Textual Growth in Isaiah 40:7-8? (Part 1)

This series of posts may be too ambitious for a blog, but I hope to present some of the key issues and the scholars involved with this textual problem and the other significant textual problems in Isaiah generally. I will present the texts in this post and the two theories used to explain the textual situation. In a second post, I will present the interpretation of Eugene Ulrich, “The Developmental Composition of the Book of Isaiah: Light from 1QIsaa on Additions in the MT,” Dead Sea Discoveries 8,3 (2001): 288-305. In a third post, I want to convey some of the main ideas in a recent article by Drew Longacre, “Developmental Stage, Scribal Lapse, or Physical Defect? 1QIsaa’s Damaged Exemplar for Isaiah Chapters 34-66,” Dead Sea Discoveries 20 (2013): 17-50. In a fourth post, I will offer my own conclusion to this textual problem. These texts are the most relevant to this problem but I will comment on all of the evidence in post four.

MT:  יָבֵ֤שׁ חָצִיר֙ נָ֣בֵֽל צִ֔יץ כִּ֛י ר֥וּחַ יְהוָ֖ה נָ֣שְׁבָה בֹּ֑ו אָכֵ֥ן חָצִ֖יר הָעָֽם ׃יָבֵ֥שׁ חָצִ֖יר נָ֣בֵֽל צִ֑יץ וּדְבַר־אֱלֹהֵ֖ינוּ יָק֥וּם לְעֹולָֽם׃ ס. The text in red is what is under consideration.

NRSV: The grass withers, the flower fades, when the breath of the Lord blows upon it; surely the people are grass. 8 The grass withers, the flower fades; but the word of our God will stand forever.

LXX: ἐξηράνθη ὁ χόρτος, καὶ τὸ ἄνθος ἐξέπεσεν, […] τὸ δὲ ῥῆμα τοῦ θεοῦ ἡμῶν μένει εἰς τὸν αἰῶνα.

NETS: The grass has withered, and the flower has fallen, […] but the word of our God remains forever.

1QIsaa: יבשׁ חציר נבל ציצ כי רוח ֑֑֑֑ נשׁבה בוא הכן חציר העם יבשׁ חציל נבל ציצ ודבר אלהינו יקום לעולם. The text in red represents a later addition to this text, perhaps written in a different handwriting by a later scribe. I provide the image from the manuscript (see here for the whole digital scroll).


As one can see, the Masoretic text is longer than the text of the LXX. Furthermore, 1QIsaa has an interesting and perhaps a mixed text. The image above shows that originally 1QIsaa had a shorter text similar to the LXX and that another hand added the longer reading of MT above the line and then continued the longer text down the left margin.

In order to describe this text and the factors involved one needs a theory which can explain these kinds of problems. Currently, the two theories on this problem and others like them are (1) 1QIsaa represents the shorter original and MT represents a growing and expanding text of Isaiah and (2) 1QIsaa and LXX represent a text that became shorter by a scribal error during the transmission of the longer text. In addition to a theory, one also needs an understanding of the character of these textual witnesses in order to describe the factors involved. For example, (1) would need to show that 1QIsaa usually has a shorter text than MT and the reason for MT’s longer text is due to intentional scribal additions. (2) would need to show that an unintentional scribal error is probable in 1QIsaa and would have to supply an equally probable solution for the rest of the shorter texts in that manuscript. In other words, a global knowledge of the textual character of 1QIsaa combined with a theory of its transmission is prerequisite to deciding between the two theories.

As we unpack this problem, I want to keep an eye on the theory which most simply explains the difference between these texts. Both theories are plausible from the outset, but which one will offer the simpler solution in the final analysis?


The Translation of the “LXX” of Job 37:18

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.

The Hexapla of Job 24:12a and the Modern Translator

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.”

St. Jerome’s Critique of Aquila’s Translation

Jerome says:

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

Does Genesis 1:1 Teach That God Created the Heavens and the Earth?

At LXX Studies we will more than likely never be the first to break a story to the public, but hopefully we will be able to offer sound reflection on some of the more newsworthy items from the perspective of the Ancient Versions.

This past week, there was some hype due to an announcement that Ellen Van Wolde has made an argument that concludes something to the tune of “the traditional view that God is creator is untenable now.”  As I understand it, a major part of Van Wolde’s argument concerns the translation of the Hebrew word ברא in Genesis 1:1.  Of course the traditional translation of this word is “to create,” but Van Wolde has suggested that the word in this context should be translated “to separate,” so that the text is not teaching the original creation of the heavens and the earth, but only the separation of the heavens from the earth (see Chris for an apt critique of Van Wolde’s analysis of the Hebrew usage).

What Van Wolde means by the “traditional view” is not clear, but one may safely assume that she has in mind the traditional exegesis of Genesis 1:1 which supports the traditional Jewish and Christian view of creation that God created the heavens and the earth ex nihilo.  I simply want to list the evidence of the Ancient Versions in order to present the traditional exegesis of this verse.

LXX: ᾿Εν ἀρχῇ ἐποίησεν ὁ θεὸς τὸν οὐρανὸν καὶ τὴν γῆν.  The text can be translated: In the beginning God created/made the heavens and the earth (see LSJ s.v. αρχη for anarthrous examples with prepositions, where the sense is still clearly definite probably because the lexical item naturally assumes only one beginning).  The LXX, then, uses a verb which does not mean “to separate,” but means “to make” or “to create” ( for the latter meaning see s.v. ποιεω A.2 LSJ).  For a thorough listing of translation equivalents for bara’ in the LXX and for what equivalents one would expect to see if the translators wanted to communicate the sense “to separate” see here.  This translation comes from the 3rd century BC (c. 280 BC) and some might wonder if the translators intended to convey that God only formed the heavens and the earth from existing matter by their use of ποιεω.  Although this is possible, only about 120 years later, we see clearly that the Jews are praying to the LORD God as ο παντων κτιστης (“the creator of all”) among other appellations in 2 Maccabees 1:24.  The Jews clearly have an understanding that God created all things in this period of their history and this theology must come from texts such as Genesis 1:1.

Aquila: ἐν κεφαλαιω εκτισεν θεος συν τον ουρανον και συν την γην.  Aquila employs the very specific verb κτιζω “to create” in order to render the Hebrew text.  Note that Aquila is not using the συν preposition in the normal sense “with” so that someone might be tempted to translate the text: God created with the heavens and with the earth or something like this.  Rather, Aquila uses συν characteristically to translate the Hebrew marker of the direct object את.

Symmachus: Not extant

Theodotion: Not extant

Vulgate: In principio creavit Deus caelum et terram.  Jerome uses creo to communicate the idea of creation, not separation.

Peshitta: ܒܪܫܝܬ ܒܪܐ ܐܠܗܐ܂ ܝܬ ܫܡܝܐ ܘܝܬ ܐܪܥܐ܂  P uses a cognate (bra’) to translate the Hebrew text in this verse.  According to the CAL in the G stem this verb may mean “to create” or “to get well.”  The former is clearly the meaning in this context given that the verb is transitive in this context.

Targum: בֲקַדמִין בְרָא יוי יָת שְמַיָא וְיָת אַרעָא׃  The Aramaic Targum uses the same cognate (bra’) as P, which has the same range of meaning in this dialect according to the CAL.

The Ancient Versions speak with one clear voice that God did indeed create the heavens and the earth, not simply the “stuff” in the heavens and the earth, and they certainly did not understand the text to mean God separated the heavens and the earth.

Therefore any argument to the contrary will have to be established firmly on the basis of the analysis of the Hebrew text of Genesis and the Hebrew Bible as a whole, which is Van Wolde’s project apparently.  However, see John Hobbins’ initial critique of Van Wolde’s analysis of the Hebrew Bible here.

HT: Charles

Determining Dependency Between Ancient Versions (part 2)

In this post, I comment on the nature of the translation of the Syriac Peshitta (S) in Qoheleth, which is an important consideration when trying to determine whether S is dependent on the Septuagint (G).

The Former Thesis

At an earlier time in research, scholars considered S to be a daughter version of the Septuagint (for this view see G.W. Anderson, “Canonical and Non-Canonical,” in Cambridge History of the Bible, ed. P.R. Ackroyd and C.F. Evans, vol. 1, (Cambridge, 1970), 158-9.), which means they considered S to be a direct translation of G.  What ailed this thesis was that it did not conform to the evidence of S itself. One is not able to read past Genesis 1:1 (ܒܪܫܝܬ ܒܪܐ ܐܠܗܐ. ܝܬ ܫܡܝܐ ܘܝܬ ܐܪܥܐ) without seeing the essential Semitic character of S and its faithful translation of the underlying Hebrew text (M or proto-M).  But however faithful S was to M, scholars still have noted that in some places S seems to be dependent on G, and the challenge was to explain this phenomenon leading to a new thesis. Continue reading “Determining Dependency Between Ancient Versions (part 2)”

Determining Dependency Between Ancient Versions (part 1)

I realize I have not been posting frequently, but I can say that I have been making progress in preparation for comprehensive exams, and I have been bringing long standing projects to a close.  One of those projects is my work on the relationship of the Peshitta of Qoheleth to the Septuagint of Ecclesiastes for Peter Gentry’s critical edition of Ecclesiastes in the Goettingen Septuaginta series.  This project began as a seminar paper for Dr. Gentry’s Introduction to the Septuagint seminar at Southern Seminary last Fall and has turned into a contribution to his forthcoming edition.  I hope in the future to publish the entire article, but it will take some time to clean it up for publication.

In the mean time, I would like to comment on some of the salient points of the work in the next few posts.

In this post I will attempt to date the sources.  In the second post I want to say something about the nature of the Peshitta as a translation of the Hebrew Bible (proto MT). In a third post, I want to work through the nature of the dependency of the Peshitta on the LXX. Continue reading “Determining Dependency Between Ancient Versions (part 1)”

מלך or βουλης in Ecclesiastes 2:12: different parent texts?

Prima facie there are significant differences between the Masoretic Text and the Septuagint, which lead one to think that the LXX was translated from a radically different text than the MT.  However, many of these differences arise from the translation technique of the LXX translator; ergo, the differences are on the level of the translation, not the level of the actual parent text or Vorlage. We have seen already in Job 3:3, and we will continue to see in Job that translation technique often accounts for the differences between the texts.

In Ecclesiastes or Qoheleth 2:12, there is a fine example of two divergent texts prima facie, but after a little digging, one finds that the LXX reading supports the consonantal text of the MT and simply represents a different vocalization of the text.  What follows is not my original work, but the work of my teacher, Peter Gentry, but of course any errors in the argument are mine. Continue reading “מלך or βουλης in Ecclesiastes 2:12: different parent texts?”

The Text of Job 3:3b

Well, I have finished reading the first two chapters of Job in the MT and the LXX with some spot readings from the Peshitta along the way.  I’m struck by differences between the LXX and MT, particularly the plus in LXX Job 2:9, which I will comment on later.  However, the differences do strike me as differences on the translation level not the textual level, which has already been observed by many (cf. the  NETS translation and Cox’s introduction to Job).

However, I’m more impressed with the simplicity of the Greek translation of the difficult Hebrew text. Job 3:3b offers a short example.  The Hebrew poetry commences in chapter 3 and with it comes a whole host of syntactical difficulties.  In 3:3b alone, the Hebrew text reads: והלילה אמר הרה גבר.  The text is translated as follows: and the night [which] he said a man was conceived.  There are two verbs in the stichos with no conjunctions to define their relationship to one another.  However, this is a common construction in Hebrew poetry called an asyndetic relative clause.  But how do we know to read the text in this way? Continue reading “The Text of Job 3:3b”