1. See below, Part IV.

2. The earliest rabbinic tradition is that these fasts were observed on the 13th, 14th, and 15th. This is explicit in Esther rabbah (8:7) and Pirkey de-Rabbi Eliezer (chap. 50), and implicit in the earlier source, Seder olam. Seder olam (chap. 29) records that Haman was hanged on the 16th. This implies that the first of Esther's two meetings with the king was on the 15th and that Esther began her fasting on the 13th. (Some manuscripts of Seder olam include an explicit statement that Esther's first meeting with the king was on the 15th.) Also, Rav (B. Megillah 15a) comments on va-ya‘avor Mordekhai (Esther 4:17): she-he‘ivir yom rishon shel pesaḥ be-ta‘anit (he transgressed the first day of Passover with a fast). Since this statement does not mention the second day of Passover, it seems to view the fasts as having been observed on the 13th, 14th, and 15th. Later rabbinic sources record a tradition that the fasts were observed on the 14th, 15th, and 16th. See, e.g., Midrash panim aḥerim, version 2, chap. 4, and Yalkut shim‘oni, para. 1056. This view is followed by Rashi. See his commentaries to Esther 4:17, and B. Megillah 15a and 16a. See also the statement by R. Sheshet at B. Megillah 16a. B. Yevamot 121b seems to view the fasts as having spanned three full days (days and nights). Compare Midrash tehillim 22:5.

3. וכאשר קימו על נפשם ועל זרעם דברי הצומות וזעקתם.

4. The three consecutive days of fasting observed by Esther and the Jews of Shushan did influence some over the centuries. For example, Sefer minhag tov (thirteenth-century Italy, author unknown) recommended fasting three days before Purim. The author wrote that ideally one should fast the three days consecutively. See Weiss, Meir Ẓevi, “Sefer minhag tov,” Ha-ẓofeh 13 (1929): 236Google Scholar. Also, a custom to fast three days in commemoration of the fasts initiated by Esther was found among Marrano women. The tradition was to fast the three days one month before Passover. See, e.g., Horowitz, Elliott, Reckless Rites (Princeton, NJ: Princeton University Press, 2006), 52–56Google Scholar.

R. Abraham Ibn Ezra, commentary to Esther 9:30, implies that the Karaites interpreted verse 9:31 to indicate an acceptance by the Jews of a fast of three days on future generations. But I have not found such a view in any Karaite source. Anan, the eighth-century founder of the sect, believed that the Jews fasted seventy days in the year of the threat, from the 13th of Nissan until the 23rd of Sivan. He also believed that verse 9:31 indicated that this period of seventy days of fasting, and fasting on both days of Purim, were accepted by the Jews as an annual obligation. See Harkavy, Avraham, ed., Zikhron la-rishonim: maḥberet sheminit … likkutey kadmoniyyot le-korot dat beney mikra ve-sifrutam (St. Petersburg, 1903), 40, 130, 133, 149, 157, and 164Google Scholar.

5. This is how R. Abraham Ibn Ezra interprets the phrase. See his commentaries on Esther 9:30 and Zechariah 8:18. But in this interpretation, the verse is suddenly switching contexts and referring to fasts not mentioned elsewhere in the book. The other problem with this interpretation is that the phrase refers to fasts associated with za‘akah (outcry). This fits the fasts of Esther 4:3 and 4:16–17. But it does not easily fit the fasts established in connection with the destruction of the First Temple, which were fasts established for mourning.

6. Admittedly, zar‘am usually means descendants. Indeed, just a few verses earlier, in verses 27 and 28, it does take on this meaning. But one can respond that, having used the word zar‘am in verses 27 and 28, the author chose to use the word again in verse 31, even though it was now not being used in its usual sense. Ve-za‘akatam (and their outcry) supports the interpretation that zar‘am means children here. Za‘akah does not sound like something accepted by future generations.

7. The fact that the letters were sent to the Jews of all the provinces suggests that the reference is to the fasts of 4:3. But the plural ha-ẓomot perhaps suggests that the reference is to the fast of three days, and ve-ka'asher kiyyemu implies that the reference is to fasts that were observed pursuant to an instruction, which also suggests these fasts.

8. Nevertheless, there are rabbinic sources that interpret verse 9:31 to imply the Jews’ acceptance of an obligation to fast annually on the 13th of Adar. R. Abraham Ibn Ezra tells us that many in his time interpreted verse 9:31 as referring to Ta‘anit Esther. Later sources that adopt this interpretation include (1) R. Abraham b. David of Posquieres, cited by the Ritba, in his commentary to Ta‘anit 10a, and by the Ran in his commentary to Rif, Ta‘anit 7a (in the pages of Rif); (2) Kol Bo, sec. 62 (but compare sec. 45); and (3) Meiri, Magen ’avot, treatise 23 (but compare his commentary to Megillah 2a). See also R. Isaac b. Abba Mari, Sefer ha-ittur, 220. It can be argued that R. Se‘adyah adopts this interpretation as well. See below, n. 84.

Presumably this interpretation understands the verse to be teaching that the Jews accepted the ẓomot of Nissan by agreeing to adopt a yearly ẓom on the 13th of Adar. This day might naturally have been chosen as the day to observe an annual fast because it was the day on which the massacre of the Jews had originally been scheduled to take place. Also, it was the day the actual fighting took place. But this is all speculation, as there is nothing in verse 9:31 to connect it to the 13th of Adar. Moreover, there is nothing in I Maccabees or II Maccabees to suggest that the holiday declared in 161 bce was declared on a day that had previously been designated a day of fasting.

The Talmud mentions verse 9:31 briefly at B. Megillah 16b. But the passage does not seem to have any bearing on our issue.

9. Maccabees, I, 7:47–49 (Goldstein, Jonathan A., trans., The Anchor Bible. I Maccabees [Garden City, NY: Doubleday, 1976], 328)Google Scholar.

10. Maccabees, II, 15:35–36 (Goldstein, Jonathan A., trans., The Anchor Bible. II Maccabees [Garden City, NY: Doubleday, 1983], 476)Google Scholar.

11. Josephus, Jewish Antiquities, XII, para. 412 (Loeb edition).

12. There may not even have been a prohibition to fast on days such as this until Megillat ta‘anit, the collection of these holidays, was authored and accepted.

13. Megillat ta‘anit is referred to in M. Ta‘anit, 2:8. According to the Talmud (B. Shabbat 13b), it was compiled by Ḥananiah b. Ḥezekiah b. Garon (first century). It is a brief list of holidays, in Aramaic. There also exists a scholium to Megillat ta‘anit. The scholium was compiled in a later period. Traditions about Yom Nikanor are found in the scholium, and in B. Ta‘anit 18b, Y. Ta‘anit 2:13 (66a), and Y. Megillah 1:6 (70c).

14. It was mentioned earlier that some authorities interpreted Esther 9:30–31 to imply the Jews’ acceptance of an obligation to fast annually on the 13th of Adar. It is not known how most of those adopting this interpretation explain the listing of the 13th of Adar in Megillat ta‘anit as a day upon which eulogizing, and by implication fasting as well, are prohibited. But in the case of R. Abraham b. David of Posquieres, there is a record of his explanation. According to the Ran (Ta‘anit 7a, in the pages of Rif), R. Abraham b. David took the position that Megillat ta‘anit never meant to prohibit fasting on the 13th of Adar. It only meant to prohibit eulogizing.

15. Or Nikanor. There are different readings.

16. References to “R. Naḥman” in the Babylonian Talmud are generally assumed to be references to R. Naḥman b. Jacob. His year of death is estimated at 320 ce (Encyclopaedia Judaica 12:773). (All references to the Encyclopaedia Judaica in this article are to the first edition, unless otherwise noted.) The story, at B. Ta‘anit 18b, records that R. Naḥman declared a fast on the 12th of Adar. The rabbis of his time objected because it was Trajan Day. R. Naḥman responded that Trajan Day was no longer in effect due to a certain tragedy that had occurred on it. The story clearly implies that the rabbis of his time held that Trajan Day was still in effect, or would otherwise have still been in effect, if not for the tragedy.

17. Although, as we learn from B. Ta‘anit 18b, some of the holidays may have been discontinued earlier. Both the Babylonian and Jerusalem Talmuds include passages recording the views of Amoraim on the issue of whether the holidays of Megillat ta‘anit were still in effect. See B. Rosh Ha-shanah 18b-19b, Y. Ta‘anit 2:13 (66a), Y. Megillah 1:6 (70d), and Y. Nedarim 8:1 (40d). All but one of the Amoraim whose views are recorded are Palestinian. Almost all of these take the position that the holidays were no longer in effect. The only Babylonian Amora whose view is recorded in these sources is Rav (or in another version: Rav Huna). The view recorded in the name of this Sage is also that the holidays were no longer in effect. Those who argue that the holidays of Megillat ta‘anit survived in Babylonia well into the amoraic period point to the fact that only one Babylonia Amora is recorded in the above sources as taking the position that the holidays were no longer in effect. See, e.g., Tabory, Yosef, Mo‘adey yisra'eil bi-tekufat ha-mishnah ve-ha-talmud (Jerusalem: Magnes Press, 1995), 320–21Google Scholar. Also, B. Rosh Ha-shanah 19b records two different traditions about whether the holidays were still in effect: ve-hilkhita batlu and ve-hilkhita lo batlu (or in the version of the Rif, Rosh, Ritba, and others: ve-hilkhita batlah and ve-hilkhita lo batlah). Perhaps there was no contradiction that needed to be resolved, just two different viewpoints about whether Megillat ta‘anit was still in effect. Moreover, according to many scholars, the term ve-hilkhita often reflects a statement made in the late amoraic, or savoraic, period. Thus, it can be argued that even in the late amoraic period and perhaps beyond, there were those in Babylonia who considered Megillat ta‘anit to still be in effect. The resolution of the “contradiction” (one statement applies to Ḥanukkah and Purim, and the other to the balance of the holidays) perhaps came in an even later period. See Halivni, David, Mekorot u-mesorot: bi'urim ba-talmud le-seder mo‘ed (Jerusalem: Beit ha-midrash la-rabbanim be-America, 1975)Google Scholar, 386 (commentary to B. Rosh Ha-shanah 18b); and Noam, Vered, Megillat ta‘anit (Jerusalem: Yad Yiẓḥak Ben-Ẓvi, 2003), 356Google Scholar, and more generally, 355–61.

18. See Brody, Teshuvot Rav Natronai, 303–304, responsum 177. In this work, Brody gathers from a variety of sources the responsa that he can attribute to R. Natronai b. Hilai. This particular responsum had been printed in the geonic collection Ḥemdah genuzah in a section of twenty-four responsa attributed to R. Natronai and R. Paltoi. For a variety of reasons, Brody concludes that all twenty-four were written by R. Natronai. That our particular responsum was composed by R. Natronai is confirmed by another work, Seder parshiyyot shel yamim tovim ve-haftarot shelahen. The expanded version of this work includes a shortened version of our responsum, with an attribution of the responsum to R. Natronai. (The original version of Seder parshiyyot is shorter and does not include any version of our responsum.) Seder parshiyyot will be discussed further in Part VI. The responsum (not the shortened version) is also found in Sefer ha-pardes, 261, and Likkutey pardes, laws of the 9th of Av. It is not attributed to any particular Gaon in these sources. It is also included in Oẓar ha-ge'onim, Megillah, 58 (sec. 216).

19. I have followed the dates for R. Natronai given by Brody in The Geonim of Babylonia, 344.

20. The Shabbat shel regel is most likely a reference to a special Shabbat in honor of the Exilarch that was observed annually on Shabbat parshat lekh lekha. Brody, Teshuvot Rav Natronai, 303 n. 5. The fast after this Shabbat was most likely instituted to atone for any sins that may have been inadvertently committed on the occasion.

21. The responsum provides one exception to this rule. On the 9th of Av, the custom was to read ki tolid banim at shaḥarit, and ve-yeḥal Moshe at minḥah.

22. The responsum does not state explicitly that the fast was on the 13th. When this responsum was printed in Ḥemdah genuzah, the text printed was בתעניות פורים. This could be understood as a reference to the Palestinian custom of three days of fasting in Adar, perhaps as a consequence of the emigration of Palestinian Jews to Babylonia. But most of the responsa printed in Ḥemdah genuzah are also found in manuscript Cambridge 474, and this manuscript reads בתענית פורים. For a description of manuscript Cambridge 474, see Solomon Schechter, “Notes on Hebrew Mss. in the University Library at Cambridge,” Jewish Quarterly Review O.S. 4 (1892): 89–101. Both Sefer ha-pardes and Likkutey pardes also read בתענית פורים. Brody has little doubt that this is the correct reading and that here we have an early reference to the practice of fasting on the 13th. The fact that the shortened version of the responsum reads either תענית פורים or תענית אסתר (see below, Part VI) confirms this reading as well.

23. R. Sar Shalom (R. Natronai's predecessor at Sura) wrote a very similar responsum, except that it contains no reference to the fast days after the Shabbat shel regel and no reference to any fast day in connection with Purim. (Perhaps this is significant. See below, n. 133.) This responsum of R. Sar Shalom is quoted in many places, the original version and an abbreviated version. See, e.g., Seder Rav ‘Amram Ga'on, ed. Goldschmidt, Daniel (Jerusalem: Mossad Harav Kook, 1971), 95–96Google Scholar, and Tur OH 566. For the additional references, see Weinberg, Refael Shmuel, Teshuvot Rav Sar Shalom Ga'on (Jerusalem: Mossad Harav Kook, 1975), 87–88Google Scholar.

24. Although R. Se‘adyah lived in Egypt for approximately the first thirty years of his life, there is convincing evidence that his siddur was composed in Babylonia. See Siddur Rav Se‘adyah Ga'on, eds. Davidson, Israel, Assaf, Simḥah, and Joel, Yissakhar (Jerusalem: Rubin Mass, 1941), intro., 22–23.Google Scholar (I will abbreviate this work hereinafter as SRSG.)

25. SRSG, 319–38. See also SRSG, 258.

26. SRSG, 319. Also, the seliḥot themselves refer to the practice of fasting. The second seliḥah for the 13th of Adar begins: tela'ot be-yom ẓomi ’asaprah (on this fast day, I will tell the story of things that made me weary), SRSG, 337. The third seliḥah also refers to the practice of fasting. See below, n. 84.

Not all of the seliḥot which R. Se‘adyah included in his Siddur were composed by him. Menaḥem Zulay believed that the three included for the 13th of Adar were composed by R. Se‘adyah. See his Ha-askolah ha-payyetanit shel Rav Se‘adyah Ga'on (Jerusalem: Mekhon Shoken, 1964), 45Google Scholar. More recently, Yosef Tobi took the position that this is probable. See his Piyyutey Rav Se‘adyah Ga'on (Jerusalem, 1982), 129 and 132. There is a geonic responsum that cites lines from the first of these three seliḥot and attributes the seliḥah to R. Se‘adyah. See Mann, Jacob, “Addenda to ‘The Responsa of the Babylonian Geonim as a Source of Jewish History,’” Jewish Quarterly Review 11 (1920–21): 465 n. 32Google Scholar. But there is no source that connects R. Se‘adyah with the second or third of the three seliḥot, even though it is reasonable to presume that he composed them.

Tobi points out that it is now known from Genizah material that R. Se‘adyah composed more seliḥot for the 13th of Adar than are included in his siddur. One such seliḥah was printed in SRSG, 423 (appendix).

27. Ginzberg, Louis, Geonica (New York: Jewish Theological Seminary, 1909), vol. 2, 67–68Google Scholar.

28. Ginzberg suggests that this word should be read shel or yom.

29. Wertheimer, Shelomoh, Sefer kohelet Shelomoh (Jerusalem: A. M. Luntz, 1899), 14Google Scholar. In this responsum, the question mentions the custom of fasting on the 13th of Adar, while the answer does not. From reading the answer, there is no reason to doubt that R. Hai accepted the premise of the question that the 13th of Adar was a fast day.

30. Ginzberg, Louis, ed., Ginzey Schechter, vol. 2 (New York: Beit ha-midrash ha-rabbanim be-America, 1929), 136Google Scholar. See also the anonymous responsum, no. 15, in Harkavy, Avraham, ed., Zikhron la-rishonim … : maḥberet revi‘it: zikhron kammah ge'onim … (Berlin: Ẓ.H. Itẓkovsky, 1887), 7Google Scholar. This responsum seems to include the same statement, but the statement is cut off mid-sentence. (This may be the same responsum, included in a different collection of geonic responsa.)

31. A few other early references to the fast of the 13th of Adar also deserve mention, even though they are not from geonic Babylonia:

Al-Biruni, a Muslim scholar of Persian origin, writing in 1000 ce, refers to the Jews’ observance of a fast on the 13th of Adar. He refers to the fast briefly, calling it “the fasting of Alburi” (Purim). See Al-Biruni, , The Chronology of Ancient Nations, trans. Sachau, C. Edward (London: W. H. Allen, 1879), 273Google Scholar. For more on Al-Biruni and his references to Jewish holidays and fast days, see Ratzaby, Yehudah, “Megillat ‘ta‘anit ẓaddikim’ be-makor ‘aravi,” Sinai 106 (1990): 1–11Google Scholar, and Elizur, Shulamit, Lammah ẓamnu: megillat ta‘anit batra u-reshimot ẓomot ha-kerovot lah (Jerusalem: Ha-’iggud ha-‘olami le-mada‘ey ha-yahadut, 2007), 105–14Google Scholar.

Yanon b. Ẓemaḥ, a paytan from Syria in the first half of the eleventh century, refers to the fast. See below, n. 130.

One of the three texts of Megillat ta‘anit batra found in the genizah includes the 13th of Adar, recording: י”ג בו תענית אסתר. See Elizur, Lammah ẓamnu, 159 n. 42. This text most likely dates to the eleventh century. (Dr. Edna Engel, The Hebrew Palaeography Project of the Israel Academy of Sciences and Humanities, private correspondence, January 1, 2009.)

A passage that refers to the practice of fasting on the 13th of Adar is found in one of the three manuscripts of the Seder Rav ‘Amram. See Seder Rav ‘Amram Ga'on (ed. Goldschmidt), 101. But Goldschmidt believes that this passage is an addition to the Seder from the time of the Rishonim. It is accepted that many additions and changes were made to the Seder in the centuries after R. ‘Amram's death (c. 875). The manuscript that included the reference to the fast was copied in 1426. (The other two manuscripts of the Seder are also late.) Goldschmidt observed that this manuscript included many unique additions, and that it reflected the traditions of Northern France. See Seder Rav ‘Amram Ga'on, 17–18.

32. The responsum is found at Midrash Tanḥuma, Bere'shit, sec. 3. (When I refer to Midrash Tanḥuma, I am referring to the standard printed Midrash Tanḥuma. The responsum is not found in the edition of Midrash Tanḥuma first published in the late nineteenth century by Solomon Buber.)

Midrash Tanḥuma also includes passages taken from the She'iltot of R. Aḥai Gaon. See Midrash Tanḥuma, Bere'shit, sec. 2, Noaḥ, sec. 4, Ḥukkat, sec. 2; and Brody, Le-toledot, 118. Based on this, scholars believe that Midrash Tanḥuma underwent final redaction in geonic Babylonia (or in an area under the sphere of influence of the Babylonian geonim). See Encyclopaedia Judaica 19:503 (2nd ed.); and Bregman, Mark, Sifrut tanḥuma-yelammedenu (Piscataway, NJ: Gorgias Press, 2003), 184–86Google Scholar. See further, below, n. 100.

33. There are many differences in the wording of the two sources, and it does not seem that the later of the two sources is copying from the earlier. She'ilta 79, in giving its source for the interpretation of yom ha-kenisah, does use the phrase peirshu ḥakhamim (the Sages explained). But it is unlikely that she'ilta 79 is referring directly to the geonic responsum. More likely, the reference is to the Sages of geonic times.

34. I will refer to the entire work as the She'iltot. I will refer to individual units by the lower case she'ilta (pl. she'iltot).

The earliest source that connects R. Aḥai with a collection of she'iltot is a responsum by either R. Sherira or R. Hai. See Brody, The Geonim of Babylonia, 207. Probably, many of the discussions in the She'iltot were written by others, and only collected or adapted by R. Aḥai. (But some of the textual witnesses to the She'iltot contain material that is later than R. Ahai.) According to the letter of R. Sherira, R. Aḥai left Babylonia for Palestine around 750 ce, when his disciple was promoted to the position of Gaon of Pumbedita instead of him. This raises the possibility that the work was composed or completed in Palestine. But there are strong reasons for believing that the work represents a collection of material from Babylonia (even allowing for the remote possibility that it was completed by R. Aḥai in Palestine). A few of these reasons are: (1) a she'ilta is a homily of a particular form that was prevalent in Babylonia and differed from Palestinian-type homilies; (2) the She'iltot is organized according to the Babylonian Torah portions (i.e., it follows the yearly cycle of portions and not the triennial cycle used in Palestine); (3) the various she'iltot in the work use phrases like lan, ’anan, and hakha when referring to Babylonia, and phrases like hatam and lehu when referring to Palestine. Brody also argues that the She'iltot, in its original form, never cites the Jerusalem Talmud or Palestinian midrashic collections. Meiri, in his introduction to M. Avot, tells us that R. Aḥai died in the year 4512. If this tradition is accurate, this would be 752 ce, only a short time after R. Aḥai arrived in Palestine. For further background to the She'iltot, see Brody, The Geonim of Babylonia, 202–215.

35. This will be explained further below.

36. B. Megillah 2a.

R. Samuel b. Isaac does not cite any verse, but he is probably alluding to a verse or verses. The verses to which he could be alluding would be 9:1–2 (which include the word nikhalu in verse 9:2), 9:16–17 (which include the word nikhalu in verse 9:16), and 8:11–12 (which include the word le-hikkaheil in verse 8:11).

The geonic responsum, when providing its interpretation of yom ha-kenisah, only includes one citation to verses. The verses it cites, in all the editions I have seen, are 9:1–2. She'ilta 79, when providing its interpretation, cites verses twice. In Appendix I, I have printed the text of the 1546 Venice edition of she'ilta 79. The citation there is to 9:16–17 in both instances. But when one looks at manuscripts, and sees how she'ilta 79 is cited in the Rishonim, the picture is different. Samuel Kalman Mirsky prepared an edition of the She'iltot based on nine relatively complete European manuscripts. Of these, eight include she'ilta 79. Four of these mention 9:2 in both citations. Of the four manuscripts that cite to 9:16–17 in the first citation, all of these, in the second citation, start with nikhalu ha-yehudim, a phrase from 9:2, even though their citation continues with 9:17. See Mirsky, , She'iltot de-Rav Aḥai Ga'on (Jerusalem: Ha-makhon le-meḥkar u-le-hoẓa'at sefarim Sura, 1960–77)Google Scholar, vol. 3, 225, notes to line 54, and see below, Appendix I, n. 147. Surely, she'ilta 79 did not cite a combination of two verses in its second citation, and some textual corruption has occurred. With regard to the way she'ilta 79 is cited in the Rishonim, most of the time its interpretation is summarized briefly without citation to any verse. But the few times a verse is cited in the sources that I have seen, the citation is to 9:2 (or to 9:1–2). See, e.g., R. Eliezer b. Joel ha-Levi, Sefer Ravyah, secs. 550 and 559; and R. Isaiah Di Trani, Tosafot Rid, commentary to Megillah 2a. But see the Genizah fragment of the latter part of she'ilta 79, referred to at Appendix I, n. 147.

37. As to what happened after the year of the threat, she'ilta 79 refers to yom ha-kenisah as something in effect bi-zeman she-beit ha-mikdash kayyam (during the period that the Temple was standing). The geonic responsum includes no such statement. But in describing what occurred in the year of the threat, the geonic responsum includes the phrase ve-gazru ta‘anit, and in describing the fast day, the geonic responsum calls it a de-rabbanan. Taken together, these passages imply that the view of the geonic responsum was that in the year of the threat a decree was enacted that bound future generations.

Neither the geonic responsum nor she'ilta 79 refers to the prohibition to fast on Yom Nikanor.

According to rabbinic chronology (e.g., Seder olam, chap. 28 and B. Megillah 11b), Aḥashverosh was the king immediately before the Darius in whose reign the Second Temple was built. In fact, Aḥashverosh is to be identified with the king the Greeks called Xerxes (whose Persian name was Khshayarsha, and whose name was recorded in Aramaic documents as חשירש,חשיארש and אחשירש‎). See First, Mitchell, Jewish History in Conflict (Northvale, NJ: Jason Aronson, 1997)Google Scholar, 166 n. 17; and Porten, Bezalel and Lund, Jerome A., Aramaic Documents from Egypt: A Key-Word-in-Context Concordance (Winona Lake, IN: Eisenbrauns, 2002), 356Google Scholar. Xerxes reigned from 486 to 465 bce, immediately after Darius I, in whose reign the Second Temple was built.

According to Herodotus (middle of the fifth century bce), the name of Xerxes’ wife was Amestris. The “is” at the end of this name is just a Greek addition (just like the “es” at the end of “Xerxes”). We have no Persian sources for her name, but can deduce from the Greek name Amestris that her name in Persian was likely based upon the consonants M, S, T, and R. Almost certainly, Esther (consonants S, T, and R), and Amestris are one and the same, despite one or two discrepancies that exist between the little information about Amestris furnished by Herodotus and certain details found in the book of Esther.

38. The text of these sections is fairly well established. I have printed a standard text. The translation is my own.

39. See the discussion below.

40. Villagers would typically enter and gather in the cities on Mondays and Thursdays for various ritual and judicial practices. See Lapin, Hayim, “Some Observations on Mishnah Megillah 1:1–3,” in Netiy‘ot le-David: sefer ha-yovel le-David Halivni, eds. Elman, Yaakov et al. (Jerusalem: Orḥot, 2005), 92–111Google Scholar.

Another way of reading the Mishnah is that the reading on the advanced dates is taking place in the villages. In this reading, the term yom ha-kenisah is used because villagers would typically gather in their synagogues (battei kneisiyyot) or in some other manner, on Monday and Thursday. See, e.g., Tosafot, Yevamot 14a, s.v. ki, and Maimonides, Mishneh torah, hilkhot megillah 1:6.

That the villagers are traveling to the city for the advanced reading is the simpler reading of the Mishnah, and is implied from the various statements at Y. Megillah 1:1 (70b); and B. Megillah 2a, 4a–b, and 19a about the villagers providing food, or water and food, to the cities.

41. Neither the geonic responsum nor she'ilta 79 states whether this advanced reading is taking place in the cities or in the villages. See Appendix III.

42. Both the geonic responsum included in Midrash Tanḥuma and she'ilta 79 stated that the fast is observed on Thursday when the 14th falls on Sunday. In the period of the Rishonim, most communities followed this practice. But the custom in Provence was to fast on Friday in such a case. See, e.g., Aaron b. Jacob ha-Kohen, Orḥot ḥayyim, hilkhot megillah u-Purim, sec. 26; and Meiri, Magen ’avot, treatise 23.

43. If the 14th fell on Tuesday or Wednesday, the fast would be observed on the 13th. These cases are not included in the responsum, but follow from the interpretation.

If the 14th fell on Monday, the responsum instructs that the reading could not be advanced. It is forced into this position by the text of the Mishnah. The text of the Mishnah also forces the position that the reading cannot be advanced if the 14th fell on Thursday, even though this case is not included in the responsum.

44. This source has yom, while B. Megillah 2a and she'ilta 79 have zeman.

45. There are rishonim who cite teshuvot ha-Ge'onim (or teshuvat ha-Ge'onim) for the ruling that when the 14th of Adar falls on Sunday, the practice is to fast on Thursday, because the recital of seliḥot and taḥanunim on Erev Shabbat would interfere with the preparations for Shabbat. See, e.g., R. Isaac of Vienna, Or zaru‘a, hilkhot megillah, sec. 367; Rosh, commentary to Megillah 2a; and Tosafot Rosh, commentary to Megillah 2a; and Ran, his own commentary to Megillah 2a, and his commentary to Rif, Megillah 2a (p. 1b in the pages of Rif).

I have not seen any such geonic responsum, and it is possible that the source is the above responsum in Midrash Tanḥuma (even though the responsum uses the word raḥamim, and not taḥanunim). If this is the case, it is surprising that the responsum is not referred to as being from Midrash Tanḥuma. Perhaps after the earliest source (R. Isaac of Vienna, or some earlier source) referred to the responsum loosely as coming from teshuvot ha-G e'onim, the later sources merely repeated this reference.

46. A similar expression is used by Pirkoy Ben Baboy (early ninth century). This will be discussed in Part VI.

47. In the next few lines, the responsum is quoting selectively from the Mishnah and paraphrasing.

48. The quotation of the responsum ends here.

49. I did not translate the phrase הואיל ומסתמא בקיאין הן, as it seems to be an erroneous reading. (Eẓ Yosef takes this position as well.) The phrase makes little sense coming from a Tanna. The statement of R. Judah is found at both B. Megillah 2a and T. Megillah 1:3. In these sources, the words immediately following bi-zeman hazeh are either הואיל ומסתכלין בה (because they anticipate it) or הואיל ומסתכנין בה (because they endanger themselves through it). See Saul Lieberman, Tosefta ki-feshutah, T. Megillah 1:3.

50. Midrash Tanḥuma, Bere'shit, sec. 3. The previous section in Midrash Tanḥuma (sec. 2) had discussed the issue of what time one who is fasting on Erev Shabbat is required to end his fast. (Sec. 2 was an abridged version of the first she'ilta in Bere'shit.) The redactors of Midrash Tanḥuma apparently decided to include our responsum here because it discussed the issue of fasting on the 13th of Adar when it fell on Erev Shabbat.

I have printed the text of the Jerusalem 1958 edition. This version of Midrash Tanḥuma was first published in 1520–1522. The third edition, published in 1563, served as the basis for all later editions. I have looked at a reprint of the 1563 edition. It is practically the same as that which I printed. The translation is my own.

51. When the 14th of Adar falls on Shabbat, the upcoming Yom Kippur would fall on Friday. Already in the time of R. Yose b. Bun (c. 300), the 14th of Adar was not being allowed to fall on Shabbat or Monday, so that Yom Kippur would not fall on Friday or Sunday. See Y. Megillah 1:2 (70b), Encyclopaedia Judaica 5:49, and Tabory, Mo‘adey yisra'eil, 28. See also B. Rosh Ha-shanah 20a. She'ilta 79 stated explicitly that the 14th of Adar no longer fell on Shabbat in its time.

52. I am following Mirsky's numbering, which is followed by Brody. These numbers are artificial, as the she'iltot can be broken up in different ways. In Mirsky's edition, she'iltot 76–79 are located where the she'iltot for va-yakhel would be, and are called she'iltot for Purim. In some of the other printed editions, these she'iltot are called she'iltot for va-yakhel (and have a different numbering).

53. Brody, Le-toledot, 186 n. 5, and The Geonim of Babylonia, 209 n. 29. Structurally, they are deficient as she'iltot. (She'iltot 76 through 78 are very short and undeveloped. With regard to she'ilta 79, see the text above.) Also, there is some variation in the manuscripts with regard to their location in the work. This suggests that they were later additions, attempted to be integrated into an already fixed work.

54. Brody, Le-toledot, 186 n. 5.

55. Even though the interpretation itself may have been older. Brody, in private correspondence, agrees that the geonic responsum was probably composed before she'ilta 79.

56. Brody, private correspondence.

57. See below, n. 100.

58. Also, nikhalu ve-‘amod ‘al nafsham at 9:16 is parallel to le-hikkahel ve-la‘amod ‘al nafsham at 8:11, and the context at 8:11 is even clearer that le-hikkahel refers to a physical gathering to fight their enemies.

59. The full statement is י”ג זמן קהילה לכל היא ולא צריך לרבויי. B. Megillah 2a. In manuscripts of the Talmud, there are variants here with regard to the last three words. I have checked the manuscripts collected by the Saul Lieberman Institute. Several have קרא instead of לרבויי. One has neither, ending with ולא צריך. Another does not have the last three words at all.

60. There is no doubt that this is the plain sense interpretation of his statement. But as further support, a statement by R. Ḥelbo at Y. Megillah 1:1 (69d) calls the 13th of Adar “yom milḥamah” in its parallel (but not identical) discussion.

Even though the rishonim (other than those following the interpretation included in Midrash Tanḥuma and she'ilta 79) agree that R. Samuel b. Isaac is referring to the 13th as a day of fighting, there is still a disagreement as to how to understand his statement. The main approaches are that of Rashi, and the Ri (R. Isaac of Dampierre; the view of the Ri is described in the Rosh and in others.) According to Rashi, R. Samuel b. Isaac is expressing the view that because the 13th of Adar was the day that all the Jews (not just the Jews in Shushan) gathered to fight, this was the day of the main miracle. Therefore, the Men of the Great Assembly did not need to find an allusion in a verse to enable them to permit the reading on this day. As to the Ri, he notes that Y. Megillah 1:1 (69d) states that if the reading is permitted on the 12th and 14th, it is illogical to hold that it cannot be read on the 13th. Therefore, one does not need to derive the permission to read it from a verse. Ri argues that R. Samuel b. Isaac is taking the same position. (But Ri's explanation does not fit the language of the statement of R. Samuel b. Isaac.) For an alternative approach to these passages, see Sperber, Daniel, Minhagey yisra'eil, vol. 4 (Jerusalem: Mossad Harav Kook, 1995), 250–52Google Scholar.

Note that the statement of R. Samuel b. Isaac is recorded twice at B. Megillah 2a (and in many of the manuscripts). First, it is recorded in a section that begins with a statement by R. Shemen b. Abba. A few lines later, it appears in a section that begins with a statement by R. Samuel b. Naḥmani. When R. Samuel b. Isaac's statement is recorded in the first section, it is written in a way that clearly suggests that it was first made in a different context (… הכא נמי ‎… כדאמר רב שמואל בר יצחק). If the statement was originally found in another context, we would need to see this original context to truly understand the statement. But the statement is not found anywhere else, and when the statement appears in the section that begins with R. Samuel b. Naḥmani, it does not occur with a ke-de-’amar or hakha nami. It has been suggested that the other context referred to was simply the section that begins with R. Samuel b. Naḥmani.

61. The severe difficulties with interpreting yom ha-kenisah as the 13th of Adar are noted by many authorities. See, e.g., R. Eliezer b. Joel ha-Levi, Sefer Ravyah, sec. 559; R. Isaiah Di Trani, Tosafot Rid, commentary to Megillah 2a; R. Naphtali Ẓevi Judah Berlin (Ha-‘emek she'elah), commentary to the above passage in the She'iltot; and R. Ḥanokh Zundel b. Joseph, Eẓ Yosef, Midrash Tanḥuma, Bere'shit, sec. 3. Nevertheless, the unusual interpretation of yom ha-kenisah is followed by some authorities. One who seems to follow it is the author of Sefer ha-mikẓo‘ot (see above, Part VI), quoted by R. Isaac of Vienna. Also, R. Isaac himself, after quoting it in the name of the Sefer ha-mikẓo‘ot and the She'iltot, does not make any attempt to disagree. It is also perhaps followed by R. Tam (see Appendix III).

Perhaps the authors of the unusual interpretation of yom ha-kenisah had a text of the Talmud that did not include the above anonymous discussion at Megillah 4a–b. Perhaps they also did not have M. Megillah 1:3 as we have it. Note that the geonic responsum included in Midrash Tanḥuma records the text of the question asked, and it can be argued that the questioner did not have R. Judah's statement at M. Megillah 1:3.

Many authorities followed the unusual interpretation of zeman kehillah la-kol hiy. Some of the earliest to do so are: (1) Sefer ha-mikẓo‘ot; (2) R. Tam, quoted by the Rosh and many others (see Appendix III); (3) R. Isaac b. Abba Mari, Sefer ha-ittur, 220; and (4) R. Abraham b. Nathan ha-Yarḥi, Ha-manhig, 1:244–45 (ed. Raphael). (Regarding the last source, the emendations suggested at 245, notes 95 and 96 are undoubtedly correct.) Interestingly, there exists a manuscript of Megillah 2a (NY-Columbia X 893 T141) in which R. Aḥai and a summary of the interpretation attributed to him are included on the talmudic page. The statement included is:

62. Margaliot, Mordekhai, “Te‘udah ḥadashah ‘al ẓom ha-ra‘ash,” Tarbiẓ 29 (1960): 339–40Google Scholar. Margaliot only published part of the fragment. The lines that concern us were published by Ezra Fleischer in his 2001 article cited below. The fragment dates from the tenth or eleventh century.

63. Fleischer, , “Hadutah-haduthu-ḥaduta—pulmus ve-shivro,” Tarbiẓ 53 (1984): 94–95Google Scholar. The article does not estimate a date for the fragment.

64. Fleischer, , “Seridim nosafim mi-kovẓei tefillah Ereẓ Yisre'eliyyim min ha-genizah,” Kibbuẓ ‘Al Yad 15 (2001): 9–11 and 33Google Scholar. This fragment dates from the eleventh century.

65. See, e.g., Encyclopaedia Judaica 15:81, based on the work of Michael Higger. See Higger, , ed., Massekhet soferim (New York; Devei Rabanan, 1937)Google Scholar.

66. See Blank, Debra Reed, “It's Time to Take Another Look at “Our Little Sister” Soferim: A Bibliographical Essay,” Jewish Quarterly Review 90 (1999): 4 n. 10Google Scholar; and Lerner, M. B., “The External Tractates,” in The Literature of the Sages, ed. Safrai, Shmuel (Assen, The Netherlands, and Philadelphia: Van Gorcum and Fortress Press, 1987), 399–400Google Scholar. Massekhet soferim makes prolific usage of the Palestinian Talmud, but follows the Diaspora custom of a two-day festival (14:16, ed. Higger). Higger had viewed references in this work to customs of the Diaspora as later glosses by a Babylonian hand.

67. The Babylonian Geonim never refer to the material in chapters 10–21. Material from these chapters is not quoted anywhere until the eleventh century, in Ashkenaz. See Blank, 4 n. 10, and 18, and Higger, 36 and 58.

68. Brody, The Geonim of Babylonia, 112; and Fleischer, , Tefillah u-minhegey tefillah Ereẓ Yisre'eliyyim bi-tekufat ha-genizah (Jerusalem: Magnes Press, 1988), 199–202Google Scholar.

69. The meaning of the phrase Nikanor ve-ḥaveirav is subject to debate. Encyclopaedia Judaica 6:914 interprets it to mean Nicanor and his men. But Hilvitz, Alter, Ḥikrey zemanim, vol. 1 (Jerusalem: Mossad Harav Kook, 1976), 373Google Scholar; and Tabory, Mo‘adey yisra'eil, 322, suggest that ve-ḥaveirav means the holiday which Megillat ta‘anit lists for the 12th of Adar. Hilvitz goes further and suggests that the true reading may be ve-ḥaveiro. (Higger's critical edition does not mention any such variant.) My friend Joshua Zakheim suggests that the meaning is “Nicanor, which is one of those Megillat ta‘anit holidays.”

The reference to Yom Nikanor here has engendered much discussion. One way to understand this passage is that it is implying that Yom Nikanor was still observed in Palestine at the time this passage was written. But the fact that almost all of the Palestinian Amoraim whose views are recorded express the view that Megillat ta‘anit was no longer in effect in their times (see above, n. 17, and especially the passages in the Jerusalem Talmud) suggests that this interpretation is not correct. In light of recent scholarship that suggests a late date for chapters 10–21 of Massekhet soferim and authorship of these chapters outside of Palestine, it is probably a mistake to read much into the Yom Nikanor reference here. The author may have just been conjecturing about why the three days of fasting in Palestine were observed after Purim. The author goes on to provide an alternative explanation for the practice.

There are sources that describe fast days that were observed in Palestine in the post-talmudic period. See, e.g., Leiman, Sid Z., “The Scroll of Fasts: The Ninth of Tebeth,” Jewish Quarterly Review 74 (1983): 174–95CrossRefGoogle Scholar; and more recently, Elizur, Lammah ẓamnu. The earliest such source (aside from a very fragmentary fifth- or sixth-century list inscribed on a synagogue wall) is a piyyut of R. Eleazar Kallir. This piyyut was first published by Fleischer, in his “Compoziẓiyyot Kaliriyyiot le-tish‘ah be-’Av,” Hebrew Union College Annual 45 (1974): 21–23 (Hebrew section)Google Scholar. Among the fast days included in this piyyut is the 10th of Nissan. Yet Megillat ta‘anit forbids fasting from the 8th through the 21st of Nissan. This suggests that Megillat ta‘anit was no longer being observed in Palestine in the time of Kallir. Moreover, on several occasions in this piyyut, Kallir does not even give the date of the fast, only specifying the month and the event. This suggests that the fast days he included were well known and may have been observed for generations by his time. Fleischer estimates Kallir's lifespan as 570–640 ce. See his “Le-pitron she'elat zemano u-makom pe‘iluto shel R. Eleazar be-Rabbi Kallir,” Tarbiẓ 54 (1985): 383–427. It remains unclear whether the fast days described by Kallir, and later by R. Phinehas b. Jacob ha-Kohen (eighth century) in his Kiddush yeraḥim, were observed by the majority of Palestinian Jewry, or only by a small segment of the community (perhaps by the leaders only). On this issue, see Elizur, Lammah ẓamnu, 25 and 230–32.

70. One can make an argument (especially if these chapters of Massekhet soferim date as late as the tenth century) that this language implies an alternative Babylonian custom of fasting before Purim, on the 13th. Scholars who make this argument include Loewenstamm, Samuel K., “Esther 9:29–32: The Genesis of a Late Addition,” Hebrew Union College Annual 42 (1971): 123Google Scholar; Hilvitz, Ḥikrey zemanim, vol. 1, 354; Elizur, Lammah ẓamnu, 158 n. 40; and Fleischer, Seridim, 10. But if this passage was composed in a place such as Italy or Byzantium, one could understand the implication differently. The passage could be implying that the three fast days were observed before Purim in the locale of the author.

71. The printed text of Massekhet soferim reads differently: להזכיר בציבור אסור עד שיעבור ניסן.

72. Massekhet soferim 17:3 and 21:1 (ed. Higger). Despite the last statement, it is possible that the custom to fast three days in Adar originated before there was any prohibition to fast in Nissan. The choice of Adar may have simply come from a desire to observe the fasts close to the time of Purim.

73. The main sources for this practice are found in Allon, Gedalyahu, “Li-yeshuvah shel baraita ’aḥat,” Tarbiẓ 4 (1933): 285–91Google Scholar. Many of these are early Christian sources that refer to Jewish practices. For example, the practice is probably alluded to at Luke 18:12 in a parable about a Pharisee who declares his righteousness by proclaiming: “I fast twice a week; I give tithes of all that I possess.” (Allon cites more explicit Christian sources for the Jewish practice as well.) See also the variant of T. Ta‘anit 2:5: yom sheni ve-ḥamishi yaḥid yoshev be-ta‘anit ẓibbur, B. Ta‘anit 12a: yaḥid she-kibeil ‘alav sheni ve-ḥamishi kol ha-shanah kulah, and the references at Y. Ta‘anit 1:6 (64c) and Y. Pesaḥim 4:1 (30d) to a custom of women to refrain from work on Monday and Thursday ‘ad de-yitfeni ta‘anita (until the fast ends). For additional sources aside from the many collected by Allon, see Elizur, Lammah ẓamnu, 160. Even though numerous sources confirm the existence of such a practice, it was probably not the tradition of a majority of Palestinian Jewry but was likely a practice of pious individuals only.

74. See, e.g., Hilvitz, Ḥikrey zemanim, vol. 1, 372; and Tabory, Mo‘adey yisra'eil, 405–406.

75. The Palestinian custom of fasting may also be alluded to in the Targum Rishon to Esther 9:31. This source includes the word דכירין in its translation: למהויהון דכירין פתגמי צומיא וצלותהון. Grossfeld, Bernard, The First Targum to Esther (New York: Sepher-Hermon Press, 1983), 72Google Scholar, translates: “that they remember the matters of the fastings and their prayers.” According to Grossfeld, this Targum was composed between 500 and 700 ce in the Galilee area.

But the Monday–Thursday–Monday fast days in Adar are not included in the piyyut of Kallir (see above, n. 69). Nor are they included in the Kiddush yeraḥim of R. Phinehas b. Jacob ha-Kohen (eighth century). But their omission from these sources would not be surprising. The three Adar fast days did not occur on the same calendar date each year, and were not designed to commemorate a historical event that took place in the particular month, unlike the other fast days included in these sources.

76. This approach to the origin of the fast is accepted by many today, in part because it is called Ta‘anit Esther. But the earliest source that referred to the fast by a name is R. Natronai, and he called it Ta‘anit Purim. See the discussion below in this section.

One can suggest alternatively that the fast on the 13th was a custom designed to commemorate the fasting by the Jews throughout the provinces (mentioned briefly at Esther 4:3). The difficulties raised with the first approach will apply to this suggestion as well.

77. To properly understand the view of Maimonides, one must first establish the correct text of Mishneh torah, hilkhot ta‘aniyyot 5:5. It turns out that an erroneous period and vav (the vav of ובי”ג‎ ) made their way into the standard printed text here, after the sixth word. (The necessary corrections have already been made in the Frankel edition. The corrected version also accords with the way the passage is quoted in Beit Yosef, OH 676.) As corrected, hilkhot ta‘aniyyot 5:5 reads:

Maimonides clearly states here that the custom of fasting on the 13th is only of recent origin, and that it commemorates a fast from the time of Haman, i.e., in Nissan. Some have wondered why Maimonides refers to Esther 9:31, a verse that includes the word zar‘am and can be interpreted to refer to a fast that was accepted for all generations. But the fact that Maimonides clearly states that the custom of fasting on the 13th is only of recent origin indicates that he does not accept this interpretation and is only referring to 9:31 to show that there was fasting in the year of the threat. That he does not cite 4:3 suggests that he believes the fast commemorates the three days of fasting initiated by Esther, and not the fasting by the Jews throughout the provinces, even though he could have stated this more clearly. He is forced to cite chapter 9 because chapter 4 does not explicitly state that the Jews of Shushan fasted in response to Esther's request.

Rashi (d. 1105) is reported to have stated that fasting on the 13th of Adar is only a minhaga be-‘alma she-nahagu ha-‘am, because Mordecai fasted at the time of Passover, for three days. (For the citation, see below, n. 139.) One can interpret Rashi to be rejecting any connection between the fast on the 13th of Adar and those observed in Nissan. But one can also interpret Rashi as believing that the fast of the 13th of Adar was meant to commemorate the fasts observed in Nissan. He was only explaining the differences to show that the fast on the 13th cannot be an obligation, only a custom.

A thirteenth-century authority who follows the approach that fasting on the 13th of Adar is a post-talmudic custom designed to commemorate the three days of fasting initiated by Esther is R. Tzedekiah b. Abraham Anav. See his Shibboley ha-leket, sec. 194.

78. One factor that may have motivated this desire was verse 9:31 of the book of Esther: “as they accepted upon themselves and zar‘am the matters of the fasts and their outcry.” See Bloch, Abraham P., The Biblical and Historical Background of the Jewish Holy Days (New York: Ktav, 1978), 230Google Scholar. This verse cries out (pun intended!) to be interpreted on a secondary level as an instruction to the generations to fast, even though the plain sense of the verse is otherwise (see above, Part I). It was noted above (n. 8) that some interpreted Esther 9:31 to refer to the fast of the 13th. But this interpretation surely arose only after the practice of fasting on the 13th had already begun.

79. Throughout the discussion in Part VI, I will be discussing the geonic responsum included in Midrash Tanḥuma and not discussing she'ilta 79, because I believe the geonic responsum is the earlier of the two. Close analysis of that source is the key to understanding the origin of the fast. Everyone who has written about the origin of the fast until now has made the erroneous assumption that she'ilta 79 was composed or edited by R. Aḥai (middle of the eighth century), and then assumed that she'ilta 79 was probably the earlier source.

80. Megillat ta‘anit would have prohibited fasting on the first twenty-one days of Nissan. But the holidays of Megillat ta‘anit were almost certainly no longer observed in Babylonia at the time the practice of fasting on the 13th of Adar commenced.

81. Admittedly a prohibition to observe a communal fast in Nissan is recorded in the twenty-first chapter of Massekhet soferim. But the tenth through twenty-first chapters of Massekhet soferim were probably composed only in the late geonic period, and not in geonic Babylonia. See Part V. I am not aware of any prohibition to fast in Nissan preceding the one found in the twenty-first chapter of Massekhet soferim.

By the early geonic period, there seems to have already been a custom in the Academies not to recite taḥanun in Nissan (and Tishrei). See Sefer sha‘arey teshuvah, ed. Fishel, Yeruḥam (Leipzig: Leopold Shenois, 1858)Google Scholar, responsum no. 243, attributed to R. Moses Gaon (early 9th cent). See also nos. 336 and 337, attributed to R. Hai.

82. It was mentioned earlier (above, n. 2) that rabbinic sources include two different traditions with regard to when the three days of fasting were observed in the year of the threat. The tradition that viewed the three days as the 13th, 14th, and 15th seems to have been the earlier one. There is no evidence that the other tradition even existed in the early and middle geonic periods.

Although the custom of firstborns fasting on the 14th is mentioned in Massekhet soferim (21:1, ed. Higger), this practice is not recorded in the Babylonian Talmud nor in any Babylonian geonic sources. See Goldschmidt, Daniel, “‘Al ta‘anit bekhorot,” De‘ot 9 (1959): 3–5Google Scholar. In the Jerusalem Talmud, there is no reference to a widespread custom of firstborns fasting on the 14th. Mention is made at Y. Pesaḥim 10:1 (37b) that R. Yehudah ha-Nasi fasted on the 14th, and one of the suggestions offered by an Amora to explain this practice is that R. Yehudah ha-Nasi was a firstborn. An alternative suggestion is offered to explain R. Yehudah ha-Nasi's practice. Mention is also made of another firstborn Sage who did not fast on the 14th. R. Eliezer b. Isaac of Worms, Sefer ha-rokeaḥ, sec. 270, writes that the custom of firstborns fasting on the 14th is mentioned in a passage in Massekhet semaḥot. This tractate may date from the tannaitic period (Encyclopaedia Judaica 14:1139). But no such passage is found in the present versions of Massekhet semaḥot. Even assuming that there was a custom of firstborns fasting on the 14th in geonic Babylonia, this would not likely have prevented the designation of the 13th of Adar as a fast day. These fast days are not full fast days, so adjacent fast days could have been observed. See, for example, the discussion below of the extended fasting that took place between Rosh Ha-shanah and Yom Kippur in parts of geonic Babylonia.

83. Also, the 13th was a significant day in its own right. It was the day originally planned for the destruction of the Jews by Haman (Esther 3:13), and was the first and main day of battle in the year of the threat.

84. Although there is perhaps one geonic source that refers to the fast as Ta‘anit Esther. This will be discussed below. It was noted above (n. 8) that an interpretation among the rishonim claims that the fast of the 13th is referred to in Esther 9:31. One might argue that R. Se ‘adyah adopted this interpretation as well. This would be a geonic source that viewed the fast of the 13th as a commemoration of the fasts initiated by Esther (or of the fasts of Esther 4:3). But in this approach, the commemoration is not of post-talmudic origin, but a rabbinic obligation dating from the biblical period. Even if R. Se‘adyah adopted this view, this interpretation surely arose only after the practice of fasting on the 13th had already begun.

The argument that R. Se‘adyah interpreted 9:31 to be referring to the fast of the 13th of Adar is based on a passage in the third of the three seliḥot that R. Se‘adyah included in his siddur to be recited on this day:

SRSG, 338. The last four words are a paraphrase of Esther 9:31. The substitution of the word le-dorotam for zar‘am shows that the passage interprets Esther 9:31 to be referring to the acceptance on future generations of the fasts mentioned in chapter 4. As this is a piyyut to be recited on the 13th, it seems reasonable at first glance to read it as expressing R. Se‘adyah's explanation for the fast of the 13th. But Palestine was the source of much of the liturgical poetry found in Babylonian geonic literature. If this seliḥah was not composed by R. Se‘adyah but merely incorporated by him, it very possibly had its origin in Palestine. If it was composed by R. Se‘adyah, it very possibly was written in the first three decades of his life when he lived in Egypt, or during the years he lived in Palestine before he moved to Babylonia and composed his siddur there (see above, n. 24). If the seliḥah was composed in Palestine or Egypt, it would reflect an explanation for the Palestinian practice, not the Babylonian.

The second of R. Se‘adyah's seliḥot for the 13th of Adar begins tela'ot be-yom ẓomi ’asaprah. One can argue based on these words that this seliḥah was composed in Babylonia, but this argument is weak.

85. Although these phrases are not found in she'ilta 79, there is nothing in she'ilta 79 to suggest that it takes a different approach.

86. SRSG, 319. That R. Se‘adyah also refers to ordinary fast days here does not detract from the interpretation that he views the origin of the fast of the 13th as parallel to the origin of the four traditional fast days. See also SRSG, 258.

87. And not the fast that commemorates fasts mentioned in the Megillah, which were observed in Nissan. It also seems unlikely that it would mean the fast that it was decided should be observed annually on the 13th in subsequent years, even though no such fast occurred in the year of the threat. (This is an argument against the claim that R. Se‘adyah interpreted verse 9:31 to imply the Jews’ acceptance of an obligation to fast annually on the 13th.)

88. See above, Part III.

89. The original is lost; the work is only known from quotations in later sources. The statement above is from Or zaru‘a, hilkhot megillah, sec. 367, in Or zaru‘a ha-shalem: hilkhot mo‘adim ve-zera‘im, ed. Marinberg, Avraham (Jerusalem: Makhon torani yeshivat Or Eẓion- Makhon Yerushalayim, 2006)Google Scholar. Sefer ha-mikẓo‘ot (in the small portion of the work that is known) includes numerous responsa from the Babylonian Geonim. Simḥah Assaf writes that it was probably composed in Babylonia or its environs, but also suggests authorship in Ashkenaz. The work is only quoted by sources from the twelfth and thirteenth centuries, and only in Ashkenaz. See Assaf, , Tekufat ha-ge'onim ve-sifrutah (Jerusalem: Mossad Harav Kook, 1955), 207–209Google Scholar.

90. From de-’amar until raḥamim, the author is paraphrasing she'ilta 79. The citation to Esther 9:31 (the last four words) is not from this or any she'ilta. The citation to Esther 9:31 is only meant to show that the purpose of a fast day is for Israel to cry out for mercy.

91. The precise meaning here is unclear. There is perhaps a textual problem.

92. In contrast, the use of the words nahagu, nohagin, and minhag in connection with the fast are common among the rishonim. See, e.g., Rashi (below, n. 139); R. Isaac b. Abba Mari, Sefer ha-ittur, 220 (one view); R. Zeraḥiah b. Isaac ha-Levi, commentary to Rif, Megillah, end of first chapter; Maimonides (above, n. 77); Naḥmanides, commentary to Rif, Megillah, end of first chapter; R. Meshullam b. Moses of Bezier, Sefer ha-hashlamah, commentary to Ta‘anit, second chapter; Rosh, commentary to Megillah, sec. 8, and commentary to Ta‘anit, sec. 24; R. Tzedekiah b. Abraham Anav, Shibboley ha-leket, sec. 194; R. Jacob b. Asher, Tur OH 676, Kol bo, sec. 45 (but compare sec. 62); R. Aaron b. Jacob ha-Kohen, Orḥot ḥayyim, hilkhot megillah u-Purim, sec. 25; Ritba, commentary to Megillah 2a; R. Meir ha-Kohen of Rothenberg, Haggahot maimuniyyot, hilkhot ta‘aniyyot 1:6; and Ran, commentary on Rif, Ta‘anit 7a (in the pages of Rif).

Aside from those who interpret Esther 9:31 to imply the Jews’ acceptance of an obligation to fast annually on the 13th of Adar (see above, n. 8), very few rishonim express the view that the fast is obligatory. One who does so is R. Ḥayyim, the son of R. Isaac of Vienna. R. Ḥayyim quotes from the Sefer ha-mikẓo‘ot, quoting the first few words above that his father had quoted. See Or zaru‘a ha-shalem: hilkhot mo‘adim ve-zera‘im, 465. R. Isaac also presumably felt that the fast was an obligation from the biblical period, as he did not express any disagreement with the statement of the Sefer ha-mikẓo‘ot.

R. Abraham Ibn Ezra views the fast as an obligation from the Sages. See below, n. 131. R. David Abudarham uses the word tiknuhu with regard to the fast, but states that this was by post-talmudic Sages. See below, n. 139. R. Abraham b. Isaac of Narbonne uses the words nitkan and kav‘u with regard to the fast (see below, n. 138), but implies that the fast was established sometime after Megillat ta‘anit was discontinued. Finally, see the view of R. Tam discussed in Appendix III.

Because of the numerous rishonim who viewed the fast as only a custom, R. Moses Isserles (OH 676) was comfortable writing ve-ta‘anit zeh ’eino ḥovah. R. Caro had written only mit‘annin be-yod-gimmel be-Adar, without clarifying the nature of the obligation. Although R. Caro titles sec. 676: Din Ta‘anit Esther, he nowhere uses the term ta‘anit esther in 676(2), the subsection in which he discusses the fast of the 13th. He only uses the term in 676(3), the subsection in which he describes the custom to fast three days: yesh mit‘annim gimmel yamim zeikher le-ta‘anit Esther. It can be argued that he concluded (after collecting various opinions in his Beit Yosef) that fasting on the 13th was based on the derivation from nikhalu and was not a commemoration of the fasts initiated by Esther. See also OH 550(4) where he does not call the fast Ta‘anit Esther, but Ẓom Purim. But see Sperber, Minhagey yisra'eil, vol. 1, 169 n. 1, for a different explanation for the use of this name for the fast.

93. See above, n. 31. He was writing in Arabic, which lacks a letter for “p.” It is clear from his Chronology of Ancient Nations that he had an extensive knowledge of the Jewish calendar and its holidays. See Ratzaby, , “Megillat ‘ta‘anit ẓaddikim’ be-makor ‘aravi,” Sinai 106 (1990): 3Google Scholar. The Chronology of Ancient Nations was written by Al-Biruni in Jurian (an area southeast of the Caspian Sea), but he seems to have lived most of his life before this in Khwarizm (Greater Iran, now Uzbekistan).

Early rishonim who refer to the fast as Ta‘anit Esther include R. Abraham Ibn Ezra (1089–1164), commentary to Esther 9:30 (this commentary was written in Italy, see Uriel Simon, Four Approaches to the Book of Psalms, 260 n. 18); R. Abraham b. David of Posquieres (c. 1125–1198), quoted by the Meiri, Magen ’avot, 154; and R. Eliezer b. Isaac of Worms (1165–1230), Sefer ha-rokeaḥ, sec. 240. Many rishonim refer to the practice of fasting on the 13th of Adar without giving it a name. Also, one of the three texts of Megillat ta‘anit batra found in the Genizah refers to the fast as Ta‘anit Esther. See above, n. 31. This text probably dates from the eleventh century. The Rosh refers to the fast as Ta‘anit Esther in his paraphrase of the view of R. Tam. See Appendix III. But this does not mean that R. Tam used the term himself. Massekhet soferim, which most likely dates from the ninth or tenth century, refers to the fast as sheloshet yemey ẓom Mordekhai ve-Esther. But this work was most likely composed in a community under Palestinian influence. See above, Part V.

94. The earlier version of Seder parshiyyot is included as a section in Halakhot pesukot (and in the Hebrew adaptation of this work, Hilkhot reu). It is shorter, and does not include any version of R. Natronai's responsum. See Sassoon, Saliman, ed., Sefer halakhot pesukot le-Rav Yehudai Gaon (Jerusalem: Makor, 1971), 449–64Google Scholar. I am going to assume that the material in the expanded version was composed in Babylonia or its environs, although it is possible that this material was composed elsewhere, such as in North Africa.

95. The three manuscripts are Vatican 142, Milan (Ambrosiana C116 Sup.), and De Rossi 1089. The Vatican 142 manuscript is estimated to have been copied in North Africa in the eleventh century. See Elizur, Lammah ẓamnu, 45. Seder parshiyyot is included in this manuscript as a section in the Halakhot gedolot. An edition of the Halakhot gedolot based on Vatican 142, and including the chapter Seder parshiyyot, was published by Azriel Hildesheimer at the end of the nineteenth century. (Most likely, the expanded version of Seder parshiyyot and other material were added to Halakhot gedolot in North Africa, at the end of the ninth or beginning of the tenth century. See Brody, Le-toledot, 136 n. 21.) The Milan (Ambrosiana C116 Sup.) manuscript was copied in Italy, c. 1400. Seder parshiyyot is also included in this manuscript as a section in the Halakhot gedolot. For a description of this manuscript, see Elizur, Lammah ẓamnu, 45. (A descendant of Hildesheimer published an edition of Halakhot gedolot based on this manuscript, but he did not include the Seder parshiyyot section.) The De Rossi 1089 manuscript was copied in Italy in the late thirteenth century. See Richler, Benjamin, Hebrew Manuscripts in the Biblioteca Palatina in Parma (Jerusalem: Jewish National and University Library, 2001), 193Google Scholar. It was published by Horowitz, Chaim M. in his Beit nekhot ha-halakhot o toratan shel rishonim (Frankfurt am Main: Slobotzky, 1881)Google Scholar. Naḥman Danzig refers to a Genizah fragment of the expanded version of Seder parshiyyot. See his Mavo le-sefer halakhot pesukot ‘im tashlum halakhot pesukot (New York and Jerusalem: Beit ha-midrash la-rabbanim be-America, 1999), 299 n. 100. But I am not aware of any Genizah fragment of Seder parshiyyot that includes the shortened version of the responsum of R. Natronai.

96. Vatican 142 and De Rossi 1089.

97. Milan.

98. Seder parshiyyot is the kind of source that lent itself to being changed by copyists, who might conform the text to reflect the customs in their locale. See Danzig, Mavo, 189 n. 69. A change of name (without a change of a custom) would seem to be an even more likely occurrence.

That the Palestinian custom was understood to commemorate the fasts initiated by Esther probably caused the Jews in Palestine, Egypt, and North Africa to assume that this was the basis for the fast of the 13th as well and to call that fast Ta‘anit Esther. This would have influenced the North African copyist of Vatican 142 (or the earlier Palestinian, Egyptian or North African source he was copying from) to change Ta‘anit Purim to Ta‘anit Esther. With regard to the De Rossi manuscript, this manuscript was copied in Italy in the late thirteenth century. By that time, the term Ta‘anit Esther was probably in use there. (R. Abraham Ibn Ezra used the term Ta‘anit Esther in his commentary to Esther 9:30, and this commentary was written in Italy. Also, the term Ta‘anit Esther is found in sec. 194 of the Shibboley ha-leket of R. Tzedekiah b. Abraham Anav. R. Tzedekiah does not use the term himself, but quotes from Minhagot ḥakhmei Magenza, which used the term.) Also, the text of the De Rossi manuscript largely follows the text of Vatican 142.

99. It has been overlooked because those who analyzed the origin of the fast always assumed that she'ilta 79 was the first explicit reference to fasting on the 13th. But the geonic responsum included in Midrash Tanḥuma is probably earlier (see above, Part IV), and is the source that deserves the closer analysis. A key passage, וכבוד עדיף שבת מאלף תעניות (honoring the Shabbat is more important than a thousand fasts), is found in the geonic responsum included in Midrash Tanḥuma, and not in she'ilta 79.

100. Midrash Tanḥuma includes three passages from the She'iltot, an eighth-century source (one of which is at Bere'shit, sec. 2, right next to our passage at sec. 3). See above, n. 32. Also, Midrash Tanḥuma, Noaḥ, sec. 3, has much in common with a section of the letter of Pirkoy ben Baboy, a source from the late eighth or early ninth century (see below, n. 103), and with a text that B. M. Lewin published and called a new version of this letter. According to Brody, this text published by Lewin is most probably an earlier polemical letter, not by Pirkoy, composed perhaps in the middle of the eighth century. See Lewin, , “Mi-seridey ha-genizah,” Tarbiẓ 2 (1931): 394–96 and 400–403Google Scholar, Brody, Robert, Pirkoy ben Baboy ve-toledot ha-pulemus ha-penim-yehudi (Tel Aviv: Tel Aviv Univ., 2003), 18–30Google Scholar, and Danzig, Naḥman, “Bein Ereẓ-Yisra'eil le-Bavel: dappim ḥadashim mi-ḥibbur Pirkoy ben Baboy,” Shalem 8 (2008): 5 n. 13Google Scholar. Midrash Tanḥuma also includes a brief passage, at Noaḥ, sec. 1, which reflects an anti-Karaite polemic. See Bregman, Sifrut tanḥuma-yelammedenu, 185. This passage probably does not date from the eighth or early ninth centuries, but is later.

Scholars have pointed to additional passages in Midrash Tanḥuma that may date from the geonic period. But Bregman points out (280 n. 136) that such passages may not be reflected in the manuscripts and may only be later additions to particular versions. (See, for example, his comments on Midrash Tanḥuma, Meẓora, sec. 8.) This is in contrast to the passages mentioned above, which are recorded in the major textual witnesses to Midrash Tanḥuma and are evidence of when and where the work was redacted.

101. For a discussion of the issue of fasting on Shabbat throughout the centuries, see Gilat, Yiẓḥak D., “Ta‘anit be-Shabbat,” Tarbiẓ 52 (1983): 1–15Google Scholar.

102. The fact that the responsum does not illustrate seven scenarios, but only the scenarios of the 14th falling on Friday, Saturday, Sunday, and Monday, also suggests that the main motivation for its interpretations was related to Shabbat and Erev Shabbat.

103. This was a polemical letter written to the Jews of North Africa and Spain, instructing them that Palestinian customs should not be followed. Pirkoy, a Babylonian Jew, tells us that he was a disciple of someone named Rava who was a disciple of R. Yehudai. (R. Yehudai was head of the academy at Sura from approximately 757–61 ce.) Pirkoy writes that many of the Palestinian customs originated as emergency measures during times of persecution, or were customs resulting from ignorance. It was only in Babylonia that accurate traditions were preserved. For further background to the letter, see Brody, The Geonim of Babylonia, 113–117, and more recently, Danzig, “Bein Ereẓ-Yisra'eil le-Bavel,” 1–32.

104. “And because they fast on Yom Tov and on Shabbat and on the first two days of Rosh Ha-shanah and on the Shabbat between Rosh Ha-shanah and Yom Ha-kippurim.” See Oẓar ha-ge'onim, Yom Tov, 19, sec. 41.

105. Pirkoy does not indicate whether the Palestinian practice he was criticizing was the practice of the majority of Palestinian Jewry.

A practice of yeḥidim (rabbinic scholars, see Ta‘anit 10a) fasting for the entire ten days of repentance seems to be referred to at Leviticus rabbah 30:7:

Leviticus rabbah was probably composed in fifth-century Palestine (Encyclopaedia Judaica 11:147). A similar passage is found at Pesikta de-Rav Kahana, sec. 27. (In Pesikta de-Rav Kahana, most manuscripts read ha-kesheirim instead of ha-yeḥidim. See ed. Mandelbaum, vol. 2, 412.) Pesikta de-Rav Kahana also probably dates to fifth-century Palestine. See Pesikta de-Rav Kahana, eds. Braude, William G. and Kapstein, Israel J. (Philadelphia: Jewish Publication Society of America, 1975), intro., xlviGoogle Scholar. (Parallels to the above passages are at Midrash Tanḥuma, Emor, sec. 22; Kohelet rabbah 9:7; Midrash ha-gadol, Emor, 658; and Yalkut shimoni, Emor, sec. 651.) By the time of Pirkoy, the majority of Palestinian Jewry may have been following the practice of fasting for the entire ten days. Support for this might be found in the responsa at Oẓar ha-ge'onim, Yom Tov, 24, secs. 47 and 48. Both of these record a practice of fasting for the entire ten days of repentance, and the practice does not seem to have been limited to rabbinic scholars. The authors of these responsa are unknown, but some have argued that they are from Palestine (see below, n. 109). (These responsa recommend that kol yisra'eil follow the practice of fasting the entire ten days.) Sec. 48 also quotes the passage referred to above from Leviticus Rabbah. In some versions of the responsum, the word beinonim is used instead of ha-yeḥidim.

106. R. Yehudai was a leading scholar at Pumbedita who eventually became head of the academy at Sura. He served in this position for three and a half years, approximately 757–61 ce. (Probably these were the last years of his life.) The statements attributed to him could have been made while he was at Pumbedita, or while he was the head of the academy at Sura. During the period that he was head of the academy at Sura, his brother was the head of the academy at Pumbedita.

107. See Oẓar ha-ge'onim, Yom Tov, 21, sec. 42. This is a section from Halakhot keẓuvot, a work that seems to have been written in Italy in the ninth century. In this section, the issue of fasting on the Shabbat before Yom Kippur is discussed, and a decision opposing it is attributed to R. Yehudai. The section also includes statements such as: (1) whoever fasts on the Shabbat before Yom Kippur profanes the Shabbat and violates the commandments of zakhor and shamor and (2) the only time fasting on the Shabbat is allowed is on Yom Kippur and in the case of ta‘anit ḥalom (and the allowance for a ta‘anit ḥalom is defined very narrowly: only in a case where one dreams that the government has decreed against him and placed him in chains or prison). These statements may reflect the views of R. Yehudai, or may only reflect the views of the author. Danzig, Mavo, 475, suggests that the source for the statement in Halakhot keẓuvot that R. Yehudai opposed fasting on the Shabbat before Yom Kippur and the statements of others that R. Yehudai opposed fasting on the first two days of Rosh Ha-shanah (see the note immediately below) was merely the letter of Pirkoy ben Baboy.

108. See Oẓar ha-ge'onim, Yom Tov, 18, sec. 39. On the custom of fasting on Rosh Ha-shanah, see Gartner, Ya‘akov, “Ta‘anit be-rosh ha-shanah,” Ha-darom 36 (1973): 125–62Google Scholar.

109. There is no specific evidence that a custom to fast for the entire ten days of repentance existed in Babylonia as early as the eighth century. But R. Natronai (head of the academy at Sura, 857–65) approved fasting on the last nine days of the ten, and described it as an old custom. He first writes that one is not allowed to fast on the first day of Rosh Ha-shanah, but continues:

See Brody, Teshuvot Rav Natronai, 308–09, responsum 182; and Oẓar ha-ge'onim, Yom Tov, 23, sec. 45. (In the version quoted by the Rosh, commentary to Rosh Ha-shanah, end of the fourth chapter, the responsum reads nohagin rabboteinu instead of nahagu rabboteinu ha-rishonim.) There is also a responsum that records that the custom bi-mekomeinu u-vi-yeshivot (in our locale and in the Academies) is to fast all ten days. See Oẓar ha-ge'onim, Yom Tov, 24, sec. 47. Weinberg, Teshuvot Rav Sar Shalom Ga'on, 39, suggests that the responsum was written by R. Zemaḥ b. Solomon (ninth century, Babylonia). (In Ḥemdah genuzah, sec. 129, it is included in a section of responsum attributed to a R. Zemaḥ.) But this responsum may be Palestinian, and from a later time period. See Gartner, 126, and Lewin's comments to sec. 48, which is a responsum similar to sec. 47.

R. Sar Shalom, who was R. Natronai's predecessor at Sura, is recorded by R. Eliezer b. Joel ha-Levi as having been opposed to fasting on Rosh Ha-shanah. See Sefer Ravyah, sec. 537. However, he may still have permitted fasting on the other nine days, in accordance with the custom mentioned by R. Natronai.

R. Hai (head of the academy of Pumbedita, d. 1038) is recorded as opposing fasting on the Shabbat before Yom Kippur, but allowing fasting on both days of Rosh Ha-shanah. See Oẓar ha-ge'onim, Yom Tov, 22, sec. 43. (In another source, he is recorded as opposing fasting on both days of Rosh Ha-shanah as well; ibid., 23, sec. 44.)

The custom of fasting for the entire ten days of repentance continued in some locales in the time of the rishonim. See the articles by Gartner and Gilat referred to above. See also R. Moses Isserles, OH 581(2).

110. See above, n. 104.

111. The fasts of Megillat ta‘anit batra originated in Palestine, as did the practice of fasting on Monday and Thursday. Palestine was an area with a tendency toward asceticism.

Pirkoy quotes a statement by a Palestinian Amora recorded at B. Berakhot 31b:

(Pirkoy only quotes up to dino). He implies that this statement served as a basis for the Palestinian practice to fast on Shabbat. But he explains that permissible fasting on Shabbat is limited to cases where life and death are at stake, such as the case of one who had a dream that the government decreed against him, and now fears for his life.

Scholars have also noted the above statement by the Palestinian Amora at B. Berakhot 31b and suggested that it was the cause of, or evidence of, a Palestinian practice to fast on Shabbat in the amoraic period. See, e.g., Gilat, 6–7; and Gartner, 145.

B. Shabbat 11a includes a statement by R. Yosef permitting ta‘anit ḥalom on Shabbat. Therefore, the Babylonian geonic tradition limited the statement at B. Berakhot 31b to ta‘anit ḥalom. See, e.g., Oẓar ha-ge'onim, Berakhot, oẓar ha-peirushim, 44, sec. 135 (R. Hai). See also Oẓar ha-ge'onim, Berakhot, nisfaḥim le-oẓar ha-peirushim, 40, sec. 133; Tosafot Berakhot 31b; and Rashba, commentary to Berakhot 31b. (See also Solomon Buber, Midrash tehillim, intro., 69.) Maḥzor Vitry, 234 (ed. Hurwitz), records that there is another version of the statement at Berakhot 31b. In this version, the statement explicitly refers to ta‘anit ḥalom (kol ha-yosheiv be-ta‘anit ḥalom be-Shabbat … ).

112. Oẓar ha-ge'onim, Yom Tov, 19, sec. 41. Halakhot keẓuvot had also used strong language in opposing the practice of fasting on the Shabbat before Yom Kippur. See above, n. 107. But as mentioned above, the words used may not have been those of R. Yehudai.

113. Oẓar ha-ge'onim, Yom Tov, 20, sec. 41. The similarity between this passage and the passage in the geonic responsum in Midrash Tanḥuma was noted long ago by Jacob Mann when he first published this passage. See his “Les ‘chapitres’ de ben Bâboï et les relations de R. Yehodaï Gaon avec la Palestine,” Revue des Etudes Juives 70 (1920): 141 n. 7. See also B. Shabbat 30a (final two lines).

114. B. Pesaḥim 68b records that Mar the son of Ravina fasted the whole year, aside from Shavu‘ot, Purim, and Erev Yom Kippur. On the simplest level, this means that he fasted every Shabbat. Of the rishonim who accept the interpretation that he fasted on Shabbat (not all do), some understand its permissibility as based on his being occupied with prayer or study. See, e.g., the commentaries of R. Eliezer b. Joel ha-Levi, Sefer Ravyah, sec. 876; and R. David Bonafed (commentary to Pesaḥim 68b). Many others interpret his fasting on Shabbat as for a ta‘anit ḥalom.

Danzig suggests, based on a weak inference, that the author of the Halakhot gedolot (9th century) may have permitted fasting on Shabbat even outside the context of ta‘anit ḥalom. See his Mavo, 119. One can make the same claim about the position of the Seder Rav ‘Amram. See Seder Rav ‘Amram Ga'on, ed. Goldschmidt, 97, sec. 60.

115. As noted, the issue of fasting on Shabbat also came up in the context of ta‘anit ḥalom. B. Shabbat 11a included a statement by R. Yosef permitting ta‘anit ḥalom on Shabbat. It is possible that some in geonic Babylonia might have interpreted this permission broadly. (See perhaps she'ilta 1, Bere'shit, which seems to give permission to fast to one who merely sees a milta de-ẓa‘ara and has a ḥalma bisha.) The authors of the geonic interpretation of M. Megillah 1:1–2 could have been responding to this and implying that permission to fast on Shabbat in the case of ta‘anit ḥalom should be severely circumscribed. This would be consistent with the views of Pirkoy and the author of Halakhot keẓuvot, which both may reflect the view of R. Yehudai.

The standard edition of the Tur includes a statement that R. Amram and R. Kalonymus held that ta‘anit ḥalom is forbidden on Shabbat. But in the manuscripts of the Tur and the early printed editions, the statement is recorded in the name of R. Kalonymus only. See Tur, OH 568, Mosdot shirat Devorah edition (Tel Aviv: Deborah, 2000)Google Scholar.

116. A prohibition of fasting on Shabbat is implied fairly clearly in T. Ta‘anit 2:7. Also, M. Ta‘anit 2:10 perhaps implies such a prohibition. M. Ta‘anit 4:3 records that ’anshei ma‘amad did not fast on Shabbat. B. Eruvin 41a includes an explicit statement by Ravina prohibiting such fasting. (Statements in the Jerusalem Talmud prohibiting fasting on Shabbat are quoted by Gilat, 1.) Nevertheless, an additional source for a prohibition to fast on Shabbat would have had some value. The other sources can be read as prohibiting fasting in certain contexts only, or as general prohibitions that allowed for the possibility of limited exceptions. Also, in rabbinic argumentation, it is natural to attempt to marshal as many sources as possible to support one's position.

117. Fasting on Erev Shabbat had long raised issues in halakha. See, e.g., M. Ta‘anit 2:10 and 4:3, Tosefta Ta‘anit 2:7 and 2:8, B. Eruvin 40b–41b, and she'ilta, Bere'shit, 1. M. Ta‘anit 4:3 records that ’anshei ma‘amad did not fast on Erev Shabbat out of kavod Shabbat, but the prevailing view in mishnaic and talmudic times (and the view of the above she'ilta) was that it was permissible to fast on Erev Shabbat. The only issue concerned the time of the end of such a fast. The geonic responsum, in interpreting M. Megillah 1:1–2 to prohibit fasting on Erev Shabbat, would be advocating a new, more stringent approach.

118. It also provides an answer to the question discussed in Appendix II: If Purim was one of the two holidays of Megillat ta‘anit that were never discontinued, and there was a prohibition to fast on the day before it, how could a fast on the 13th ever have arisen? The second approach has an easy answer. In describing what occurred in the year of the threat, the geonic responsum included in Midrash Tanḥuma included the phrase ve-gazru ta‘anit, and in describing the fast, called it a de-rabbanan. Taken together, these passages imply that the geonic interpretation took the position that in the year of the threat a decree was enacted that bound future generations. If the practice to fast on the 13th arose as a result of this interpretation, the practice would have easily overridden any possible prohibition to fast. The practice would have been viewed as the resumption of an ancient obligation, based on a biblical verse. A few sources allude to this answer. See, e.g., Rosh, commentary to Ta‘anit, sec. 24, and Tur, OH 686. (As to the effect of Megillat ta‘anit on this ancient fast, perhaps it was thought that the fast was suspended during the centuries that Megillat ta‘anit was in effect.)

But there are several other easy answers to the question raised in Appendix II (see particularly answers 1, 2, and 3). Therefore, one cannot point to the fact that the question is easily answered in the second approach as evidence that it is the correct approach.

119. Based on B. Shabbat 11a, the Babylonian Geonim were in agreement that some type of fasting for ta‘anit ḥalom on Shabbat had to be permitted.

120. See above, Part III.

121. See above, n. 109.

122. Ibid.

123. But admittedly, it does use the phrase ve-gazru ta‘anit to describe what occurred in the year of the threat, and describes the practice of fasting as a de-rabbanan. The import of these statements is that in the year of the threat a decree was enacted that was binding on future generations. In this way, the responsum does imply that there at least should have been a practice to fast on the 13th in its time. If one wants to read the responsum as implying not only that there should have been, but that there actually was a practice of fasting on the 13th in its time, a response could be that the interpretation originated earlier, before the responsum was reduced to writing. By the time the responsum was reduced to writing, the practices it implied were already in existence as a result of the oral interpretation. This would also be a response to any evidence that the written responsum included in Midrash Tanḥuma dates from after the time of R. Natronai.

124. The root כנס appears in the context of Esther's command to the Jews of Shushan to fast (in Nissan). She instructs Mordecai: “Go, gather (כנוס‎) all the Jews found in Shushan, and fast for me … (Esther 4:16). This helps support the interpretation of yom ha-kenisah as a fast day. I would like to thank Rabbi Richard Wolpoe for this observation.

125. Also, perhaps the authors of the unusual interpretations had a text of the Talmud that did not include the anonymous (late?) discussion at Megillah 4a–b which understood yom ha-kenisah as Mondays and Thursdays, and perhaps they did not have M. Megillah 1:3 as we have it. (The statement of R. Judah at M. Megillah 1:3 strongly suggests that the term yom ha-kenisah at 1:1–2 refers to Mondays and Thursdays, but does not absolutely mandate this interpretation.)

126. In this approach, one can still accept the suggestion that the authors of the unusual interpretations were motivated by opposition to practices of fasting on Shabbat and Erev Shabbat that were widespread in their time. One can believe that the motivation behind the unusual interpretations was twofold: opposition to practices of fasting on Shabbat and Erev Shabbat, and a desire to justify a preexisting custom to fast on the 13th.

127. One can alternatively suggest that it began as an enactment from the geonic period, without a claim by the geonim that it had ancient origins.

128. One can also suggest that the fast came to be treated as a rabbinic obligation from the biblical period based on the interpretation of 9:31 that eventually arose, if this interpretation arose as early as the geonic period. It can be argued that there is evidence for this interpretation in R. Se‘adyah. See above, n. 84.

129. This explanation was already discussed at length (the first approach).

130. See Esther 3:13. The only source in the period of the geonim that perhaps alludes to this approach is a piyyut composed by a Syrian paytan from the first half of the eleventh century, Yanon b. Ẓemaḥ. This piyyut (based on the fasts of Megillat ta‘anit batra) includes one line regarding the 13th of Adar:

See Zulay, Menaḥem, “Shivrey ẓelilim,” Sinai 28 (1950–51): 167–69Google Scholar; and Elizur, Lammah ẓamnu, 95. But the Jewish community of Syria is usually considered to have been within the ambit of the Jewish community of Palestine. Thus, I am hesitant to rely on a paytan from Syria to help explain the origin of a custom that originated in Babylonia. It is also very possible that the line in this piyyut has nothing to do with the suggestion I made in the text. The term ‘amad Haman suggests that the author of the piyyut believed that Haman was still alive on the 13th of Adar and still trying to destroy the Jews. See similarly some of the later versions of Megillat ta‘anit batra quoted by Elizur at 159 n. 42 (ק‎-9, ת‎-5, ק‎-18) and her reference to Sefer ha-tadir.

131. Even though the Jews had by now been given the right to fight and destroy their enemies (Esther 8:11–13), and were being assisted by the governors of the provinces and were feared by the masses (Esther 8:17, 9:2 and 9:3), the 13th of Adar must still have been a traumatic day. R. Abraham Ibn Ezra writes (commentary to Esther 9:30): ולפי דעתי חכמינו ז“ל קבעו התענית יום הפחד (in my opinion, the Sages established this fast on the day of fear.) Probably he is adopting this approach. In the second recension of his commentary to Esther, Ibn Ezra writes very similarly: וחכמים ז”ל תקנו התענית ביום הפחד. He adds that perhaps there was a tradition that the Jews fasted on the 13th in the year of the threat. See Zedner, Joseph, Va-yosef Avraham (London: D. Nutt, 1850), 34Google Scholar. Simon, Uriel, Four Approaches to the Book of Psalms (Albany: SUNY Press, 1991), 154Google Scholar, refers to a manuscript with a third commentary on Esther by Ibn Ezra. I have not seen this manuscript.

Worth mentioning is the explanation for the origin of the fast set forth in Siddur kol Ya‘akov/The Complete ArtScroll Siddur, 855 (italics are mine):

132. See above, n. 31.

133. The earliest dateable reference to the fast of the 13th is found in a responsum of R. Natronai. It was mentioned earlier (above, n. 23) that R. Sar Shalom, R. Natronai's predecessor at Sura, wrote a responsum very similar to the one by R. Natronai, except that it included no reference to the fast. This might suggest that there was no practice of fasting on the 13th at the time of R. Sar Shalom. A less drastic interpretation is that the difference between the two responsa helps define the period when the practice of fasting on the 13th began to spread.

One might point to the lack of a reference to the fast in the Halakhot gedolot (a work compiled around the middle of, or perhaps the early, ninth century) as evidence that there was no practice to fast in Babylonia then. But this argument is weak. Halakhot gedolot may have been compiled outside of Babylonia (see Brody, The Geonim of Babylonia, 228) or perhaps inside Babylonia but outside of the Babylonian yeshivot, where the fast may have first arisen.

134. R. Aaron b. Jacob ha-Kohen (thirteenth–fourteenth century), Orḥot ḥayyim, hilkhot megillah u-Purim, sec. 25. This idea is also found in the related work, Kol bo, sec. 45.

135. R. Abraham b. Isaac of Narbonne (twelfth century, father-in-law of R. Abraham b. David of Posquieres), Sefer ha-eshkol (ed. Albeck), vol. 1, 137, citing unnamed others. He himself disagrees with this view. See below.

136. R. Mordekhai Jaffe (sixteenth century), Levush, OH 686. R. Jaffee's explanation is paraphrasing Job 2:10.

137. Maggid mesharim, parshat va-yakheil. This suggestion of the Maggid mesharim is quoted in the Kaf ha-ḥayyim, OH 686. The Maggid mesharim also suggests this as the reason for the custom of the firstborn and others who fast on the day before Passover. The Maggid mesharim was composed by R. Joseph Caro (Encyclopaedia Judaica 5:200).

138. R. Abraham b. Isaac of Narbonne, Sefer ha-eshkol, vol. 1, 137:

Finally, an interesting suggestion made by Sperber, Daniel, in the Journal for the Study of Judaism 16 (1985): 280–81Google Scholar, also deserves mention. (He repeated the suggestion with minor variations in Minhagey yisra'eil, vol. 1, 176–77.) Sperber suggested that when the statement of R. Samuel b. Isaac is quoted at B. Megillah 2a, only a portion of the original statement is being quoted. He suggested that what follows the statement of R. Samuel b. Isaac in she'ilta 79, a citation to a verse and the phrase מאי קהילה יום תענית, were also part of the original statement. Sperber noted that when this statement was first recorded at B. Megillah 2a, it is recorded with the following language: ke-de-’amar Rav Shemuel bar Yiẓḥak … hakha nami …, suggesting that the statement was initially made in another context, and allowing for the possibility that B. Megillah 2a was not quoting the full statement. R. Samuel b. Isaac was a Babylonian Amora of the late third and early fourth centuries ce. If the full statement of R. Samuel b. Isaac were as Sperber hypothesized, this would be evidence that the fast was kept in Babylonia as early as his time, or would be a statement by an Amora which could have motivated people to start fasting on the 13th. (Sperber suggests only the former. I am adding the latter.) But the practice of fasting on the 13th is not included in the Halakhot gedolot, and its inclusion in Seder Rav ‘Amram is only from the time of the rishonim. This suggests that the practice did not exist as early as the third or fourth century. It also suggests that it did not arise in response to a statement by a third- or fourth-century Amora, but arose later. Also, the full statement of R. Samuel b. Isaac quoted in the Talmud is י”ג זמן קהילה לכל היא ולא צריך לרבויי‎. There are different readings in the manuscripts with regard to the last three words (see above, n. 59), but the second part of the statement is there in some form in all but one manuscript. If the statement of R. Samuel b. Isaac had a second half, one cannot read she'ilta 79, which does not have any statement beginning with ve-lo ẓarikh, as the full statement of R. Samuel b. Isaac. The same problem exists if one attempts to claim that the geonic responsum included in Midrash Tanḥuma records the full statement of R. Samuel b. Isaac. Rather, both the geonic responsum in Midrash Tanḥuma and she'ilta 79 offer an interpretation that has expanded upon the original statement of R. Samuel b. Isaac.

139. That the fast is geonic or post-talmudic in origin has been set forth explicitly by several of the rishonim. See, e.g., (1) Haggahot maimuniyyot, hilkhot ta‘aniyyot 1:6: mi-divrey ha-Ge'onim; and (2) R. David Abudarham, Sefer Abudarham, 282: ha-ḥakhamim ha-’aḥaronim tiknuhu ’aḥar ḥatimat ha-talmud. See also the statement of Maimonides quoted above, n. 77 (ve-nahagu kol yisra'eil bi-zemanim ’eilu …). Also, Rashi implies that the fast is post-talmudic. In a responsum quoted in Maḥzor Vitry, he writes that this fast is not from the Torah or from divrey Soferim, but is a minhaga be-‘alma she-nahagu ha-‘am. See Maḥzor Vitry, ed. Aryeh Goldschmidt (Jerusalem: Mekhon oẓar ha-poskim, 2004), 372. (See similarly, Maḥzor Vitry, ed. Hurwitz, 210.) This responsum of Rashi's is quoted elsewhere in Rashi literature at Siddur Rashi, sec. 345; Sefer ha-pardes, sec. 204; and Teshuvot Rashi, sec. 128. It is also included in Beit Yosef, OH 686. R. Caro's source for it was Shibboley ha-leket, sec. 194.

Many rishonim use words such as nahagu, nogahin, and minhag when describing this fast. See above, n. 92. These words often imply post-talmudic origin. But most of the rishonim who used these words probably believed that Megillat ta‘anit was cancelled prior to the conclusion of the amoraic period. There would have been room, in their view, for a custom of fasting on the 13th to have arisen during the talmudic period. One interesting reference cited in n. 92 is Naḥmanides. He writes that there was no fasting on the 13th in the days of rabboteinu, because it was Yom Nikanor. He does not specify whether by the term rabboteinu he means the entire amoraic period or only part of it.

140. See above, Part VI (the possible exception being Seder parshiyyot, where a question exists about whether the original reading is Ta‘anit Esther or Ta‘anit Purim). The earliest undisputed references to the fast as Ta‘anit Esther are: (1) manuscript Vatican 142 of Seder parshiyyot, estimated to be from North Africa in the eleventh century; (2) a text of Megillat ta‘anit batra found in the Genizah, which probably dates from the eleventh century (see above, n. 31); and (3) the commentary to Esther 9:30 of R. Abraham Ibn Ezra (which was composed in Italy). As suggested in Part VI (n. 98), the name Ta‘anit Esther for the fast of the 13th may have originated in Palestine, Egypt, or North Africa, influenced by the Palestinian custom of three days of fasting, which did commemorate the fasts initiated by Esther. Also, if the true origin of the fast lies in the second approach, an approach based on unusual interpretations of nikhalu and yom ha-kenisah, it is not surprising that an alternative “folk” understanding of the origin of the fast eventually arose. The name Ta‘anit Esther would be the result of this folk understanding.

141. This was the first printed edition.

Mirsky prepared an edition of the She'iltot utilizing nine relatively complete European manuscripts (eight of which include our passage). See his She'iltot de-Rav Aḥai Ga'on, 5 volumes (1960–77). I will mention some of the variants he noted. Mirsky made very little use of the Genizah material. (He intended to devote a separate volume to the Genizah material, but died before he was able to do so.) Dozens of Genizah fragments have been discovered. Of the limited Genizah material that has been published or described to date, the only fragment of she'ilta 79 that I am aware of is the one described by Shraga Abramson mentioned below.

142. The text I printed is citing 9:16–17 here, except that it skips the last twelve words of verse 16 after nafsham, and continues with the first five words of verse 17. In the case of the four manuscripts that cite to 9:16–17 here, they skip the last fifteen words of verse 16, and include the first eight words of verse 17. (Mirsky's notes indicate that these four also include the word gam at the beginning of verse 17. But there is no such word at the beginning of verse 17.)

143. Six of the eight manuscripts read הכניסה. Two of the manuscripts read הכנסה.

Mirsky notes (vol. 3, 223) that four of the eight manuscripts include the additional words יום שיני וחמישי after the phrase ומאי יום הכניסה. These additional words would make the definition of yom ha-kenisah more consistent with the Mishnah and Talmud. However, this she'ilta repeats its interpretation of yom ha-kenisah later, and Mirsky does not note any such variants there. Moreover, the Genizah fragment referred to by Abramson seems to be consistent with the present text there; see below, n. 147. There is no question that this she'ilta defined yom ha-kenisah in the way that it did, as the geonic responsum included in Midrash Tanḥuma takes the same approach. Moreover, this she'ilta and its unusual definition of yom ha-kenisah are cited by many rishonim (whether they agree or disagree). Also, none of the manuscripts used by Mirsky come from earlier than the fourteenth century. Surely, the additional words in the four manuscripts noted by Mirsky originate with an addition by a copyist attempting to bring at least one of the statements in this she'ilta in line with the traditional understanding of yom ha-kenisah. (These same four manuscripts also include a passage stating that when the fast falls on Shabbat, the villages and cities advance the reading to yom ha-kenisah. See Mirsky, vol. 3, 225.)

144. Mirsky notes that seven manuscripts have the additional word בט”ו here. The other manuscript has ובחמשה עשר. It seems fairly clear that the omission in the printed edition is an error.

145. Mirsky notes that all eight manuscripts have additional words here: בי”א בי”ב. It seems fairly clear that the printed edition is in error.

146. Six of the eight manuscripts read הכניסה. Two read הכנסה.

147. The citation in the printed edition above starts with nikhalu ha-yehudim from verse 9:2, but continues with words from 9:16 and 9:17. Four of Mirsky's eight manuscripts cite 9:2 here, without any reference to 9:16–17. The other four cite a combination of 9:2 and 9:17. Abramson, Shraga, in his ‘Inyanot be-sifrut ha-ge'onim (Jerusalem: Mossad Harav Kook, 1974), 361Google Scholar, mentions a relevant Genizah fragment. He describes it as starting with kakh peirshu ḥakhamim yom ha-kenisah and continuing until le-’aḥeihem she-be-kerakhim. He does not print the fragment but describes it in a general way as matching the printed edition.

148. I have inserted a space here to indicate that a new topic is being discussed, not relevant to our topic.

149. The disagreement between them was that R. Yose held that fasting on the day following such a holiday was prohibited as well.

150. In disputes between Rav and Samuel, the halakha follows Rav in the case of prohibitions. (It follows Samuel in monetary matters.) See B. Bekhorot 49b, B. Niddah 24b, and Yad Mal'akhi, secs. 147–51. In disputes between R. Yoḥanan and Samuel, the halakha follows R. Yoḥanan. See B. Eruvin 47b and Yad Mal'akhi, sec. 616.

151. Of course, for those who believed that the fast of the 13th was referred to at 9:31, this question did not arise.

152. See above, n. 118. As stated there, some rishonim allude to such an approach.

153. See, e.g., R. Zeraḥiah ha-Levi, commentary on Rif, Megillah, end of first chapter; R. Isaac b. Abba Mari, Sefer ha-ittur, 220–21; R. Abraham b. Isaac of Narbonne, Sefer ha-eshkol, vol. 1, 136–37; R. Eliezer b. Joel ha-Levi, Sefer Ravyah, secs. 550 and 559; Tosafot Ta‘anit 18a, Naḥmanides, commentary on Rif, Megillah, end of first chapter; Ran, commentary on Rif, Ta‘anit 6b–7a (in the pages of Rif), Tur, OH 686; R. Aharon ha-Kohen of Lunel, Orḥot ḥayyim, hilkhot megillah u-Purim, sec. 25; Kol bo, sec. 45; Rosh, commentary to Megillah, sec. 8 and commentary to Ta‘anit., sec. 24; Mordekhai, commentary to Megillah, sec. 776; Ritba, commentary to Megillah 2a, Rosh Ha-shanah 19b, and Ta‘anit 10a and 15b.

154. It must be mentioned that the Jerusalem Talmud takes the position that there was a prohibition to fast on the day before Purim even before the advent of Yom Nikanor. See Y. Ta‘anit 2:13 (66a), and Y. Megillah 1:6 (70c). Presumably, such a prohibition would have survived the discontinuance of Megillat ta‘anit.

155. This approach is based on a statement made by R. Ashi at Ta‘anit 18b in an analogous case.

156. See, e.g., (1) R. Judah b. Barzilai Al-Bargeloni (late eleventh–early twelfth century, Spain), author of Sefer ha-ittim, cited in Naḥmanides (commentary to Rif, Meg., end of first chapter) and many others; (2) R. Ephraim Ibn Avi Alragan (late eleventh–early twelfth century, N. Africa), author of a commentary on the Rif, cited in Naḥmanides, ibid.; (3) R. Meshullam of Bezier (late twelfth–early thirteenth century, Provence), Sefer ha-hashlamah, commentary to Ta‘anit., 2nd chapter; and (4) Ritba, commentary to Ta‘anit 10a and 15b, and to Rosh Ha-shanah 19b. One other authority who takes this approach, in a responsum, is R. Abraham b. Isaac of Narbonne (twelfth century). See Elizur, Lammah ẓamnu, 115 n. 2. This responsum may not yet be published. It is found in a manuscript that includes geonic responsa, interspersed with responsa by the Rif and his pupils and by Provençal sages. The geonic responsa included in this manuscript have been published by Simḥah Emanuel. See his Teshuvot ha-ge'onim ha-ḥadashot (Jerusalem: Mekhon Ofek, 1995)Google Scholar. Emanuel shows that the author of the Shibboley ha-leket used this manuscript. The author of the Shibboley ha-leket refers to the responsum of R. Abraham b. Isaac without mentioning the name of its author, but instead uses the phrase: sha'alu le-Ge'onim z”1… . See Shibboley ha-leket, sec. 278 (included in Oẓar ha-ge'onim, Ta‘anit, 17, sec. 33). The attribution of this responsum to the “Ge'onim” has misled many, as pointed out by Elizur. See her Lammah ẓamnu, 115 n. 2. But it is not uncommon among rishonim to use the term Ge'onim to mean early rishonim. (The responsum of Abraham b. Isaac is also referred to in Tanya rabbati, sec. 62, a work closely related to the Shibboley ha-leket. See Encyclopaedia Judaica 2:938. It is attributed to Ge'onim there as well.) I could find no evidence of an actual geonic view that the halakhah followed R. Simeon b. Gamliel.

157. His view is only known from summaries by others.

158. See their commentaries to B. Megillah 2a.

159. Megillah, sec. 1.

160. שהכל מתאספין לתענית אסתר ובאים בני הכפרים לעיירות.

The Rosh mentions the view of R. Tam elsewhere in his commentaries: commentary to Megillah, sec. 8, and commentary to Ta‘anit, sec. 24. He does not include this language there. Nor does Tosafot Rosh, commentary to Megillah 2a. R. Tam states that the reading, when advanced, takes place in the cities. This is not inconsistent with either the geonic responsum or she'ilta 79, as neither states clearly where this reading takes place. Both refer to the 13th as a day on which Jews are yoshev be-ta‘anit or yoshvin be-ta‘anit, but this is an idiom and does not necessarily imply that one stays in one's locale. She'ilta 79 includes more explanatory statements than the geonic responsum and some of the statements in she'ilta 79 can perhaps be read to imply that the reading, when advanced, takes place in the cities. Also, the people who knew how to read the Megillah would have lived mainly in the cities.

161. OH 686. The Tur is the son of the Rosh. The Tur does not state here that he is describing the view of R. Tam, but that is clearly what he is doing.

162. Comm. to Megillah, sec. 776, in Mordekhai ha-shalem, ed. Rabinovitz, Meir (Jerusalem: Makhon Yerushalayim, 1997)Google Scholar.

163. See, e.g., (1) Ḥiddushey ’anshey shem, in the name of R. Meir of Rothenberg (thirteenth century), commentary to Rif (Megillah 4a, in the pages of Rif): ר“ת בשם השאילתות;‎ (2) Ran, commentary to Rif (Megillah 1b, in the pages of Rif): וכן מצא בשאלתות; and (3) Amarkal (in Nathan Coronel, ed., Ḥamishah kunteresim, 25): וכן מצא ר“ת בשאילתות. Amarkal is an anonymous work written by a pupil of the Rosh.

164. B. Megillah 2a: zeman kehillah la-kol hiy.

165. Comm. to Megillah, sec. 8: הוי תענית אסתר תקנת חכמים אף קודם שבטלה מגילת תענית.

166. This tradition is found in many sources. See, e.g., Mekhilta of R. Ishmael and Mekhilta of R. Simeon b. Yoḥai to Exodus 17:9 and 17:12; and Midrash Tanḥuma, Beshalaḥ, 27. This tradition is followed by Rashi in his commentary to Exodus 17:10. For additional sources, see Menaḥem M. Kasher, Torah shelemah, Beshalaḥ, chapter 17, secs. 62 and 97. Many rishonim cite this tradition as being in the Talmud, Ta‘anit, although it is not in our printed editions.

167. R. Tam does not say this, but the reading of the Megillah every year can be construed as an engagement with Amalek's descendant. In rabbinic tradition (B. Megillah 13a), “Haman, the son of Hammedata, the Agagite” (Esther 3:1) teaches that Haman was a descendant of the Agag who was king of Amalek at the time of Saul.