Renderings of paronymous infinitive constructions in OG Exodus and implications for defining the character of the translation

producing a literal translation. He generally follows the Hebrew text’s word order, but is creative in his choice of Greek equivalents, often showing sensitivity to context and Greek idiom. This translation approach differentiates this translator’s approach somewhat from those responsible for other portions of the Pentateuch.

Several recent monographs and articles discuss Hebrew infinitive absolute usage with paronymous finite verbs (Callaham 2010;Goldenberg 1971;Kim 2009), as well as the Septuagint (LXX) translation of this construction (Sollamo 1985;Tov 1999, andmuch earlier Thackeray 1908). The analysis of Sollamo (1985) focuses on the LXX Pentateuch. Tov's study (1999) covers the entire LXX and uses the CATSS database. Neither study had access to Wevers (ed. 1991 [Wev ed ]). Evans (2001:129) gives brief attention to the translation of this construction. Krause (1993) examined New Testament examples of finite verb + cognate participle constructions. More recent articles by Sollamo (2012) and Harper (2016) show the merit of considering how the translators of discrete segments of the LXX have treated this Hebrew construction. The Exodus translator's (G) treatment of pre-posed paronymous infinitive absolute constructions (PPIA) has not received specific attention. This article seeks to remedy that deficit. By evaluating G's various strategies used to render PPIAs, the author gains important insight into his mode of working, affirms his intent to produce an acceptable Greek translation and gains insight into translation choices made in specific contexts. This article does not discuss contexts in Exodus where IA's function differently: 8:11(LXX15); 12:48;13:3;20:8, 12;30:36;32:6;33:7;36:7 (Callaham 2010:236-243). Discussion continues about whether one or more translators are responsible for Exodus 1-34 and Exodus 35-40. All of the examples of PPIAs in Hebrew Exodus occur in Chapters 1-34 and so this issue is not germane to this investigation.
Today we have the complete list of 507 PPIAs (Callaham 2010:236-243) in the Hebrew text (MT) provided by Callaham and also the full list for the Pentateuch of the corresponding Greek equivalents in Lee (2018:302-304). Both scholars classify each occurrence according to form and function. The articles by Sollamo (1985:111) and Tov (1999:247-252) categorise the equivalents for This article gives insight into the world of 3rd century BCE Alexandrian Judaism by analysing one aspect of the Greek translation of Exodus and provides a detailed evaluation of the way the translator managed to express the essence of the Hebrew text of Exodus while reflecting to some degree the form of the Hebrew text. No previous study only analyses this translator's treatment of Hebrew paronymous infinitive absolute constructions in Greek Exodus. This research contributes to the preparation of a commentary on Greek Exodus in the Society of Biblical Literature Septuagint Commentary Series. Using the Göttingen text of Greek Exodus prepared by John William Wevers, it evaluates how this translator rendered each occurrence of a paronymous infinitive absolute construction in Hebrew Exodus, defining the primary modes employed and also seeking to explain variations from these norms. The primary method incorporates a close exegetical reading of both the Hebrew and Greek texts and comparing their texts, in order to discern how the translator treated this Hebrew idiom and illustrating his translation approach.
PPIAs used by G. Both depend upon Rahlfs edition (ed. 1965) of the LXX for their analysis. Their tables (see Tables 1 and 2) show minor variations, some of which reflect different evaluations of textual variants. Because Sollamo does not list the occurrences in Exodus for G's use of dative noun + cognate verb, I cannot explain the difference between her calculation (25) and mine (23). The category 'Adverbs, Paraphrases' includes 8:28(24MT); 15:1, 21; 22:13(MT12), but not 34:7. She may include it under the rubric 'aliter'. Tov (1999:250) Tov (1999:251, fn. 11) indicates that the translator uses a single finite verb as a rendering at: 2:19; 5:23; 12:9; 21:19, 36; 22:2, 4, 5, 11, 13; 23:5 (Hebrew text chapter and verse numbering), for a total of 11 instances. As noted in the next paragraph, the collocation in 12:9 probably is not a PPIA + finite verb construction and does not belong in this list. At 21:19, G renders ‫ירפא‬ ‫ורפא‬ with the noun καὶ τὰ ἱατρεῖα and so it does not fit this category. At 22:4, the translator employs two different finite verbs to render the PPIA construction ‫בידו‬ ‫תמצא‬ ‫אם-המצא‬ (ἐὰν δὲ καταλημφθῇ καὶ εὑρεθῇ ἐν τῇ χειρὶ αὐτοῦ), and so this example also does not fit this category. Tov does not include 31:15 in his list because he follows Rahlfs' text θανατῳ θανατωθησεται. Wev ed , however, accepts the reading of B 55 txt θανατωθήσεται as original, because he regards θανατῳ, marked by an asterisk in Arm, to be a hexaplaric addition (Wevers 1992:249 Rahlfs-Hanhart's edition (eds. Rahlfs & Hanhart 2006) that Wev ed regards as textually suspect (11:9; 22:20(MT19); 23:20). These contexts have no corresponding PPIA construction in the MT. At 11:9 πληθυνων πληθυνω is supported by B 58-82 f-246 392 120-128′ 76′ 130 799 = RA). Wevers (1992:243) evaluates the presence of πληθυνων as 'a dittograph and only the addition of nunation makes it a possible reading. It has no basis in MT…'. At 22:20(MT19) only B* reads θανατῳ ολεθρευθησεται. The variant at 23:20 shows the influence from the parallel text in 19:5. Tov (1999:251, fn.11) seems to include 12 ‫מבשל(‬ ‫)9:ובשל‬ in his statistics, but this Hebrew construction is an adjective + pual participle collocation, according to Houtman (1996:179) and Koehler and Baugartner (2001:164), and does not belong   Pre-posed paronymous infinitive absolutes normally precede the verb and G reflects this word order. Exceptions only occur where G chooses a non-verb equivalent for the infinitive absolute (IA) at 11:1 (σὺν παντὶ ἐκβαλεῖ ὑμᾶς ἐκβολῇ) and perhaps at 34:7 (καὶ οὐ καθαριεῖ τὸν ἔνοχον). A significant number of textual witnesses at 11:1 read εκβολη εκβαλει υμας (A M O′-72-29 C′ʼ b d t 121 68′ 18 55 76′ Lat cod 101 Aeth Syh). The preceding σὺν παντί is not feminine and so cannot modify the dative feminine noun ἐκβαλεῖ. Wevers (1992:171) suggests that G did not want to have two distinct, dative singular nominals following one another and so he inserted the verb between them to prevent misunderstanding. This unusual word order disregards the PPIA's structure. This makes this reading the lex difficilior and thus probably original. Callaham (2010:5) notes that 'Biblical Hebrew employs repetition of a word or its root to reinforce the significance of a word or to apply some kind of stress'. He argues that PPIA + yiqtol verb forms generally emphasise the modality of the cognate verb and tend to occur in discourse rather than narrative. Waltke and O'Connor (1990: ). Later in the article, I discuss interpretational issues in specific contexts that may contribute to G's translation decisions.
The PPIA construction in Biblical Hebrew has no counterpart in Greek idiom. The Greek infinitive cannot function in this manner. If G is motivated to prepare a target text that accommodates Greek convention and also reflects the presence of PPIAs, he has to select non-isomorphic Greek collocations. Usually (28 times) he retains superficial semantic similarity with his Hebrew parent text by selecting two, Greek cognate terms.
This fact needs to be balanced by his use of a single verb or noun (21:19) in 10 contexts to translate PPIAs. Either he discerned no particular nuance expressed in the PPIA, or if he did, he did not think it necessary to express it in his target text through some repetitive structure. Tov (1999:251) says that 'in some cases the translator may have known a shorter Vorlage, but in most cases different translation techniques must be presumed'. There is no evidence apart from the LXX text that G possessed a parent text different from the MT in these cases. In OG Exodus 31:15, the majority of the Greek textual tradition reads θανατῳ θανατωθησεται, reflecting ‫מות‬ ‫,יומת‬ but Wev ed accepts as original θανατωθήσεται witnessed by B 55txt (Wevers 1992:249; cf. earlier discussion).
In Sollamo's opinion (1985:108-109), the PPIA in such instances 'has not been translated at all' and G 'deliberately employed only one equivalent for the whole paronomastic construction'. However, the lack of an explicit equivalent for the infinitive absolute in these cases may not be a failure of translation (unless the translator followed a very literal translation process) but may in fact demonstrate a rather skillful translation, rendering sense, rather than adhering to strict isomorphism. If this is the case, then G did not think it necessary in every instance to reflect the PPIA in his translation by incorporating a related noun, participle or adverb. Sometimes he regarded the single, finite verb as an adequate rendering. Explanations for this usage might vary. In 2:19 and 5:23, G may use aorist finite verbs to maintain parallelism with the use of aorist tense forms in the previous clause. At 23.5 similar factors might influence the use of the single future tense form. G renders all occurrences of ‫ישלם‬ ‫שלם‬ with a single, future tense form, primarily using a form of ἀποτίνω ('pay restitution'; 21:36; 22:6(MT5); 22:14(MT13)). However, at 22:3(MT2), he selects ἀνταποθανεῖται to render the Hebrew construction. This text requires a thief or victim, whoever is the subject of the verb, to make restitution. The Greek rendering changes this outcome to the death penalty. Korytko (2022:136) argues that this corporal punishment refers 'to what happens to the night burglar and not the victim of the theft who acts in retaliation the following day', indicating that its roots lie 'in Graeco-Egyptian legal traditions'. G seems to be responsible for this interpretation. G uses a single, future passive tense form to render the PPIA construction.
Lastly, at 31:15, θανατωθήσεται has an imperatival force and does not require additional emphasis to communicate the application of a death penalty for violating the Sabbath. The full formula occurs in 31:14 (θανάτῳ θανατωθήσεται ‫יומת‬ ‫)מות‬ in the case of people who profane the Sabbath (ὁ βεβηλῶν). Individuals who do ἔργον on the Sabbath ἐξολεθρευθήσεται… ἐκ μέσου τοῦ λαοῦ αὐτοῦ (31:14b). In these two verses (31: 14-15), the Hebrew text requires the execution or banishment/put to the ban of those who transgress the Sabbath prescriptions. Having used the full formula to describe execution in 31:14a, perhaps G did not think it necessary to repeat the formal pronouncement of the death penalty in 31:15 when describing the consequences of virtually the same offence.  Sollamo (1985:103) affirms that 'the most literal method of translating these cases is to use a participle with a finite form of the same verbal root'. In Tov's opinion (1999:249), this translation equivalence is 'probably as close as the translators could come within the possibilities of the Greek verbal system'. Thackeray (1908:599) argues that 'where [a participle] is used in the Pentateuch an attempt is often made to render it more classical by varying the verb…or by using the simple and compound verb'. According to my analysis, G employs this translation strategy four times using cognate forms (3:7; 4:14; 22:17(MT16); 23:24b) and three times using non-cognate forms (21:5; 22:22b(MT23b); 23:4). In G's perception, the PPIAs give emphasis to the action expressed in the verb and he reflects that by repeating the verbal form as a pre-posed participle. G perhaps did not know of a noun cognate with certain verbs (e.g. ἀνανεύων ἀνανεύσῃ (22:17(MT16)) and συντρίβων συντρίψεις (23:24b)) and so he could not easily use a dative of instrumentality as he often does, to represent the PPIA construction. Such a case also might be made for ἰδὼν εἶδον (3.7) and λαλῶν λαλήσει (4.14).
He employs a present participle (4:14; 22.17(MT16); 23:24b) or an aorist participle (3:7; 21:5; 22:22b(23b); 23:4) with future and aorist tense forms. The tense form of the participle reflects G's understanding of the relationship between the action of the participle and the action of the finite verb in a specific context and is not dependent on the form of the Hebrew infinitive absolute. G's variation in tense forms of participles in such constructions diverges both from Tov's observation (1999:251) that 'almost exclusively the participle of the present tense…is used', to render a PPIA in various voices, and also Thackeray's statement (1908:599) that 'instances of the bald use of the present participle and finite form of the same verb,…are not frequent until we come to Deuteronomy which has nine of them'. G shows no particular bias for either participial tense form. G probably knew that the use of the Greek participle with the finite form of the same verb is 'to say the least, distinctly unidiomatic' (598) and given his general goal to produce a readable Greek text, he limits his use of this translation strategy.
Rather the participle reports an action that subsequently fails in some way (e.g. 'seeing they were seeing in vain'  (MT18)). ἀποκτείνω in Greek legal documents refers to homicide or onthe-spot killing and can refer to killing through the execution of a death sentence but does not describe 'the formal death penalty' (Korytko 2022:53-54). τελευτάω describes untimely death that occurs as life unfolds, namely 'the end' that a person experiences. G does not use forms of θανατόω/ ἀποθνήσκω in such contexts to describe the formal death penalty, even though Greek legal diction uses it in this way (155). As Korytko demonstrates (47-57), G's variation in his choice of verb to render ‫יומת‬ ‫מות‬ reflects G's interpretation of the verb construction in specific contexts and in accord with legal diction found in Classical Greek writers and Ptolemaic documents.  also shows that the dative θανάτῳ refers to 'untimely death' when modifying a verb of death. When it qualifies τελευτάω, it refers to an untimely death that happens in the normal course of life, and when it modifies ἀποκτείνω, it refers to untimely death that occurs by specific killing such as homicide or manslaughter, but these collocations do not refer to the death penalty (60-62). Although the formulation θανάτῳ θανατόω does not occur in Classical Greek or Ptolemaic legal documents to prescribe a death penalty, it does parallel other legal formulae such as θανάτῳ + ζημιόω and θανάτῳ + ἔνοχος + εἰμί (47-56). In the legal prescriptions of Exodus 21-23, when ‫יומת‬ ‫מות‬ refers to a death sentence appropriately adjudicated and executed and G perceives that this sentence for the crime under discussion fits within the Ptolemaic legal framework, he selects θανάτῳ + θανατόω as the equivalent. (e.g. 21.12, 15). In all of these cases, G uses various formulations that incorporate θανάτῳ to retain serial and quantitative fidelity with his parent text but not necessarily semantic equivalence.
The Greek construction in 21:22 ἐπιζήμιον ζημιωθήσεται could be interpreted in several ways. ἐπιζήμιον, a neuter singular form of the adjective ἐπιζήμιος (τὸ ἐπιζήμιον 'a fine' [LSJ, 633. I.2]), could function as the subject of the passive verb, with the sense 'a fine shall be fined'. Alternatively, as Wevers suggests (1990:333), the person involved in the altercation, who is not the husband of the woman, is viewed as the subject and the text means 'he shall be fined a fine', with ἐπιζημίον functioning as the direct object. The verb also could function as an impersonal passive construction, as the Hebrew niphal verb form might suggest, with ἐπιζήμιον functioning as a cognate accusative used in the sense 'there shall be a fine that brings loss' (ἐπιζήμιος 'bringing loss upon, hurtful, prejudicial' [LSJ, 633. I]). The MT points the finite verb as a niphal imperfect, and according to Koehler and Baumgartner (2001:859), this form of the verb means 'to be paid for'. Houtman (2000:160, 170) translates it as 'a fine shall be required'. If the Hebrew text functions as the arbiter of meaning intended by the translator, then the impersonal verb construction modified by the cognate accusative is probably intended.
καὶ οὐ καθαριεῖ τὸν ἔνοχον (34:7) could be a third example of an accusative transformation. The Hebrew construction is a negative PPIA construction ‫ינקה‬ ‫לא‬ ‫,ונקה‬ with the negative particle ‫לא‬ preceding the finite verb form. Waltke and O'Connor (1990:583 §35.2.2e) state that in PPIAs 'a negative particle, where needed, is normally placed before the finite verb'. G renders it as καὶ οὐ καθαριεῖ τὸν ἔνοχον, using a finite verb + object. This is a unique translation strategy for a PPIA in Exodus. It is possible that G read ‫ונקה‬ as a w e qatal formation and rendered it with the future καθαριεῖ. He then construed ‫ינקה‬ as a nominal form, translating it with an articulated adjective functioning as the object of the verb (τὸν ἔνοχον). However, this is speculation. Within the Greek tradition, some textual witnesses place τον ενοχον before the verb. However, given the Greek textual evidence (τον ενοχον καθαριει is supported by B 15′ f -129 30′ 318′ 55 426 799 Cyr IV 420 VI 944 Lat Hi Ezech VI 18 Co), Wevers suggests (1990:171) that the transposition has hexaplaric origins. G renders piel forms of ‫נקה‬ with καθαρίζω (20:7; 34:7).
Numbers 14:18 repeats the declaration made by Yahweh in Exodus 34:7, and in that context OG Numbers does reflect the PPIA construction explicitly as καὶ καθαρισμῷ οὐ καθαριεῖ τὸν ἔνοχον. This rendering, which also includes τὸν ἔνοχον, raises the question whether G intended his rendering καὶ οὐ καθαριεῖ τὸν ἔνοχον to reflect the PPIA construction, or whether he in fact has rendered it with a single verb. If G chose the second option, then this example belongs to the category of 'transformation by a single verb'. Le Boulluec and Sandevoir (1989:338) state that 'la figure étymologique du TM,…est absent de la LXX' and define καθαρίζω in this context as 'declarer pur'. In their view, τὸν ἔνοχον has no specific equivalent in the Hebrew text.
Based on a detailed review of each of these cognate noun + verb constructions, I would suggest that G has used several different strategies that may express a note of intensification. The most frequent strategies are his use of dative pre-posed cognate nouns to express manner or means or accusative preposed cognate nouns to express a cognate accusative. Secondly, the form of verb G selects usually is a causative formation (e.g. contract verbs [simplex contract verb forms include ἐκδικέω, ζημιόω, θανατόω, καθαιρέω, κακόω, λιθοβολέω, τελευτάω; compound contract verb forms include ἐκδικέω, λιθοβολέω); or verbs ending in -ιζω (e.g. ἐνεχυράζω, καθαρίζω, ὁρκίζω, φερνίζω; also δοξάζω with ἐνδόξως]), or a compound verb whose prefix may have an intensive significance (e.g. εἰσακούω, ἐκβάλλω, ἐπισκέπτομαι, κατατοξεύω). Occasionally G selects verbs that have none of these characteristics, primarily because these represent his default renderings for the corresponding Hebrew verb (e.g. ἀκούω, ἀποκτείνω, λέγω, ὁράω In three contexts, G employs an adverb preposed before a finite verb as the equivalent for PPIA constructions. Tov (1999:248)  The plural form here perhaps has the sense 'costs associated with treatment at a medical clinic'. G carefully distinguishes the fine paid 'other than for the lost work' (πλὴν τῆς ἀργίας) and 'the costs associated with the doctor's clinic' (καὶ τὰ ἰατρεία). πλήν + genitive functions as a preposition in this context. According to Koehler andBaumgartner (2001:1274.2), the absolute use of the piel verb form ‫רפא‬ means 'to pay the costs of healing'. G has transformed the two independent clauses in 21:19b into one clause whose verb is modified by an adverbial prepositional phrase (πλὴν τῆς ἀργίας αὐτοῦ) and a direct object (καὶ (ascensive) τὰ ἰάτρεῖα). He expresses the sense of the Hebrew text but in a form that accommodates the Greek language.
Another anomalous rendering occurs at 22:13(MT12) where G renders the protasis ‫יטרף‬ ‫אם-טרף‬ as ἐὰν δὲ θηριάλωτον γένηται, employing γίνομαι + predicate adjective. The Hebrew finite verb form is niphal, expressing a passive sense. This section gives instructions for restitution in cases where a borrowed animal is injured, stolen or killed by wild animals. 22:13(MT12) deals with the last situation, and the protasis introduces the circumstances, namely 'if it evidently has been mauled [by wild animals],…' According to LSJ (799) and TLG, this is the first attestation of the adjective θηριάλωτος in Greek literature. It is a compound form created from θήρειος ('of wild beasts' [LSJ,799]) and ἀλωτός ('liable to capture or conquest' [LSJ,75]). G renders the sense of the Hebrew collocation but takes the initiative to offer an interpretative translation that accommodates the Greek language. Tov (1999:250)  We should regard G's choice of a single finite verb (nine times) as an intentional translation, not as a failure to translate the PPIA construction, because in those contexts he perceives no need for an additional lexeme to communicate the meaning of the PPIA construction. Tov (1999:255) notes that G and OG Isaiah are the translations that most frequently use a single finite verb to render a PPIA. The goals of these translators were somewhat similar in their desire to represent the meaning of the Hebrew parent text in a form that accommodated Greek idiom.
If Korytko's analysis (2022:61-63) of the renderings of the ‫מות‬ ‫יומת‬ construction is correct, then G's variation in his translation choices of the finite verb used with the dative θανάτῳ demonstrates his awareness of the subtle ways that writers in Classical and Ptolemaic Greek communicated death-producing actions in legal and other contexts. Lexical precision in the target text seems important to G.
Even though G has preferred modes of rendering PPIA constructions, he is not wedded to them. Sometimes he employs two different finite verbs connected by καί or condenses and transforms two Hebrew clauses into a single clause, representing the PPIA construction with a noun, or uses an adjectival neologism predicatively with γίνομαι. What leads G to choose participle + finite verb or dative cognate noun + verb, or simply a finite verb as the rendering in a particular context remains elusive. Renderings with θανάτῳ seem to reflect various legal expressions. Conversely, he may render ‫ישלם‬ ‫שלם‬ consistently with a single finite verb because paying fines reflects a legal judgment whose action may inherently be emphatic or in some sense final. Modifying a verb with a cognate dative noun expresses instrumentality but may also be a concise way to translate the PPIA construction that fits Greek convention, and thus G tends to choose it as a translation. His renderings of PPIAs in 8:28(MT24); 21:19; 22:4(MT3), 13(MT12); 34:7 reveal the extent to which this translator is willing to transform Hebrew syntax in order to communicate his intended meaning.
His rendering in each context has to be evaluated carefully to discern why he adopted it. His translation choice does not usually reflect a parent text that differs from the MT, either with regard to the presence or absence of a PPIA construction.
In some cases, G may have read the consonantal text differently from the MT's pointing indicates (e.g. 34:7).
Although Tov's findings (1999:253) indicate that G's translation choices for PPIAs show considerable similarity with OG Genesis, G nonetheless demonstrates an unusual degree of flexibility in his renderings in comparison with the rest of the Greek Pentateuch. He focuses primarily on producing a Greek text that is readable, sensible and somewhat contextualised for Greek speakers, who cannot reference the Hebrew parent text. Communicating the meaning of the Hebrew text as he discerned it has priority over literal, isomorphic renderings that reflect Hebrew syntax or specific lexemes. Word order is an exception in that his Greek translation usually shows serial fidelity with the MT's word order.