later scribe, in his turn, did not realize IVman from his spoiler (i.e. Satan); cp. ”.deniamer llits ytluciffid eht taht 40. Louis inzberg, “Randglossen zum hebräischen Ben Sira,” in Orientalische Studien: Theodor Nöldeke zum Siebzigsten Geburtstag, ed. Carl Bezold (Giessen: Töpelmann, 1906), 322. He makes a connection between a61 .taB .b ni noitacifitnedi eht dna ‫חותף‬ ‫ (הא‬21:9 bor .gT ot osla gnirrefer ”htaed fo legna eht“ ‫ מלאך המות‬dna ‫ יצר הרע‬neewteb weN eht gnippaa ,.de ,rezuR egreS ees ;)‫יחתףאנש מן עלמא ומן יתיבניה מן יאמר ליה מה את עביד‬ , rewish and sisTestament3 Earlc ihristian nritings as a nitness for iewish Biblical Exege The Text of the Targum Stec, Christian Perspectives Series (Leiden: Brill, 2007); David M. .)4991 ,llirB :nedieL( 02 Ur A ,noitidE lacitiri dna noitcudortnV na 3boi fo 41. Conleth Kearns, The Expanded Text of Ecclesiasticus: Its Teaching on the Future Life as a Clue to Its Origin, DCLS 11 (Berlin: De Gruyter, 2011), 94. See also Roland E. Murphy, “cēṣer in the Qumran Literature,” Bib 39 (1958): 335. 42. Di Lella, The Hebrew Text of Sirach, 119–125. 43. The verbal form of the root .92.21 IIIX aHQ1 ;4:05 ;12:5t riS ;21:9 bor ni sraeppa ‫חתף‬ The nominal form is also attested in Prov 2t:28. The word is attested in Aramaic in ”spoort sih“ ‫ גדודיו‬werbeH eht setalsnart ti erehw )‫) (חתפוהי‬21:91 bor =( 2 II borgTQ11 ,)‫ ܚܛܦ‬cairyS eht dna( ‫ חטף‬ciamarA eht fo mrof-ib a si drow ehT .)21:9 bor .gT osla ees( .9:01 sP dna 12:12 gdur ees 44. Di Lella and Skehan, The Wisdom of Ben Sira, 267. 45. Smend, Die Weisheit des Jesus Sirach, 19. 46. See Paul Zumthor, Toward a Medieval Poetics, trans. P. Bennett (Minneapolis, MN: University of Minnesota Press, 1992). The word mouvance, translated as mutability by Bennet, is used by Zumthor to design 47. nate “the high degree of instability in medieval text traditions qualifying this instability as an ‘interplay between variant readings and reworkings’r” (Stefan ränicke and David roseph Wrisley, “Visualizing Mouvance: Toward a visual analysis of variant medieval text traditions,” Digital Scholarship in the Humanities 32 [2017]: 106). See also Paul Zumthor, “Intertextualité et mouvance,” Littérature 41 (1981): 8–16. 48. roseph Bedier, “La tradition manuscrite du Lai de L’rombre: Réflexions sur l’rart d’réditer les anciens textes,” Romania 54 (1928): 161–196; 321–356. 49. Talmon, “The Textual Study of the Bible: A New Outlook.”