What a discourse unit can teach us about the relationship between the Mishnah and Tosefta: an examination of the parallelism between the compilations in the halakhic give-and-take conversation

Abstract

This article presents an examination of the parallelism between the Mishnah and Tosefta in one discourse unit – the halakhic give-and-take conversation. It aims to show that a description of discourse units found in both compilations can contribute to the discussion of the relationship between the two compilations and the status of the Tosefta in regard to Mishnah. In the examined corpus of halakhic give-and-take conversations from the Mishnah and Tosefta from three orders, for only 16 conversations in the Tosefta (14%) was there found a parallel conversation in the Mishnah, and in most cases the parallels are not identical. The structural and linguistic comparison between these 16 parallel conversations showed that the conversations in the Tosefta contain more exchanges as well as more complete exchanges, and that the language in the Tosefta seems less redacted and earlier compared with the language in the parallel conversations in the Mishnah.

1. The relationship between the Mishnah and Tosefta in Talmudic and linguistic studies

The question of the relationship between the Mishnah and the Tosefta and of the status of the Tosefta concerned Jewish scholars already in the Middle Ages, as well as scholars of Talmudic literature from the period of the Enlightenment up to the present. Various views have been presented in the context of the discussion on these questions, reflecting different traditional, historical or literary approaches.¹

Houtman (1996, vol. 1: 1–2) distinguishes three broad streams in the research in regard to this issue: (1) the predominant stream, which regards the Tosefta as a nondependent corpus that derives its value from its relationship to the Mishnah; (2) the stream that regards the Tosefta as a collection of beraitot, and regards it and the Mishnah as two branches of the same tree; (3) the stream that regards the Tosefta as the oldest remnant of the Palestinian Mishnah. Brody (2015) describes the shift in trend of the research in relation to this issue: at first scholars emphasized the characterization of the Tosefta as dependent on the Mishnah and supplementing it, while in the last several decades scholars have given greater emphasis to the preservation of ancient material in the Tosefta.

The later trend in the research can be found in the writings of different scholars, some of whom will be noted here. Hauptman (2001, 2005) maintains that the Tosefta is both a commentary on an earlier text, a forerunner of the Mishnah, which she calls “urMishnah”, and also the basis of the Mishnah. Houtman (1996) claims that the Tosefta is a kind of Talmud on the Mishnah, that is readable without the Mishnah; this is because both of them drew their material from an older composition, which was redacted first in the Mishnah, and then reproduced in a different manner by the compilers of the Tosefta. Feliks (2004) also supports the view that the Tosefta contains a tradition that predates the Mishnah, and in his view, the direction of development was as follows: Tosefta > Mishnah > Talmuds.

Friedman (2013a) has conducted a detailed investigation of Mishnah–Tosefta parallels. He claims that there are many halakhot in the Tosefta that are primary to their counterparts in the Mishnah, but in his view, this view cannot be applied to both compilations as a whole. Brody (2015) agrees with Friedman that in some cases, the Tosefta preserves sources, but rejects the tendency to look for an explanation for every case of parallel halakhot. Brody takes the view that the Tosefta was redacted following the Mishnah subject to the order of the debate in it, that it is an “anthology and collection” of sources, in which, unlike in the Mishnah, it is difficult to find traces of active redaction, and that the Tosefta contains additions and commentaries on the Mishnah, as well as earlier halakhot, from which certain mishnayoth were redacted.

Most of the scholars who presented their views on the subject of the relationship between the Mishnah and the Tosefta examined parallel texts in the two compilations and sometimes in other compilations as well.²

Friedman (2013b: 5) characterizes the halakhot of the Tosefta which relate to the Mishnah to one of the following three situations: 1) “literary dependence” – the Tosefta responds to a halakha included in the Mishnah, which it explicates or broadens; 2) the Tosefta transmits an independent halakha that is absent from the Mishnah but is linked in some fashion to halakhic or aggadic topics discussed in the same tractate of the Mishnah; 3) the entire halakha is found both in the Mishnah and in the Tosefta, in its order and style, except for smaller or larger differences of wording that characterize each of the two parallels.

Friedman (2013a) focuses in his work on the third situation, and claims that the Tosefta preserves the earlier form of the halakha, and the parallel in the Mishnah was revised and altered, apparently by R. Judah ha-Nasi (“the Patriarch”) (Friedman 2013c).

According to Friedman, at least a significant proportion of the parallels of the Mishnah and Tosefta are “edited parallels”, i.e., independent parallels created as a result of a free transmission of a common and ancient source. The Tosefta as a complete compilation is a collection arranged after the Mishnah; to it were added additions and an interpretation of the Mishnah, on the one hand, along with ancient halakhot from which R. Judah ha-Nasi redacted the Mishnah. Friedman calls the Mishnah a “compilation”, whereas the Tosefta is a “collection”, arguing that in the Mishnah the text is more linguistically processed and updated. Friedman assumes that the parallel in the Tosefta came first and that the parallel version in the Mishnah was redacted from it.

The linguistic research contains relatively few references to linguistic differences between the Mishnah and the Tosefta.

Nathan (1984: 324–45) examined the Erfurt manuscript of the Tosefta and compared it with the Kaufmann codex of the Mishnah and, according to her findings, there are almost no differences in the grammatical systems between the Mishnah and the Tosefta, and the differences are concentrated mainly in the areas of the lexicon and syntax (of which only a small part was examined). Bar-Asher (2009, vol. 1: 43) mentions Nathan's observations and states: “This issue – the linguistic distinction between the various layers of the Tannaitic literature – awaits a comprehensive, in-depth and thorough study” (p. 43, originally in Hebrew).

In his studies, Braverman (1985, 1993 and 1995) explored the differences between the Mishnah and the Tosefta, presenting syntactic differences between them; conjunctive elements and emphatic and supplementary elements, which are more common in the Tosefta than in the Mishnah; as well as differences in specific expressions (Braverman 1985). He also refers to lexical differences between the two compilations (Braverman 1985, 1993), as the cases where the noun in the Mishnah is in Hebrew, whereas in the Tosefta the noun is foreign.

Braverman argues that the linguistic and stylistic differences between the Mishnah and the Tosefta stem from the difference in how the compilations were redacted. The Mishnah largely represents the language and style of its redactor (but not in the more ancient mishnayoth and in the unique language of the Tannaitic scholars), while the Tosefta represents the different dialects and linguistic styles of various Sages.

Feliks (2004) presents numerous tools to determine which of the parallel texts in the Mishnah and Tosefta came later, some of them are of a linguistic nature.

The review in this section appears to indicate that a linguistic study comparing the Mishnah to the Tosefta is definitely called for, and that a linguistic study of this kind may advance research into the relationship between these two compilations. I will present three studies from recent years that addressed an analysis of structures and discourse units in the Mishnah and Tosefta.

In her study about the structures of the tetralemma and trilemma in the Mishnah and Tosefta, Tubul-Kahana (2011) shows that in most cases, the tetralemmas and trilemmas of the Mishnah reflect a more careful and masterful redaction of the structures as opposed to those of the Tosefta. She reveals differences between the two compilations, and finds that in the parallels from the two compilations, the Mishnah is usually more redacted than the Tosefta. From the findings of her study, she concludes that this analysis can reinforce the conclusion that the Tosefta preceded the Mishnah.

In a further study (Kahana [Tubul] 2019), Tubul-Kahana describes numerical sayings in the Mishnah and Tosefta, that is structures of a list of items with a shared feature, with a heading that presents the subject and the number of items in the group. She reveals that in numerical terms, there are more sayings attributed to the Mishnah compared to the Tosefta. She views this difference as further support for the conclusion that the Mishnah is more tightly redacted than the Tosefta.

Also, in my study on the בראשונה (at first) discourse unit (Shemesh 2008b: 70–7), a number of differences were found between the Mishnah and the Tosefta, which may show the distinction between the two compilations, and these are reflected in different components of this unit.

At the centre of these studies are structures or discourse units in the Mishnah and Tosefta that are of a linguistic nature. I believe that further linguistic studies that analyse structures and discourse units in both compilations will be able to advance the linguistic study of Mishnaic Hebrew as a whole, and that an in-depth study of the linguistic features of the structures and discourse units can contribute to the general discussion of the relationship between the Mishnah and Tosefta and the status of the Tosefta.

It should be clarified that these linguistic studies differ from the Talmudic studies described in the above review in two respects: first, the Talmudic studies focused on examining parallels from the Mishnah and Tosefta without a common denominator shared by the texts under examination, whereas the linguistic studies proposed here would first describe the linguistic features of structures and discourse units, and only afterwards move on to a comparison between what is found in the two compilations; and second, unlike the Talmudic studies that examine various differences between the Mishnah and Tosefta, the proposed linguistic studies focus on linguistic aspects, and thus can further our knowledge of the various linguistic phenomena present in Mishnaic Hebrew. Linguistic differences between the Mishnah and Tosefta were raised in Braverman's studies, as noted in the above review, but he did not examine them in the context of the linguistic structures or discourse units shared by the two compilations.

The concept “discourse unit” was presented by Enkvist (1978) in order to describe a unit which is larger than a sentence and a part of a text, or a unit which is a separate text in itself.³

2. An examination of the parallelism between the Mishnah and Tosefta in the halakhic give-and-take conversation

In this section, I will present an examination of the parallelism between the Mishnah and Tosefta in a single type of discourse unit – the halakhic give-and-take conversation. The examination will focus on both structural and linguistic comparisons between the parallel conversations (in subsections 2.4.2–2.4.3). The aim of this examination is to draw conclusions from the study of this discourse unit regarding the relationship between the Mishnah and Tosefta and the status of the Tosefta in regard to the Mishnah.

The halakhic give-and-take conversation is a discourse unit that presents the views of sages regarding a halakhic issue. Occasionally, the presentation of the views is followed by a debate between the parties regarding their views. In such a case, a halakhic give-and-take conversation takes place. The conversation is composed of one or more exchanges; each exchange contains the words of the addressor and the addressee, or solely the words of the addressor.⁴

The halakhic give-and-take conversation in the Mishnah is the subject of my book Halakhic Give-and-Take Conversations in the Mishnah (Shemesh-Raiskin 2022). This survey involves 190 halakhic conversations in the Mishnah, which includes 240 exchanges. Later, I collected a corpus of the halakhic give-and-take conversations that appear in three orders of the Tosefta: Zera‘im, Moe‘d and Nashim. This corpus involves 118 conversations that include 172 exchanges.

The various sections of this section will describe the parallelism between the Mishnah and Tosefta in halakhic give-and-take conversations.

2.1. The inventory of the halakhic give-and-take conversations in the Mishnah and Tosefta included in the examination

To examine the parallelism between the Mishnah and Tosefta in halakhic give-and-take conversations, of the total of 190 conversations in the Mishnah, only those conversations found in three orders were counted: Zera'im, Moed and Nashim – corresponding to the same three orders examined in the Tosefta, as stated. It was found that these orders in the Mishnah contain 79 conversations, compared to 118 conversations compiled from these orders in the Tosefta. The quantitative disparity between the number of conversations in the two compilations – 79 compared to 118 – is at a ratio of 1:1.5. This ratio is compatible with the ratio between the two compilations in terms of the number of words they contain – 1:1.6.⁵ In other words, the disparity between the number of halakhic give-and-take conversations in the Mishnah and Tosefta is consistent with the disparity between the sizes of the two compilations.

2.2. The extent of the overall parallelism between halakhic give-and-take conversations in the Mishnah and Tosefta

At the time of the collection of the corpus of halakhic give-and-take conversations from the three orders in the Tosefta, the extent of the parallelism of these conversations to conversations in the Mishnah was examined. In other words, an effort was made to identify a corresponding passage in the Mishnah, and if the corresponding passage in the Mishnah was found – to see if a conversation also appears in passage in the Mishnah, and if a conversation indeed appears – an effort was made to see if the conversation in the Mishnah parallels the halakhic give-and-take conversation in the Tosefta.⁶ Even if the search did not produce identification of all the corresponding passages and conversations in the Mishnah, those identified by this examination make it possible to assess the extent of the parallelism between the Mishnah and Tosefta regarding this particular discourse unit.

Figure 1 shows the parallelism to the Mishnah from the examination of the corpus of halakhic give-and-take conversations in the Tosefta.

Figure 1. The parallelism to the Mishnah in the corpus of halakhic give-and-take conversations in the Tosefta

As shown in Figure 1, the conversations included in the corpus of halakhic give-and-take conversations in the Tosefta can be divided into three types:

1. 1. About one-half of the conversations in the Tosefta (47% – 56 conversations) have no parallel in the Mishnah;

2. 2. For 39% of the conversations (46), a parallel was found in the Mishnah, but without a conversation;

3. 3. For 14% of the conversations in the Tosefta (16), a parallel conversation was found in the Mishnah, which in most cases is not identical to the conversation in the Tosefta.

About half of the conversations in the Tosefta are included in the first type of conversation, those for which no parallel passage in the Mishnah was found. In other words, this is the most frequent situation in the discourse unit of the halakhic give-and-take conversations. The conversations of this type can be attributed to the second type presented by Friedman (2013b: 5, as noted in Section 1 above), in which the Tosefta transmits an independent halakha that is absent in the Mishnah, but is linked in some fashion to topics discussed in the same tractate of the Mishnah.

The conversations in the Tosefta for which there is a parallel discourse unit in the Mishnah that does not include a conversation (type 2) will be briefly described in Section 2.3. Section 2.4 will discuss the conversations in the Tosefta for which there is a parallel conversation in the Mishnah (type 3).

2.3. Halakhic give-and-take conversations in the Tosefta that have a parallel without a conversation in the Mishnah

As noted in Section 2.2 above, a large proportion of the halakhic give-and-take conversations in the Tosefta corpus of conversations was found to have a parallel in the Mishnah that does not contain a conversation – 39% (46 conversations – type 2 of conversations there).

An examination of the parallelism between the two compilations in this type of conversations found that in most (32 out of 46 conversations = 70%), a passage appears in the Tosefta that begins with a presentation of halakhic views followed by a halakhic give-and-take conversation between the interlocutors, whereas in the parallel in the Mishnah, the presentation of the views is not followed by a conversation.⁷ For example, the two corresponding passages from the Mishnah and the Tosefta presented in Citation 1⁸ begin with the presentation of views regarding the law that applies to a sukkah that is higher than twenty cubits:

The presentation of views is almost identical in the two compilations (in the Tosefta, the word למעלה [more] is added for clarification). In the Tosefta (but not in the Mishnah), this is followed by a conversation between the holders of the views R. Judah and the sages, which includes two exchanges. In the first, R. Judah presents a מעשה story aimed at bolstering his halakhic opinion that a sukkah of this kind is halakhically valid, because the story describes a case whereby a sukkah of this kind was erected and elders entered it and not one of them protested to the woman who owned the sukkah about its height; and the sages in response offer an explanation for the case as it is described, in order to argue that this case cannot serve as an example for the type of sukkah in question. In the second exchange, R. Judah asks a rhetorical question in response to the sages’ explanation, and claims that the story actually can serve as an example because that woman had seven sons who were disciples of sages who sat in the sukkah.

It should be clarified that sometimes in this type of parallels from the Mishnah and Tosefta, there can be differences between the passages in the part of the presentation of the views too. For example, in the corresponding passages in Citation 2, the presentation of views in the Mishnah and Tosefta is not identical, followed by a halakhic give-and-take conversation only in the Tosefta:

Both passages discuss the question: from when one may, at the end of the seventh year (i.e. shmittah year), buy arum from a person who is suspected of not observing the seventh year, without fearing that the arum was grown during the seventh year? The question is formulated similarly in both passages (the Tosefta adds the words מכל מקום [“under any circumstances”] at the end of the question to make it clear that this also involves buying from someone suspected of not observing the seventh year). Following this question, the views on this matter are presented: In the Mishnah, the views are of R. Judah and of the sages, whereas in the Tosefta, only R. Judah's view is presented; in other words, the two passages differ with respect to the part that presents the views too. The presentation of the views in the Tosefta is followed by a halakhic give-and-take conversation between R. Judah and R. Yosé, which is absent from the Mishnah. R. Judah tells a story to bolster his opinion that it is permitted to eat arum immediately upon the conclusion of the seventh year, which describes the eating of arum at the end of Sukkot following the completion of a seventh year⁹ authorized by R. Tarfon; and Rabbi Yosé asks a rhetorical question that expresses his opposition to inferring from the story that this is permitted, following which he makes a claim regarding another festival in which the event occurred – the end of Passover, based on his own personal knowledge as one who was present at the event. Differences in the part of the presentation of views can be found also in Citation 8 below, as will be discussed in the explanation of this citation in subsection 2.4.2.

When examining conversations of this kind – those in the Tosefta that have a parallel in the Mishnah but without a conversation – we find that the Tosefta includes an addition that does not appear in the Mishnah. Conversations of this kind in the Tosefta appear to belong to the situation that Friedman (2013b: 5) calls “literary dependence”, i.e. a situation in which the Tosefta responds to a halakha included in the Mishnah, which it explicates or expands on (as detailed in Part 1 above).

2.4. Halakhic give-and-take conversations in the Tosefta that have a parallel conversation in the Mishnah

As presented in Section 2.2 above, for a small proportion of the halakhic give-and-take conversations in the Tosefta – 14% (16 conversations – type 3 of the conversations there) – a parallel was found in the Mishnah that contains a parallel conversation, which is generally not identical to the conversation in the Tosefta. These conversations can be ascribed to the third situation presented in Friedman (2013b: 5) – the situation of literary parallels in the Mishnah and Tosefta, as noted in Section 1 above.

This section will focus on an examination of the 16 parallel halakhic give-and-take conversations in the Mishnah and Tosefta.

2.4.1. The extent of the parallelism between the halakhic give-and-take conversations in the Mishnah and Tosefta

An examination of the halakhic give-and-take conversations in the Mishnah and Tosefta found that three situations of parallelism between the conversations can be distinguished:

(1) The parallel conversations are identical: Of the 16 parallels in the Mishnah and Tosefta, in only one conversation (6.25%) were the two passages almost identical:

This passage begins with the presentation of the views of the House of Shammai and the House of Hillel on the matter of bringing the dough-offering and gifts given to the priest on a festival, and the presentation of views is dissimilar in the two passages. It is followed by a halakhic give-and-take conversation that is identical in the two compilations, which includes a single exchange in which the House of Shammai argues an analogy while the House of Hillel rejects it. The two parallel conversations in the Mishnah and Tosefta are identical, the only difference between them is that the introductory pattern in the first part of the exchange in the Tosefta is אמרו בית שמיי, whereas in the Mishnah, there is the addition of a complement (אמרו בית שמיי לבית הלל).

(2) The parallel conversations are partially similar: In half of the parallel conversations in the Mishnah and Tosefta (8 out of the 16 conversations in the Tosefta corpus = 50%), one of the parts of the exchange in the conversation is identical or similar in both compilations.

For example, the corresponding passages in Citation 4 begin with the formulation of a law related to the drawing of water from a water-channel located in a shared courtyard, and it is formulated in the Mishnah more briefly; this is followed in the Tosefta by the presentation of R. Judah's view, which is opposed to what was stated previously, and which does not appear in the Mishnah.¹⁰ This in turn is followed by a conversation between R. Judah and אמרו לו, which is not identical in the two parallels:

In this conversation, in the first part of the exchange, R. Judah presents a story about a water-channel that passed from a city in the Upper Galilee called Abel, from which the sages allowed water to be drawn on the Sabbath without a special partition; the story is presented to bolster R. Judah's view (presented, as noted, only in the corresponding passage in the Tosefta), according to which the courtyard wall is considered a partition, and consequently the water-channel does not require a special partition. This part is similar in both passages. In the second part of the exchange, אמרו לו present an explanation for the sages’ dispensation in the case described in the story, which is the reason why they refuse to accept the story as evidence to bolster R. Judah's view; the explanation is that the water-channel was not the size of the one that is the subject of this discussion and consequently, the law for it is different. This part is formulated differently in the two passages, but the intention of the content is identical. The corresponding conversations in Citation 4 are then similar for both parts of the exchange, with differences between them in content and language (for the linguistic differences between the two passages in both parts, see subsection 2.4.3 below).

Further examples of partial similarity between the parallel conversations can be seen in the two corresponding passages in Citation 7 below, in which the first part of the exchange (the only one in the Mishnah and the first in the Tosefta) is almost identical: הואיל וטהורה אסורה לזרים ואף טמאה אסורה לכהנים, מה טהורה עולה אף טמאה תעלה (“Since […] and also […]”), and in the Tosefta –טהורה אסורה לזרים וטמיאה אסורה לכהנים […] (for more on these particles, see subsection 2.4.3 below); and in Citation 10 below – in the second part of the exchange (the only one in the Mishnah and the first in the Tosefta): ¹¹.בחדשים אנו בושים/ן אלא שאתם מגלגלין עלינו את הי(י)שנים

(3) The parallel conversations are different: Fewer than half of the parallel conversations in the Mishnah and Tosefta (7 out of the 16 conversations in the Tosefta corpus = 43.75%) are different from one another in their content, as will be shown in these two citations.¹²

In the corresponding passages in Citation 5, views are presented regarding two types of figs (white figs and Persian figs), whose fruits have different laws regarding the seventh year because they take three and two years (respectively) to produce fruit. R. Judah's view, which appears in an identical form in both passages, is that the laws of the seventh year apply to the Persian figs upon the conclusion of the seventh year because their fruits take two years to grow. This is followed by a conversation in the two passages containing a partial exchange, in which a different view from אמרו לו is presented:

The arguments presented in the two parallel conversations in this citation are different in content. In the Mishnah, אמרו לו argue: לא אמרו אלא בנות שוח; in their view, the sages ruled that the law of the seventh year applies only to the white figs but not to the Persian ones. However, in the Tosefta, they argue: הרי הן אצלך בטבריה, ואינן עושות אלא בנות שנתן; that R. Judah's view is influenced by where he lives, whereas in Tiberias, the fruits grow within a single year, and consequently, his view is not accepted.

In Citation 6, an identical presentation of views regarding fruits bought from the money of the second tithe that became blemished appears in the corresponding passages, followed by a different conversation between אמרו לו and R. Judah:

The conversation in the Mishnah includes one exchange: in the first part, there is an a fortiori question by אמרו לו, who infer from the second tithe regarding fruits bought with second tithe money, and in the second part R. Judah's argument appears, which opposes the a fortiori argument because of the extra stringency that should be taken in the case of fruits. In the Tosefta, the conversation is different and includes two exchanges: in the first, אמרו לו ask a rhetorical question regarding the sage's view, which is intended to express opposition to the stringency regarding the secondary fruit in comparison to the money itself, and in response, R. Judah reinforces his view from another case in which even greater stringency is taken with regard to the secondary than with the primary; and in the second exchange, אמרו לו maintain that their view can be proved from the case the sage presented, by means of an analogy.

This subsection described the extent of the overall parallel between the parallel halakhic give-and-take conversations in the Mishnah and Tosefta. The examination showed three situations of parallelism: a situation (1) of identical conversations was found for only one conversation (6.25%); a situation (2) of partial similarity between the conversations was found in half of them (50%), when one part of the exchange appears in an identical or similar form in the parallels; and a situation (3), in which the parallel conversations are different from one another in content, was found in fewer than half of the parallel conversations in the Mishnah and Tosefta (43.75%). The division into three situations shows that there is little similarity between the parallel conversations in the Mishnah and Tosefta. This finding is even more striking when we consider that the proportion of these parallel conversations from the entirety of the conversations in the corpus of halakhic give-and-take conversations in the Tosefta is small. As shown in Section 2.2 above, only 14% of all the conversations (16 out of a total of 118) are included in this type of conversation in the Tosefta, and the rest of the conversations in the Tosefta do not have a parallel in the Mishnah, or if there is one, the parallel in the Mishnah does not contain a conversation.

An examination of the discourse unit of the halakhic give-and-take conversation shows a low degree of parallelism between the Tosefta and Mishnah. It would appear that only after the other discourse units in the Mishnah and Tosefta are examined, and the findings obtained regarding the extent of parallelism between the two compilations for these discourse units will it be possible to compare the full extent of parallelism between the Mishnah and Tosefta to the findings described here regarding the extent of the parallelism between the two compilations for the specific discourse unit that was examined. Only then will it be possible to understand the nature of the findings regarding the overall parallelism between the parallel halakhic give-and-take conversations in the Mishnah and Tosefta.¹³

2.4.2. Structural comparison between parallel halakhic give-and-take conversations in the Mishnah and Tosefta

This subsection will conduct an overall structural comparison between the halakhic give-and-take conversations in the Mishnah and Tosefta, and subsection 2.4.3 will conduct a linguistic comparison between them. The comparisons will relate, as noted, to the parallel conversations from both compilations, i.e. to the 16 conversations out of the 118 in the corpus of Tosefta conversations (14% of the total number of conversations).

The structural comparison found that in more than one-third of the parallel conversations in the Mishnah and Tosefta (6 of the 16 conversations = 37.5%), no difference was found between the conversations in terms of their structure, i.e. a structure containing the same number of exchanges appears in both conversations. For example, one exchange appears in the two almost identical conversations in Citation 3 above, and the same is true for the two similar conversations in Citation 4 above; and one partial exchange appears in two conversations that are different from one another in Citation 5 above.

Nevertheless, in most of the parallel conversations – 62.5% (10) – : (1) the number of exchanges included in the conversations in the Tosefta is greater than the number of exchanges in parallel conversations in the Mishnah, or (2) the exchanges in the Tosefta are complete compared to the partial exchanges found in the parallel conversations in the Mishnah, as will be shown below.

(1) A larger number of exchanges in the conversations in the Tosefta compared to the Mishnah can be seen in Citation 6 above as well as in Citation 10 below,¹⁴ in each of which there is one exchange in the conversation in the Mishnah, whereas in the parallel conversation in the Tosefta there are two exchanges (the second of which is partial). In Citation 7, there is one exchange in the conversation in the Mishnah, whereas in the Tosefta there are three exchanges:

In the corresponding passages in this citation, the negotiation begins with the presentation of the views of the House of Shammai and the House of Hillel, after which the halakhic give-and-take conversation between them regarding their views begins. In both conversations, there is a parallel exchange, which is the only one in the Mishnah and is the first in the Tosefta: its first part is almost identical and contains an a fortiori argument presented by the House of Hillel; the second part is different, but both conversations contain a rejection of the a fortiori argument. The Tosefta contains a continuation of the conversation, which includes two additional exchanges: in the second exchange, the House of Hillel presents evidence from another case, which contradicts the words of the House of Shammai in the first exchange, and the House of Shammai responds to what was said: and in the third, partial exchange, the House of Hillel responds to the words of the House of Shammai.

(2) Complete exchanges in the conversations in the Tosefta compared to partial exchanges in the Mishnah can be seen in Citation 8:

This citation begins with a presentation of views: of the Tanna Kama (“the first Tanna”) and of R. Judah in the Mishnah, and of R. Meir and of sages in the Tosefta. This is followed by a halakhic give-and-take conversation: in the Mishnah it contains a partial exchange, which is almost identical to the first part of the exchange in the Tosefta; whereas in the Mishnah it is attributed to אמרו לו, in the Tosefta it is attributed to R. Meir (and see Section 2.3 above regarding the differences between the compilations in the part that contains the presentation of views). In the Tosefta, a second part of the exchange appears, and it is said by אמרו לו.

No cases were found of a halakhic give-and-take conversation that was longer in the Mishnah than its parallel conversation in the Tosefta. As noted in this subsection, the most common situation (found in 62.5% of the parallel conversations) is that the conversations in the Tosefta have a longer structure than their counterparts in the Mishnah, i.e. the conversations in the Tosefta contain more exchanges or complete exchanges; the least common situation (37.5%) is of an equal number of exchanges in the parallels in the two compilations.

2.4.3. Linguistic comparison between the parallel halakhic give-and-take conversations in the Mishnah and Tosefta

The parallel halakhic give-and-take conversations in the Mishnah and Tosefta were compared to one another not only in terms of their structure, but also in linguistic terms. Three types of linguistic differences were found between the corresponding parts in the parallel conversations in the Mishnah and Tosefta, as enumerated below:

1) Connectors and question particles in the Mishnah that are absent from the parallel conversations in the Tosefta

An examination of the parallel conversations found that in two conversations in the Mishnah there appear connectors and question particles that are absent from the parallel conversations in the Tosefta.

In Citation 7 above there is a similarity between the first part of the exchange in the Mishnah and Tosefta (the only one in the Mishnah and the first in the Tosefta), which is said by the House of Hillel. Their argument is made up of two parts. The first part presents an introduction that clarifies the similarity between the items that are presented – a clean heave-offering may not be eaten by non-priests and an unclean heave-offering may not be eaten by priests. The second part presents an a fortiori argument מה טהורה עולה אף טמאה תעלה: just as a clean offering is made void if it falls into a hundred measures of hullin, thus an unclean offering is made void if it falls into a hundred measures of a clean offering. In the first part two connectors appear in the Mishnah that are absent from the parallel conversation in the Tosefta: in the Mishnah –הואיל וטהורה אסורה לזרים ואף טמאה אסורה לכהנים […] whereas in the Tosefta – טהורה אסורה לזרים וטמיאה אסורה לכהנים […]. The use of the causal particle הואיל ו (since) in the beginning of the introductory words and the use of the additive particle אף (also)¹⁵ in the beginning of the second co-ordinate clause in the Mishnah appears to make the logical connections between the components of this part clearer: הואיל ו clarifies the causal relationship between the introductory words in the first part and the a fortiori argument in the second part; and אף clarifies the additive relationship between the two types of offerings presented in the introductory words.

In Citation 8 above there is similarity between the Mishnah and Tosefta in the first part of the exchange (which is the only part in the partial exchange in the Mishnah), and in it too a particle appears in the Mishnah that is absent from the parallel conversation in the Tosefta: in the Mishnah, the question begins with the question particle כיצד (how) –על עצמו אינו נאמן, כיצד יהא נאמן על שלאחרים?, whereas in the Tosefta, the question is formulated without the question particle at the beginning – על עצמו אינו נאמן, יהא נאמן עלי?. This rhetorical question, which is asked by אמרו לו in the Mishnah and by R. Meir in the Tosefta, presents a difficulty regarding the view of the sages that one who accepts the hospitality of an ‘am ha'areṣ (one who is suspected of not properly giving tithes) is trustworthy; literally, it means: “How is it possible to rely on a person who eats at the table of an ‘am ha'areṣ and trust that he will not feed others the food of the ‘am ha'areṣ without first taking a tithe?” (See more details on the question in Part 2 below; and see a further linguistic difference regarding the question in these conversations in terms of the pronouns there). The use of the question particle כיצד (how) in the Mishnah clarifies the rhetorical nature of the question, which raises a query about the previous view and comes to contradict it.¹⁶ In the Tosefta, this exchange includes a second part in which אמרו לו base themselves on what they relate was the practice מימיהם של בעלי בתים לא נמנעו מלהיות אוכלין זה אצל זה…, and argue that the situation described in the story bolsters their view that one who accepts the hospitality of an ‘am ha'areṣ may be considered trustworthy in the matter of tithes, even though it is forbidden in the first place to accept hospitality from a person who is suspected of not observing tithes.

It should be noted that this situation of a rhetorical question formulated in the Tosefta without a question particle at the beginning is found in another conversation in the Tosefta, which has a different parallel conversation in the Mishnah, and which does not contain a rhetorical question. In Citation 6 above in the first part of the first exchange, אמרו לו are opposed to the view of R. Judah because it is impossible to redeem fruits bought with second tithe money that became unclean; they ask him a rhetorical question: תחמיר בטפילה יתר מן העקר?, in other words, should you rule more strictly in regard to secondary fruits than in regard to the second tithe money, which is the primary matter, and this rhetorical question in the Tosefta does not begin with a question particle.

2) A personless form or an indefinite form in the Mishnah compared to a parallel first-person form in the parallel in the Tosefta

In two parallels, there appears in the Mishnah a personless form or an indefinite form as opposed to a first-person form in the conversation in the Mishnah.

In Citation 8 above, in the first part of the exchange, the corresponding conversations are different in terms of the grammatical persons used in them (further to the difference in terms of the question particle, as presented in number 1 above). In the Mishnah, the words of אמרו לו use the indefinite form – על עצמו אינו נאמן, כיצד יהא נאמן על שלאחרים? (“[…] how should he be trustworthy concerning that of others?”), whereas in the Tosefta the parallel question, asked by R. Meir, contains a singular first-person form that expresses indefiniteness – ?על עצמו אינו נאמן, יהא נאמן עלי (“[…] should he be trustworthy concerning me?”). In this rhetorical question R. Meir raises some doubt regarding the possibility of trusting a person who does not behave in a trustworthy fashion and eats at the table of an ‘am ha'areṣ, and relating to him as someone who can be trusted in the matter of tithes and that he will not feed others from the food of an ‘am ha'areṣ, without first taking tithes. When R. Meir asks, יהא נאמן עלי?, he means to say that one cannot trust a person who accepts the hospitality of an ‘am ha'areṣ, i.e. this person cannot be trusted by others; but whereas in the Mishnah, the indefinite form אחרים (others) is used, in the Tosefta, the singular first-person pronoun עלי (me) is used, which does not refer to the addressor R. Meir.¹⁷

In Citation 4 above, in the first part of the exchange, R. Judah presents a story of a water-channel that passed from the city of Abel, from which the sages allowed water to be drawn on the Sabbath without a special partition. In his words in the conversation in the Mishnah, an indefinite form of the verb appears – שהי!ה! ממלין ממנה על פי הזקנים בשבת (“did they draw […]”) (i.e. שהיו ממלאין¹⁸), whereas in the parallel conversation in the Tosefta the form appears in the first-person plural –והיינו ממלאין […] (“and we would draw […]”). In the Mishnah, R. Judah tells a story that relates to the practice in the city of Abel, and the indefinite verb form relates to the people in general (היו ממלאין); in the story in the Tosefta, on the other hand, R. Judah tells the story based on his own personal knowledge, and he uses the first-person plural form, which includes him too (היינו ממלאין).

It should be noted that the use of the first-person form can be found in another conversation in the Tosefta, which has a different parallel conversation in the Mishnah in which there is no corresponding pronoun use:

In both compilations, first the views of R. Eliezer and R. Joshua are presented (and in the Mishnah, also the view of R. ‘Aqiba) regarding the right of a levir to annul the vows of a childless brother's widow.¹⁹ After the presentation of the views, non-identical conversations appear in the two compilations between R. Eliezer²⁰ and R. ‘Aqiba. The first part of both conversations contain an a fortiori argument in the pattern “מה אם… אינו דין ש…?” (“what if… is it not logical that…?”), in which the sage infers from the fact that it is permitted to annul the vows of a woman who is betrothed to the case under discussion here – that a levir may annul the vows of the childless widow he would marry as part of a levirate marriage. In response, in the second part of the exchange, R. ‘Aqiba rejects the a fortiori argument using the pattern of rejection of an a fortiori argument ‘לא, אם אמרת ב… ש… תאמר ב… ש…?’ (“No, if you have said in… that… will you say in… that…?”), arguing that this is not a case of a fortiori as was claimed. In the Tosefta, in both arguments, the sage uses the first-person singular form: in the first part of the exchange, in R. Eliezer's a fortiori argument –ומה אם אשה שאין לי בה חלק עד שלא באת לרשותי, משבאת לרשותי ניגמרה לי, אשה שיש לי בה חלק עד שלא באת לרשותי, משבאת לרשותי אינו דין שנגמרה לי? (“Now in the case of a woman in whom I have no part before she enters my domain, once she enters my domain, she is wholly in my power, in the case of a woman in whom I have some part before she comes into my domain, once she enters my domain, is it not logical that she should be wholly in my power?”); and in the second part, in the corresponding words of R. ‘Aqiba that reject the a fortiori argument: לא, אם אמרת באשה שאין לי בה חלק עד שלא באת לרשותי […]? (“No, if you have so stated matters in the case of a woman in whom I have no part before she comes into my domain […]”). The forms of the personal pronouns and the possessive pronouns in first-person singular, which are bolded in these arguments, do not refer to the addressor, i.e. R. Eliezer or R. ‘Aqiba, but rather relate to them as representing all people, i.e. they are used as an alternative to the indefinite pronoun. However, in the parallel conversation in the Mishnah, in the formulations of the parallel – but not identical – arguments, there is no such use of the first-person forms; the arguments related to a man who is engaged or to the levir: in R. Eliezer's argument – מה אם אשה שקנה הוא לעצמו הרי הוא מפר נדריה, אשה שקנו לו שמים אינו דין שיפר נדריה? (“Now in the case of a woman whom he acquired for himself, lo, he annuls her vows, a woman who is acquired for him by Heaven, is it not logical that […]?”), and in R. ‘Aqiba's response – לא, אם אמרת באשה שקנה הוא לעצמו שאן לאחרים בה רשות, תאמר באשה שקנו לו שמים שיש לאחרים בה רשות? (“No, if you have so stated the rule in regard to a woman whom he has acquired for himself, the fact is that others have no claim on her. But will you say the same in the case of a woman acquired in his behalf by Heaven, in whom others have a claim?”). Bolded forms appear in both these arguments, which are third-person forms, and in the second argument the form אחרים (others) also appears (which is described above in the discussion of Citation 8); the third-person forms and the form אחרים express indefiniteness.

Feliks (2004) presents the use of the indefinite form among the tools used to determine which of the parallels came later, and the use of the first-person form among the tools used to determine which of the parallels was earlier (pp. 81–3). This difference between the parallel conversations – the personless form or the indefinite form in the Mishnah as opposed to the use of the first-person form in the parallel in the Tosefta – can be explained in light of Feliks's determination, and thus to argue that the first-person form in the Tosefta came earlier than the personless or the indefinite forms in the parallels in the Mishnah. Perhaps there is room to examine this use of the first-person forms to express indefiniteness and see if it is characteristic of the vernacular.

3) More precise and detailed language in the Tosefta compared to the parallel in the Mishnah

In two cases, linguistic precision relating to factual details regarding place and measure, which is absent from the parallel in the Mishnah, was found in a parallel conversation in the Tosefta. Both can be found in Citation 4 above.

In the first part of the exchange in Citation 4, in the words of R. Judah, the Mishnah contains a general reference to the presence of a water-channel in the city of Abel – מעשה באמ!ת! של אבל,²¹ which is formulated by means of an unbound genitive construction with של (of). However, in his words as they are cited in the Tosefta, the precise course of the water-channel is noted as going from Abel to Sepphoris – מעשה באמת המים שהיתה באה מאבל לצפורי. After the bound genitive construction אמת המים (water-channel) (compared to the unbound genitive construction אמה של אבל [“the water-channel of Abel”] in the Mishnah)²² a verbal relative clause appears that clarifies that this refers to the course of the channel (שהייתה באה מ… ל…) and also adds the destination – Sepphoris.

In the second part of this exchange, אמרו לו present an explanation for their refusal to accept as evidence the story about the water-channel because in their view, this water-channel was of a different size from the one discussed in the previous context. In the parallel in the Mishnah, the explanation relates to the size in general – מפני שלא היה בה כשיעור (“because it was not of the requisite size”), whereas in the Tosefta, the exact size is noted – מפני שלא היתה עמוקה עשרה ורחבה ארבעה (“because it was not ten handbreadths deep and four handbreadths broad”).

Seemingly, the factual precision in these cases could be attributed to the level of the content, rather than the language. It was nevertheless decided to include this difference in the linguistic comparison of the parallel conversations in this subsection, both because it involves a linguistic expression in the formulation of the parallels as well as because it may express a stylistic difference between the Mishnah and Tosefta.

In addition to the precision in the factual details noted here, precision related to details in terms of the course of the conversation was found in another conversation in the Tosefta. In Citation 10, the corresponding passages contain a conversation between Rabban Gamaliel and his interlocutors, and they are presented by another sage – R. Judah:

In the corresponding exchange in the conversations (the only one in the Mishnah and the first in the Tosefta), in the first part, Rabban Gamaliel's interlocutors present a rhetorical question in the Mishnah compared to an analogy argument in the Tosefta, and the aim of both is to present a difficulty regarding the law under discussion – in the context of the law that applies to property that a betrothed woman received after her betrothal. The second part of the exchange is identical in the two parallels, and Rabban Gamaliel rebukes those asking the question why they are raising a difficulty in the law concerning the woman's old property from the time of her betrothal when there is a difficulty in a different law cited later regarding new assets that the woman received after her marriage: בחדשים אנו בושים/ן אלא שאתם מגלגלין עלינו את הי(י)שנים_.

Following this exchange, a further – partial – exchange is presented in the Tosefta, in which another sage – R. Ḥanina b. ‘Aqabya – presents a different opinion regarding the words of Rabban Gamaliel that appeared in the second part of the previous exchange. Instead of Raban Gamaliel's rebuke (which was presented in the previous exchange), R. Ḥanina b. ‘Aqabya presents an argument from Rabban Gamaliel that serves to reject an a fortiori argument, whose intent is to argue that it is impossible to infer from the law that applies to a married woman regarding the law that applies to a betrothed woman. Thus, the Tosefta contains the addition of a detail regarding the course of the conversation – the presentation of an alternative argument from another sage to the words of Rabban Gamaliel presented in the previous exchange, and this addition is absent from the parallel in the Mishnah.

Difference (3) – the use of more precise and more detailed language in the halakhic give-and-take conversations in the Tosefta compared to the more condensed language in the parallel conversations in the Mishnah – can be related to the findings of Feliks (2004: 81), who enumerates the brevity of language and the condensation of the content among the tools he used to determine whether a particular parallel came later than another parallel. Accordingly, it can be explained that the briefer and more condensed language in the conversations in the Mishnah reflects a later parallel, whereas the more precise and detailed language in the parallel in the Tosefta reflects an earlier language. Also in difference (2) – the personless form or an indefinite form in the conversations in the Mishnah compared to the use of the first-person form in the parallel conversations in the Tosefta – it was explained that the forms in the Tosefta appear earlier than their counterparts in the Mishnah, in accordance with the distinctions made by Feliks (2004). And regarding difference (1) – connectors and question particles in the Mishnah that are absent from the parallel conversations in the Tosefta – this may be considered an expression of the fact that in the Mishnah the language is more extensively redacted than the less redacted language of the Tosefta: the Mishnah expresses the content by means of connectors and question particles, whereas in the Tosefta, the content is not reflected linguistically and there is a preference for asyndetism. The signs of linguistic redaction found in the Mishnah express a later language compared to the earlier language of the Tosefta, as maintained by Friedman and Braverman, among others, and noted in Section 1 above.

3. Summary and conclusions

This article examined the parallelism between the halakhic give-and-take conversations in the Mishnah and Tosefta. It sought to show that a review of a discourse unit found in both compilations can advance the description of various linguistic phenomena and also contribute to a discussion of the relationship between the Mishnah and Tosefta.

The examination included a corpus of halakhic give-and-take conversations from the orders of Zera‘im, Moe‘d and Nashim – 79 conversations from the Mishnah and 118 from the Tosefta. An examination of the extent of the overall parallelism between the conversations showed that there is a parallel conversation in the Mishnah for only 16 conversations that appear in the Tosefta – 14% – which in most cases are not identical; and the rest of the conversations in the Tosefta have no parallel in the Mishnah (47%), or the Mishnah contains a parallel, but one that does not contain a conversation (39%).

Of the 16 parallel halakhic give-and-take conversations, it was found that for only one conversation, the two parallels were almost identical (6.25%), in half of the conversations, one part of the exchange in the conversation was similar or identical (50%), and the rest of the parallel conversations in the two compilations differ from one another in their content (43.75%).

It can be seen both from the extent of the overall parallelism between the conversations as well as from the description of the similarity between the parallel conversations in the total number of conversations, that there is little similarity between them. At the same time, it seems that the similarity in the halakhic give-and-take conversation may be fully assessed only in comparison to findings that will be obtained from an examination of the other discourse units in the Mishnah and Tosefta.

The 16 parallel halakhic give-and-take conversations in the Mishnah and Tosefta were compared structurally and linguistically.

The structural comparison between the parallel conversations showed that in most of them (62.5% = 10 out of the 16 conversations) the number of exchanges included in conversations in the Tosefta is greater than the number of parallel exchanges in the Mishnah, or that complete exchanges appear in the conversations in the Tosefta, whereas partial exchanges appear in the conversations in the Mishnah; the rest of the parallel conversations have the same number of exchanges (37.5% = 6 conversations). The corpus did not contain any cases of conversations in the Mishnah being longer than their counterparts in the Tosefta.

In the linguistic comparison of the parallel conversations, three differences were found: (1) connectors and question particles that appear in the Mishnah but are absent from the Tosefta parallels, (2) a personless form or an indefinite form in the Mishnah as opposed to a first-person form in the Tosefta; (3) more precise and detailed language in the Tosefta as opposed to the parallel in the Mishnah. These differences reflect less redacted and earlier language in the Tosefta as opposed to more redacted and later language in the Mishnah. This different linguistic nature of the Mishnah and the Tosefta also emerged from the findings of the three studies described above.

These findings, which emerged from the comparison between the Mishnah and the Tosefta regarding a discourse unit found in both compilations – the halakhic give-and-take conversation – should be examined in light of findings on additional discourse units examined in the two compilations. It is my hope that linguistic descriptions of additional discourse units will be conducted for both compilations, and that they will contribute to an understanding of the linguistic characteristics of the discourse units as a whole, and of the relationship between the Mishnah and the Tosefta in particular.