Once printing became inexpensive and printed books ubiquitous, Jews increasingly came to pray by reading words printed on a page (or today, sometimes a screen). Even when most prayed from memory, they knew that the official version of their prayers resided in a written text, perhaps an elegant manuscript housed in the local synagogue. A very strong sense of correct and incorrect words existed, especially in medieval Central Europe. But evidence for written prayers in the rabbinic tradition is exceedingly sparse from before c. 1000 CE, especially before the introduction of cheaper writing materials like paper. Are we correct to presume that early rabbinic liturgy had a textual tradition coherent with the models we know, including from their predecessors at Qumran? Did early rabbis have a system of memorized, orally transmitted yet fixed texts, or did their worship operate in a more free-form system? When and where did the verbal worship of the rabbinic elite spread to other social groups and geographic areas? These questions, with answers not yet fully resolved in my mind, have dominated my current investigations into early rabbinic liturgy.
The earliest rabbinic texts about liturgy were redacted in the early third century CE Mishnah from extant oral traditions. They presume a structured system of daily and holiday prayers. While drawing on some earlier models, they seem mostly to have emerged as a response to the liturgical chasm created by the Roman destruction of the Jerusalem Temple in 70 CE. These prayers remain the core of the Jewish prayer book. However, rabbinic literature incorporated traditions about the actual language of these prayers only where disputes arose, leaving us snippets of evidence but no coherent picture from which to reconstruct these core prayers (occasionally, there is a text for a more peripheral passage).[1] Much of our imagination about this period thus rests on how we query these texts and which snippets we choose to emphasize. We therefore need to be cautious about our own presumptions and the edifices that we can and often do build on their shaky foundations.
Another challenge is the growing recognition over the past decades that the rabbinic texts themselves are literary constructions, where supposedly carefully preserved oral traditions have been worked and reworked to fit the cultural and literary contexts of the texts that we have received. Texts about liturgy are particularly vulnerable to this process of reworking, since scribes and editors frequently “corrected” them to fit their known prayers. Therefore, even consulting manuscripts (generally preserved only from c. 1000 CE on) does not always reliably tell us even what the early medieval talmudic editors redacted, let alone the original form of a tradition in the name of rabbis from c. 100 CE! But occasionally there are real insights from this and from paying attention to parallel versions of texts and their variants. What appears in the standard printed version of the Babylonian Talmud, a text of great authority in today’s Jewish world, is frequently not sufficient for reliable historical answers. Nevertheless, reading these texts with a critical methodology does produce useful insights.
One category of text that is particularly productive are narratives of actual prayer situations, especially times when something has failed to operate as it should. These raise questions about any widely known and accepted texts for the prayers. Let me give an example.
Midrash Leviticus Rabba 23:4 (Marguliot ed.,[2] redacted 400-500 CE) identifies three people who stand out “like a lily among thorns” (Song of Songs 2:2). In each case, when a quorum of ten gathers for a different liturgical occasion (daily prayer, a wedding, or a post-funeral meal),[3] only one person, the “lily,” knows how to recite the prayers. The midrash follows these three abstract examples with a more specific story, set among rabbis in the early second century CE.
Rabbi Elazar Ḥisma went to a place, and they said to him, “Does Rabbi know how to lead the public recitation of Shema?”[4] He answered, “No.” “Does Rabbi know how to lead The Prayer?”[5] He answered, “No.” They said to him, “Do they call you Rabbi without basis?” He became embarrassed and went to Rabbi Akiva with a pale face. [Rabbi Akiva] asked, “Why is your face pale?” He told what had happened. [Rabbi Akiva] asked, “Does Rabbi want to learn?” He answered, “Yes.” After he had learned, he returned there, and they asked him, “Does Rabbi know how to lead the public recitation of Shema?” He answered them, “Yes.” “Does Rabbi know how to lead the Prayer?” He answered them, “Yes.” They responded, “Rabbi Elazar can now speak (ḥaseim)!” Therefore, they called him Rabbi Elazar Ḥisma.
The midrash adds, about a fourth-century rabbi, “Rabbi Yonah would teach his students even the wedding and mourning blessings. That is to say, they should be competent in all things.”
This story is best read in conjunction with a more famous passage from Mishnah Berakhot 4:3-4a that points to a debate among the leading rabbis c. 100 CE about the nature of The Prayer’s text. We read:
Rabban Gamliel taught: Each person recites eighteen benedictions daily.
Rabbi Yehoshua said: [No,] a digest of eighteen.
Rabbi Akiva said: If his prayer flows fluently, he recites eighteen, and if not, a digest of eighteen.
Rabbi Eliezer said: One who makes his prayer fixed fails to make his prayer supplicatory.
The Mishnah, itself an early third century CE composition, suggests that promulgating the new obligation being imposed by the rabbis is not a trivial matter. Rabban Gamliel decrees that The Prayer be recited universally daily. We must note, though, that rabbinic “universalism” not infrequently refers only to the rabbinic class. Some contemporary scholars, most notably Ezra Fleischer, posit that such universal participation was not possible without a fully defined text: Rabban Gamliel must have disseminated a fully scripted composition.[6] However, the other voices in the Mishnah suggest other possibilities, which are confirmed by our story. Rabbi Elazar, a member of exactly this circle of rabbis, obviously did not personally recite daily a fixed text. If he prayed according to the rabbinic system at all, he clearly did not consider his personal prayers adequate for functioning as a prayer leader and helping others to fulfill their obligations to pray. Embarrassed, he consulted Rabbi Akiva, who, in the Mishnah, advocates a flexible system of prayer, at least for the individual. There are also traditions that Rabbi Akiva was an ecstatic worshiper, who compromised and moderated his ecstasy when with others (Tosefta Berakhot 3:5). What does Rabbi Elazar learn from him? We can’t know whether Rabbi Akiva taught him a text or skills for fluent composition of his own innovative prayers according to the rabbinic pattern. With these new abilities, he was able to return to the place of his embarrassment and redeem his status as a rabbi and the authority of the rabbinic class.
However, we must be cognizant of the fact that this story appears in a fifth-century CE compilation from the Land of Israel. Its context denigrates those without basic liturgical skills as “thorns,” and it celebrates Rabbi Elazar’s transformation from an embarrassed “thorn” to a “lily.” In conclusion it cites the fourth-century Rabbi Yonah who preemptively taught his students the necessary liturgical skills before they attained rabbinic status. Thus, by his time and certainly by the redaction of the midrash, there was an expectation that rabbis know how to pray fluently and lead public liturgies.
Does this mean that by the fifth century CE rabbinic leaders were operating with a fully fixed prayer text, or just a fixed pattern of prayer into which the prayer leader and individual worshipers might insert their own language? The answer may lie somewhere between the two. It also seems to have differed for the rabbinic communities in Babylonia and the Land of Israel. By the time of the midrash, we know that a vast creative tradition of liturgical poetry was beginning to emerge in the Land of Israel, which was open to substituting increasingly elaborate poetic versions for all but the structural markers of a growing number of central prayers. This poetry frequently took as its topic the theme of the day, as defined by the calendar or the lection. Thus, there was not a presumption of a fixed prayer text, only of a fixed framework into which one could substitute new language.
There is a risk involved in free-form prayer: will it be acceptable to God? This concern may indeed lie behind the tradition of beginning The Prayer with Ps. 51:17, “Eternal God, open my lips that my mouth may voice Your praise,” and concluding it with Ps. 19:15, “May the words of my mouth and the meditations of my heart be acceptable to You, O Eternal, my Rock and my Redeemer.” In other words, these verses ask first for divine help in choosing the words, and then that the words chosen have indeed pleased God. This tradition is transmitted in both Talmuds in the name of Rabbi Yoḥanan (d. 279 CE), a leading rabbi in the Land of Israel. The Jerusalem Talmud (Berakhot 4:4, 8a) cites it when discussing Rabbi Eliezer’s objection to fixed prayer in the Mishnah cited above. The voices collected in the Talmud discuss different models for ensuring a lack of fixity in The Prayer: adding a new element to it, totally new prayers each time, or just avoiding rote recitation. The dissident voice, though, is that of Rabbi Zeira, a Babylonian rabbi spent time in the Land of Israel, who is afraid that innovation in prayer will cause him to err.
It seems that Babylonian rabbis were significantly uncomfortable with freely voiced prayer. The Babylonian Talmud is troubled by this tradition of beginning The Prayer with Psalms verses because the verses interrupt the juxtaposition between the preceding blessing and The Prayer.[7] They do not read this verse as seeking inspiration before praying a significantly unscripted liturgy, but rather they justify it as an expansion of the liturgy’s script itself (Berakhot 4a). Their parallel discussion of Rabbi Eliezer’s teaching in the Mishnah (Berakhot 29b) reworks the Jerusalem Talmud’s tradition, eliminating its debate about which verses best express a call for divine help in composing the moment’s texts, and concluding with Rabbi Zeira’s hesitation about deviation from a fixed text, here in elaborated form: “I am certainly capable of innovating in my prayer, but I am afraid that I will become confused.” Indeed, we know that the insistence on correct prayer texts takes on momentum precisely in Babylonia. There, liturgical poetry ceased to replace statutory texts and instead came to augment them, if it was recited at all.
As it was the Babylonian voice that became dominant in the Jewish world, those formed in its traditions have tended to read the earlier texts through its presumptions. However, particularly stories of prayer contexts, when considered critically and with an eye to their editing and reediting in later contexts (where this can be documented) opens up the possibility of a significantly more complex picture in which early rabbinic prayer, at least within rabbinic circles, was thematically organized but far from fully scripted.
Ruth Langer is Professor of Jewish Studies in the Theology Department at Boston College and Interim Director of its Center for Christian-Jewish Learning.
[1] For a detailed discussion of this question, see: Ruth Langer and Richard S. Sarason, “Re-Examining the Early Evidence for Rabbinic Liturgy: How Fixed Were Its Prayer Texts?,” in On Wings of Prayer: Sources of Jewish Worship; Essays in Honor of Professor Stefan C. Reif on the Occasion of his Seventy-fifth Birthday, Nuria Calduch-Benages, Michael W. Duggan, and Dalia Marx, eds., Deuterocanonical and Cognate Literature Studies 44 (Berlin: DeGruyter, 2019), 203-231, https://doi.org/10.1515/9783110630282-014.
[2] The standard printed edition does not include the wedding benedictions, but this version is well-attested in the manuscripts and parallels.
[3] The birkat ḥatanim, i.e., the seven wedding blessings recited in the ceremony and then at subsequent festive meals for a week (b. Ketubot 8a), and the birkat avelim, a more obscure ritual, perhaps recited to comfort the mourners at post-funeral meals (Tosefta Berakhot 3:22-3).
[4] Pores al Shema refers to a public, perhaps responsive recitation practiced by the early rabbis (Mishnah Megillah 4:5-6, b. Sotah 30b), but later generations apply the term to other rituals.
[5] HaTefillah is the classical rabbinic name for the central complex of their liturgy, recited at all three services daily, consisting originally of eighteen juxtaposed benedictions on weekdays.
[6] For an English summary (and critique) of Fleischer’s theory, see my “Revisiting Early Rabbinic Liturgy: The Recent Contributions of Ezra Fleischer,” Prooftexts 19:2 (1999): 179-194. See also the follow-up, “Controversy,” in 20:3 (2000): 381-387.
[7] There was c. 200 CE still some flexibility in the ordering of the elements of rabbinic liturgy, and the principle that The Prayer must be juxtaposed to the blessing(s) following Shema emerges to clarify this structure (Tosefta Berakhot 1:2).