Michael Law has posted a short, but significant piece on Jerome’s use of Aquila as a source of the Vulgate. In this field, there are few scholars who post on these significant matters, so I thought I would draw more attention to it. Thanks for sharing, Michael!
This post is a response to Ed Babinski’s questions on my “About” page. He raised a significant question regarding the differences between Brenton’s translation and the NETS (New English Translation of the Septuagint) at Job 37:18. His question also gave me the opportunity to look ahead at a verse that I will have to prepare for my dissertation, since this verse is sub asterisk (※).
He asked two questions, 1) what accounts for the disparity between the English translations of the Septuagint, and 2) how does the “LXX” translate Job 37:18? I will answer them in order.
The Problem with Brenton’s Translation
Brenton differs with NETS in two instances in this verse: he has a verb in v. 18a, where NETS has a noun and he reads the word “mirror”, where NETS has “appearance.” I have not done an exhaustive study of Brenton’s translation, but my experience with it is that he sometimes translates the Hebrew text at the expense of faithfulness to the Greek of the LXX. This verse seems to be an example of this phenomenon in both places. In the first example, the LXX has στερεώσεις, which is a plural noun from στερεωσις, “a making firm, a making solid” (see LSJ) or as NETS has it “solidifications.” I prefer NETS here because the noun is a -σις noun and it usually indicates an abstract noun, thus a solidification. The Hebrew text has a Hiphil verb from רקע, and Brenton seems to translate this word instead (see BDB). In the second example, LXX has ὅρασις “appearance”, while the Hebrew has ראי “mirror”, which is listed in both BDB and KB. This word does seem to be a genuine hapax legomenon. Brenton has read the Hebrew text again, for the Greek word does not mean “mirror” (see LSJ). NETS has translated the Greek text more faithfully here, and there is a reason for the Greek translation to which we now turn.
The Reading of the “LXX”
The second question regarding the Greek translation of ראי brings us to the main issue. Most significantly, this reading does not come from the (O)ld (G)reek translator (1-2 centuries BCE), but from the Jewish reviser Theodotion, who completed his work in the 1st century CE (many argue for a late second century date, but see Peter J. Gentry, The Asterisked Materials of the Greek Job for a contrary and persuasive proposal). The text is under the asterisk (※), marking a text present in the Hebrew but absent in the OG, and in this case the Syro-hexapla and Catena Ms 740 preserve the attribution to Theodotion. Thus, when attempting to discover the translation technique, one needs to understand Theodotion’s use of ὅρασις, not necessarily the technique of the OG. Here is the available evidence of ὅρασις for Th (the references correspond to Field):
מראה: Ezek 1:13, 16; 10:22; Dan 8:16, 26 Th.
רוהּ: Dan 3:25 Th
חזון: Dan 9:24; 10:14; (7:13?) Th.
חזוי: Dan 2:28 Th.
Is 22:1 Sym and Th for חזיון.
Ezek. 13:16 Sym and Th for חזון.
Ezek 1:27, 28; 10:10 Aq and Th מראה.
Amos 5:26 Th for סכות perhaps for שכה “to observe, behold.”
Problematic readings occur in Dan 4:7, 7:2, 8:2, and so have been left out of this analysis.
The plethora of evidence allows one to leave aside the double attributions or those instances where Th is joined by Aq and Sym as the author of the lemma, since there are clear cases where only Th uses ὅρασις for Hebrew מראה and חזון.
In Job 37:18, Th must be reading ראי “mirror” as a form of מראה “appearance.” Perhaps, he was unaware of the meaning of this hapax legomenon in the Hebrew Bible. A comparison of the rest of the Versions confirms this, since none of these Versions have an equivalent for “mirror”:
Aquila: also uses ὅρασις “appearance.”
Symmachus: uses an infinitive ὀφθῆναι “to appear.”
Vulgate: qui solidissimi quasi aere fusi sunt.
Peshitta: No equivalent for the reading in question or it has interpreted the reading with “to support simultaneously.”
The Versions all take an interpretive stance, which is almost certainly because of the hapax in the Hebrew text. The only other possibility is to posit that modern Hebrew lexicography is wrong in this case and to posit that the Hebrew word really does mean “appearance” on the basis of Th, Aq, and Sym.
These translators are attempting to render the Hebrew text in front of them in a quite literal way. I doubt we can discern their cosmology based on their rendering of the text. They want to be faithful to the Hebrew text and its message from what I can tell. Perhaps Ed can now enlighten us as to what he thinks is happening in this text regarding ancient cosmology.
Regarding the consultation of other Versions in his own translation work, Jerome says,
Sed de Hebraeo transferens, magis me Septuaginta Interpretum consuetudini coaptavi: in his dumtaxat, quae non multum ab Hebraicis discrepebant.
(But when translating from the Hebrew, rather I joined myself to the knowledge/experience of the 70 translators [i.e. the Septuagint]: precisely in the places which were not differing much from the Hebrews.)
With the launch of the NIV update, some scholars are wondering whether the editors should have scrapped some of the readings of the previous NIV all together. It is interesting that Jerome had a very common sense criterion for the use of other Versions in translation: when they don’t differ much from the Hebrew. It stands to reason, therefore, that if modern scholars can trace gross mistranslations back to previous versions, which do not convey the Hebrew sense well, that they should indeed attempt a new translation of the source text. This principle seems consonant with the Reformation, and Luther and Tyndale seemed to follow it well. What is the church to gain by preserving bad or worse misleading renditions of the original text?
I have not posted on the LXX in a while, but I wanted to draw attention to a problem at which I had to look for my dissertation on the Hexapla of Job and show how this work is relevant to modern translations of the Bible.
A sampling of modern translations of this verse is insightful. For example, the ESV says, “From out of the city the dying groan”. The word of interest is “the dying”, and it is also the translation of the NIV. Interestingly, the KJV(!), HCSB, and the NASB have “men” in their translation. So what is happening beneath the surface?
Here is a listing of the MT and the Versions:
MT: מֵ֘עִ֤יר מְתִ֨ים׀ יִנְאָ֗קוּ
LXX: οἳ ἐκ πόλεως καὶ οἴκων ἰδίων ἐξεβάλλοντο
Sym: ἐκ πόλεως ἄνδρας ἐποίησαν στενάξαι (retroverted from Syh .ܣ. ܡܢ ܡܕܝܢܬܐ ܠܓܒܪ̈ܐ ܥܒ̣ܕܘ ܠܡܬܬ̣ܢܚܘ ܀)
Th: ἐκ πόλεως ἄνδρες κεκράξονται (retroverted from Syh ܬ. ܡܢ ܡܕܝܢܬܐ ܓܒܪ̈ܐ ܢܩ̣ܥܘܢ܀)
Vulgate: de civitatibus fecerunt viros gemere
Pesh:ܡܢ ܓܘ ܩܪܝܬܐ ܡ̈ܝܬܐ ܢܬܐܢܩܘܢ
A few observations are in order. First, the consonantal text of M may be pointed to mean “men” or “the dying”, but the vocalization of M clearly indicates that the former is in view. Now, a modern scholar may dispute this vocalization and claim that M’s vocalization was a relatively new invention; however, this vocalized reading is very old, if not the original reading, since it is also the reading of Theodotion, Symmachus and Jerome’s Vulgate. The LXX does not have either of these readings, but the translator seemed to understand people who are alive, not dead, but this version cannot always be trusted for this type of information in the book of Job. Second, the Pesh certainly has “the dead” or “the dying”. Third, and on a more pedantic note, Jerome seems to have used Symmachus for his translation, since we know he used Sym from time to time and these versions are the only two interpreting the Hebrew verb as a causative “from the city they made/caused men to groan”, when in fact the word is not causative. Thus, the Hexapla actually preserves the earliest evidence of the reading of the Hebrew Bible in this verse.
Now, let me highlight the problem of the ESV, which is the version I use most regularly and I find it to be a good translation overall. The translation, “the dying”, is put into the text without any explanation. As far as the reader knows, this translation is a faithful rendering of the Hebrew text. The problem, however, is that it is a faithful rendering of the Syriac Peshitta and perhaps one other insignificant Hebrew ms., and the English translator did not note the other Versions and witnesses to the text in this instance. What should have happened in this case?
The translator probably should have maintained the principles of the ESV and translated the Hebrew text (codex L) faithfully, and if he disagreed with the reading, he should have put all of the evidence in a footnote below the text. Or he could have put “the dying” in the text and left a footnote indicating that the reading of the Hebrew and the rest of the Versions is “men”.
The reader of the ESV wants to know what the original Hebrew text says, nothing more or less. In Job 24:12a, M and the oldest Versions have preserved the original vocalization. The text should say “From the city, men will groan.”
Jerome writes to Augustine concerning the nature of the Septuagint of the 4th century (i.e. after Origen’s Hexapla) as follows:
Because, however, in other letters you ask, why my former translation has asterisks and obelisks noted in the canonical books, and afterwards I published another translation without them –I say with your pardon— you do not seem to me to understand, because you have inquired (about them). For this first translation is from the seventy translators and, wherever the marks are, that is the obelisks, it is shown, that the Seventy said more than is contained in the Hebrew; however, where there are asterisks, that is stars which light the way, the reading was added by Origen from the version of Theodotion. And in that (former) translation we translated from the Greek, in this place from the Hebrew itself, we expressed what we were understanding, preserving more importantly the truth of the sense than the order of the words now and then. And I am amazed how you do not read the books of the Seventy in their pure form, as they were published by the Seventy, but rather as emended (emendatos) by Origen or rather corrupted by the obelisks and asterisks, and you do not follow the translation of a Christian man, especially when he (Origen) transferred these additions, which have been added from the edition of a man, a Jew and a blasphemer, after the Passion of Christ. Do you wish to be a true friend of the Septuagint? You should not read these (additions), which are under the asterisks; on the contrary, erase/scrape them from the chapters, so that you might show yourself to be a true patron. If you do this, you will be forced to condemn all the libraries of the churches. For scarcely one or another manuscript/book will be discovered, which has not such additions.
Epistula CXII, 19 Ad Augustinum
However, Aquila, a proselyte and contentious translator, who has attempted to translate not only words (uerba) but also the etymologies (etymologias) of the words (uerborum), is rightly rejected (proicitur) by us. For who is able to read and to understand χευμα (that which is poured), οπωρισμον (vintage), στιλπνοτητα (brightness) [words from Deut 7:13], for grain and wine and oil [Deut. 7:13], in so far as we are able to read “pouring” (fusionem) and “harvesting of fruit” (pomationem) and “shining,” (splendentiam)? Or because the Hebrews not only have αρθρα (connecting word, the article), but also προαρθρα (prefixes), so that he κακοζηλως (in bad style) may interpret both syllables and letters and he may say συν τον ουρανον και συν την γην [Gen. 1:1; Aquila renders the marker of the direct object in Hebrew with συν, even though this rendering has no acceptance in Greek or Latin], which no Greek and Latin dialect accepts? We are able to take his precedent of the matter from our discussion. For how many words are spoken well among the Greeks, which, if we translate according to the word, do not resound in Latin, and from a region, where they are pleasing among us, if equally the words are altered with respect to the arrangement, then among the Greeks they will displease.
Epistula LVII, 11
It has been a while since I have posted anything on this site, but here is a snippet from some of the material I have been reading for hexaplaric research.
Regarding translation technique Jerome says:
For I not only admit, but I declare freely that I express not word for word (uerbum e uerbo) but sense for sense (sensum de sensu) in translating from the Greek (except from the holy scriptures, where even the order of the words is a mystery.)
Ad Pammachium de optimo genere interpretandi Epistula LVII, 5
I put off a solution even of such a little question, so that my critics may question and understand that the words (uerba) of scripture must not be examined (consideranda), but the sense (sensum).
Ad Pammachium de optimo genere interpretandi Epistula LVII, 10