With these words, Rabbenu TamFootnote ² begins his introduction to Sefer ha-yashar, declaring that his work will be devoted to “straightening out” the teachings and text of the Talmud. However, the book accomplishes much more than the goals set in this declaration. While it contains numerous treatments of the correct text of talmudic passages, it also incorporates fragments of commentary, halakhic monographs, and numerous responsa. Its attention to textual matters is thus a very small portion of the book as a whole.

In contrast, the introduction to the work can stand alone as a wide-ranging manifesto in which the author discusses the proper methods and the exigencies of how to treat the text of the Talmud, in an era when the dissemination of a work was dependent on manuscripts, the habits of copyists, and the conduct of its readers. In this article, I will focus on just one element of the vast wealth of material that awaits those who study this introduction, namely, Rabbenu Tam's criticism of those who emend the text of the Talmud itself: “It is not enough that [those who emend the text] corrupt texts that seem to be glosses, but they even [corrupt] the words of the Amoraim and Tannaim themselves; this is inconceivable for anyone who fears God.”Footnote ³

For the purposes of this article, the most important aspect of this passage is R. Tam's distinction, articulated as an offhand philological remark, between “texts that seem to be glosses” (ha-nir'in perush) and “the words of the Amoraim and Tannaim themselves.”  The conscious differentiation between the dicta of talmudic sages and “glosses”—interpolations that explain, translate, or interpret a word or phrase—did not originate in R. Tam's academy; his grandfather, Rashi, preceded him in this regard.Footnote ⁴ Nor did it end with him; in the hands of the later Tosafists, this technique developed into a method that reached its fullest expression in the academy of Naḥmanides.Footnote ⁵ However, whereas R. Tam's predecessors maintained that after identifying a gloss as a late addition it should be emended or deleted from the text, R. Tam's view was that the student has no right to excise those “texts that seem to be glosses.” His words imply a forgiving attitude toward such deletions ex post facto, but that the prohibition on performing such excision—with a sharp knife or metal scraper—remains in force.

A long-standing scholarly tradition portrays Rabbenu Tam as a figure who did not live up to his own standards with regard to these emendations. Thus, for example, Isaac Hirsch Weiss wrote:

So, too, in his introduction to Sefer Ra'avyah, Avigdor Aptowitzer wrote:

Ephraim Elimelech Urbach summed up in brief:

The present study will demonstrate that Rabbenu Tam was, in fact, perfectly consistent in mediating the tension between his sweeping and vehement opposition to effacing the text as it was and his recognition of the inevitable necessity of occasionally making emendations in the text of the Talmud. Moreover, as we will see, his opposition to deleting anything from the text itself was rooted in his view that many of the accreted glosses and suggestions of students and copyists, from the era of the Amoraim to the medieval era, calcified and became part of the canonic text of the Talmud itself.

There is no doubt that R. Tam's opposition to excising snippets that were identified as exegetical additions produced complications. On one hand, he did not, as stated, permit the emendation of the body of the text, but on the other hand, he took the exegetical liberty of explaining what he identified as “the words of the Amoraim and Tannaim themselves” differently from the glosses to the Talmud text. Faced with such situations, R. Tam developed a way to both preserve the text while also identifying segments as glosses, thus legitimating the disregard of their substance and even their halakhic implications.

Let us demonstrate this with a passage from B. Ḥagigah 3b:

According to R. Elazar, the halakhah transmitted to Moses at Sinai (halakhah le-Moshe mi-Sinai) was that those who dwell in the regions of Ammon and Moab are commanded to separate the poor man's tithe during the Sabbatical year. A rationale is offered for this “halakhah transmitted to Moses at Sinai,” namely, that the Transjordanian regions of Ammon and Moab were sanctified in the “first consecration,” that is, the conquest of Joshua, but not in the “second consecration” that transpired at the time of Ezra's migration. The first consecration was “not consecrated for the future,” so there is no obligation whatsoever to tithe the produce of these regions. Nevertheless, the sages imposed an obligation to give the paupers’ tithe, to provide for the welfare of the poor.

This explanation also implies that in those places that Joshua did not conquer, that is, regions that are neither in nor adjacent to ʾEreẓ Yisraʾel, there is no obligation to give tithes. This implication, however, contradicts Rabbenu Tam's basic position that, according to Halakhah (albeit not as practiced), terumah and tithes must be given even outside of ʾEreẓ Yisraʾel, regardless of the halakhic significance of the “first” and “second” consecrations.Footnote ⁹ To escape the implications of this passage, R. Tam wrote:

According to R. Tam, the explanation of the law transmitted to Moses at Sinai, beginning with “Why?” does not represent the words of R. Elazar himself, but the words of the “author of the Talmud.” R. Tam found support for his claim in a mishnah in tractate Yadayim, where the law appears without this explanation. That being the case, R. Tam preferred his own understanding of the tannaitic law that “in Ammon and Moab one separates the poor man's tithe in the Sabbatical year” to the one that originated in the academy of the “author of the Talmud.” There is no erasure of text, but there certainly is a differentiation between the basic halakhah and its explanation, and the exploitation of this division in order to dispute both the theoretical basis and the halakhic conclusions of a talmudic passage.

Another example: B. Ketubbot 31b addresses the appropriate punishment for stealing on the Sabbath. The principle that drives the passage is kim leh be-derabbah mineh (the more severe suffices for him). According to this principle, one who committed two transgressions with one act is subject only to the more severe punishment. A baraita that appears in this passage asserts, “One who steals a wallet on the Sabbath is obligated [to repay], because he was already liable for stealing before he violated a prohibition punishable by stoning. If he was dragging it out, he is exempt [from payment], for the violation of the Sabbath and the theft are simultaneous.” When one commits a theft on the Sabbath by dragging something from the owner's property into the public domain, the violation of the Sabbath (transporting something from the private domain to the public domain) and the act of theft (bringing the stolen object into the thief's possession) occur at the same time. In accordance with the principle of kim leh be-derabbah mineh, only the capital punishment incurred by Sabbath desecration applies.

However, the Talmud has difficulty finding a situation in which these two transgressions are violated simultaneously, for if the thief dragged the stolen object into the public domain, he has not acquired it, because there is no acquisition in the public domain. Alternatively, if he dragged it into a private domain, he has not desecrated the Sabbath. To resolve the problem, Rav Ashi proposes the following: “He linked his hands below three [handbreadths] and received it, and in accordance with Rava, for Rava said: A person's hand is considered like four [cubits] by four [cubits of land].” According to Rashi, Rava's words, “a person's hand,” were said with regard to putting things down on the Sabbath. R. Ashi uses this principle to assert that just as with respect to Sabbath violations a person's hand is considered a space where things can be placed, so too, with respect to acquisition “it is as though he lifted it up,” in Rashi's words.

The Tosafists found Rashi's explanation to be problematic.Footnote ¹¹ According to them, there should be no connection between the definition of a space for the purposes of the laws of the Sabbath and that of a space that enables acquisition. Moreover, the fact that a person's hand acquires on his behalf is a trivial principle that does not need support from Rava's statement, which was initially said with regard to the laws of the Sabbath. To resolve these problems, R. Tam, ignoring part of the text, suggests (ad loc.), “R. Tam's version does not have, ‘for Rava said: A person's hand is considered,’ but only, ‘in accordance with Rava’; nothing more.” According to R. Tam, the original passage referred only to Rava and did not mention any substantive statement. Only at a later stage were Rava's words, “A person's hand is considered like four by four,” introduced. In R. Tam's view, this explanatory reference is a mistake, and the passage actually had a different statement of Rava's in mind.Footnote ¹² As in the earlier example, here, too, R. Tam differentiated between the basic talmudic assertion, “in accordance with Rava,” and its interpretation, which included an erroneous reference. For our purposes, it is significant that, in contrast to the first example, in which R. Tam identified the source of these additions as the “author of the Talmud,” here R. Tam mentions neither the author of the explanation nor when he lived: Was it one of the “authors of the Talmud” or a copyist from a later era?Footnote ¹³

A third example: In B. Nedarim 19a, the Talmud compares the law stated in the Mishnah, “Uncertainties regarding vows are [treated] stringent[ly]” with another tannaitic source that states, “Liquid about which there is uncertainty whether it became unclean—is [itself] unclean. But as for rendering something else unclean—it [remains] pure. These are the words of Rabbi Meir, and Rabbi Elazar would say likewise.” The source of this passage is the Tosefta (T. Teharot 5:10), and R. Tam wondered why the author of this passage asked a question based on the Tosefta when he could have cited a mishnah (Teharot 4:9) that states simply: “Liquid about which there is uncertainty whether it became unclean—is [itself] unclean.” In a brief comment in Sefer ha-yashar, R. Tam posits that the shorter mishnah in Teharot, not the corresponding Tosefta passage, appeared in the “original” Nedarim passage. He explained how the longer Tosefta passage entered the discussion as follows: “Disciples wrote the baraita of the first chapter of Pesaḥim, where it is taught thus, in order to challenge this passage, and scribes mistakenly wrote it inside [the text].”Footnote ¹⁴ That is, R. Tam argues that a passage in B. Pesaḥim (16a) addresses the longer Tosefta passage, which ascribes the rulings to R. Elazar and R. Meir. In contrast, the Nedarim passage at the center of our discussion had no need for that additional information; for its purposes, it quoted only the mishnah that states “liquid about which there is uncertainty whether it became unclean—is [itself] unclean.” Students who noticed the similarity between these subjects and passages copied the Tosefta passage and the ensuing discussion in B. Pesaḥim, in the margins of their Nedarim texts, to serve as a sort of reference for the purposes of their own studies. The students’ books, which contained only the original laconic mishnah in Teharot in its main text, and which contained the Pesaḥim passage, including its citation of the Tosefta, in its margins, reached the hands of scribes who attempted to integrate the marginalia into the main text they were copying. Thus, the lengthier, more complicated Tosefta passage made it into the text of B. Nedarim, replacing the shorter line from the Mishnah, which was perfectly adequate for the purposes of that discussion.

Unlike the first two examples, this one is not about interpretive clauses that were added to an earlier dictum. However, like our earlier examples, here, too, R. Tam uses literary and philological tools to identify segments that were added to the Talmud at a later stage. Rabbenu Tam identified a marginal note, a gloss, of those who copied and studied the passage over the course of the generations. It is not attributed to the “author of the Talmud” as the earlier examples; this time, R. Tam attributes the additions to copyists of Talmud manuscripts, who merged what they found in the margins of the books they copied into the main part of the books they produced.Footnote ¹⁵

A fourth and final example: One of R. Tam's most dramatic and controversial innovations was based on this type of reading. Most medieval halakhists maintain that any amount of ḥameẓ (leavened grain, which is prohibited on the festival of Passover), even a trace amount, renders a mixture forbidden. R. Tam, however, rules permissively regarding a mixture that contains ḥameẓ, as long as the ḥameẓ component in the mixture is less than one-sixtieth of the total. This permissive ruling relies on R. Tam's identification of the words “in any amount, in accordance with Rav,” in the sentence, “The law is: ḥameẓ, at its time [of prohibition], whether [in a mixture] with its kind or with [things that are] not its kind, is forbidden in any amount, in accordance with Rav” [Pesaḥim 30a], as words that were added to the talmudic dictum by someone familiar with the compilation Halakhot Gedolot, which was attributed by medieval French scholars to the eighth-century Yehudai Gaon. As Rabbenu Tam put it:

Rabbenu Tam's demonstration is as follows: Halakhot Gedolot states that since the Talmud's discussion makes no assertions about the quantity that renders the mixture forbidden, it is clear that the mixture is rendered forbidden by any quantity of ḥameẓ, even the tiniest amount. In the words of Halakhot Gedolot: “The fact that [the Talmud] assigned no measure means that [it is forbidden] in any amount, in accordance with Rav.” R. Tam derived from this passage that “‘in any amount, in accordance with Rav’ was not in the books [of the Talmud]” and that “apparently Rabbi Yehudai ruled thus, and as a result, this is how things were written in the new books.”

Rabbenu Tam's argument, and all the more so his subsequent ruling, provoked a major controversy at that time. (Simcha Emanuel has made an important contribution to our understanding of one side of this controversy.)Footnote ¹⁷ For our purposes, the important point is that in every instance surveyed, R. Tam identified units that appeared in his Talmud as being later addenda to the primary text of the Talmud. Once he made this differentiation, he did not hesitate to offer new interpretations of the talmudic dicta themselves, and, as we saw in the last case, even to render novel halakhic rulings on the basis of his new interpretation, while ignoring the interpretation or ruling that appears in the extant copies of the Talmud.

In all of these cases, there is no indication that R. Tam emended the text of the Talmud. According to R. Tam, he was following the policy of his grandfather, Rashi:

It emerges from all of these examples that in Rabbenu Tam's view it is legitimate for a scholar to suggest emendations of the talmudic text as long as he does not tamper with the text of the Talmud itself.

This very issue was the central point of R. Tam's polemic against Rabbi Meshullam of Melun in the 1150s.Footnote ²¹ Close scrutiny of R. Tam's words shows that his anger at R. Meshullam was mainly because of his suspicion that the latter had emended the text of the Talmud itself:Footnote ²² “For when they don't know, they mislead the world, and when they are confounded, they emend the books. But you are worse than them, for they write ‘HG’ [= hakhi garsinan, which means “this is our reading”] before their emendations, so their work is discernable. It is not so with you; rather, you do not indicate your emendations, and others rely on you and are misled. Moreover, they emend one or two words, whereas you [emend] twenty-two.” Rabbenu Tam's central claim was that R. Meshullam emended the books themselves, to the point that the reader cannot distinguish the original text from the emendations. It is true that R. Tam also criticized the flippancy with which R. Meshullam and others like him emended the Talmud, but the main thrust of his polemic was directed at emendations made in the main text of exemplars of the Talmud itself. R. Meshullam responded,Footnote ²³ “Yet I am steadfast that I have never erased from a book. Rather, I study your variant, and then I show them my variant, and some write it in a notebook.” R. Meshullam thus understood—and may even have agreed with—R. Tam's primary opposition, which was to emendation of the talmudic text without leaving indication of having done so. This comports with the examples addressed above.

In each of the examples cited, R. Tam identifies a part of the text that was added at a later stage. However, these additions are attributed to different eras and can therefore shed light on R. Tam's thinking about the genesis of the talmudic text.  In the fourth example above, the source of the words in the Talmud, “in any amount, in accordance with Rav,” were identified by R. Tam as originating from Halakhot Gedolot, under whose influence “error-prone exegetes wrote.”Footnote ²⁴ Likewise, in the third example, about “uncertainties regarding vows are [treated] stringent[ly],” R. Tam identified the addition as a student's marginal gloss that was later integrated into the main text.Footnote ²⁵ His interpretation, like the one discussed before it, reflects an awareness of an existing Talmud, into which errors and emendations creep as part of the processes by which written material is transmitted and studied.

In contrast to these cases, when R. Tam addresses the first case, “Ammon and Moab separate the poor man's tithe during the Sabbatical year,” he attributes what he identifies as an addition to the “author of the Talmud” (baʿal ha-Talmud).Footnote ²⁶ This expression, which appears frequently in R. Tam's writings, refers to one of the sages responsible for the particular talmudic passage—sometimes its redactor and sometimes the Amoraim named in the discussion. Thus, in his treatment of B. Shabbat 101b, R. Tam wrote:Footnote ²⁷ “The mishnah is not thus. Rather, it is typical of the Talmud to expand on it [according to] what it knows of its interpretation … and the author of the Talmud, within this expanded interpretation, teaches us.”

Taken together, we see that Rabbenu Tam identified additions to the Talmud from several different eras and thus that he viewed these additions as part of a prolonged process that characterized the production of the Talmud in its entirety—beginning in the first generations of Amoraim, whose explanations became an integral part of the Talmud, and extending all the way to the medieval era, to the addenda of students and copyists who tinkered in the margins and tampered with the main text. R. Tam related to such addenda again at the end of his introduction to Sefer ha-yashar, where he wrote:Footnote ²⁸

The “additions that they made in order to change the face of the matter in the interpretation of Halakhah” are, in my opinion, those same additions that we surveyed above—additions made to the Talmud because students did not understand the passage. Yet Rabbenu Tam does not object to these because they had already penetrated the text of the Talmud and became part and parcel thereof. One may suggest an alternative interpretation to these additions, and even deny that the interpretations implied by these additions are authoritative. However, according to R. Tam, one may never “object to them,” that is, emend or delete them, “because they were not deleted from the Talmud.” Even a segment that was added to the Talmud at a later stage is canonized, as it were, from the moment it is absorbed into the written talmudic text, and it cannot be erased from the text of the Talmud.Footnote ³¹

A responsum of R. Isaac of Dampierre, the nephew and disciple of R. Tam, offers insight into how R. Tam dealt with passages that he would not delete but believed should be emended, “I examined R. Tam's [copy of] the order of Kodashim, and I saw that it is written therein ‘the law is in accordance with the sages,’ and he made horns [karnayim] to indicate that this is an error, and that ‘the law is in accordance with the sages’ should not be part of the text.”Footnote ³²

Rabbenu Tam thus tampered with the text of his own books, but this tampering, the marking with “horns,” allowed for clear differentiation between the text that was initially in the book and R. Tam's suggested emendation. This is consistent with his remarks to R. Meshullam of Melun seen above. It stands to reason that R. Tam, who objected, in many and varied places, to any tampering with the text found in books, applied this method consistently. The attestation of R. Isaac can be generalized: in every case where we find attestations that R. Tam emended his books, the meaning is not that he deleted any part of the text, but that he made a discernable change, by using “horns” and other graphic symbols that prevent one from ignoring the problematic, rejected text, yet enable a future reader to preserve the old version alongside the suggested emendation.

Another attestation of R. Isaac is similar: “In a book containing [the orders of] Yerushalmi Zeraʿim and Moʿed that belonged to Rabbenu Jacob [= R. Tam], it is written, at the beginning, that R. Asi said, etc., as I have written. It seems to be the writing of R. Solomon, the son of R. Jacob. At the end of the passage, our master [=R. Tam] wrote, in his own handwriting, which is familiar to me: ‘This is the custom everywhere.’”Footnote ³³ Regardless of whether R. Solomon was the copyist of his father's entire Yerushalmi volume or only glossed specific words, the key point is that when R. Tam added something to this volume, R. Isaac could identify that it was in a different handwriting.

Thus, R. Tam cautioned against changing the text of the Talmud itself without leaving any trace of what was there before. His instruction was very simple: Alternative suggestions may appear in the margins of manuscripts of the Talmud, as long as they are distinct from the older version, whose clear and unambiguous presence will persist. To that end, R. Tam used the aforementioned “horns”; it stands to reason that he indicated additional suggested emendations in a similar way. The “horns” should therefore be viewed like parentheses in printed texts, and it is reasonable to presume that other symbols indicated other insights.

The many instances in which there is documentation that R. Tam emended or erased the Talmud text should therefore be seen as instances in which he “glossed” the text of his Talmud in a way that preserves the old text alongside the newly proposed version. These are not deletions that erase all trace of what came before, but glosses that say what they say while enabling the rejected version to persist in the sight of the glossator.Footnote ³⁴  This policy may be implicit in R. Tam's letter to R. Meshullam: “Do not be angered by my words, for I said to our masters, face to face, that it should be decreed against some of them that they not convene a yeshiva. For when they don't know, they mislead the world, and when they are confounded, they emend the books. But you are worse than them, for they write ‘HG’ before their emendations, so their work is detectable. It is not so with you; rather, you do not indicate your emendations, and others rely on you and are misled.”Footnote ³⁵ Suggested emendations are undesirable, but sometimes they are unavoidable. What R. Tam forbade completely, and what he deemed unconscionable, was making a proposed textual emendation into a part of the text itself, not merely a suggestion in a commentary or in the margins. Only when the old text is completely effaced and a new one takes its place does R. Tam use the bans of Rabbenu Gershom and attack R. Meshullam and those like him.

The words from R. Tam's introduction to Sefer ha-yashar, “And this stands to reason: If a person does not know the Halakhah, he should write its interpretation as he sees fit, if he wishes. However, he should not delete from books,” are clear.Footnote ³⁶ Suggested emendations, when warranted, were precisely delimited and indicated by various graphic symbols. Their role could be ascertained only once the wellsprings of commentary had been dredged.

The scholarly tradition briefly summarized above—the tradition of Weiss, Aptowitzer, and Urbach—became enthralled, in my opinion, by the turbulent, fascinating, self-contradictory, and somewhat aggressive figure of R. Tam. Indeed, there is no doubt that the figure of Rabbenu Tam has all the components that make for a compelling character. Ultimately, however, a figure whose impact on the history of Halakhah was so significant and decisive cannot be explained solely through a description of his psychological traits.Footnote ³⁷ In this article, along with my article on R. Tam's aforementioned polemic against R. Meshullam of Melun,Footnote ³⁸ I attempt to pierce the fog of R. Tam's stormy personality and address the substance of his debates. It is precisely this substance that is lacking in these earlier indispensable studies.