The Best Way to Study Ancient Texts

The best way to understand the craft of the ancient scribe and the circumstances which resulted in differences in detail from one manuscript to the next is to work directly from high resolution photographs of the manuscripts themselves. Much can only be learned from studying the manuscripts as opposed to a modern print edition based on them. A quiet joy awaits those who read an ancient text as it was transmitted in antiquity.

It is splendid that it is now possible to read the Aleppo Codex online. The Dead Sea Scrolls Electronic Reference Library is a glorious achievement, though I wish it were possible to upload an image for discussion in a post without violating copyright. A. M. Ceriani’s Translatio Syro Pescitto Veteris Testamenti ex codice Ambrosiano (Milan: Angeli della Croce, 1876-1881) is available online. So is Constantine von Tischendorf’s Bibliorum Codex Sinaiticus Petropolitanus (Leipzig: Giesecke & Devrient, 1862) (go here). Besides Codex Ambrosianus 7a1 for the Peshitta and Codex Sinaiticus for the Septuagint, it would make things much easier if high resolution photographs of Codex Leningradensis for MT were available online, of Codex Alexandrinus and Vaticanus for the Septuagint, and of Codex Amiatinus for the Vulgate. The Aleppo Codex site is pioneering in this respect, and might serve as a model for other biblical codex sites.

The realization of high-quality electronic publications is expensive. Speaking as a student of antiquity, I don’t care who is elected President, so long as she or he initiates an “Ashurbanipal Project” that would finance and coordinate the electronic publication of ancient texts.

High resolution photographs of Hebrew Ben Sira are not available online, with the exception of an image of the top right hand corner of a page from ms. B made available thanks to the good offices of the University of Cambrige Taylor-Schechter Genizah Research Unit. The image is not the best – the lighting is insufficient – but even so, it is a window into the world of the Hebrew text. If you click on the link and take a look at the text, a number of things will catch your eye:

(1) The parchment is lined with the tip of a sharp object (the lines are white). Letters were placed in a row a short distance below the lines.

(2) The words sometimes run together, with little or no spacing between them.

(3) The text is arranged stichographically. This is not as obvious as it might be, because the image reproduces a sequence of “a” stichoi, the “b” stichoi are beyond the range of the image. The manuscript’s layout is such that the poetic structure of the text is respected. For an image of the whole page, go here.

(4) The text contains alternative readings in the margin.

Here is the text in the image, Ben Sira 39:29-30 according to ms. B. The left hand continuation of the same page is added in on the basis of the image of the whole page and discussions in the secondary literature; a translation follows:

אשוברדרעודבר       גם אלה למשפט נ[בראו]

חית שן עקרבופתן     וחרב ֯נקמות ֯להח֯רים [רשעים]

כ֯ל אלה לצורכם נברא֯ו    והמה באוצר֯ ולעת יפקדו

 

Left margin of middle line: חר[ב נו]קמת, לה[חריב], and פט

Right margin of last line: גמ and נבחרו

Left margin of last line: בא[וצ]רו לעת

 

Fire and hail, misfortune and pestilence,

      these too were created for judgment,

fanged beast, scorpion and viper,

     and the sword of vengeance, to exterminate the wicked.

All these were created for a purpose*

     and they are in repository, and for a time they are appointed.

 

*lit. for their (intended) purpose

Left margin of middle line: avenging sword, to desolate, and the firstb [?]

Right margin of last line: also and are selected

Left margin of last line: in his repository; for a time]

                                           Ben Sira 39:29-30

 In Beentjes’ edition of the book of Ben Sira in Hebrew, the mem in גמ is normalized to a mem sofit. Normalizations, of course, are best avoided. The same edition omits the marginal readings to the left of the middle line, said to “have been lost as the result of a lacuna.” This is odd: details on the photograph of the whole page are difficult to make out, but the secondary literature regularly reports the remains of marginal readings as indicated above. The importance of having high resolution photographs to work with is illustrated by the disagreement. Furthermore, the Cairo Genizah mss have deteriorated somewhat since 1896, which is why first facsimiles and first publications of the relevant materials must also be consulted whenever possible.

The long ו in נבחרו visible in the image is noteworthy. It is not expected – though not impossible to do, technically – for unusual letter shapes to be reproduced in transcription. But it’s helpful to note the occurrence of letter shapes that might lead to mistakes. In this case, the waw looks a lot like a nun sofit.

Ben Sira 39:29-30 is a garden of text critical delights. Here is the text as it appears in Codex Sinaiticus online, its stichographic presentation respected, the capital letters reduced and accents and spaces added between words according to convention:

πῦρ καὶ χάλαζα καὶ λειμὸς καὶ θάνατος

πάντα ταῦτα εἰς ἐκδίκησιν ἔκτισται

θηρίων ὀδόντες καὶ σκορπίοι καὶ ἔχις

καὶ ῥομφαία ἐκδικοῦσα εἰς ὄλεθρον ἀσεβεῖς[1]

Here is a retroversion into Hebrew:

אש וברד ורעב ודבר

כל אלה לנקמה נבראו

חית שן ועקרב ופתן

וחרב נוקמת להחרים רשעים

Text which differs from the Hebrew of ms. B as reconstructed above is highlighted in blue. The heuristic value of retroversion as a means to understanding the text-critical value of a translation is uncontested, but how can we be sure that Hebrew variants which never existed are not manufactured in the process?

We cannot be sure. Even so, statements of a probabilistic nature on a case-by-case basis are not only possible, but are in need of formulation if text-criticism is a discipline like any other that formulates hypotheses based on all the available evidence. For example:

(1) The supposition that λιμὸς (spelling now normalized) καὶ θάνατος reflects רעב ודבר is based on the observation that λιμὸς and θάνατος often reflect רעב and דבר respectively in translation elsewhere. רעב fits better than רע amid the specific evils of the list. The omission of ב before ו could be an aural error (conflation of two bilabials), or due to a scribal lapse in copying.

(2) Familiarity with texts in Hebrew as attested in Qumran and medieval sources suggests that the conjunction ו was not uncommonly added into originally asyndetic constructions in the course of transmission. We cannot be sure that the translator of Ben Sira into Greek read a ו before רעב and עקרב in his Vorlage, but we can say that it would not be surprising if he did, and that the asyndetic constructions attested in ms. B are probably original.

(3) The best way to approach גם vs. כל in 39:29, as often, is to pose the “what if?” question both ways. What if גם is original and כל is secondary? What would account for the change? In my view, כל amounts to a syntactical simplification. גם requires a reading of the text which extends back to 39:28. The paired – and original – occurrences ofכל אלה in 39:27 and 30 are part of a larger structure which extends back to 39:26; a כל אלה in stich “b” of 39:29 is out of line, literally, with respect to the paired occurrences just mentioned. Note the color-coded textual components (the following is an eclectic text based on an analysis of all witnesses):

39:26-27

[ראש כל צרך ]ל[חיי] אדם   מים ואש וברזל ומלח

[חלב חטים ח]לב ודבש   דם ענב יצהר ובגד

כל [אלה לטוב]ים ייטיבו   כן לרעים לרעה נהפכו

39:28

[יש רוחות לנקמה נו]צרוו   [ובאפם הר]ים יעתיקו

[ובעת כלה חילם ישפכו  ורוח עושי]הם יניחו

39:29-30

אש וברד רעב ודבר    גם אלה למשפט נ[בר]או

חית שן עקרב ופתן   וחרב נקמות להחרים ר]ש]עים

כל אלה לצורכם נבחרו   והמה באוצרו לעת יפקדו

 

כל אלה in 39:29 instead of גם אלה  requires a reading of the text that does not extend beyond the line in which it occurs. That makes it the easier, and less likely reading.

But let’s assume, since it’s important to play the argument both ways, that כל is original and גם is secondary in 39:29. What would account for the change? Ifגם אלה לצורכם נבראו represents the original continuation of the text two lines down in 39:30, one might suggest that the change to גם in 39:29 anticipates its occurrence in 39:30.

But it seems more likely that כל אלה לצורכם נבחרו is original to 39:30, in vertical parallelism with כל אלה לטובים ייטיבו in 39:27. In both 39:26-27 and 29-30, a list of items is picked up by a fronted anaphoric כל אלה clause in which the items’ function is clarified. If this is the case, גם is original and כל is secondary in 39:29, and the reverse is true in 39:30.

(4) The best way to approach חרב נקמות vs. חרב נוקמת (the probable marginal reading in B based on the reconstructed Vorlage of G) is in the same way. What if חרב נקמות is original and חרב נוקמת secondary? What would account for the change? That’s easy: חרב נוקמת  is an existing biblical phrase (Lev 26:25), to which the diction of Ben Sira would have been assimilated. חרב נקמות, on the other hand, follows a known pattern, but is not attested. It is not clear why an original חרב נוקמת would be changed to חרב נקמות. It seems likely, it should be noted, that from the very beginning (early second century bce) נקמות would have been spelled with a ו to mark the feminine plural. חרב נקמות is more probably the original reading.

(5) When extant Hebrew Ben Sira contains lacunae, it is sometimes possible to restore it with a degree of certainty. In 39:29, נ[בר]או is restored on the basis of ἔκτισται and ܐܬ݂ܒܖ݁ܝ. (The relevant page of Codex Ambrosianus 7a1 is here.) The Greek verb is the translation of choice for ברא in Ben Sira: 16:26; 38:4; 39:28; 39:29 (ex hypothesi); 40:10; exceptions: 5:14 (ἐστὶν); and 15:14 (ἐποίησεν). The cognate Syriac verb renders ברא without exception: 3:16; 5:14; 15:14; 16:26; 34:13; 38:4; 39:28; 39:29 (ex hypothesi); 43:14. In 39:30, ר[ש]עים – others read differently[2] - is restored on the basis of ἀσεβεῖς and ܖ̈ܫܝܥܐ. Elsewhere, τοῖς ἀσεβέσιν translates לרשעים whereas ἁμαρτωλοὺς translates רעים in 12:6; cf. also 41:5, 7; note that רע is never translated by ασεβ- in Ben Sira. The Syriac cognate of רשעים at this point suggests that its Vorlage also read רשעים. The retroversion proposed by Segal and Vattioni - רעים - is not well-founded. The retroversion proposed here - רשעים - was already suggested by Barthélemy-Rickenbacher (ad רשעים).

(6) It is likely that the Vorlage of the Greek read לנקמה, not למשפט, given that in Ben Sira κριν- nouns and verbs consistently translate משפט and שפט , and εκδικ- nouns and verbs, נקם (verb and noun) and נקמה. But לנקמה is probably secondary, an assimilation to נקמה two lines up in a similar construction.

(7) Both M and B attest to two lines, not one, in 39:30. Greek and Syriac Ben Sira lack an equivalent altogether. There are many examples in Ben Sira in which the versions based on Hebrew Vorlagen are shorter by a line or more, and the text in question does not appear to be a redactional plus. Agreements between the Greek and Syriac in this respect, or lack thereof, deserve a careful synoptic study.

(8) נבחרו vs. נבראו in the second line of 39:30: נבראו appears to be an assimilation to context – the נבראו of two lines up - cf. (6). Given the widespread reduction of ח to not much more than a placeholder for a vowel, like א, aural confusion could have been a concomitant factor in the posited change from נבחרו to נבראו.

This post may seem long, but I have actually skipped over details of text-critical interest for the sake of brevity. My most important methodological point is the following: it really isn’t possible to do serious text-critical work without examining high-resolution photographs of the relevant manuscripts. I thank Jeremy Corley for helpful comments. Remaining errors are my responsibility alone.

References

 

Dominique Barthélemy and Otto Rickenbacher, Konkordanz zum Hebräischen Sirach mit syrisch-hebräischem Index (Göttingen: Vandenhoeck & Ruprecht, 1973); Pancratius C. Beentjes, The Book of Ben Sira in Hebrew (VTSup 68; Leiden: Brill; Atlanta: SBL, 2006); Jan Liesen, Full of Praise: An Exegetical Study of Sir 39,12-35 (JSJSup 64; Leiden: Brill, 2000 [not available to me]); Solomon Schechter, “A Fragment of the Original Text of Ecclesiasticus (39:15-40:7),” Expositor 4 (5^th series, 1896) 1-15; Solomon Schechter, Charles Taylor, and Arthur Ernest Cowley, The Wisdom of Ben Sira: Portions of the Book Ecclesiasticus from Hebrew Manuscripts in the Cairo Geniza Collection Presented to the University of Cambridge (Cambridge: Cambridge University Press, 1899); Moshe Z. Segal, ספר בן סירא השלם (1^st ed.; Jerusalem: Bialik Institute, 1953; the latest edition, 1997, was not available to me); Francesco Vattioni, Ecclesiastico: Testo ebraico con apparato critico e versioni greca, latina e siriaca (Testi 9; Napoli: Istituto Orientale Napoli, 1968); Yigael Yadin, מגלת בן סירא ממצדה, ErIsr 8 (1965) 1-45 [= 1965¹]; idem, The Ben Sira Scroll from Masada (Jerusalem: Israel Exploration Society and the Shrine of the Book, 1965) [= 1965²]; Joseph Ziegler, Sapientia Iesu Filii Sirach (2^nd ed.; Septuaginta 12/2; Göttingen: Vandenhoeck & Ruprecht, 1980).

For a printer-friendly version of this post, go here.

---

[1] Corrected Sinaiticus has ἐκδιωκοῦσα; the addition of the ω assimilates the text to 30:19, as Ziegler notes, and is an inner-Greek corruption. 

[2] Against the judgment of Vattioni, and Barthélemy-Rickenbacher, Beentjes reads [ ]בם in M at this point. I was not able to control Yadin 1965¹ and 1965² for this post. The importance of having high resolution photographs to work with and access to original photographs and/or facsimiles is illustrated by the disagreement.

Comments

[Eric Sowell]

I have spent quite a bit of time in the last eight months in digital images of NT Greek manuscripts. It is great for giving someone perspective on what variants and scribal practices were really like. Variants, critical texts, scribal errors and whatnot are all so much less nebulous ideas when you see these things for yourself in different manuscripts. We just need more digitized manuscripts!

[JohnFH]

Thanks, Eric. You make the point very well.