From protection to passover: possible greek influences

1 From Protection to Passover: Possible Greek Influences1 Alan Flashman ‫ פ‬originally carried only one As Cohen2 has pointed out, the Hebrew verb meaning, that of protection. At some point this meaning became lost to Jewish exegesis, being replaced with the meaning to pass over, in Hebrew pressure to conflate homonyms, in this case ‫פ‬I with . Scholars presume a ‫פ‬II, the latter meaning to limp, as responsible for this transformation. No historical hypothesis to account for such a major revolution in the understanding of a verb of such centrality to Jewish traditional practice has been proposed. The terminus ad quem for this change is presumed to be the third century B.C.E. according to the evidence given by LXX to Ex. 12:233. Further evidence is brought from Philo4. The actual term Passover occurs first in Josephus in Antiq. ii.313 where he states that the meaning of the Pesach Holiday is that the Lord passed over the homes of the Hebrews in Egypt: ὅ ῦ ἔ αί Αἰ α ὰ ὸἔ ᾽ὑ ε ί ἐ α έ ὕ ά ι , ό ὴ ύ α ᾽ἐ ό ὴ ἑ ί ὴ ά ὴ ἡ έ α ὁ α α ῦ , ὸ αὐ ῶ ὑ ε ὰ . (my emphasis added). This same translation occurs in Aquilas. This meaning of Passover was accepted in the following centuries by the Rabbis, and became dominant to the exclusion of the original 1 The author gratefully acknowledges the critical comments of Profs. Emanuel Tov and Tsipora Tal Shir which led to significant alterations. Prof. Chaim Cohen provided immeasurable assistance, guidance, support and friendship. The author takes full responsibility for the final proposals in this paper. 2 See ωhaim ωohen , " χdditional ‘False Friends’" (Heb.) in Avi Hurwitz Festschrift, Jerusalem: Hebrew University Press, 2008, pp. 188-191. Comp. S. E. Loewenstamm , The Evolution of the Exodus Tradition. Jerusalem: Magnes Press,1992, pp. 197-202. 3 S. Liberman, Greek and Hellenism in Jewish Palestine (Heb), 1962, pp. 187-8 4 Colson, Loeb Classics, Philo, Vol, 7, Trans and Ed. F. H. Colson, Cambridge, MA: Harvard, 1937, p. 27. 1 2 meaning. Thus the transition is presumed to be based on purely linguistic pressure and its occurrence spread over at least six centuries! The purpose of this paper is to reexamine the evidence from the Greek sources, to suggest the possible contributions of the Greek in LXX and elsewhere in influencing the specific choice of as a "solution" to the homonymic problem, and to place this radical metamorphosis within a historic perspective. To the reader of Greek, Josephus' and Aquilas' choice of verb is nothing less than astonishing. The verb they use, π α , could never translate the verb occurrences in Ex. 12 in its concrete meaning. π α ‫ פ‬in its three as a verb of concrete action always denotes passing over a threshold into. I note here two examples from the most widely read classics. When Hector crosses the Achaean bulwarks and enters into the Achaean camp in Iliad 12:468 , π α Hesiod’s Works and Days (verse 828) π is his verb. As the penultimate word in α means to trespass. Josephus himself in the Bellum used the term eight times exclusively in the original concrete meaning: i.151; iv. 127,182; v. 199, 301; vi. 125, 126, 395. Notably in vi. 125, 126 he has Titus reprimanding John for trespassing into the Temple that was surrounded by Greek notices that the boundary not be passed over5. The same Greek term π α translates in LXX II Sam (II Kings) 22:30 with the same concrete meaning, and may provide the link between a general verbum movem and . What made such an apparently problematic translation possible? Already by the time of Plato6 the verb π α came to have an abstract meaning as well, which is to In Antiq. xv. 410 the same sign is quoted without this term, rather forbidding entrance ( ἰ έ α ). The Greek text of the extant Temple Warning (in Istanbul and in the Rockefeller Museum) forbids entrance with a synonym ( π υ α ). 6 Republic 528d 5 2 3 skip. A related term, πα α (see below) also came to have an equivalent meaning. Here we have contemporary proof positive. Philo, at the very beginning of Moses7 writes about things that might be skipped over in an abstract way, and uses the two verbs, πα α and π α in two adjoining sentences with identical meanings.8 Thus, as an abstract verb, it would have been suitable to replace ‫ פ‬of Exodus 12 with π α , and so “pass over”, in the abstract meaning of to skip, was born. In LXX π α assumes this abstract meaning but once, in Micah 7.189. However, to the natives of the Land of Israel, the concrete Homeric and Hesiodic associations would have been far too remote to arouse discomfort with this choice consistent with contemporary colloquial Greek. For these natives π verb α ‫פ‬, and so the “second” meaning of could be considered the meaning of the ‫פ‬, with the verb that also translates , could resolve the problem of homonymy. But the question of the historical context for this linguistic transition still requires attention Here we turn to the interesting treatment that the root noun ‫ פ‬is given in LXX. The ‫ פ‬is never translated in the Septuagint. The Aramaic form ‫ פ‬is consistently transcribed as πα α, except in Paralipomenon (Chronicles) where it is transcribed as 7 Mos. i. 3-4. The identical pairing occurs in Spec. Leg. i. 162-163. It is interesting to note the pairing of these two verbs, in their concrete meaning, in LXX Job 9:11. The case of Ben Sira 20:7 poses an interesting intermediate example, where the silence of the wise man awaiting the proper moment is contrasted with the fool whose relationship to the opportunity is π α . The English renderings varied from reading the verb in its concrete meaning as speaking too early to reading its abstract meaning as speaking too late (skipping the opportune moment). Manuscript C reading ‫י‬ ‫י‬ ‫י י‬ makes it clear that at this dateable moment π α was used as a translation in its concrete meaning. See Pancratius Beentjes, The Book of Ben Sira in Hebrew, Supplements to VETUS TESTAMENTUM 68, 2006, p. 98. There is perhaps some irony in π α translating ‫י‬, the direct negative of the meaning of πα (see below). 9 T. Muraoka, A Greek-English Lexicon of the Septuagint, Louvain: Peeters, 2009. s.v. pp. 697-8. 8 3 4 φασε 10 ‫ פ‬occurs on only three occasions, all in Exodus 12 where . However, the verb the original Pesah story is recounted, and each occurrence is translated. In the first and ‫ פ‬is translated with the same Greek term, third verses (v.13; v.27) means literally to shelter, to put a roof ( πα which π ) over them.11 In the middle verse, Ex. 12:23 two verbs appear in the original Hebrew passage. Where the Lord passes into Egypt in order to smite the Egyptian firstborn, the Hebrew is employed. Where the Lord performs some action with regard to the doors of the Hebrews, the verb ‫ פ‬is employed. Here LXX uses the same future indicative form of the verb παρερχομα to translate both Hebrew verbs: α παρε εύ ε α ἐπ' ἀ φ ῶ ὀ α ἰ This verb, πα πα ῶ α ῖ ἰ α Aἰ υπ υ α ὄ α παρε εύ ε α ἰ α ῶ πα αἷ α ἐπ α α α α ἀφή α (my emphasis added) α is the consistent translation for the verbum movem LXX, and its forms are used to translate the various other meaning of tense and transgression12. In v. 23 as a translation for pass over. But why should πα φ ᾶ throughout , such as past ‫פ‬, it would mean pass by13, not α translate the verb ‫ פ‬in this verse while several For example, II 30:2. This form appears once in Jer. 31:7 (LXX 38:8). See Muraoka, s.v. πα α, p. 538. The alternative φα does not appear in the Lexicon. 11 Muraoka, s.v. p. 623: "to provide protective covering or shield for". This verb was not chosen lightly. It recurs in LXX at important moments where another verb might be expected. For example, when the infant Moses is hidden (Ex.β:β), at the Splitting of the Red Sea (Ex. 1η:β), εoses’ epiphany at Horeb (Ex. γγ:ββ), the Curtain's relation to the Ark (Ex. 40:3), 11 times in Psalms, and at crucial points in Maccabees II when Judas is miraculously protected (10:30; 13:17) and in both Maccabees I 3:3, 6:37, 9:38 and Maccabees III 3:27,29; 5:6 and 6:9. Parenthetically, the translation of the verb ‫ פ‬in Ex. 2:2 with this same πα verb that translates ‫פ‬ could shed light on the ancient reference to the ‫ פ‬sacrificial lamb, and it’s later replacement the " ‫" פי‬ as “ ּ‫ '” פ‬a term otherwise difficult to explain. (ωompare the alternative somewhat forced suggestion offered by the illustrious, albeit innocent of Greek erudition, Rabbi Chaim Soloveitchik of Brisk, quoted in J. B. Soloveitchik, Hagadah Siah HaGRI"D (Hebrew), Ed. Y. A. Lichtenstein, Jerusalem, 1995, p. 81, n. 68.) 12 Muraoka, s.v., p.534. First meaning, " to move past without taking notice of". 13 Muraoka, Ibid. Second meaning, citing both parts of the verse " pass by beside." 10 4 5 verses earlier and later a different verb is employed? There exists no textual evidence for an alternative reading of ‫ פ‬in this verse that could solve our problem.14 While lexical variants abound in LXX, a distinct change in the meaning of a verb as it appears in closely related verses is very unusual, especially in the Pentateuch. The speculative range for accounting for the alteration is broad. At a minimum, it is possible that the pressure to solve the homonymic problem was felt already in the third century B.C.E. Verse 23 seemed to invite the use of the verbum movem because it already introduced it with the of the movement of the Lord into the Egyptian area. This would assume that the LXX audience was aware that the same Hebrew verb ‫ פ‬was given the two different meanings it had already come to possess15. ‫ פ‬in v. 23 was due A more extreme speculation would be that the rendering of mainly to Greek literary considerations. The double use of πα α makes for a very good literary reading in v.23. Using the same verb to mean pass into and to pass by would appeal to the literary taste of the Greek audience. Πα α is a verb especially suited for such double meanings. On the other hand, to translate the πα ‫ פ‬of v. 23 with could create two possible problems. The first problem is that the verb used with the preposition homes. Since πα ‫ פ‬is to refer to the two-dimensional doors of the Hebrews’ taken concretely suggests a three-dimensional activity of roofing over, it would produce a less attractive Greek alternative. Σ πα tends to the meaning of protective enclosure, to the point of meaning clothing or skin. One could say, the Glasson, in Loewenstamm, op. cit., p. β00, suggests an “inadvertent” copyist error due to attraction of the second verb to the first. 15 S. Liberman, op. cit., pp. 187-8. 14 5 6 literalism of the translation would be apparent, since this would be a clumsy Greek choice of words. The second problem could have been of a theological nature. The Septuagint consistently translates the Tetragrammaton as υ (Lord) and ‫י‬ . By the time of Philo, at least, it was obvious to Greek readers that υ the demanding and punishing aspect of God, while - as referred to referred to the merciful and protecting aspect16. (Rabbinic theology adopted the opposite understanding.)17 The verb ‫ פ‬is used in v. 23 juxtaposed to υ Greek reader to find υ have the same υ πα , unlike in the two neighboring verses. For a performing a sheltering function might read badly, while to smite the Egyptians and pass by the Hebrews – with the same α - reads very well indeed18. An even more extreme speculation would suggest that the LXX audience was already innocent of knowledge of the Hebrew text, and that LXX served as a replacement rather than an exegetical translation19. According to this line of thinking, the literary 16 For example Mos. II 99. Compare Urbach, The Sages,(Heb.) , Jerusalem: Magnes Press, 1969, p. 106, n. 22. Tov is certainly correct in pointing out that LXX does not employ such a distinction with any consistency ( Emanuel Tov, "Theologically Motivated Exegesis Embedded in the Septuagint, in in Emanuel Tov, The Greek and the Hebrew Bible. Supplements to VETUS TESTAMENTUM #72, 1995, pp. 260-1). However, it seems plausible to me that such a distinction might have had a basis in at least some passages from the outset, and that the use of this distinction grew to become more systematic over time. The present passage, thick with theological overtones, could have been one of the original examples where the distinction between the two divine epithets bore significance at an early stage. 17 See Urbach, op. cit., p. 400. 18 It is possible that πα and πα α created an accepted literary form of a complementary set of verbs. The meaning would be that that which is protectively sheltered can then be safely passed by. The two verbs work together with such a meaning in LXX in the theologically and dramatically charged verse Ex.33:22: ἡ α ' ἂ παρέ ῃ υ ἡ α α ή ἰ ὀπ π α α επά ω υ ἐπ ἕ ἂ παρέ ω. It is worth noting that Jerome preserves the play between these two verbs. The Vulgate translates both the verb (Ex. 12:12,23 transibo, transibit) and the verb ‫( פ‬Ex. 12:13,27 transibo, transivit) , and sometimes the noun ‫( פ‬Ex.12:11 phase id est transitus Domini; Ex 12:27 victima transitus Domini) with the Latin equivalent of πα α , transeo. It is precisely in v. 23, where LXX reads πα α for the verb ‫ פ‬that Vulgate uses the Latin equivalent of none other than π α , viz. transcendet. 19 See Mireille Hadas-Lebel, Philon d’Alexandrie, trans A. Gilead (Heb.),2006, ch.3. Compare Harry Wolfson, Philo, 1947, ch.1. It is worth adding that the Letter of Aristeas claims that the translation was 6 7 choice of πα α would not have been noticed in Alexandria as a translation of ‫פ‬ at all. According to this speculation, LXX may have created the verbum movem option which later became attractive as a solution for the homonymic problem among Hebrew speakers20. The place of this πα α translation in LXX may influence our understanding of a later rabbinic text. R. Yoshia is quoted in Mekhilta as suggesting an alternate reading of the Hebrew ‫י‬ ‫י‬ 21 ‫י‬ ‫ פ‬in place of " . ‫י‬ ‫פ‬ ‫פ‬: ‫י‬ ‫י פ‬ ( " ‫י‬ ) ‫י‬ ‫ 'י‬: ‫י‬ ‫ופ חתי‬ ‫י‬ Liberman (above) sees this as an expression of the Rabbinic understanding of the verb ‫פ‬, that is an exegesis. With due deference to Liberman's extraordinary authority, in accord with the more extreme speculation regarding LXX, I here offer an alternative reading. According to this understanding, R. Yoshia does not mean literally to replace the ‫ פ‬verb in all three instances with Bible with ‫פ‬, a verb that does not appear at all in the Hebrew as its middle letter. That would entail changing the very name of the holiday. Rather, I suggest that R. Yoshia is playfully suggesting that in order to get mean pass over, you have to replace the Hebrew verb ‫ פ‬to ‫ פ‬with a Greek verb that means made for King Ptolemy accords with the notion that the translation was not meant to have any interaction with the original. There the Hebrew language is referred to as entirely foreign and inaccessible to the Greek world (section 11).The rabbinic memory of “the things the Sages changed [when translating] for King Ptolemy” accords with this as well. τn this see Emanuel Tov, "The Rabbinic Tradition concerning the 'Alterations' Inserted into the Greek Translation of the Torah and Their Relation to the Original Text of the Septuagint," in Emanuel Tov, The Greek and the Hebrew Bible. Supplements to VETUS TESTAMENTUM #72, 1995, pp. 1-20. 20 A possible parallel in which lexical variants of one Hebrew word is influenced by proximity to a Greek word could be found in Dt 28:59. There the Hebrew ‫י‬ is rendered first as αυ α and then as π in the same verse. The first and much less literal rendering could well have been influenced by ὸ αυ α ὸ in the preceding verse. Of course, this parallel would be valid only if an alternate reading in the Hebrew anlage is excluded. 21 Mechilta D'Rabbi Ismael, ed. H.S. Horovitz & I. A. Rabin, Jerusalem: Bamberger & Wahrman, p. 24 ll. 3-5. LXX for in Cant. 2:8 is π . 7 8 to pass by, ‫פ‬. R. Yoshia transports us from the ironical suggestion of ‫ פ‬to the ‫ פ‬means. I am suggesting that R. Yoshia is meaning of pass over, which is not what referring to just the process of translating back from the Septuagint, in v. 23, where πα α could indeed translate the verb ‫ פ‬. R. Yoshia could then have been hinting at the replacement of the verb in the Septuagint, and its later Palestinian replacement with its synonym π α which means as well. I propose, with due tentativeness, that his jest22 was made regarding verse 23, but once the gist of the play on the Greek was lost, his bon mot was moved up to the first instance of ‫ פ‬in verse 13. 23 One more Greek verb requires discussion here. During this same period, this third verb form in Greek became prevalent for referring to the Pesah holiday itself. Philo makes reference in two places to this Holiday bearing the name prefix α is added to the common Greek verb for going, α α α α24. When the , we get a verb that always means to cross over a boundary. That is how it is used in the Septuagint consistently.25 This usage for the name of the holiday referred not to a translation of the verb ‫ פ‬, which actually refers to the 14th of Nisan when the Pesah sacrifice was offered in Jerusalem, but to the Biblical (Festival of Unleavened Bread), which denoted the consequent days of the holiday. For the community residing in Egypt, calling the holiday which celebrated leaving Egypt as ‫י‬ α α α would be the equivalent of referring to it elsewhere ‫( י י‬The Exodus from Egypt). Philo uses this meaning for a homily in which 22 Liberman gives one example of the Rabbinic affinity for the bon mot in a different context in Tosefta KiFeshuta Sheviit p.517 l. 19. 23 One difficulty with this reading is the greater affinity α than α to the stepping aspect of ‫פ‬. An additional difficulty may be a phonetic consonance ‫ פ‬of and ‫פ‬. It is intriguing to note that the same R. Yoshia makes a similar suggestion employing the ‫י‬ formulation also for in Ex. 12.17, suggesting or playing with the LXX α φυ ἐ α . Mechilta D'Rabbi Ismael p. 33 ll. 1-2. 24 Spec. Leg. ii. 145, De Mig.25. 25 For example Num. 32:5,7,27,29,30,32. Muraoka, s.v α α ., p 148. 8 9 the subject of the verb α α is not the Lord, but the Israelites! There is no reason to think anyone ever translated the verb ‫ פ‬as α α .26 However, it would seem likely that the Jews in The Land of Israel were fully aware of this name for the Holiday, and the possible connection of the Holiday with a verb rooted in α could have entered popular consciousness. If we then remove LXX and Philo from the change in the meaning of ‫פ‬, we have a change that can be dated to the late first century C.E., the period of the destruction of the Second Temple. The two background Greek usages from Alexandria, πα for ‫ פ‬in v. 23 and α α the ultimate choice of υπ α in the popular name for the Holiday, could have influenced α in the First Century C.E. In addition to the homonymic problem, it could be that the Pharisees27 began to interpret the Biblical story in a somewhat less concrete fashion, and with a more consistent theological line. When Ex. 12:23 is read in the Hebrew, with ‫ פ‬meaning to protect, it would appear that two forces Contra Colson's note that "Philo consistently uses α α α or α α = πα α" in De Spec. Leg. and the longer note in the Appendix, p. 27. (Loeb Classics, Philo, Vol, 7, Trans and Ed. F. H. Colson, Cambridge, MA: Harvard, 1937.) I suggest rather that Philo confused the name of the Holiday with the name of the Sacrifice, which are one in colloquial Hebrew (Aramaic), since this provided the allegorical homily he was seeking (in De Mig.). I do not think that he would have then used α α to translate the verb ‫פ‬, as I do not believe that he was aware of the occurrence of the verb ‫ פ‬in the Hebrew original. Philo's etymologies are almost entirely limited to proper names. Here is one striking example of Philo's utter ignorance of the Hebrew verb in the original text. Philo consistently defines the meaning of the term ‫ י‬as "one who sees God." The Hebrew source of this derivation is of course Gen. 32:28, but in LXX the verb ‫ י‬, translated υ α , bears no clear relation to the new name. Philo had to rely upon the later verse, Gen. 32:30, where Jacob's choice of the name of the place ‫ פ י‬refers to seeing God, ‫ י‬- ‫ י י י‬which in LXX reads α . Only if one reads the passage in question solely in the Greek with no knowledge of the Hebrew verbs, would one conclude that v. 30 rather than v.28 provides the etymology of the name. J. W. Earp collected all the sources in "Index of Names", Philo, Vol. X, p. 334. Contra Loewenstamm, op.cit. p. 200. 27 The assumption here is that the later rabbinic adoption of this understanding suggests its origin with the Pharisees, with whom Josephus claimed to have been affiliated (Life 10-12). An alternative hypothesis could be that the passing over rather than sheltering meaning of the Holiday could be connected with a premonition and then later reaction to the loss of the Temple. Such anxious concerns were expressed prior to the destruction of the Temple, for example (assuming historical credibility) in NT Mt 24.1-2. 26 9 10 are at hand. One force is the smiting force that afflicts the unprotected Egyptians. The other force is the protecting force, which protects the Hebrews from this first force. The Pharisees could have found such an allusion troubling to a more strict and abstract monotheism that they were attempting to consolidate. I speculate that the Pharisees could have noted an alternative solution in the Septuagint, taking πα α for ‫ פ‬as an exegesis. Here only one force is at hand in the verse. However, the word πα α to designate this central action that gives the Holiday its name would have been entirely unsuitable, because this verb commonly connotes the transgression of commandments, as I mentioned above. Another verb was needed with a similar meaning. Thus we can speculate that three Greek elements may have contributed to the seemingly Semitic problem of the ‫ פ‬homonyms. (1)The neighboring Alexandrian could have prompted the choice of a verb with a similar α πα α α α basis. (2)The presence of α in LXX Ex.12:23 may have suggested or reinforced a verbum movem solution. (3)The synonymic abstract meaning of πα α and υπ α could come together as a literary, philological and theological "resolution". If we regard only the written evidence at hand in precise chronological order, a more restricted hypothesis could be that Josephus is personally responsible for the change. We would place him in Rome, two decades after the Destruction, presumable with Greek sources (LXX and perhaps Philo) and his (imperfect) memory of Hebrew sources. When he wishes to explain the meaning of very verse he paraphrases, uses πα verb ‫ פ‬he has LXX which in v. 23, the α . Perhaps he did not recall exactly where the ‫ פ‬appeared in the Hebrew original. He recalls the Alexandrian epithet (or sees and only partially comprehends the Philonic text). He conflates α α with πα α, 10 11 takes them in the abstract meaning of to skip (he mentions the people of Israel, not their homes) and comes up with υπ α as an abstract equivalent. We then take this error into Palestine (one of the presumed audiences for whom Antiq. was intended, with the same purpose as Bellum), where evolving Rabbinic theology prefers the one Divine force rather than two forces reading of v. 23. Thus Aquilas adopts it now as an exegetical translation. The homonymic conflation then gradually comes to support the new resolution. While this hypothesis offers a narrative of seductive simplicity, this must not be confused with historical plausibility. It must be recognized that few problems in ancient philology have such precise solutions. This "resolution" became consensual, I hypothesize, only very gradually and perhaps only as the people and their leaders became resigned to the irreparable loss of the Temple. In the Mechilta p. 39 ll. 3-4 ‫ פ‬is taken to mean protection, without any discussion. Clearly the transition to the meaning of passing over was still far from consensus. It is possible that the experiential meaning of the Festival became over the centuries, and perhaps more so in the Diaspora, more akin to escaping destruction, that is, surviving, than to being graced with Divine protection. The Biblical Pesah holiday was after all mainly a Temple holiday, and was transmuted into a home festival with tremendous grief and longing for restoration. The very transition of the meaning of the holiday from protection to passing over would thus fit into the historical context of the readjustment of Jewish religious discourse in the period during and after the loss of the Temple in Jerusalem. 11