Here is the "version 2.0" updated draft of my article on Philo's text of Gen 2.18. It has attempted to take into consideration Al  Pietersma's earlier objections [on Ioudaios] in a general way, and can serve as a base for Al to expand further on his position. I have worked through the Philo materials once more with Al's comments in view, and must confess that I find Al's attempt to drive a wedge between Philo's outlook on the issues under discussion and the syntax of the text with which Philo identifies his opinions to be quite unconvincing. Whatever one thinks of Philo's conclusions, he is an interpreter tied to his texts in his own "intentionality" and selfconsciousness, to the extent of being driven to his symbolic interpretations by a literalistic initial approach to the text. This is clear again and again, it seems to me, and applies fully to the material at hand. Thus I am convinced that Philo read the singular verb "I will make" in Gen 2.18. Whether that represents an "original" Septuagint reading is to me less clear, and the problem is not made any simpler by looking more closely at the "history" of the similar passage in Gen 1.26.

Bob Kraft, UPenn


Philo's Text of Genesis 2.18 ("I will make a helper") [version 2.0]

Background: At Gen 2.18 the Cambridge Larger LXX (Brooke-McLean) lists as a variant to ποιήσωμεν "ποιησω Phil-codd 1/3"; the recent Gttingen edition of Genesis by Wevers lists no variant to this lemma.

Summary: The evidence strongly supports the probability that Philo originally read ποιήσω ("I will make") rather than the text that has survived in all LXX witnesses ("let us make"). Whether this Philonic reading represents the original or oldest recoverable LXX text is more difficult to determine.

The Cohn-Wendland edition of Philo includes three closely related passages in which the ποιήσωμεν phrase in LXX Gen 2.18 is cited in the main text: LegAlleg 2.1,5,9 (C-W 1, pp.90-92). Sure enough, the ποιήσω variation is noted in two (not only one, as B-M suggests) of the passages (LegAlleg 2.5,9), but the textual situation there is even more curious than I had anticipated. [To the best of my knowledge, none of the papyri or other textual materials for Philo that have appeared subsequent to the C-W edition have any bearing on these passages.]

If C-W are correct in what they report, only three Greek MSS of this work have been preserved (their codices M A P, but P is defective in 2.9), and they ALL agree in reading ποιήσω in these two passages. The plural verb ("let us make") represents an editorial retroversion based on the Armenian version of this Philonic work. In the third passage (LegAlleg 2.1), which provides the basic extended lemma for this section, all the cited witnesses agree in reading the plural (MAP Arm). In addition, MS M also gives a title to the work that refers to the "let us make a companion" lemma (A has a different title, P and Arm have no title). Thus it could be argued, based on the frequently invoked principle that in such situations, extended lemmas are more likely to be "corrected" than the more atomized comments that follow, that (1) Philo actually read the singular throughout, not the plural, and that (2) the lemma in 2.1 has been harmonized to the more familiar LXX text in the transmission of the Greek and (3) all three passages have become plural for the Armenian translation (either because the Greek vorlage already had been harmonized, or in the act of translation into Armenian, or in the transmission of the Armenian, or some combination of those alternatives) for the same reason.

Is this wishful thinking in an overenthusiastic quest for variations? How might one evaluate such a possibility? A TLG check of Philo's works reveals that he does not refer to these words from Gen 2.18 elsewhere in the preserved Greek (see below on QuGen). But wherever he cites the similar material from Gen 1.26, which contains the plural verb fortified by the plural pronoun ("let us make man in our..."), he ALWAYS comments on the use of the plural, which for him explains that God is responsible for what is good in man, while what is less than good can be blamed on God's helper(s)! -- see the passages listed below and also Fug 68ff. So he is painfully aware of the presence of the plural in that passage. Further, in LegAlleg (books 1 and 2), where our problematic passages appear or function in the background, Philo again and again emphasizes the unity and singleness of Deity. That he would cite "let us make" in such a context without explaining it is very unlikely, it seems to me, given his hermeneutical conduct elsewhere. The dependence of Philo's allegorizing on a literalist approach to the surface text is everywhere apparent ("this is what it says, but it surely can't mean that"), and should not be overlooked in this discussion.

But there is more "hard evidence." In his treatise on Creation (OpfMund) 72-75, where Philo takes characteristic pains to explain the plural forms in Gen 1.26, he emphasizes that "it is only [ἐπὶ μόνης] with reference to the creation of ἄνθρωπος/man that God says 'Let us make'" (Creat 75). Does this encourage the idea that Philo had a creation text pertaining to man's "helper" that also contained the plural verb? Not impossible, but on the surface of things, not very likely either. After all, the evidence strongly suggests that we are in a traditionally sensitive subject area here, even before Christian apologetes exploited these passages in defense of their ideas of divine plurality.

But yet more. In Confusion of Tongues 168-179 Philo again faces the problem of the plural verbal form for Divine activity and he provides a list of this linguistic phenomenon in the scriptures: Gen 11.7, 1.26, 3.22. No mention is made of Gen 2.18. Thus we have an argument from silence, but a fairly loud one since it fits into the pattern observed elsewhere and would be a surprising omission if Gen 2.18 were relevant, given the intensity with which the subject matter is approached in all these passages and in the apologetic tradition at large.

Based on the currently available evidence, then, my tentative conclusion would be that Philo's text of Gen 2.18 did indeed read "I will make" (ποιήσω) and not "let us make" (ποιήσωμεν) at the occasion of his writing LegAlleg, Confusion, and Creation. That he had the singular text form in front of him for Questions in Genesis 1.17-18 as well, where the existing edited texts of this textually complex material give two introductory lemmas containing the plural form for Gen 2.18, is perhaps equally probable insofar as his preserved comments about those lemmas take no notice of the supposed plurality. (QuGen on 1.26 is not preserved.)

Other pieces of homework need to be done before very much more can be said judiciously about the textual situation: Do the Armenian MSS of Philo evidence any variations in the LegAlleg passages? In the QuGen lemmata? Does the Gen 2.18 material receive attention in any confrontational contexts in other early authors -- that is, was it seen as a problem or perhaps as a solution (e.g. for Christian viewpoints on deity) elsewhere in the early Greek (or other) traditions?

The next question, then, is whether Philo here represents the oldest recoverable form of the LXX for this reading in Gen 2.18? If so, the entire LXX MS tradition became assimilated to Gen 1.26 fairly early in the game. That is not impossible, but it might be a tough sell. Reasons must be explored for such an assimilation (desire for consistency? Unconscious harmonization?), although evidence from the period in which it would have taken place is extremely scarce.

Or could it be that Philo's text (and presumably that of his community of listeners) had already been harmonized to the proto-MT of Gen 2.18, despite the fact that this phenomenon left no impact on the later LXX MSS? That's also tough to defend, but not impossible. It assumes that the singular formulation that has survived in the Hebrew textual tradition of Gen 2.18 was sufficiently ancient and sufficiently influential to modify the pre-Philonic Greek tradition appropriately.

One might also argue that there was variation within the Hebrew text tradition of Gen 2.18 from pre-Philonic times (note the plural verb in the Vulgate, which normally follows MT; also in some Peshitta manuscripts? And in the Samaritan? more details are desirable here!), and that the Greek situation is thus even more complicated than suggested above. It is not at all inconceivable that the two Greek readings in question, the singular and plural forms in Gen 2.18, both go back to the earliest period of the development of the Greek text. Of possible relevence here is the rabbinic tradition reflected in Sopherim 1.9, etc., that accuses the Greek translators of changing the Gen 1.26 text to read "I will make man in an image and likeness." Extant Greek witnesses do not attest the singular verbal form in Gen 1.26, although the possessive pronoun is lacking in at least one patristic citation of that verse. [Do any versions have the singular verb in Gen 1.26?] It is probably safe to assume that the rabbinic tradition reflects the vague recollection of contention about this passage, perhaps even variation within the Hebrew -- it would be understandable, for example, if in the early Christian period an attempt was made to deflect the Christian apologetic use of this passage by removing the plural reference, or even in the pre-Christian period with reference to possibly problematic interpretations such as in Philo [check Alan Segal and "Two Powers in Heaven" materials]. Paradoxically, however, the plural verb is preserved in the "Masoretic" Hebrew text and defended in the rabbinic tradition at Gen 1.26 (so also Philo), while an issue is not made of the confusion in Gen 2.18. Strange, it seems, from our later perspectives.

We certainly have only a minuscule slice of the textcritical pie for these materials, and openness to the various possibilities while we await more careful analyses of the data would seem to me to be in order. The completion of the CATSS data bank of LXX/OG textual variants will be a significant step in that direction. Meanwhile, I suggest that at the very least, the variant form of the verb in Gen 2.18 based on the Greek MSS of Philo should be restored to the textual apparatus.


NOTE: For purposes other than the immediate IOUDAIOS discussions, this posting constitutes a published critical note, and may be quoted and referred to in accord with normal scholarly conventions. Its publisher is IOUDAIOS, originally on 20 May 1993, with revision 2.0 on 10 June 1993, and its author Robert A. Kraft of the University of Pennsylvania, who retains copyright and the right of revision and reissue under his own name (including hardcopy). 

 [Text recovered by Max Bolno and Greek font inserted by RAK 13 September 2013]