1. Some examples are cited in Charlesworth, James H., “The Concept of the Messiah in the Pseudepigrapha,” in Aufstieg und Niedergang der Romischen WeltII. 19.1, ed. Wolfgang, Haase (Berlin and New York, 1979), p. 189, n. 4.Google Scholar

2. Neusner's, JacobMessiah in Context(Philadelphia, 1984) argues at length for the relative insignificance of the Messiah in most early rabbinic works.Google Scholar

3. Tanhumaed. Buber, II, p. 43 and parallels. See Ginzberg, L., Eine Unbekannle Judische Sekte(New York and Pressburg, 1922), p. 334 (hereafter cited as Sekte) = Monatsschrift fur Geschichte und Wissenschaft des Judentums58 (1914): 412 (hereafter cited as MGWJ) = An Unknown Jewish Sect(New York, 1976), p. 234 (hereafter cited as Sect).Google Scholar

4. Sektepp. 335–336 = MGWJpp. 413–414 = Sectp. 235.Google Scholar

5. Josephus, Antiquities20.5.1.

6. Ibid. 20.8.6.

7. Sibylline Oracles5.256–259. See Teeple, H. M., The Mosaic Eschatological Prophet(Philadelphia, 1957), pp. 10–11 (and note the references on pp. 29–31 concerning the exodus as a prototype of the final redemption). Cf. also G. Vermes, Jesus the Jew(New York, 1973), p. 98.Google Scholar

8. Be-Midbar Rabbah11:3; Shir ha-Shirim Rabbah2:22; Ruth Rabbah5:6; Pesikta Rabbati15, ed. Friedmann, p. 72b (cf. esp. n. 63 there); Pesikta de-Rav Kahanaed. Buber, p. 49b. See also Sektep. 335 = MGWJp. 413 = Sectp. 234.

9. Or at least he said so. See Antiquities20.8.6.

10. Matt. 2:14–15. The fact that the plain meaning of Hosea 11:1 refers to the exodus means that Matthew's citation of that verse strengthens rather than weakens the typological interpretation.

11. Matt. 4:2. This, of course, is a miracle, but not a redemptive one.

12. Antiquities20.8.6; War2.13.4. On the typology of Moses, see Teeple, Mosaic Eschatological Prophetpassim; S. Isser, The Dositheans: A Samaritan Sect in Late Antiquity(Leiden, 1976), pp. 131–142; Vermes, Jesus the Jewpp. 97–98, and esp. his references in n. 61.

13. B. Sanhedrin99a; Pesikta Rabbati1, p. 4a.

14. “The Messiah of Ephraim and the Premature Exodus of the Tribe of Ephraim,” Harvard Theological Review68 (1975): 1–16. A Hebrew version of the article had appeared in Tarbiz40 (1971): 450–461, and has been reprinted in Heinemann's Aggadot ve-Toldoteihen(Jerusalem, 1974), pp. 131–141. References here will be to the version in HTRwhere the summary of earlier theories appears on pp. 1–6.Google Scholar

15. Ginzberg, , Sektepp. 336–339 = MGWJpp. 414–417 = Sectpp. 235–238. The rabbinic sources about the Ephraimites are noted by Ginzberg and discussed by Heinemann, “Messiah of Ephraim,” pp. 10–13.Google Scholar

16. Parteipolitik der Hasmonāerzeit im Rabbinischen und Pseudoepigraphischen Schrifltwn(Vienna and New York, 1927), p. 107.Google Scholar

17. “Messiah of Ephraim,” p. 4. In the Hebrew, “and the like” was the stronger “and acts of salvation” (), which reflects Aptowitzer's assertion more closely. Whether neutral acts, which are neither redemptive nor sinful, would be recapitulated is left ambiguous.

18. The second reference in B. Sukkah52a. On the problems of dating the earlier reference on that page, see Klausner, J., Ha-Ra'ayon ha-Meshihi be-Yisrael(Jerusalem, 1927), pp. 318–319.Google Scholar

19. einemann, , “Messiah of Ephraim,” pp. 6–8.Google Scholar

20. Ibid., pp. 10–13. Heinemann attributes special significance to this last source (B. Sanhedrin92b and elsewhere); I have downplayed it somewhat for a reason that will soon become evident.

21. “When [Messiah son of David] saw that Messiah son of Joseph was killed, he said before God, ‘Master of the Universe, I ask you only for life'” (B. Sukkah52a). The point was made by Klausner, Ha-Ra'ayon ha-Meshihip. 318.

22. “Messiah of Ephraim,” p. 8, n. 31.

23. Precisely this conviction is attested in sources from a much later period; see Kasher, M., Ha-Tekufah ha-Gedolah(Jerusalem, 1969), pp. 428–431.Google Scholar

24. Though there are imperfections in the analogy, one cannot help but think of the Zionist reevaluation of the ma'pilimof Numbers 14:40–45 in Bialik's Metei Midbar.

25. See nn. 8–13 above and cf. n. 17.

26. See n. 21 above. The same can be said about the possibly tannaitic source a bit earlier in Sukkah52a.

27. Let me make it clear that I consider Heinemann's point about the likely attitude toward the Ephraimites in the post–Bar Kokhba period to be extremely useful but not absolutely indispensable for a defense of Ginzberg. A weaker defense might maintain that a condemnatory and a neutral attitude toward the Ephraimites coexisted in the pre–Bar Kokhba period and that the latter (which saw them as mistaken calculators) produced the typological figure of Messiah son of Joseph. One might even regard the severe condemnation in the Mekhiltaand elsewhere as a later development–a reaction to the Bar Kokhba revolt by one (minority) faction that was so concerned to prevent a repetition of this disaster that they were indifferent to the implication for R. Akiva's reputation. Nevertheless, I agree with Heinemann to the extent that I cannot imagine this as a majority view. (For a new typological explanation that does not persuade me, see Raphael Patai's suggestion that Messiah son of Joseph dies because Moses died short of the promised land [The Messiah Texts(New York, 1979), introd., p. xxxüi]. Shimon Toder's “Mashiah ben David u-Mashiah ben Yosef,” Mahanayim124 (1970): 100–112, came to my attention after this article was completed. Though it contains no reference to Ginzberg, it maintains the typological origin of Messiah son of Joseph and notes that the attitude toward the Ephraimites in the Aggadah is not uniformly negative.

28. B. Sanhedrin97b. On rabbinic opposition to calculations, note the material assembled by A. H. Silver, A History of Messianic Speculation in Israel(Boston, 1959), pp. 195–206.

29. See the middle of B. Sanhedrin97b, and note Klausner's emendation of R. Simlai to Rabbi Ishmael (Ha-Ra'ayon ha-Meshihip. 272).

30. B. Sanhedrin97a-b; B. Avodah Zarah9a.

31. B. Avodah Zarah9b.

32. Ibid.

33. B. Sanhedrin97b.

34. Ibid. Because of a misreading of three rabbinic passages dealing with the durationof the messianic age, Silver presents three other dates for the time of its advent; see his Messianic Speculationpp. 19–20, #3 (and contrast his correct reading of analogous material on p. 14, #2), and pp. 25–26, #1 and 2. Silver's misreading was endorsed by Yehudah Even Shmuel, Midreshei Ge'ullah(Jerusalem and Tel Aviv, 1954), introd., p. 42; the proof-texts cited in these passages, however, rule out this interpretation. A rabbinic statement which could be considered typological describes Balaam's speeches as taking place at the midpoint of world history; though some medievals cited this as a messianic calculation (and the proof-text tends to support such a reading), it may tell us only when the world will end. See J. Shabbat6:9, fol. 8d, and cf. A. Halkin's introduction to Maimonides' Epistle to Yemen(New York, 1952), p. xüi. For what may be another typological calculation with details unclear, see the last statement in section 21 of the introduction to Eikhah Rabbati.

35. Whatever Iranian influences may have affected this calculation (see the reference in E. Urbach, Hazal: Pirkei Emunot ve-De'ot[Jerusalem, 1969], pp. 610–611 = The Sages: Their Concepts and Beliefs[Jerusalem, 1975], p. 678) cannot be allowed to overshadow the straightforward relationship with the days of creation. Cf. the associated talmudic statement (B. Sanhedrin97a) about a six-thousand-year period followed by a one-thousand-year “Sabbatical” destruction.

36. The discussion of this point in Neusner's Messiah in Contextp. 180, creates the impression that the only duration assigned to the sojourn in Egypt by Scripture is 430 years (Exod. 12:40).

37. Die Eschatologie der jüdischen Gemeinde im neutestamentalichen Zeitalter(Tūbingen, 1934), p. 144.

38. Or 3829. See the Ba'al ha-Ma'or's comments on Avodah Zarah9b (= fol. 2b of the Rif), s.v. amar R. Huna.In either case, the last official year of the Temple is considered 3828, and 3829 is the first year of destruction; hence, the four hundredth year remains 4228. The years 3828 and 3829 are 68 and 69 C.E. according to the current Jewish calendar; nevertheless, the common view that the rabbis misdated the destruction of 70 c.E. by one or two years is mistaken, because their calendar differed by a year or two from the one that became standard among medieval Jews. See the Ba'al ha-Ma'or, loc. cit., and E. Frank, Talmudic and Rabbinical Chronology(New York, 1956). This affects other rabbinic dates as well and means, for example, that the eighty-fifth jubilee is not 441–490 c.E., as scholars routinely indicate, but 442–491 or 443–92.

39. Silver, Messianic Speculation(p. 26), apparently oblivious of the Talmud's comment, also considers 4231 as the four hundredth year of the destruction, since in the current Jewish calendar it is “c. [this little letter deserves notice] 470 c.E.” In a puzzling passage, Urbach cites the talmudic remark about a three-year discrepancy between the four hundredth year and 4231, and in the first sentence of text following this footnote says that 4231 is identical with that year (Ijazalp. 613 = Sagesp. 682). Perhaps he is tacitly suggesting a new understanding of the talmudic statement which would take it to mean that there is a three-year difference in calculating the four hundredth year; he does not, however, say this explicitly, and it is not, in my view, a tenable reading of the passage.

40. Tosafot Avodah Zarah9b, s.v. le-ahar;Isaac Abravanel, Yeshu'ot Meshiho1812, p. 10b. Abravanel explains 4228 (= 400 years after the destruction) in a similar fashion as a majority of the eighty-fifth jubilee in sabbatical units. (A typographical error in this edition of Yeshu'ot Meshihohas changed.)

41. See the references in n. 8. The discrepancy between 1290 and 1335 determines that the Messiah will be hidden forty-five days. Though Rashi on Dan. 12:12 understandably interprets this midrash as a reference to forty-five years, its plain meaning resists such an interpretation. For forty-five days, not years, in this context, see also the apocalyptic midrashim in Even Shmuel, Midreshei Ge'ullahpp. 43, 81, 104, 195. Some of the apocalypses also take the reference to “time, times, and half a time” in Dan. 7:25 and 12:7 in the literal sense of three and a half years; see Midreshei Ge'ullahpp. 103 and 470, and R. Bonfil's plausible suggestion in his “'Hazon Daniel' ki-Te'udah Historit ve-Sifrutit,” Sefer Zikkaron le-Yizhak Baer(= Zion44 [1979]), p. 146. It should also be noted that had the rabbis taken these days as years, they would have been forced to delay the redemption unbearably. Indeed, their failure to use Daniel as an important basis for calculations may result precisely from the fact that they regarded the numbers there as references to events taking place within the final messianic process; such numbers cannot be useful in predicting when the process itself will begin.

42. Even Shmuel maintains, as I do, that the number 4231 is also based on the fourhundred- year period of exile, but he accounts for the three-year delay by a rather uncomfortable expedient. He argues that what begins after 4228 is the seven-year period during which the Messiah will come; and “after three years of this seven-year period have elapsed, normal life cannot continue” (Midreshei Ge'ullahintrod., p. 45).

43. So too Silver, Messianic Speculationp. 26, and Urbach, Ifazalp. 613 = Sagesp. 682. Though I remain skeptical, it is worth recording a characteristically brilliant explanation proposed by Gerson Cohen when I was his student at Columbia; 4291, he suggested, may constitute a sabbatical unit of years for each commandment (613 X 7). An elaborate but unpersuasive effort to account for this date was made by Even Shmuel in his introduction to Midreshei Ge'ullahp. 46. The setting up of the abomination of desolation in Daniel 12:11, he says, must have been taken as the establishment of the city of Rome, and from that point we must wait 1290 days (= years). The traditional date of the founding of Rome is 753 B.C.E., and this corresponds to 3008 A.M. (Even Shmuel [p. 54, n. 49] regarded this Hebrew equivalent, given in a late Jewish source, as approximate. In fact, it is precise; since there was no year zero, the Hebrew year 3000 = 761 B.C.E., even though the more familiar year 4000 = 240 c.E.) 3008 + 1290 = 4298, when Rome will fall. But the rabbis often spoke of the seven-year period in which the Messiah will come, and that period will therefore begin in 4291. This is ingenious, but aside from the fact that we have no early evidence that Jews used or knew the date 3008 as the beginning of Rome (cf. the end of n. 74 below), the reference in Daniel 12:11 to the removal of the burnt- offering, which can have no association with the date of the founding of Rome, would appear to make Even Shmuel's proposal impossible.

44. Midreshei Ge'ullahintrod., pp. 45–46. Baron's summary of Even Shmuel (A Social and Religious History of the Jewsvol. 5 [New York, London, and Philadelphia, 1957], p. 366, n. 28) can leave the impression that this typological reasoning about the Temples is actually attested in the ancient sources. For such a calculation in the Middle Ages, see Nahmanides, Sefer ha-Ge'uilahin Ch. D. Chavel, Kitvei Rambanvol. 1 (Jerusalem, 1963), p. 294, citing debatable evidence from section 21 of the introduction to Eikhah Rabbati.Moshe Ber suggested that the messianic hopes associated with this jubilee may have been connected with the problems of Babylonian Jewry at the time; see Sinai 48 (1961): 299–302. On this talmudic passage, cf. also I. Levi's note in Revue des Études Juives1 (1880): 110. Urbach (Hazalp. 612 = Sagesp. 680) may have a point in stressing Elijah's uncertainty about the precise year of redemption, but that surely does not mean that there is no messianic calculation here. This explicit uncertainty, however, does have an important corollary: it prevents us from assuming that the Talmud has in mind only the last year of the jubilee, despite the fact that the Testament of Moses (1:2 and 10:12) appears to point to the year 4250 A.M. as the year of redemption. The connection of that text to our talmudic passage was already made by R. H. Charles, The Apocrypha and Pseudepigrapha of the Old Testament(Oxford, 1913), 2:423, and was repeated by E. S. Artom in his commentary to 10:12 (Ha-Sefarim ha-Hiaonim: Sippurei Aggadahvol. 1 [Tel Aviv, 1965]) and by S. B. Hoenig, “Dor she-Ben David Ba,” Sefer Zikkaron li-Shmuel Belkin(New York, 1981), p. 142.Google Scholar

45. Yeshu'ot Meshiho.p.12a.

46. Midreshei Ge'ullahintrod., p. 46. Once again, Baron's summary {History5:167) can leave the impression that this is more than a conjecture.

47. B. Sotah11b; SifreiNumbers 78, Friedmann's ed., p. 19b. There seems, however, no alternative to the conclusion of the Maharsha (Sotahad loc.) that the Talmud is referring to descent through one of David's female ancestors.

48. B. Arakhin13a. I have formulated this sentence fairly strongly in light of what I think is the correct observation at the end of Tosafotad loc, s.v. Caleb.

49. Finally–a reminder that if my speculation about Caleb is rejected, the most reasonable explanation of the eighty-fifth jubilee remains the proliferation of messianic dates within that fifty-year period, and every one of those dates is typological. Needless to say, this proliferation of dates could have enhanced the suggestiveness of the passage in Joshua as well.

50. While none of the sources portrays Armilus as Prince Charming, I have reproduced one of the most elaborate descriptions from Midrash va-Yosha', Midreshei Ge'ullahp. 96. See also pp. 79, 131, 136, 320. For an English translation of some of the Armilus texts, see Patai, Messiah Textspp. 156–164.

51. Pseudo-Jonathan to Deut. 34:3, Isa. 11:4. Cf. A. Kohut, Arukh ha-Shalem(Vienna, 1878), p.292.Google Scholar

52. For , see Even Shmuel, Midreshei Ge'ullahpp. 42 and 51, and cf. his discussion on pp. 34–35, n. 12, 18.

53. Ibid., pp. 74, 79–83.

54. ed. N. Bonwetsch, Abhandlungert der Koniglichen Gesellschaft der Wissenschaften zu Göttingenphil.-hist. Klasse, n.f., vol. 12, no. 3 (Berlin, 1910), pp. 4–5, 66, 70–71, 86, and more.

55. See Bousset, W., The antichrist Legend(London, 1896);Google ScholarBerger, J., Die griechische Daniel Exegese–Eine altkirchliche Apokalypse(Leiden, 1976), pp. 103–150. I see no persuasive evidence that the Christian conception comes from earlier Jewish sources (other than Daniel itself).Google Scholar

56. For brief summaries, see Reeves, M., The Influence of Prophecy in the Later Middle Ages(Oxford, 1969), pp. 299–301, and N. Cohn, The Pursuit of the Millennium2d ed. (New York, 1970), pp. 31–34. Cf. also I. Levi, “L'Apocalypse de Zorobabel et le roi de Perses Siroes,” Revue des Etudes Juives71 (1920): 59–61.Google Scholar

57. Hitzig, F., Das Buck Daniel(Leipzig, 1850), p. 125.Google Scholar

58. K.Kohler in Jewish Encyclopedia1:296–297, s.v. Ahriman; Kohut in Arukh ha- Shalemloc. cit., and esp. in his Über die Jüdische Angelologie und Daemonologie in Ihrer Abhängigkeil vom Parsismus(Leipzig, 1866), p. 62. Kohler emphasized the gimelin the of the Targumim (see n. 51 above).

59. See, e.g., Schürer, E., Geschichte der Jüschen Volkes im ZeitalterJesu Christi(Leipzig, 1907), II, pp. 621–622; Klausner in Ha-Ra'ayon ha-Meshihi, p. 232, and in Enziklopedyah Ivrit.5:954–957; Levi, “Apocalypse de Zorobabel,” p. 59; M. Guttmann in the German Encyclopaedia Judaica3:364–366; Baron, History.5:145: J. Dan, Ha-Sippur ha-'Ivri bi-Mei ha-Beinayim: Iyyunim be-Toledotav(Jerusalem, 1974), p. 42.Google Scholar

60. Cf. Klausner, Ha-Ra'ayon ha-Meshihiloc. cit.

61. Since Romulus Augustulus had at least one competitor for his dubious distinction, and since a seventh-century resident of the Eastern Roman Empire may not have shared the perception that the Western Empire had “fallen,” we should perhaps be cautious about pressing this point too hard.

62. Zeitschrift der Deutschen-Morgenländischen Gesellschaft39 (1895): 343.

63. Sackur, Ernst, Sibyllinische Texte und Forschungen(Halle, 1898), p. 76. The pseudo- Methodian passage was noted by Bousset (Antichrist Legendp. 105), Levi (loc. cit.), and others.Google Scholar

64. Even, Shmuel, Midreshei Ge'ullahp. 195.Google Scholar

65. Ma'arikh,ed. Jellinek, A. (Leipzig, 1853), p. 15.Google Scholar

66. Levi, I., Revue des Études Juives 68 (1914): 136 = Midreshei Ge'ullahp. 387. The text of the passage is slightly corrupt, but however we emend it (see Levi's note on p. 152), it clearly says that Armilus means Levi notes other early scholars who proposed this translation, and cf. also the citation from David de Lara's Keter Kehunnahin Kohut's Arukh ha-Shalemp. 292.Google Scholar

67. B. Sanhedrin105a.

68. “Eine glückliche griechische Nachbildung des biblischen Urtypus der Feindseligkeit gegen Israel” (my translation). See Jahrbuch fur Israelilen5265 [1864/65], ed. J. Wertheimer and L. Kompert (Vienna, 1865), p. 19. The essay has recently been translated into English by I. Schorsch in H. Graetz, The Structure of Jewish History and Other Essays(New York, 1975), pp. 151–171 (notes on p. 310). Cf. also J. Levy, Chaldäisches Wörterbuch über die Targumim und einen grossen Theil des Rabbinischen Schrifttums(Leipzig, 1881), 1:66, s.v. Armilus.

69. Enziklopedyah Ivrit.5:955. All reference to the eremolaosderivation was dropped from the abridged English translation of Klausner's article in the recent Encyclopaedia Judaica.(Why is an article on a Jewish theme that appears in a general encyclopedia abridged when it is transferred to a Jewish encyclopedia?) Cf. also the brief references to this explanation in Schurer and Guttmann, loc. cit. (see n. 59 above).

70. On the frequent midrashic contrast between Balaam and Moses, see the references in Ginzberg, Legends6:125, n. 727.

71. B. Sanhedrin105a and Sotah10a; for Armilus, cf. n. 50 above.

72. Num. 23:29–30. Cf. Even Shmuel's note in Midreshei Ge'ullahp. 82.

73. See the Targum to I Chron. 1:43 and the reference in Ginzberg, Legends5:323, n. 324.

74. In this context, I think that the argument that Romulus was the founder of the cityof Rome, not all of Edom, and that Bela ben Beor's city was Dinhavah (Gen. and I Chron., loc. cit.) would be a quibble. There is an overwhelming likelihood that in the apocalyptic mentality, where Rome and Edom had merged into synonyms, Romulus would have been perceived as the first king–and symbol–of all of Edom. On the fluid midrashic tradition about the founding of the city, which ranged from the time of Esau's grandson Zepho to the time of Solomon, see Ginzberg, Legends5:372, n. 425, and 6:280, n. 11.

75. Midreshei Ge'ullahintrod., p. 51, n. 67. The midrash cited is best known for its appearance in the Passover Haggadah.

76. Targum to I Chron. 1:43. On the variety of relationships between Laban and Balaam posited in rabbinic literature, see Ginzberg, Legends5:303, n. 229, and 6:123, n. 722. See also the references in R. LeDeaut and J. Robert, Targum des Chroniquesvol. 1 (Rome, 1971), p. 42, n. 22.

77. Midrashic literature is not devoid of Greek puns. Is it beyond the realm of possibility that the famous and problematic midrashic interpretation of based in part on an understanding of vnx as both “Aramaean” and “destroyer”?

78. Let me finally propose two suggestions that may be improbable but should nevertheless be noted, (a) Balaam was the son of Beor. The root b'rrefers to an animal, and associations with the story of the she-wolf that suckled Romulus could have arisen despite the fact that b'rusually means a beast of burden, (b) I. Levi in “Apocalypse de Zorobabel” thought that Armilus' birth from a statue was a parody of the alleged virgin birth of Jesus. (Note especially the Christianized Armilus in Even Shmuel, Midreshei Ge'ullahp. 320.) Though I am skeptical, someone attracted by this theory might want to suggest a connection with the possible talmudic association between Balaam and Jesus.

79. If Gerson Cohen's reading of Abraham ibn Daud's Sefer ha-Kabbalahis correct (see his edition [Philadelphia, 1967], esp. pp. 189–222), then it is a case of typological messianism in its most striking form. For another illustration of what remains a significant approach, see Yehudah Liebes, “Yonah ben Amitai ke-Mashiah ben Yosef,” Mehkarim be-Kabbalah Muggashim li-Yesha'yah Tishby( = Mehkerei Yerushalayim be-Matiashevet Yisrael3, pts. 1–2 [1983–84]), pp. 269–311, and cf. n. 85 below.

80. “Parshanuto ha-Tippologit shel ha-Ramban,” Zion45 (1980): 35–59.

81. Ibid., p. 55.

82. The effect on such Jews would, of course, have been more limited, and it may be worth noting that the contrast between the relative messianic activism of Sephardim and the quietism of Ashkenazim in the Middle Ages is in significant measure a contrast between Jews living under Islam and those living under Christianity. In a classroom discussion of Gerson Cohen's “Messianic Postures of Ashkenazim and Sephardim,” in Studies of the Leo Baeck Instituteed. Max Kreutzberger (New York, 1967), pp. 117–156, my former student Avraham Pinsker made the interesting suggestion that Jews in the Christian world, who constantly saw themselves as rejecting the claims of a false Messiah, may have been instinctively more cautious about any involvement with messianic pretenders.

83. See n. 41 above.

84. See David Tamar, “Ha-Zippiyyah be-Italyah li-Shenat ha-Ge'ullah Shin-Lamed- He,” Sefunot2 (1958): 65–68.

85. In the Sabbatian heresy, of course, typology was mobilized once again for the same reasons that it was mobilized in Christianity, the unorthodox career of a messianic personality had to be prefigured by biblical heroes whose own careers would be subjected to subtle, innovative scrutiny.