Esther Before the King, (detail) Doré's English Bible, 1866. Wikimedia
The Fast of Esther’s 8th Cent. C.E. Origins
The book of Esther never mentions a fast on the 13th of Adar; in fact, in the late Second Temple period, it was a day of celebration. Talmudic sources never mention the fast either. How did it originate?
Published: March 1, 2023
|
Last Updated: September 7, 2026
This article is based on my “The Origin of Taʿanit Esther,” AJS Review 34.2 (2010): 309–351. For correspondence, I can be reached at mfirstatty@aol.com.
Editor’s note: For more on this festival and its importance among the early Maccabees, see Zev Farber, “The Ancient Judean Holiday: Yom Nicanor—13th of Adar,” TheTorah (2014); idem, “Chanukah: Not Judah Maccabee’s Holiday,” TheTorah (2022).
For a brief introduction to this text, see Vered Noam, “Megillat Ta’anit and Its Scholion (Commentary),” TheTorah (2014).
See discussion in Zev Farber, “The Cancellation of Megillat Taʾanit,” TheTorah (2014). The last clear evidence of a Megillat Taanit holiday being observed is in a story involving Rav Nachman, a Babylonian Amora living in Nehardea in the late third/early fourth century C.E. (b. Taanit 18b).
The commentary is traditionally attributed to R. Moses Maimonides, but this seems to be in error, and some refer to it nowadays as the Pseudo-Rambam commentary.
The Hebrew follows Moshe Weissman, פירוש הרמב"ם על מגילת אסתר [The Commentary of Rambam on Megillat Esther] (Jerusalem, 1997), 92. Another translation, by Yosef Yoel Rivlin, which has the glosses together with the text in rewritten Bible style, appears on the Project Ben Yehudah website.
He begins by noting the two options referenced above:
וטעם קיימו על נפשם ועל זרעם דברי הצומות – על דעת רבים: על יום תענית אסתר. והמכחישים אמרו כי על שלשה ימים שהתענו בניסן הכתוב מדבר, וכל ישראל ראויין להתענות כן תמיד. ולפי דעתי שחכמינו ז"ל קבעו התענית יום הפחד,
The meaning of “they accepted upon themselves and their descendants the matter of the fasts”—most people think it is a reference to Taʾanit Esther. The deniers (=Karaites) say that it refers to the three days they fasted in Nissan, and that all Jews should fast on these as well. In my opinion, it was the Sages that established the fast on the day of fear.
See also, the commentaries of R. Yosef b’R. Yosef ibn Nachmiash, Ralbag, Malbim, and Hoil Moshe, who all read this way. For modern commentators, see discussion in Jon D. Levenson, Esther, The Old Testament Library (Louisville, KY: Westminster John Knox, 1997), 131; Adele Berlin, אסתר [Esther], מקרא לישראל [Mikra LeYisrael] (Tel Aviv/Jerusalem: Am Oved / Magnes Press, 2001), 149–150. For a different possibility, see Carey A. Moore, Esther (New York: Doubleday, 1974), 96.
This seems to be Maimonides’ understanding as well (though it is less clear):
משנה תורה, זמנים, הלכות תעניות ה:ה ונהגו כל ישראל בזמנים אלו להתענות בשלשה עשר באדר זכר לתענית שנתענו בימי המן שנאמר דברי הצומות וזעקתם.
Mishneh Torah, Zemanim, Laws of Fasts 5:5 All of Israel in these times have the practice of fasting on the thirteenth of Adar in memory of the fast that they fasted in the time of Haman, as it says (Esth 9:31) “the matter of the fasts with their lamentations.”
(The text here follows the Kapih and Frankel versions; the standard printing has a textual error that confused the point.) Maimonides does not specify when this practice started, only that it is currently practiced and is based on something ancient.
Soferim was most likely composed in the 9th or 10th century, in a community under the influence of Eretz Yisrael, such as Italy or Byzantium. For some basic background on the book, including the dating, see Myron B. Lerner, “The External Tractates,” in The Literature of the Sages, First Part, ed. Shmuel Safrai, Compendia Rerum Iudaicarum ad Novum Testamentum 2 (Philadelphia, PA: Fortress Press, 1987), 367–403 [397–400]; See Debra Reed Blank, “It’s Time to Take Another Look at “Our Little Sister” Soferim: A Bibliographical Essay,” Jewish Quarterly Review 90 (1999): 1–26.
This is possibly a reference to Yom Turyanus (the 12th of Adar, another day mentioned in Megillat Taʾanit is a day people should not fast). See Vered Noam, “Yom Turyanus: The 12th of Adar,” TheTorah (2014). This does not mean that Yom Nicanor and Yom Turyanus were actively being celebrated in this period (though some scholars make this argument), only that the author was looking for an explanation for why the fast is kept after Purim instead of before.
It is an acronym based on letter values: bet is 2, for the second day of the week, heh is five (for the fifth day of the week.
This practice may have preceded the Babylonian practice of fasting on the 13th of Adar, although this cannot be proven. The custom appears again in Soferim 21:1, where it explains that the fasts are not observed in Nissan—the time they took place according to the narrative—because of the rule not to fast during Nissan:
סופרים כא:א מנהג רבותינו שבמערב להתענות שלשת ימי צום מרדכי ואסתר פרודות, ולאחר פורים, שני וחמישי ושני. ולמה אין מתענין אותן בחדש ניסן, מפני שבאחד בניסן הוקם המשכן, ושנים עשר נשיאים הקריבו קרבנם לשנים עשר יום, יום לכל שבט ושבט, וכל אחד היה עושה ביומו יום טוב; וכן לעתיד לבא עתיד המקדש להבנות בניסן, לקיים מה שנאמר, אין כל חדש תחת השמש. לפיכך אין אומרים תחנונין כל ימי ניסן, ואין מתענין עד שיעבור ניסן.
Soferim 21:1 The rabbis in the west have the practice to observe the three days fast of Mordechai and Esther separately, after Purim, on Monday, Thursday, and Monday. And why don’t they observe these fasts during the month of Nissan? Because on the first of Nissan, the Tabernacle was erected, and the twelve princes offered their sacrifices for twelve days, one day for each tribe, and each one made their day one of celebration. So too, in the future, the Temple will be built on Nissan, in order to fulfill what it says (Eccl 1:9) “there is nothing new under the sun.” Therefore, tachanun [prayers] are not recited throughout Nissan, and no fasts are observed until Nissan passes.
The Hebrew translation, as well as the publication of the original, are from Ezra Fleischer, “הדותה–הדותהו–חדותא – פולמוס ושברו” [Haduta—Hadutahu—Chedweta: Solving an Old Riddle] Tarbiẓ 53.1 71–96 [94–95].
The Hebrew translation, as well as the publication of the original Judeo-Arabic line, are from Ezra Fleischer, “שרידים נוספים מקובצי תפילה ארץ ישראליים מן הגניזה” [Additional Fragments of Land of Israel Prayer Collections from the Genizah], קבץ על יד [Kobeẓ Al Yad] 15 (2000): 1–37 [pp. 9-11 and 33]. The text was first discussed, and partially published, in Mordechai Margulies, “תעודה חדשה על צום הרעש” [A New Document on the Fast of the Earthquake] Tarbiẓ 29.4 (1960): 339–344.
The Hebrew translation, as well as the publication of the original Judeo-Arabic line, are from Fleischer, “Additional Fragments,” 33.
Robert Brody, Teshuvot Rav Natronai Bar Hilai Gaon (1994), pp. 303-304, §177. The expanded version of Seder Parshiyyot Shel Yamim Tovim ve-Haftarot Shelahen, which dates from the late 9th or early 10th century, includes a shortened version of the responsum of R. Natronai. The text exists in three manuscripts, none of which was copied in Geonic Babylonia. In one, the day is called תענית פורים “the fast of Purim” but in the other two, it is called תענית אסתר “the fast of Esther,” which is not how the Geonim refer to it. It seems likely that this reading reflects copyists altering the name to the one more familiar in their locale and time. (Here I would like to take the opportunity to acknowledge that my argument in this piece would not have been possible without access to the careful analysis and reconstruction of the Geonic sources available in Robert Brody’s many works.)
Siddur Rav Saʿadiah Gaon, eds. Israel Davidson, Simḥah Assaf, and Yissakhar Joel (1941), pp. 258 and 319-338. Although Saʿadiah lived the first part of his life in Egypt and then Israel, the siddur dates from his time in Babylonia.
Louis Ginzberg, Geonica (1909), vol. 2, pp. 67-68. The corrected reading is Ginzberg’s suggestion. (He mentions של “of” as an alternative possibility.) The responsum was coauthored with his son, Rav Hai, and written to R. Jacob of Kairouan. The responsum itself is lost, but an index to the (now lost) collection of Babylonian Geonic responsa records the opening lines of each.
Shelomoh Wertheimer, Sefer Kohelet Shelomoh (1899), p. 14.
Louis Ginzberg, Ginzey Schechter, vol. 2 (1929), p. 136.
The Hebrew is from the Kaufmann manuscript (as transcribed on AlHaTorah.com) and the English from Alyssa Grey’s translation in The Oxford Annotated Mishnah, ed. Shaye J. D. Cohen, Robert Goldenberg, and Hayim Lapin (Oxford: Oxford University Press, 2022), 1.722.
Many scholars believe that Midrash Tanhuma underwent final redaction in geonic Babylonia (or in an area under the sphere of influence of the Babylonian geonim). See, e.g., Mark Bregman, ספרות תנחומא-ילמדנו: תיאור נוסחיה ועיונים בדרכי התהוותם [Tanhuma-Yelammedenu Literature: Studies in the Evolution of the Versions] (Piscataway, NJ: Gorgias Press, 2003), 184–186. Notably, this responsum does not appear in the Buber edition.
This responsum also appears in a revised version in a supplementary section of the She’iltot (§79). The numbering follows the Samuel K. Mirsky’s edition of the She’iltot (1960-77). (These numbers are artificial, as the she’iltot can be broken up in different ways.) This sheilta is part of a section of four she’iltot for Purim that contemporary scholars see as later additions. See Robert Brody, לתולדות נוסח השאליתות [The Textual History of the Sheiltot] מקורות ומחקרים [Texts and Studies] 7 (New York/Jerusalem: The American Academy for Jewish Research/Bialik Press, 1991), 186 [n. 5]; idem, The Geonim of Babylonia and the Shaping of Medieval Jewish Culture (New Haven: Yale University Press, 1998), 209 [n. 29]. Structurally, they are deficient as she’iltot. 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. This sheilta is especially problematic since, after the first few lines in Aramaic, the balance is almost entirely in Hebrew. Comparison between the language of the responsum in Tanchuma and she’ilta #79 shows that the Tanchuma text is earlier, as it is organized and concise, and reflects an attempt to record an official interpretation of the Mishnah while the she’ilta takes for granted an already established explanation upon which it is commenting.
This same baraita appears in the Tosefta with noticeable differences:
תוספתא מגילה א:ג אָמַר רַבִּי יוֹסֵה: אֵימָתַי? בְּמָקוֹם שֶׁנִּכְנָסִין בַּשֵּׁנִי וַחֲמִישִׁי, אֲבָל מָקוֹם שֶׁאֵין נִכְנָסִין לֹא בַשֵּׁנִי וְלֹא בַחֲמִישִׁי, הוֹאִיל וּמִסְתַּכְּנִין בָּהּ, אַל יִסְתַּכְּנוּ בָהּ אֶלָּא בִּזְמַנָּהּ.
t. Megillah 1:3 Rabbi Yose said: “When [does the law allowing earlier reading apply]? In a place where they gather on Mondays and Thursdays, but in a place where they don’t gather on Mondays and Thursdays, since they are endangered (by gathering) in any case, they should only endanger themselves at the proper time.”
Notably, the Talmudic statement as we have it differs slightly in wording, and more importantly, greatly in context. The passage is arguing that we learn from the word זמניהם “their times” (Esth 9:31), written with an extended plural, that the Megillah may be read not just on its proper day, but on other days. The Talmud then argues that the minimum plural is two, but the Mishnah adds three extra days (11th, 12th, and 13th). To this, the Talmud answers:
בבלי מגילה ב. כדאמר רב שמואל בר [רב] יצחק: שלשה עשר זמן קהילה לכל היא, ולא צריך לרבויי.
b. Megillah 2a As Rav Samuel son of [Rav] Isaac said: “The thirteenth was the time everyone gathered, and it is unnecessary to have a biblical proof that it would be included.”
Clearly the Talmudic passage has no connection to the Gaon’s claim that this day is what the Mishnah referred to as “the day of assembly.”
Already in the time of R. Yose b. Bun (c. 300), hundreds of years before this responsum, it was impossible for Purim to fall out on a Shabbat according to the Jewish calendar (j Megillah 1:2 [70b]; also b. Rosh Hashanah 20a). See discussion in, Joseph Tabory, מועדי ישראל בתקופת המשנה והתלמוד [Jewish Festivals in the Time of the Mishnah and Talmud] (Jerusalem: Magnes Press, 1995), 28. She’ilta #79 stated explicitly that the 14th of Adar no longer fell on shabbat in its time. This suggests that the responsum is not describing a contemporary practice of pushing back the fast day, but is envisioning something taking place in Mishnaic times.
In addition, the Talmud itself (b. Megillah 4a-b) understands yom ha-kenisah as a reference to Mondays and Thursdays.
The fact that the responsum does not illustrate seven scenarios, but only illustrates 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.
Otzar ha-Geonim, Yom Tov, p. 20, sec. 41. 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 757-761 C.E.) Pirkoy writes that many of the customs in Israel originated as emergency measures during times of persecution, or were customs resulting from ignorance. It was only in Babylonia that accurate traditions were preserved.
Some fasted specifically on the Shabbat before Yom Kippur (Shabbat Shuva), as a sign of repentance.
Indeed this reading connects with the most common explanation for the fast in traditional sources, quoted ubiquitously in the name of Rabbeinu Tam (Jacob ben Meir, ca. 1100–1171), who paraphrases this Geonic reading and argues that the Jews fasted on the day of the attack and that the fast of Esther commemorates this fast. See, e.g., Rosh, Halakhot, Megillah 1:1; R. Yerucham ben Meshullam, Toldot Adam veChavah, 10:1; R. Joshua Boaz, Shaltei Gibborim, Megillah 1b.
The book itself is lost, and we know it mostly from quotes in halakhic works from the German rishonim (12th/13th cent.). This quote was preserved in Isaac of Vienna’s Or Zarua, “Laws of Megillah” §367.
The fast of the 13th was already known in some areas as Taʿanit Esther by the 11th century.