Journal of Hebrew Scriptures - Volume 20 (2020) - Review

John Screnock presents us with a revision of his doctoral dissertation completed at the University of Toronto. He seeks to demonstrate that:

[W]hen translation technique is isomorphic and source oriented, the process of translation bears many affinities with transmission. We find in translation the elements that are at work in scribalism and the types of changes made in the transmission of the text. As a result, the OG should be used more liberally in textual criticism, should be used as an example of scribalism, and should be considered as evidence in all branches of textual studies. (p. 184)

To that end, he employs data from the Old Greek (OG) translation of Exodus 1–14.

One of the significant contributions Screnock makes is to demonstrate similarities between the work of translators and scribal transmission. He employs the concepts of interlingual and intralingual translation to show that where the OG demonstrates consistent isomorphism, and discounting transformations related to translation technique (p. 94–95), the resultant text reflects the same kinds of textual variants that scholars have observed in extant Hebrew textual traditions of Exodus 1–14, i.e., MT, SP, 4QpaleoGen–Exod¹, 4QGen–Exod^a, 4QExod^b, 2QExod^a. His analysis of scribal error, linguistic variant, and sense variant (which he defines and describes on pp. 95–103) shows that “excepting the obvious language differences between OG and the Hebrew, there is little difference in terms of what we see, how often we see it, and where in the text we find it” (p. 6). A fundamental premise that drives his research is that “the OG is not a disconnected and interpretive rendering of its source. In the broader picture, we must recognize that all tradents—scribe and translator alike—engaged in both preservation of their source texts and production of new material” (p. 17). In this review I will focus on three elements specifically. I will address the issue of isomorphism; engage the proposition that the translator creates a “mental Hebrew text” in the translation process which should be treated as a distinct Hebrew tradition; and interact with his interpretation of some data from OG Exodus, i.e., what constitutes an example of isomorphism.

Screnock states that “much of the translation found in the OG, and OG Exodus in particular, is best characterized as source oriented” (p. 30). “Isomorphism” describes how this source oriented translation process becomes visible in the target text, demonstrating a translation strategy of preservation. Screnock defines the particular type of isomorphism “characteristic of OG translation” in these terms: “almost every morpheme in the Hebrew is represented by a corresponding morpheme in the Greek” (p. 28). While he applies this characterization to the entire OG Pentateuch, research has revealed that the degree and nature of the isomorphism expressed among the five books of the Greek Pentateuch, and within each at specific points, varies considerably, with OG Genesis and OG Exodus especially active in accommodating their translation to the target language.

For example, Lemmelijn's overall judgment is that “OG Exodus is target oriented.”[1] This assessment does not necessarily invalidate Screnock's general premise, but it does mean that careful attention has to be given to the kinds of accommodations the translator of OG Exodus selects and what this means for our use of this OG text in establishing specific details of the source text behind the translation. Screnock critiques Lemmelijn's assessment, referencing OG Exodus's translation of 1:14: “to take one example, the movement of a pronoun slightly alters word order, but the remaining 20 or so morphological items occur in exactly the same order in the Greek and Hebrew” (p. 43). This is true when it comes to word order, but it overlooks other features in the Greek that demonstrate accommodation to the target language. These would include no representation of the repeated ב to mark prepositional phrases; the use of κατά plus the accusative to represent את כל־עבדתם; the use of the adjectivizing article with the prepositional phrase τοῖς ἐν τοῖς  πεδίοις; and the use of imperfect Greek verb forms to render wayyiqtol verb formations, along with shifts in number. While it can be argued that these changes do not affect in essence our ability to reconstruct some specific linguistic formations in the presumed source text, these variations do show the degree to which accommodation to the target text does occur. We observe serial fidelity, but not strict isomorphism. Placing Greek Exodus in the continuum between source-oriented and target-oriented translations remains challenging.

Screnock's perspective on isomorphism leads him to conclude, for example, that OG Exodus's rendering at 1:17 καθότι συνέταξεν αὐταῖς (rendering כאשר דבר אליהן in MT, SP, and 4QGen–Exod^a) reflects a Piel form of צוה (p. 123). He cannot confirm whether this “variant” was in the OG's Vorlage or “in the translator’s mental text,” but he is clear that “the scribe/translator substituted צוה for דבר because of the command given in the near context in 1:16” (p. 123). He concludes that while context might have influenced the translator's selection of συνέταξεν, “it is just as likely to have occurred through text-critical processes in the Vorlage or in the translator's mental text” (p. 124). However, the fact remains that no evidence in the Hebrew textual tradition exists for such a text. Given this translator's penchant for varying lexical equivalences in order to accommodate his target text, it is far more probable, in my opinion, that this is entirely due to the translator's work. Screnock does not note that this equivalence occurs three other times in OG Exodus (9.12; 12.35; 31.33). Did the translator have or generate the same alternative source text in these three contexts as well? Nor does he reference Martha Wade's argument “that this verb in Ex means ‘give detailed instructions.’”[2]

Screnock assumes that the “rigid isomorphism…used throughout much of the OG” (p. 74) characterizes OG Exodus. One of his examples involves the first use of the Hebrew idiom ומלאת את ידם (Exod 28:41), which the translator renders as καὶ ἐμπλήσεις αὐτῶν τὰς χεῖρας. Screnock regards this rendering as an example of “isomorphism that results in unintelligible Greek, but “a knowledgeable reader…would know the Hebrew it represented” (p. 75). Screnock may be correct in his analysis. However, consider the following. The next clause in 28:41 provides some explanation (καὶ ἁγιάσεις αὐτούς). The translator's rendering of this expression in Exod 29 does employ τελειόω/τελειώσις consistently. πληρόω is used in 32:29 to render the final occurrence of this idiom in Exodus. Does the fact that the translator uses three different renderings for this Hebrew idiom mean that he generated a different mental Hebrew text in each case where these lexical alternatives occur? The Greek rendering in 28:41 is intelligible at some level in Greek, but it does not convey the idiomatic sense of the Hebrew in this context. How he treats this idiom in the following chapter suggests that the translator is not committed to “rigid isomorphism” but is working with a more sophisticated translation strategy.

Screnock cites several types of evidence that enables us to discern this “mental Hebrew text,” including “the presence of simple scribal errors in the OG” (p. 76) and “extended isomorphic pluses in the OG” (p. 77). He also references recent research into the role that orality and memory may play in a translator's work. However, because we have no records in which the translators describe their translation process, much of this remains conjecture and as Screnock himself notes various possible explanations can and are used to explain differences between extant Hebrew texts and the form of the OG of Exodus (p. 88). Screnock argues that most of these differences “stem from a change to the Hebrew text, whether in the Vorlage or in the mind of the translator” (p. 89). Even in cases where the OG seems to be interpretive, if it is essentially isomorphic “we can locate its changes at the level of the mental text” (p. 89), because these kinds of changes can be documented in the scribal practices evidenced in extant Hebrew texts. He asserts that such retroverted texts should be regarded as legitimate “Hebrew” texts alongside the extant Hebrew texts. What is the difference between the conjectures formed as “mental text” in the minds of modern scholars, based on the OG, and the conjectured “mental text” that Screnock seeks to identify in the mind of the OG translator? Why is one more cogent or “valid” as a textual witness than the other? In the introduction to Notes on the Greek Text of Exodus, J. W. Wevers cautioned that we should “be distrustful of texts which are extant only in the minds of scholars eminent though they may be.”[3] Screnock seems to suggest a return to the kind of retroversion about which Wevers cautioned.

Screnock's hypothesis about a “mental Hebrew text” is interesting. Obviously ancient Greek translators of the Pentateuch engaged in some type of mental process as they worked from well-known Hebrew textual traditions. Whether they formally created first a “mental Hebrew text” and then a “mental Greek text” before putting in writing their proposed Greek translation is unclear. They may have done this more deliberately in contexts where the Hebrew was difficult or where they wanted to accommodate the Greek text more fully to target language conventions. However, it is unclear that they would have considered it necessary to work in this complex manner with every text unit. The very isomorphism characteristic of OG, so critical to Screnock's argument, might lead us to consider that with many text units the translator spends very little time thinking about the translation because he was employing stereotyped renderings. However, if an OG text does demonstrate variation from extant Hebrew texts, perhaps here is where the translator took more care in choosing his rendering. But at this point we come to the perennial problem—do these differences between extant Hebrew texts and the resultant OG translation result because the translator's source text (whether written or ‘mental’) was different, or because the translator was aware of various ways to read the consonantal text, or because the translator wanted to make sure the target text expressed the “correct” meaning of the Hebrew text? Or for some other reason we cannot deduce? Does Screnock's method really help in resolving these questions regarding specific text units? It might suggest that in some contexts the translator did have a different source text, but this is where traditional methods of translation technique have to bring some guidance to the decision.

We do not have space to engage Screnock’s analysis of every OG text in Exodus 1–14. One example is the “extended isomorphic plus” found at 3:12 (εὶπεν δὲ ὁ θεὸς Μωυσῇ λέγων ὅτι Ἔσομαι μετὰ σοῦ / ויאמר כי־אהיה עמ ך). The plus in the OG identifies the subject and the indirect object, as well as marking the quotation with λέγων. In Screnock's view, if the translator was interested in accommodating to the target text, he would have used more idiomatic Greek, particularly with respect to the collocation λέγων ὅτι, which Screnock interprets as “strange Greek” (p. 78). To my knowledge λέγων ὅτι occurs only twice in Greek Exodus. At 2:22 the narrator indicates that Moses named one of his sons Γηρσάμ λέγων Ὅτι πάροικός εἰμι ἐν γῇ ἀλλοτρίᾳ (גרשם כי אמר גר הייתי בארץ נכריה). In this context generally Ὅτι is understood to be causal. 3:12 is the other context; and could it be that the translator similarly intended λέγων ὅτι (כי) to convey a causal sense? In the previous verse Moses has asked, “Who am I…?” that he should lead Israel. In v. 12 perhaps Yahweh responds, “Because I will be with you….” Wevers does not punctuate it this way in his edition of OG Exodus, but it is worth noting that the translator in no other contexts uses this construction. If this hypothesis is correct, then the translator has generated an intelligible Greek text. With respect to the plus ὁ θεὸς Μωυσῇ it is worth noting that in the two surrounding verses (vv. 11, 13) the responses of Moses are introduced by καὶ εἶπεν Μωυσῆς πρὸς τὸν θεόν. The plus in v. 12 creates parallelism and informs the reader who is speaking to Moses (see vv. 14 and 15). Of course, as Screnock argues the translator could have used more “acceptable Greek,” but the construction in the text conforms stylistically to the Greek literary context. Did the translator have to have a “mental Hebrew text” in view in order to produce such a translation?

Screnock too quickly in the case of OG Exodus assumes that pluses, minuses or word order changes are due to a different Hebrew text or mentally constructed Hebrew text, rather than the work of the translator as he begins to create his target text. Is a “mental text” of the original a required step, or is a “mental construction of the meaning of the original text” what is developed and then expressed in the target language? What gives rise to the specifics of a “Hebrew mental text” as hypothesized by Screnock? Certainly the OG translation remains a significant witness to Hebrew textual traditions of the Jewish Scriptures in the third century BCE. Screnock offers new perspectives for our understanding of the translators’ processes, as well evaluating the translator’s data and employing it for understanding, and at times reconstructing, its source text.

Larry Perkins, ACTS Seminaries