Two Types of 'Mah ha-Tzad': Conceptual Construction

With God's help

Meisharim – 5764

Introduction

A. The approach of Maimonides and Rabbi Yosef Migash to the Mishnah in Shevuot 42b.

B. Conceptual construction in the bailee's oath.

C. Conceptual construction in the labors of the Sabbath.

D. The difference between the rule of 'mah ha-tzad' and these two cases.

E. Implications and practical ramifications.

F. Primary categories and derivatives on the Sabbath and in damages.

G. Additional examples: the significance of conceptual construction.

H. A brief logical discussion.

Appendix

Introduction

One of the hermeneutic rules by which the Torah is interpreted is 'mah ha-tzad' (or: 'ha-tzad ha-shaveh'; below[1] it will be explained that there may be a difference between these two).[2] The rule of 'mah ha-tzad' serves to derive a new legal rule from two known source cases. In this article I wish to point to a certain process in halakhic logic which, formally, resembles the rule of 'mah ha-tzad', yet in fact differs from it, and perhaps cannot even be subsumed under that heading at all. Below I shall call this process 'conceptual construction'.

The structure of the article is as follows. First I will present two examples of conceptual construction, one from the laws of oaths (below, sections A-B) and the other from the laws of the Sabbath (section C). I will then compare it to the rule of 'mah ha-tzad' among the hermeneutic rules of Torah interpretation, and show several implications of the differences between them (sections D-E). I will then discuss a particular implication of our distinction for the meaning of the concepts of 'primary category' and 'derivative' in Jewish law (section F). After that I will discuss additional contexts in which conceptual construction appears, which will further clarify its character (section G). I will conclude with a brief logical analysis of conceptual construction as defined here (section H).

A. The approach of Maimonides and Rabbi Yosef Migash to the Mishnah in Shevuot 42b

The dispute between Rabbi Meir and the Sages

It is taught in the Mishnah (Shevuot 42b):

Rabbi Meir says: There are things that are in the ground and yet are not like land, and the Sages do not agree with him. How so? 'I entrusted you with ten vines laden with grapes,' and the other says, 'There were only five.' Rabbi Meir requires an oath, and the Sages say: Anything attached to the ground is like the ground.

And in the Gemara there Rabbi Yosei bar Hanina explains the dispute as referring to grapes ready to be harvested (ibid., 43a):

Rabbi Yosei bar Rabbi Hanina said: Here they disagree about grapes that are ready to be harvested. Rabbi Meir holds that they are regarded as harvested, and the Rabbis hold that they are not regarded as harvested.

And so too in the parallel passages.[3]

We thus have a tannaitic dispute as to whether grapes ready to be harvested are regarded as already detached or not. The practical consequence discussed by the Mishnah concerns an oath, since it is established that one does not swear concerning land. The dispute between Rabbi Meir and the Sages therefore is whether one swears regarding these grapes, since they are considered detached, or not, since they are treated as land.

In the parallel sugyot (Gittin 39a and Sanhedrin 15a), it is explained that according to Rabbi Meir we distinguish between grapes that still need the ground and those that no longer need the ground. The tannaitic dispute concerns only grapes that no longer need the ground (see Ritva on our sugya, and Maggid Mishneh, Hilkhot To'en Ve-Nit'an 5:4, who raise the possibility of three gradations, but this is not the place). When the grapes still need the ground, everyone agrees that they are attached.

The medieval authorities on the passage raise several difficulties from sugyot in which it appears that grapes ready to be harvested are considered detached according to everyone, and they answer in various ways (see Tosafot s.v. 'kevetzurot', and Rashba, Ritva, and the other medieval authorities here and in the parallels). Their common denominator is that they understand the dispute between Rabbi Meir and the Sages as a sweeping dispute touching on a number of issues (and some of these are explicit in the parallel sugyot, which connect this tannaitic dispute to other discussions as well). In light of that, they point to the exceptions that appear in the conflicting sugyot as isolated cases.

The approach of Maimonides and Rabbi Yosef Migash in explaining the dispute between Rabbi Meir and the Sages

By contrast, Rabbi Yosef Migash and Maimonides in the Commentary on the Mishnah interpret the Mishnah differently. According to them, the dispute pertains only to the laws of bailees, and only there do the Sages say that the grapes are not considered detached. In all other halakhic contexts, however, everyone agrees that grapes ready to be harvested are considered detached.

This explanation appears in Rabbi Yosef Migash's novellae here, where he explains that only for the purpose of bailees are the grapes not considered detached, since the bailee has no right to harvest them. As stated, in other contexts the grapes are considered detached according to everyone.[4]

At first glance, however, their approach is difficult in light of the continuation of the Mishnah (see above). There it states that Rabbi Meir and the Sages disagree in a case where the plaintiff claims that he entrusted ten grapevines laden with grapes, and the defendant claims that only five were entrusted to him. Such a case does not involve the bailee's oath, but rather the ordinary oath of admission of part of the claim, for the defendant is not admitting that he is a bailee of the additional five. He is denying the very existence of the safekeeping. The bailee's oath is relevant when there is a known bailee and he is sued for negligence, which he denies (that is, he claims he was not negligent and the deposit was lost through circumstances beyond his control). Here, by contrast, the discussion concerns the very question whether he is a bailee at all, that is, whether the additional five vines were ever entrusted to him. In such a case it seems clear that the bailee's oath is inapplicable. Several later authorities indeed noted this difficulty (see, for example, Hiddushei Ha-Ilui Mi-Meitshet, sec. 64, and the later authorities cited below).

If so, according to the explanation of Rabbi Yosef Migash and Maimonides that the entire dispute between Rabbi Meir and the Sages is limited to the laws of bailees, it is very difficult to understand why the Mishnah explains their dispute by means of a case in which the defendant is actually required to take the oath of admission of part of the claim, a case unrelated to the laws of bailees.

One could ostensibly say that although we are dealing here with an oath of admission of part of the claim, nevertheless, since in the final analysis the claim arises from an allegation of safekeeping, the grapes are still not deemed detached. The reason would be that according to the plaintiff's own claim, the defendant is a bailee, and therefore from his standpoint the grapes are not deemed detached; accordingly he cannot demand that the defendant swear concerning them.

This indeed seems to be Maimonides' intention in the Commentary on the Mishnah there. He writes:

The dispute between Rabbi Meir and the Sages concerns grapes that are ready to be harvested. And the law follows the Sages. Specifically if he gave them to him for safekeeping. But regarding sale and purchase, the laws of overreaching, and admission of part of the claim, if the main claim was not in the context of safekeeping, the legal rule is that they are like movables.

It is explicit in Maimonides that even with respect to the oath of admission of part of the claim, if the claim is that they were given to him for safekeeping, then although the oath is apparently not the bailee's oath (but the oath of admission of part of the claim, as explained above), the rule is that the grapes are not deemed detached. This also appears to be the correct reading of Rabbi Yosef Migash. It is highly plausible that the emphasized addition in Maimonides' words is meant to answer the later authorities' question from the continuation of the Mishnah.

Maimonides speaks specifically about the laws of bailees

However, in the laws of Hiring, the wording of Maimonides seems otherwise. He writes there (2:4):

One who gives another something attached to the ground to guard, even if they were grapes ready to be harvested, they are like land in the law of bailees.

His wording implies that he treats the grapes as land only in 'the law of bailees,' and not in every situation involving a bailee. In a case of admission of part of the claim in a bailee's claim, as in our Mishnah, we are not dealing with the law of bailees but only with a situation of a claim based on the allegation that he is a bailee. Maimonides' wording here implies that in such a case the grapes are considered detached. If so, it appears from his wording that only with respect to the bailee's oath itself (which, as stated above, our Mishnah is not discussing) are the grapes not deemed detached.

What thus emerges from Maimonides' wording in the Mishneh Torah is that he interprets the Mishnah differently. Apparently he explains the dispute between Rabbi Meir and the Sages in terms of the laws of bailees, and not in terms of a claim against a bailee, and therefore he must explain our Mishnah as dealing with the bailee's oath and not with the oath of admission of part of the claim.

This also seems evident from the fact that he places this law in the laws of bailees and not in Hilkhot To'en Ve-Nit'an, as he should have done if we were indeed dealing here with the laws of claims and admissions. It should further be noted that in Hilkhot To'en Ve-Nit'an (5:4) Maimonides cites the case of admission of part of the claim that appears in the Mishnah, but there he does not indicate at all that it is a claim of safekeeping. There too he rules like Rabbi Meir, that the grapes are like movables. If so, this seems not to be a ruling based on what the Mishnah states explicitly, but perhaps only inferred from it (that is, according to Maimonides the Mishnah really deals with the bailee's oath, but one may infer from it laws regarding the oath of admission of part of the claim in other contexts, and those are what Maimonides brings in Hilkhot To'en Ve-Nit'an).

There is much to discuss concerning Maimonides' ruling in the dispute between Rabbi Meir and the Sages, and many have already dealt with it, but this is not the place.[5]

What matters for our purposes is that Rabbi Yosef Migash and Maimonides understand the dispute between Rabbi Meir and the Sages as pertaining specifically to the laws of bailees, and not to every claim that arises from safekeeping. In light of Maimonides' wording in the Mishneh Torah, it seems that this is also his meaning in the Commentary on the Mishnah: when one comes on the basis of a claim of safekeeping, this is truly a matter of the laws of bailees, and not merely a factual situation of safekeeping (as we suggested above). According to this, it may be that this is also the correct explanation of Rabbi Yosef Migash's view, namely, that the dispute pertains specifically to the laws of bailees, as Maimonides' wording here implies.

It likewise seems that this is how the later authorities understood Maimonides, namely, that the dispute regarding the grapes depends on actual bailee law, and not on the mere situation of a claim arising from safekeeping, as will now be detailed.

Tosafot Yom Tov on the Mishnah asks against Maimonides why the Mishnah framed the case as one of admission of part of the claim, since according to Maimonides' ruling the bailee's oath does not require such an admission. This clearly shows that he assumes the Mishnah is dealing with the bailee's oath, and he asks why a case of admission of part of the claim is needed, given that in practical law a bailee swears even when he denies everything.

Rabbi Akiva Eiger, in his glosses there, objects that the Mishnah is not speaking of the bailee's oath at all, but of the oath of admission of part of the claim (as explained above), and he leaves the words of Tosafot Yom Tov as requiring great explanation.

Rabbi Akiva Eiger there says that another question should be asked on Maimonides, and this is our own question above: why does the Mishnah in fact present a case obligating the oath of admission of part of the claim (and not, as Tosafot Yom Tov asked, why they chose such a case in the context of the bailee's oath), rather than a case where the oath obligation is the bailee's oath, given that according to Maimonides only in that case do Rabbi Meir and the Sages disagree. So too asks the Shakh (sec. 95, subsec. 9). From his wording it is clear that he understood this to be the intent of Tosafot Yom Tov as well, but the wording of Tosafot Yom Tov itself clearly suggests otherwise.

It should be noted that the Shakh rejects Maimonides' view in practical law on the basis of this question: if the disagreement really concerns only the bailee's oath, why does the Mishnah not present the dispute in a case of the bailee's oath?

It stands to reason that Tosafot Yom Tov's assumption that the Mishnah does in fact deal with the bailee's oath is based precisely on the question of the Shakh and Rabbi Akiva Eiger. Because of the interpretation of Maimonides and Rabbi Yosef Migash regarding the dispute between Rabbi Meir and the Sages, we are compelled to understand that the Mishnah is dealing with the bailee's oath and not with the oath of admission of part of the claim. What was unclear to Tosafot Yom Tov was only why in such a case one needs a situation of admission of part of the claim.[6]

It thus appears that all these later authorities understood Maimonides to be speaking about the laws of bailees, and they were not satisfied with the possibility we raised above, that Maimonides and Rabbi Yosef Migash are speaking merely about a factual situation involving a bailee and not about bailee law. The likely reasons are the meaning of Maimonides' wording in the laws of Hiring (as noted above), and the very form of their objections. Presumably they would likewise interpret Maimonides' wording in the Commentary on the Mishnah, namely, that when the plaintiff comes with the claim that the defendant is a bailee, this is deemed the bailee's oath. Yet this explanation is extremely difficult, for after all there is here an oath of admission of part of the claim. Hence these later authorities remain puzzled by Maimonides' position. This will be clarified in the next section.

The aspect of heilekh

The medieval authorities in Shevuot (42b, 43a) asked, with regard to Rabbi Meir and the Sages, why they discuss exemption from an oath on the basis that grapes ready to be harvested are like land; one could derive the exemption from the rule of heilekh ('here it is'). For here he admits land and grapes attached to land, and therefore this is heilekh (according both to Rabbi Meir and to the Sages). If so, he should be exempt from an oath even if the grapes are considered detached (as Rabbi Meir understands), since he is admitting land and grapes attached to it.

See Rashba and Ritva here, who cite the need to establish the case such that at least some of the grapes had been harvested (and see there two directions in explaining the matter, but this is not the place).

However, when we examine Hilkhot To'en Ve-Nit'an (5:4), and certainly chapter 2 of the laws of Hiring, it appears that Maimonides does not indicate at all that the grapes had been harvested. His wording clearly implies that the grapes are still attached to the vines. Rabbi Yosef Migash too, in his commentary on the Mishnah and Gemara, does not address this issue at all. If so, the medieval authorities' question would seemingly apply to them: why is he not exempt from the oath because of heilekh?

Interim summary

Before turning to a proposal for the view of Maimonides and Rabbi Yosef Migash, let us summarize. Their approach faces two difficult questions:

1. Why does the Mishnah not frame the dispute in terms of the bailee's oath, but instead in terms of the oath of admission of part of the claim, even though, according to Maimonides and Rabbi Yosef Migash, there is no dispute there at all? On the other hand, if this is the bailee's oath and not the oath of admission of part of the claim, then in practical law (at least according to Maimonides) admission of part of the claim is not required at all. Beyond this, as we have seen, the wording of the Mishnah itself clearly indicates that it is not speaking of the bailee's oath (contrary to what Tosafot Yom Tov seems to have assumed).
2. Why is he not exempt from the oath because of heilekh?

A direction toward a solution

From these two difficulties it emerges that, according to Maimonides and Rabbi Yosef Migash, one must say that the Mishnah is indeed speaking of the bailee's oath. For this reason the problem of heilekh did not trouble them either. The exemption from an oath by reason of heilekh applies only to the oath of admission of part of the claim, whereas for the bailee's oath we do not find such an exemption.

One must still understand what specific case is involved. If the plaintiff is demanding payment for the bailee's negligence and the bailee claims unavoidable accident, which is the ordinary case of the bailee's oath, then admission of part of the claim is unnecessary in order to incur the oath (and this is the question of Tosafot Yom Tov). Moreover, the Mishnah does not seem to describe such a case, for it sounds as though they are disputing the very safekeeping itself-whether five vines were entrusted to him or ten (and this is the assumption behind the question of the Shakh and Rabbi Akiva Eiger against Tosafot Yom Tov and Maimonides).

Thus, as emerges from the wording of the Mishnah, we must say that the dispute is not about negligence in safeguarding the vines, but about the number of vines entrusted to him for safekeeping. Yet this would apparently be not the bailee's oath but the oath of admission of part of the claim. In that case, however, all the earlier difficulties return.

B. Conceptual construction in the bailee's oath

The Mishnah does indeed deal with the bailee's oath

It therefore appears that the Mishnah truly is dealing with the bailee's oath. As we have seen, this follows clearly from Maimonides' approach, and so the later authorities understood him. According to this, it is also well explained why there is no exemption from the oath because of heilekh, as stated above.

As we noted, if this is indeed the interpretation of the Mishnah, then Maimonides' wording in the Commentary on the Mishnah must be read accordingly as well. When he writes that there is an oath if the claim was not made in the context of safekeeping, the meaning is that if it was made in the context of safekeeping, then this is the bailee's oath, and therefore the grapes are not deemed detached-and not, as we explained above, merely a factual situation of safekeeping. Here we explain it as 'the law of bailees,' in accordance with his wording in the laws of Hiring. We have also seen that the later authorities understood him this way (there is no reason to assume that his words in the Mishneh Torah contradict his words in the Commentary on the Mishnah). According to this, it may also be that for Rabbi Yosef Migash this is the correct explanation of the Mishnah (as is well known, Maimonides often follows his view).

It still requires explanation, however, how such an oath, in which the very deposit is under discussion, can be the bailee's oath.

How the obligation of the bailee's oath arises here

It appears that the explanation is as follows.

Let us consider a case of admission of part of the claim in a loan claim. When Reuven claims one hundred from Shimon, and the latter denies everything, he is exempt from a Torah oath. But when he admits fifty, he is liable to the oath of admission of part of the claim. The simple explanation is that a situation has been created in which it is known that there was a loan, and the dispute between the parties concerns only its amount. The one who admits part of the claim thus admits that there was a loan, but he claims that the loan amounted to fifty, while the plaintiff claims that it amounted to one hundred. According to this understanding, it stands to reason that the defendant's legal status is that of someone who says 'I have paid,' rather than that of someone who denies the claim. It must be emphasized that his actual claim is, of course, a denial of the loan itself with respect to the additional fifty. But since we know that there was a loan, the Torah teaches that our legal treatment of his claim must be as though he were claiming that it had been paid, and not that there had never been any loan at all. That is, he stands in a defensive posture, rather than the other side being 'the extractor from his fellow,' and therefore we do not exempt him without an oath. This is the accepted legal and conceptual basis for the oath of admission of part of the claim.

In light of this, let us now examine the case described in the Mishnah. Reuven claims that he entrusted ten vines to Shimon for safekeeping, and Shimon admits five and denies five. If so, everyone agrees that there was here a contract of safekeeping; only the bailee claims that he received only five vines for safekeeping, while the depositor claims that he deposited ten. According to the logic we presented regarding admission of part of the claim in a loan, we should say here too that the defendant is known to be a bailee according to everyone, and the legal status of his claim-that the remaining five were never entrusted to him at all-is as though he were claiming that they were lost through unavoidable circumstances (which parallels the claim of payment in the law of loans). That is, he resembles a bailee who seeks to free himself from his obligations, and not someone who entirely denies the claim.

If so, in light of this innovation derived from the oath of admission of part of the claim, we treat him as a bailee seeking to free himself from his obligations. But now a second datum enters: from the Torah passage of the bailee's oath we learn that a bailee who raises an exempting claim (rather than a claim of complete denial) cannot be released from his obligations except by an oath. This is the Torah's innovation in the section of the bailee's oath. It follows that in a case where he admits part of the vines, he is liable to the bailee's oath-not to the oath of admission of part of the claim-with respect to the remainder.[7]

We thus learn from two known source cases-admission of part of the claim and the bailee's oath-a new oath obligation in the case before us. This obligation is a type of bailee's oath.[8]

It would seem, however, that in such a case he should also be liable to the ordinary oath of admission of part of the claim with respect to the rest, since after all he does admit part of the claim. Yet with respect to the oath of admission of part of the claim there is in fact no oath obligation here, because this is a case of heilekh (according to the Sages, who hold that the grapes are not deemed detached and are therefore like land), as the medieval authorities noted above. As remembered, according to Maimonides and Rabbi Yosef Migash it does not appear that the grapes were harvested. Therefore what remains here is only the obligation of the bailee's oath regarding the rest.

Thus, in a case where Reuven claims that he deposited ten vines with Shimon, and the latter admits only five, Shimon would incur both the bailee's oath and the oath of admission of part of the claim. In the case of land, where he is exempt from the oath of admission of part of the claim because it is heilekh, only the bailee's oath remains.

Resolving the later authorities' questions

This resolves all the later authorities' questions (Tosafot Yom Tov, Rabbi Akiva Eiger, and the Shakh). According to Rabbi Yosef Migash and Maimonides, the Mishnah actually deals with the bailee's oath. On the other hand, it is clear that this is only because there is a situation of admission of part of the claim; for otherwise, if the defendant denied everything, he would be exempt even from the bailee's oath. The obligation to swear the bailee's oath arises only because he admits that he received some of the vines for safekeeping. That is why the Mishnah frames the case as one of admission of part of the claim.

This also explains why there is no heilekh exemption here, for, as stated above, the heilekh exemption applies only to the oath of admission of part of the claim and not to the bailee's oath.[9]

A remark on Maimonides' own view

In Maimonides himself, however, it does not seem that he ruled this way in practical law. In Hilkhot To'en Ve-Nit'an (5:4) he requires an oath regarding grapes that no longer need the ground even when the case is not framed as one involving a bailee (there he cites the Mishnah's case, not necessarily as a bailee's claim). If so, it seems that in such a case he does not apply the rule of heilekh at all, and not because the matter is treated as the bailee's oath.[10] There also remains some difficulty in that Maimonides nowhere explicitly states the great innovation that in such a case the obligation is the bailee's oath. It should also be noted that there is still some difficulty in the Mishnah itself: why did it not frame the case in the ordinary manner of the bailee's oath, since that would suffice to sharpen the dispute between Rabbi Meir and the Sages regarding grapes-whether they are considered detached or not? One could, however, say that the Mishnah wished to teach us this additional innovation as well, namely, that in such a case an obligation of the bailee's oath is generated.

Earlier we saw that in Rabbi Yosef Migash one can indeed interpret the Mishnah this way. More than that: even according to the understanding of those later authorities in Maimonides, the Mishnah must be dealing with the bailee's oath, and therefore it seems reasonable to say that he too agrees that in our sugya the Mishnah should be interpreted this way (except that in practical law he splits the ruling like the Sages, as mentioned in the note above).

An explicit source in Maimonides for our proposal

After writing all this, I subsequently noticed that in the laws of Borrowing and Deposit Maimonides writes as follows (5:6-7):

If someone deposited with him produce that was not measured, and he mixed it with his own produce and did not measure it, he is negligent. The owner of the deposit says, 'There was such-and-such an amount,' and the bailee says, 'I do not know'—he must pay without an oath, for he has obligated himself to payment and does not know how much he owes. It follows that he is obligated in an oath and cannot swear. And this is how my teachers, Rabbi Yosef Ha-Levi and his teacher, ruled.

And so too any bailee who became obligated to pay and says, 'I do not know how much money I must pay,' while the owners say, 'We know, and it was worth such-and-such,' they collect without an oath…

And how do we know that this is the law? Imagine that someone deposited with him a purse full of gold coins and he was negligent with it, and the owners say, 'There were two hundred dinars in it,' and the bailee says, 'Certainly there were dinars in it, but I do not know how many.' This is equivalent to one who is claimed for two hundred, admits part, and says, 'As for the rest, I do not know'—he is obligated in an oath and cannot swear, and therefore pays, as will be explained.

And in halakhah 7 there is a case in which the bailee says 'I do not know' with respect to the whole deposit, and there the plaintiff swears and collects. Maimonides writes:

There was an incident involving a man who deposited a tied sack with his fellow, and he was negligent with it. The depositor says, 'It contained gold ornaments, pearls, and the like,' and the bailee says, 'I do not know; perhaps it contained dross or sand.' The Sages said: the owner of the deposit shall swear and collect… And why does the owner of the deposit swear here? Because the bailee is not obligated in an oath. For even if he admitted and said, 'I am certain it was full of dross,' while the depositor says, 'It contained pearls,' the bailee takes a hesset oath and is exempt, just like one against whom wheat was claimed and who admitted barley, and so in all similar cases. And the fundamental principles will be explained in the laws of claims and admissions.

It emerges from Maimonides' words that a bailee who denies, on grounds of uncertainty, part of the value of the deposit is liable to an oath. But if he denies, again on grounds of uncertainty, the entire value of the deposit, the plaintiff collects without an oath by the rule of mitokh-that one who is obligated to swear and cannot do so must pay. However, if the bailee denies and raises the possibility that there was dross there, he is entirely exempt from an oath, for this is like a case where wheat was claimed and barley admitted. In such a case the plaintiff does not collect, since the rule of mitokh does not apply to the defendant.

From the opening of his words in the first halakhah it appears that an oath obligation is generated even though he does not admit even anything at all.[11] One could say that his meaning is that there was some admission, yet it still remains clear that his admission is not of something subject to measure and number, and therefore he should have been exempt from an oath. So too object the Raavad and other medieval authorities there.[12]

In light of what we explained above, it follows that the oath imposed on a bailee who denies the value of the deposit is the bailee's oath, and not the oath of admission of part of the claim. It seems that this can be said even if he did not admit even a perutah, for after all he admitted that there had been a deposit there-even if it was worthless-and, as we have seen, once it is known that there was a deposit, the defendant's status is that of one who claims unavoidable loss; accordingly he cannot be exempt without the bailee's oath.

This nicely explains why an oath obligation is generated here even though he admitted nothing at all.

It also explains the objection of the Raavad and the other medieval authorities against Maimonides, namely, that the admission here is not one of measure, number, or weight, and therefore he should have been exempt from an oath. According to our explanation of Maimonides, the answer is as Hazon Ish wrote (Hoshen Mishpat 12:8): in the bailee's oath there is no requirement at all that the claim involve measure, number, and weight. Such a requirement was stated only for the oath of admission of part of the claim.[13]

And the same likewise explains the question the medieval authorities asked from the rule of 'two silver pieces'. For in any case-even if his admission amounts to a perutah-it is still not clear that he admits an amount of two silver pieces, so why is he liable to an oath? According to our proposal, this is well understood, for the view of Rabbi Yosef Migash (in his novellae to Shevuot 40a, also cited by Rashba there) and Maimonides (Hiring 2:8, and To'en Ve-Nit'an 3:6) is that the bailee's oath has no requirement of two silver pieces.

According to this, however, one must consider why Maimonides writes in halakhah 7 that when one claims gold and the other says that perhaps there was dross there, he is exempt on the basis that this is a case where wheat was claimed and barley admitted. If this is indeed the bailee's oath, we do not find such an exemption there.

According to what we explained above, however, there is no difficulty. We saw that the bailee's oath in such a case is the creation of a new oath obligation from the logic of the oath of admission of part of the claim together with the Torah's innovation in the passage of the bailee's oath. Hence, where no obligation of admission of part of the claim is generated-for example in our case of wheat claimed and barley admitted-the new oath obligation is likewise not generated.

At first glance, one could object: we do find that with respect to heilekh (as explained above), and likewise with respect to the rule of an item measured and weighed and the rule of two silver pieces (as we saw in Maimonides' view here), the fact that there is an exemption from the oath of admission of part of the claim does not exempt from the bailee's oath. If so, why does the exemption from admission of part of the claim in the case of wheat claimed and barley admitted also exempt from the bailee's oath?

It seems that this too is no difficulty. The rule of wheat claimed and barley admitted is not like the rule of heilekh, nor like the rules of measure and weight and two silver pieces. In wheat claimed and barley admitted, the basis of the exemption is that with respect to the wheat it is as if he denied everything, and thus there is a substantive exemption from the oath of admission of part of the claim. In light of our explanation of the law of admission of part of the claim, it follows that in such a case no situation is created in which it is known to us that there was a safekeeping contract regarding the wheat. If so, it is obvious that in such a case he is not regarded as advancing an exempting claim, and the legal logic of admission of part of the claim is not renewed at all. Therefore the new oath obligation is likewise not generated here. By contrast, the exemption by reason of heilekh (at least according to some of the medieval authorities, as we noted above), and the exemptions of measure and weight and two silver pieces (as seems clear conceptually), are specific exemptions within the laws of admission of part of the claim. Their rationale is not that in such a case there is no indication that there was a loan, or that there was a contract of safekeeping. Therefore even where we have heilekh, or a claim lacking measure and weight, or lacking two silver pieces, the defendant is still considered one who says 'I have paid'; it is only by scriptural decree that he is exempt from the oath of admission of part of the claim. Hence in these cases he will still be liable to the new oath that we have learned about.[14]

According to this, it is also explained why Maimonides places these laws in the laws of Borrowing and Deposit, which belong to the laws of safekeeping, and not in Hilkhot To'en Ve-Nit'an, which deal with the laws of claims and oaths (admission of part of the claim and the like). According to our proposal, this is very well understood, since we are dealing here with the bailee's oath. When Maimonides writes at the end of the passage that the essential principles will be clarified in Hilkhot To'en Ve-Nit'an, his intention is that the laws of the bailee's oath here are derived from the laws of the oath of admission of part of the claim, in that where one is exempt from the latter the obligation of the bailee's oath is likewise not generated here. Hence Maimonides adds that the laws of claims and of admission of part of the claim themselves will be explained in Hilkhot To'en Ve-Nit'an. Here he mentioned them only in order to show their implications for the laws of bailees.

It should be noted that no clear source is given for these words of Maimonides. According to our proposal, however, his source is pure and simple: our Mishnah. The reason he did not cite our specific case is that it is unnecessary, since this innovation is already known. From our Mishnah he brings only the conclusion that grapes are not deemed detached for purposes of safekeeping in general.

And if our proposal is correct, and the source of Maimonides' view that in such a case the oath is the bailee's oath can indeed be our Mishnah, then it follows that Maimonides did in fact bring the innovation of our Mishnah, and he did so in the above laws of Borrowing and Deposit. This resolves our earlier question as to why Maimonides did not record this innovation in his code.

One final point to note here is that the source for this law is Rabbi Yosef Ha-Levi, as Maimonides explicitly writes in halakhah 6, and as we saw above, he is also the source for Maimonides' entire approach in our sugya. The medieval authorities who discussed these words of Maimonides in the sugya of a pledge (Shevuot 43a) disagreed with him because they did not understand that an obligation of the bailee's oath could be generated here, and therefore they explained everything in terms of the oath of admission of part of the claim. This fits their general approach, since they did not accept his interpretation of our Mishnah and disagreed with him there as well.

Summary: conceptual construction

In summary, we have seen that there is a possibility of generating an obligation of the bailee's oath on the basis of a situation of admission of part of the claim in a safekeeping claim. This is an oath obligation built on the basis of two 'primary categories,' or two source cases, as follows: 1. the logic of the oath of admission of part of the claim, which establishes that admission of part of the claim points to the existence of a contract as alleged by the plaintiff, thereby weakening the defendant's claim. Therefore the defendant's claim is legally considered an exempting claim and not a denial. 2. the obligation that the Torah places upon a bailee to swear when he advances an exempting claim-that is, the obligation of the bailee's oath.

This is one example of what I shall call here 'conceptual construction.' In such a construction, we build a new concept on the basis of two known concepts, each of which contributes part of the new concept. In the next section I will try to present an additional example of this kind of logic, and afterward we will discuss its relation to the rule of 'mah ha-tzad'.

C. Conceptual construction in the labors of the Sabbath

'Mah ha-tzad' among the primary categories of damage and the primary categories of labor

At the beginning of tractate Bava Kamma (2a, 3a), the Gemara discusses the relationship between primary categories and their derivatives. The Gemara determines that with regard to the Sabbath 'their derivatives are like them,' whereas concerning damages the Gemara concludes that some of them are like them and some are not.

Against this background, let us note a puzzling phenomenon regarding the relation between primary categories and derivatives in the laws of the Sabbath. In the context of damages, the Mishnah at the beginning of Bava Kamma teaches that there are derivatives learned by 'mah ha-tzad' from two primary categories: 'the common factor in them…' And in the Gemara (6a) several examples are brought, such as one's stone, knife, and burden that he left on the top of his roof, and they fell in an ordinary wind, and they were not declared ownerless, and they caused damage. This is learned by 'mah ha-tzad' from a pit and fire. See there for several additional examples.

By contrast, with regard to the Sabbath we do not find any derivation of 'mah ha-tzad' at all-that is, derivatives learned from two primary categories together by way of 'mah ha-tzad'.

One might perhaps attribute this to the fact that on the Sabbath the derivatives are like their primaries, and therefore every derivative is learned from the primary to which it is similar; but that is not really an explanation, only an indication. Perhaps the reverse is true: because there is no derivation of 'mah ha-tzad,' therefore their derivatives are like them.

This therefore requires explanation: why do we not find, with regard to the Sabbath, the derivation of a derivative from two primary categories?[15]

An example of conceptual construction in the labors of the Sabbath

Yet we do find one case that seems as though such a thing exists in the laws of the Sabbath as well. The Jerusalem Talmud states (Sabbath, chapter 'Kelal Gadol,' halakhah 2): One who spits and the wind disperses it is liable for winnowing, and anything that depends on the wind is liable for winnowing.

This was likewise ruled in practical law by Or Zarua (part 2, sec. 59), Rokeach (sec. 62), and the Rema (Orah Hayyim, end of sec. 319). The Rema writes there: One who spits into the wind on the Sabbath, and the wind scatters the saliva, is liable for winnowing.

Some commentators explained that this means the wind breaks the saliva up into small parts (see, for example, Korban Ha-Edah on the Jerusalem Talmud there).[16] But it has been noted against this (see Chayei Adam in Nishmat Adam, rule 15, letter 1; Rabbi Akiva Eiger, responsum 20; and also Menuhat Ahavah, part 2, chapter 7, note 13) that our Talmud proves that the whole idea of winnowing is the separation of food from waste, similar to selecting, for the Gemara (Sabbath 73b) asks: Isn't winnowing the same as selecting and the same as sifting?

For that reason, some emended the text and explained that one is liable because of throwing, not because of winnowing. See the glosses of Erekh Lehem by Maharikash (end of sec. 319) and the responsa Yehaveh Da'at (part 6, sec. 25).

However, the Bi'ur Halakhah cites, in the name of Rabbi Menasheh of Ilya, one of the disciples of the Vilna Gaon, the following explanation (end of sec. 319, s.v. 'mefazer'):

And in the book Alfei Menasheh he explained that the intention of the Jerusalem Talmud is a case of carrying something four cubits in the public domain by means of the wind, and it is given by way of example. That is: just as in winnowing, although the wind assists him he is nevertheless liable, so too in the case of one who spits, where the movement is caused by the wind, he is also liable, and this is correct.

It thus appears that Rabbi Menasheh of Ilya understands the Jerusalem Talmud to obligate specifically where one spits in the public domain (and presumably also from the private domain to the public domain and vice versa). In this respect he is like those later authorities who explain the liability in terms of throwing. But unlike them, he explicitly reads the Jerusalem Talmud as saying 'winnowing' and not 'throwing,' and he explains that the labor of winnowing serves as the model for spitting.

According to his words, then, spitting is a derivative of two primary categories: winnowing, and carrying four cubits in the public domain.[17] From the labor of winnowing we learn that prohibited labor performed by means of the wind is considered as though the person did it, and therefore carrying four cubits in the public domain-which is certainly a prohibited labor-is prohibited even when it is done by the wind.[18] It should be noted that this also seems to be the conclusion of the Mishnah Berurah itself, as appears at the end of his remarks.

This case resembles what we saw in the laws of oaths. There we created a new oath obligation by combining two different obligations. Here we create a new labor by combining two different labors. Both are examples of conceptual constructions-that is, the building of a new concept out of two prior concepts, each of which contributes, or constitutes, part of the new concept.

D. The difference between the rule of 'mah ha-tzad' and these two cases

The rule of 'mah ha-tzad': the example of 'immediate and for future generations' from the Sifra

The rule of 'mah ha-tzad' is called in the baraita of Rabbi Ishmael (brought at the beginning of the Sifra) 'binyan av' (and the medieval authorities disagreed whether this means from one verse or from two verses).[19] It too describes a situation of two 'primary categories,' or two source cases, from which a third rule-a derivative-is learned. In that sense there is a parallel between it and the two examples we saw above. At first glance, then, we seem to have here a logical mechanism that serves as the prototype for those two examples. Let us now examine it in greater detail.

In order to follow the process of an ordinary 'mah ha-tzad' derivation, let us take the example given in the above baraita (at the beginning of the Sifra) as an example of a binyan av from two verses:

How does a prototype construction from two verses work? The section of the lamps is not like the section of sending away the impure, and the section of sending away the impure is not like the section of the lamps. The common factor between them is that they were stated with the command 'command' for immediate application and for future generations; so too anything stated with 'command' applies immediately and for future generations.

In fact, a step is missing here that appears in most cases of 'mah ha-tzad' in the Talmud, namely, the attempt to derive from each source case alone and the rejection of that derivation because of a distinctive feature present in each source case. The Raavad, in his commentary there to the Sifra, fills in what is missing and explains as follows:

The Temple lamps are more stringent, for they are frequent and are part of the inner service, and one cannot learn from there to something that is not part of the inner service. Sending away the impure is more stringent, for it involves karet, and one cannot learn from there to something that does not entail karet. But we learn from the common factor of the two, for in both it says 'Command the children of Israel,' and they apply immediately and for future generations.

Here we see the basic mechanism of 'mah ha-tzad': the attempt to derive from each of the source cases fails because that source case has a distinctive feature not present in the derivative-that is, a refutation. But the second source case does not have that feature, and the rule nevertheless applies there. It is therefore proven that this feature is not what causes the rule. And the same is true in the opposite direction with respect to the second source case. The conclusion is thus that the learned feature stems solely from the common side of the two source cases, and their two distinctive features are irrelevant to the rule under discussion. Hence, in the conclusion, we derive from that common side-which also exists in the derivative under discussion-that the rule in question applies there as well.

Thus, in the final analysis, one does not learn from the two primary cases as such, but only from the common factor that exists in both. So too explained the authors of works on the hermeneutic rules-Halikhot Olam, Middot Aharon on the Sifra, Sefer Ha-Keritut, and others-in discussing 'ha-tzad ha-shaveh' (see Encyclopedia Talmudit, entry 'binyan av'). More than that: in the conclusion one no longer needs both source cases at all, and one of them-whichever it be-would suffice. The second is required only in order to negate the relevance of the unique feature of the first source case.[20]

Its relation to conceptual construction

Such a logical structure differs from the conceptual construction we saw in the two examples above. There we learned from each primary case a specific component of the derivative. The combination of the two creates the derivative in its fullness. Here we find a different structure: the common feature shared by the two primary cases teaches the derivative. Each primary case serves only to remove the relevance of one of the other's distinctive features, and in the end we are left only with the common side-the feature found in each of the two primary cases. With respect to the rule being learned, neither source case contributes anything beyond the other. They are equal in this regard; we needed both only in order to realize that only their common feature-and not any of their distinctive features-causes the rule.

If so, one may say that the rule of 'mah ha-tzad' is a derivational rule. We need both source cases in order to complete the logical move. Yet the need for them is not essential. In the final reckoning, once we have ascertained that the distinctive features are irrelevant to the rule being learned, one can derive from either source case alone, since each contains the common factor that causes the rule. The derivation requires the two source cases, but the conclusion itself does not require them at all.

By contrast, in the two examples of conceptual construction above, the learned rule is not merely a derivative of the two primary cases in the sense of a derivation, but a concept actually built and composed out of both of them. Each part of it is learned from one of them, and therefore even at the end of the process we still require both in practice in order to build it-and not merely that each one is needed to remove a barrier, that is, to block a refutation against deriving from the other, as in the rule of 'mah ha-tzad'.[21]

Thus, in the two examples discussed above, it seems that we are not dealing with the rule of 'mah ha-tzad' as we know it in the Talmud. We seem to have here a different mechanism: the construction of a concept out of two other concepts, with the resulting derivative composed of both together.

The example of a derivative from two primary categories of damage

However, when we turn now to the example from Bava Kamma 6a mentioned above, this way of looking at things becomes problematic. The discussion there concerns liability for compensation when a person placed his stone, knife, and burden on the top of his roof, they subsequently fell in an ordinary wind, he declared them ownerless, and they caused damage. As stated, the Gemara learns this from a pit and fire, and the process is as follows: a pit proves it. What is distinctive about a pit? That no other force is involved in it. Will you say the same of these, where another force is involved? Fire proves otherwise. What is distinctive about fire? That it typically moves and damages. A pit proves otherwise. And the rule returns.

The formal structure is that of 'mah ha-tzad,' and so it is presented in the Mishnah at the beginning of Bava Kamma, in the Gemara, and by the commentators there. Yet if we examine it more closely, we will see that there is also here a conceptual construction very similar to the examples given above.

One could certainly understand the process in Bava Kamma 6a as built of two stages, each based on one of the two source cases. In the first stage, the stone, knife, and burden fell from the roof in an ordinary wind, and this liability we prove from fire, whose movement by an ordinary wind does not exempt one from payment. That is, we learn that a damaging force that goes with the wind still obligates the owner to pay. Afterward, when they reach the ground, a different primary category of damage is created here: a pit. The fact that it was created by an ordinary wind is, as we have already seen, irrelevant, because the primary category of fire taught us that this does not exempt him from liability. At this stage the second primary case, the pit, enters and teaches that my property lying in the public domain and causing damage-even though it does not ordinarily move and damage-obligates me to pay.

Thus, although this is presented as 'mah ha-tzad,' as is evident from the wording of the Mishnah and the Gemara, even here we are apparently constructing the new damager out of a sequential combination of two source cases. The process here does not look like the mere removal of refutations and derivation from a common factor. More than that: in this example as well, it seems that even at the conclusion one cannot dispense with either of the two source cases, and in that too it resembles conceptual construction and not the ordinary rule of 'mah ha-tzad,' as explained above.

On the other hand, as noted above, the Mishnah at the beginning of Bava Kamma presents this as a derivation from 'the common factor.' That is, the determining element is the logical structure, not the conceptual construction. The logical structure is that we prove from the two primary cases that the traits of being stationary and involving another force are not traits that exempt from liability for damage. It therefore becomes clear that the common factor-namely, that all these are one's property and that their safeguarding is incumbent upon him (as is explicit in the closing words of the Mishnah at the beginning of Bava Kamma)-is what obligates him to pay. It is from that common factor, which exists in each of the two source cases, that we learn the case of the stone, knife, and burden. This is already a regular derivational process of 'mah ha-tzad'.

The conclusion is that the example in Bava Kamma can be presented both as a derivational rule of 'mah ha-tzad' and as a process of conceptual construction.

Again: 'mah ha-tzad' and conceptual construction

It therefore seems that we must examine whether the rule of 'mah ha-tzad' and conceptual construction are not simply two aspects of the same phenomenon. In other words: can the other examples also be presented in both ways? More specifically: can our two examples be presented as a derivational 'mah ha-tzad'? And can the 'mah ha-tzad' of the Sifra be presented as a conceptual construction? Let us now try to examine this.

Let us begin with the example of the bailee's oath, and try to formulate it as a derivational process of 'mah ha-tzad': let us derive an oath obligation in the case of one who admits five vines from the law of admission of part of the claim. One may object: what is distinctive about admission of part of the claim is that there is an exemption of heilekh; if so, here too he should be exempt on the basis of heilekh. The bailee's oath proves otherwise, for there there is no exemption by reason of heilekh. One may then object: what is distinctive about the bailee's oath is that it applies where it is already known that he is a bailee, whereas here he denies the very safekeeping. Admission of part of the claim proves otherwise, for there too he denies the very loan, and yet he is liable to an oath. And the rule returns.

Similarly with regard to spitting in the public domain, we could formulate it in parallel fashion: let us derive that one who spits in the public domain is liable from winnowing. One may object: what is distinctive about winnowing is that it separates waste from food. Throwing proves otherwise, for it does not separate, but merely transfers the object in the public domain. What is distinctive about throwing is that no other force is involved in it. Winnowing proves otherwise, for another force is involved in it and one is still liable. And the rule returns.

It should be noted that describing these two examples as 'mah ha-tzad' is problematic and somewhat forced. The need for the second primary case does not stem only from the fact that there is an exemption of heilekh in admission of part of the claim; for even if there were no heilekh exemption, we would still need the bailee's oath in order to impose the bailee's oath in addition to the oath of admission of part of the claim. And vice versa. It therefore stands to reason that these are not examples of 'mah ha-tzad' at all, but of a different logical process, in which the new concept is built from the sequential combination of its two predecessors.

By contrast, it seems that the 'mah ha-tzad' of the Sifra cannot really be presented persuasively as conceptual construction. An attempt to do so might appear formally similar, but it is clear that in essence it is not a process of conceptual construction at all, but a purely derivational process.

Viewed broadly, it seems that the difference between these three cases and the 'mah ha-tzad' of the Sifra depends on the question whether we are trying to create a new halakhic concept, with the laws branching out from the very fact that such a concept has been created, or whether we are trying to learn an application of a certain halakhic principle to an existing concept-without creating a new halakhic concept. In the Sifra's example of 'immediate and for future generations,' it is clear that no concept is being created; rather, a halakhic principle that applies to the two teaching commandments is being applied to other commandments. Therefore such a process can apparently be formulated only as a derivational 'mah ha-tzad,' and not as conceptual construction. Perhaps for this reason the Sifra chose precisely this example in order to illustrate the rule of binyan av from two verses.

By contrast, in Bava Kamma we are trying to create a new halakhic concept-or a new primary category of damage-and the laws (the payment obligations) will then apply to it automatically, by virtue of its very definition as a primary category of damage. Therefore there the process can be presented also as conceptual construction. It should be noted, however, that as we have seen, in that case too one can, though somewhat awkwardly, present the process as a purely derivational process of 'mah ha-tzad'.

It should further be noted that in our two examples one cannot conclude with 'the common factor…,' because the two primary cases have no common factor. True, as we have seen, one can also present the derivation as a derivational process and not only as a conceptual construction, yet even that derivational process is not an ordinary 'mah ha-tzad.' It too is structured in such a way that, even at the end, it requires both teaching source cases.

In the Bava Kamma example, by contrast, one can apparently conclude the 'mah ha-tzad' by speaking of the common factor of the source cases-that they are your property and their safeguarding is upon you. Therefore that case is presented as a derivational process of 'mah ha-tzad'. For our purposes, what matters is to point out that it also has clear characteristics of conceptual construction.[22]

Summary

To summarize: we have seen two types of learning from two source cases:

Conceptual construction, which creates a new halakhic concept, while the laws branch out from the very definition of that concept. We have seen that this process can be presented both as conceptual construction and as a derivational process, but in this case even the derivational process requires both teaching primary cases.

'Mah ha-tzad,' which introduces a new rule with respect to an existing subject. This can be presented only as a derivational process-a hermeneutic rule by which the Torah is interpreted-and not as conceptual construction. As we have seen, in this case the derivational process ultimately does not require both primary cases.

With respect to our two examples, we have seen that the more plausible view is to treat them as conceptual construction rather than as a derivational process of 'mah ha-tzad'. The example from the Sifra, by contrast, is specifically a derivational process and not conceptual construction.

As for the examples in Bava Kamma 6a, however, there is room for hesitation. As stated, the Mishnah there presents them as 'mah ha-tzad,' and so the medieval authorities understood them. Yet the Gemara concludes each derivation with the words 'and the rule returns,' and not with the words 'the common factor.' Accordingly, there may perhaps be room to understand them too as a form of 'mah ha-tzad,' in which case there are indeed two types of 'mah ha-tzad,' as above. Or perhaps this is nothing other than conceptual construction, in which case one can distinguish between it and 'mah ha-tzad,' which is only a purely derivational process. In the next section we shall see that the medieval authorities probably disagreed on this very point.

E. Implications and practical ramifications

The dispute between the Rosh and 'the great authorities' concerning derivatives from two primary categories of damage

At first glance, a derivational process that requires both primary cases is nothing other than conceptual construction. If so, it follows that when we are trying to create a new halakhic concept, there are not two possibilities but only one: conceptual construction. To understand why this is not so, let us examine the dispute between the Rosh and 'the great authorities' in the sugya of Bava Kamma 6 mentioned above. There we will see a practical ramification of the difference between conceptual construction and 'mah ha-tzad'.[23]

The Rosh, in the sugya there, discusses the question of the legal status of derivatives learned from two different source cases. Which exemptions, if any, apply to them? The Rosh writes (Bava Kamma, sec. 1):

And some of the great authorities wrote that one is liable only for what is liable in both, and exempt from damage to vessels and from causing the death of a person, like a pit, and from concealed items, like fire. For since they come through 'mah ha-tzad,' we assign them the lesser law of the two.

And some were uncertain about the matter.

But it seems to me that they have all the laws of a pit… It is a full-fledged pit; rather, we could not derive it from a pit alone because it does not fully resemble a pit-from the outset either because it was not made to cause damage, or because another force is involved in its action, or because his actions did not cause it, or because it was made on his own property. It was therefore necessary to bring a proof from the other forms of damage, to show that these refutations do not distinguish them from a pit in a way that prevents liability.

Therefore these too do not leave the category of pit because of these refutations, and it is a full-fledged pit, and all the laws of a pit apply to them, its leniencies and its stringencies. And we do not assign them the leniencies of fire and foot…

The Rosh here cites a dispute among medieval authorities concerning the examples that appear in the Gemara there-of derivatives learned from two different source cases. The question in dispute is which specific laws apply to these derivatives. For example, to the case of one's stone, knife, and burden learned from a pit and fire, one might assign the laws of a pit (which is exempt with respect to vessels and the death of a person), and/or the laws of fire (which is exempt with respect to concealed items).

It appears from the Rosh's words that there are here three views among the medieval authorities:[24] there are some of the great authorities who hold that the derivative receives the distinctive laws of both teaching source cases together; there are others who are uncertain; and the Rosh himself holds that the derivative receives only the distinctive laws of a pit.

At first glance, this is a general dispute in the laws of 'mah ha-tzad': how to define the new rule learned from two primary cases. But, as we shall immediately see, it is not altogether clear that this is not instead a unique dispute.

The root of the dispute

If we are indeed dealing here with a regular process of 'mah ha-tzad,' then according to the standard description-that we remove obstacles and then derive from the common factor-it would seem to follow that the derivative should receive none of the distinctive laws of the primary cases. It is not fully like either of them, but is learned only from the factor common to both. The reason is that the distinctive laws stem from the distinctive features of each primary case, and the common factor by itself generates no distinctive rule at all. For if it did, that rule would apply in all primary cases, and would therefore no longer be a distinctive rule.

Thus, the case of one's stone, knife, and burden learned from a pit and fire should have been exempt with respect to concealed items and with respect to vessels, since the common factor does not obligate those. The liability or exemption in each of those cases stems from the unique character of each of the primary categories here, and that unique character does not characterize the derivative. Hence the derivative should have been exempt by all the exemptions of both primary cases.

There is, however, room to argue the reverse: that the derivative should specifically have been liable in every case, because the distinctive exemptions-not the liabilities-that the primary cases have are what derive from the unique nature of those teaching primary cases, a nature not shared by the derivative. If so, the derivative has no such distinctive exemptions. According to this view, the common factor obligates in every case, and each distinctive primary case is exempt-rather than liable-because of its distinctive nature.

This dilemma depends on whether the distinctive rule is that a pit is liable for concealed items, or that fire is exempt from concealed items. Alternatively: whether the distinctive rule is that a pit is exempt with respect to vessels, or that fire is liable with respect to vessels.

This dilemma is formal-that is, it follows from the very logical structure of 'mah ha-tzad'. But conceptually it seems more plausible that the exemption, rather than the liability, is what is distinctive. The reason is that the exemption exists only in one primary category of damage, whereas the liability exists in all the others. Hence it seems more plausible that the exemption is the distinctive feature, not the liability. Accordingly, the common factor should obligate in every case, and not confer any of the exemptions.

By contrast, if this is indeed a case of conceptual construction-that is, a new primary category of damage built from two existing primary categories, pit and fire-then the new primary category actually contains a combination of pit and fire, and not merely as derivational premises. In that case, it follows conceptually that it should have both exemptions of the two primary cases. Here too there is room for formal disagreement-perhaps the exemption depends on its having only its own characteristic and not that of its fellow, so that the composite derivative would not have the exemption-but conceptually the view just stated seems more plausible.

It therefore appears that the great authorities understood the case as conceptual construction, and for that reason they held that the derivative has both exemptions of the two primary cases. The Rosh, by contrast, understood that in practice we learn primarily from a pit, while fire merely removes a derivational obstacle-that is, we are dealing with a derivational process of 'mah ha-tzad'. Even so, there is clearly an innovation in the Rosh's words. His innovation is that there are cases in which the derivational rule of 'mah ha-tzad' is not 'the common factor.' That is, we do not learn from the common factor but from one of the sides. The second source case serves only to remove a derivational obstacle, as explained above.

As for the medieval authorities whom the Rosh cites as uncertain, it is unclear whether their uncertainty lay between these two positions, or whether a third possibility stood before them: that the derivative would have no exemptions at all.[25] That conclusion would follow from understanding the process as a derivational 'mah ha-tzad,' but without the Rosh's above innovation. In any event, these authorities agree in principle with the Rosh that we are dealing here with a process of 'mah ha-tzad' (except that one may wonder whether there is a dominant primary case, or whether both have equal status-that is, whether we derive from one 'side' or from the common factor).

We should note here that in our discussion above (at the end of section B) we saw that in the conceptual construction of the bailee's oath there are features of the source oaths that pass over to the derivative (as in the case of wheat claimed and barley admitted, which exempts both from the oath of admission of part of the claim and from the derivative oath), and there are features that do not pass over (such as the exemption of heilekh, and the rule requiring measure and number, where one is exempt from the oath of admission of part of the claim but not from the derivative oath).

There we explained this in terms of the nature of the exemption: whether it removes the teaching feature-that is, the legal conclusion that there was a contract, whether of loan or of safekeeping-or not. In the context of damages it is harder to see matters this way, since the exemptions appear to be scriptural decrees,[26] and therefore the discussion conducted here is more formal and less substantive.

According to our earlier analysis, it may be that the dispute between the Rosh and the great authorities can be explained on that plane: everyone agrees that in damages this is conceptual construction, and they disagree only whether the distinctive exemptions are substantive (as in the case of wheat claimed and barley admitted), or not (as in heilekh, and the requirement of measure, number, and weight). Below we shall see from the Rosh's own position that it may be that he does not accept the very process of conceptual construction at all.

The dispute among the commentators: is 'mah ha-tzad' always a binyan av?

Let us now note another point that bears on the distinction between 'mah ha-tzad' and conceptual construction. As is well known, there are two types of 'mah ha-tzad': one begins from similarity-that is, a binyan av from one verse-to the first teaching source case, and after a refutation is raised, a second source case ('proves otherwise') is brought in to remove the refutation. The second begins from an a fortiori inference from the first teaching source case, and then continues in similar fashion.

Now the medieval authorities and later scholars disagreed whether 'mah ha-tzad' of the second type, after the derivational process from the two primary cases is complete, has the status of an a fortiori inference, or whether in the final analysis it takes on the status of a binyan av-or, in other words, of 'the common factor' (see, for example, Encyclopedia Talmudit, entry 'binyan av', notes 59-61).

In Halikhot Olam (gate 4, chapter 2) and Middot Aharon (chapter 2, section 9), it is understood that at that stage it is a binyan av-that is, like an ordinary 'mah ha-tzad' of binyan av. By contrast, the Shelah (Torah She-Be'al Peh section, the rule of a fortiori inference), and as cited by Middot Aharon there in the name of Rabbi Yosef Almushnino, held that it remains an a fortiori inference.[27]

It seems that this dispute is parallel to that between the Rosh and the great authorities. If in the end one learns from the common factor, then it is indeed a binyan av. For although with respect to each of the teaching source cases there is initially a relationship of a fortiori inference, that relationship is refuted because of the distinctive features of each of the primary cases. Hence in the end we are left only with a derivation from the common factor, and that is a binyan av derivation.

But if the basic derivation is from one of the primary cases, and the refutation of the a fortiori inference is what is repelled by the 'proves otherwise' of the second primary case, then what remains in the end is the derivation from the first primary case, which was made by an a fortiori inference.

From the perspective of our topic, one should note that in conceptual construction it is clear that the derivation that remains at the end of the process is an a fortiori inference. The reason is that each of the primary cases builds half of the derivative, and with respect to that half it teaches by an a fortiori inference. The 'refutation' raised against it is that it lacks the second half, but that half is built from the second primary case, also by an a fortiori inference. In other words, there seems to be no room for that dispute of the medieval authorities here; according to everyone, what we have here would be an a fortiori process.

According to our explanation above of the view of the great authorities-that the example in Bava Kamma is conceptual construction and not a regular derivational 'mah ha-tzad'-it is obvious that this question is irrelevant to the dispute among the authors of the hermeneutic rules. The great authorities did not intend to state a general principle about 'mah ha-tzad,' but only about contexts in which 'mah ha-tzad' appears as conceptual construction. In other contexts, where we are dealing with a derivational rule, they may well agree with the Rosh that the features of the principal teaching source case pass over to the derivative, and that ultimately one learns only from it. If so, it may be that according to everyone the final derivation in the type of binyan av discussed here-which is a derivational 'mah ha-tzad,' and not conceptual construction-would be an a fortiori inference and not a binyan av.

F. Primary categories and derivatives on the Sabbath and in damages

On the Sabbath there is, if anything, only conceptual construction

Let us return now to the question why we do not find a derivative from two primary categories in the laws of the Sabbath, whereas in damages we do find such a thing.

First, we must note that in the example of spitting we do find conceptual construction in the laws of the Sabbath. It is true that this does not seem to be accepted by most decisors, since we saw that some reject the Rema as practical law. Nevertheless, according to Alfei Menasheh (and the Mishnah Berurah, who agreed with him) and the Rema, such a logical process exists in the laws of the Sabbath. Moreover, according to the view of Rashi (Bava Kamma 60a), who understands-unlike the Rosh there, who disagrees with him-that the case of winnowing with the assistance of the wind is a general principle in all the labors of the Sabbath, it would appear that we would have to say so even without the Jerusalem Talmud. According to Rashi, this is a principle that necessarily emerges from the words of Rav Ashi in the Babylonian Talmud; in fact, that itself is what Rav Ashi says there. If so, according to Rashi not only is the prohibition of spitting not contradicted by the Babylonian Talmud, but the principle from which it follows-namely, conceptual construction-is explicitly stated in the Babylonian Talmud itself.

Thus, from Rashi's words it follows that an action performed by means of the wind-which is ordinarily treated as indirect causation-will incur liability in all the labors of the Sabbath. In the simple sense, his intention is precisely a case of this kind: performing some labor with the help of the wind, where the intended result is achieved. If so, he goes even further than Alfei Menasheh. According to Rashi, it appears that there would be such derivatives in many Sabbath labors together with winnowing, and not specifically only in throwing and carrying four cubits in the public domain. The conclusion is that, at least according to Rashi, there is an appearance of a derivative of two primary categories on the Sabbath as well, except that it arises from a process of conceptual construction and not from the derivational rule of 'mah ha-tzad'.

In damages, the medieval authorities disagreed on this

Conversely, in damages we do indeed find 'mah ha-tzad'; however, according to the great authorities cited by the Rosh, this is conceptual construction and not ordinary 'mah ha-tzad'. If so, conversely, it seems that in damages too there is no ordinary derivational 'mah ha-tzad,' but only conceptual construction.

The conclusion that emerges according to the view of the great authorities cited by the Rosh is that whenever the Gemara enumerates a set of primary categories (and therefore of derivatives), it intends a complete set, each component of which is necessary. If we were to learn other derivatives through the derivational rule of 'mah ha-tzad,' it would turn out that in fact we do not need this set, each member of which has unique characteristics, since what obligates-payment in damages, or stoning for the Sabbath-is only the common factor. Therefore in such situations a derivative by means of a derivational hermeneutic rule ('mah ha-tzad') is not conceivable; only conceptual construction is.

'Mah ha-tzad' will appear in Jewish law only when the two teaching source cases are not members of a complete set of 'primary categories,' or of some other complete taxonomy. One can make a 'mah ha-tzad' only when the two source cases come from two different domains, and therefore do not claim to cover an entire field. Only then can one derive from their common factor a halakhic principle regarding another case.[28] In such cases, the goal is generally not to create a new concept, but to derive a halakhic principle.

When the teaching source cases are members of a set, that set includes all the components that exist in Jewish law in that area. In such a case every component of the set is essential, and it is impossible to say that it represents only the common factor of all of them-for if that were so, there would be no point in constructing the set, and we ought merely to present the common factor.

It should be noted that the Gemara (Bava Kamma 5b) concludes that in fact the primary categories of damage are indeed redundant, since all of them could have been learned from a pit plus one other category. This is exactly in line with our claim that a set enumerated by the law cannot represent only a common factor, for otherwise we would not need the whole set and most of its items would be redundant. Yet according to that, it is truly unclear why the Torah wrote all the primary categories of damage. To that the Gemara answers explicitly that it was only for their specific laws-that is, for their distinctive exemptions, which do not depend on the common factor but on their unique features. See there in the medieval authorities, and also Hiddushei Ha-Griz at the beginning of Hilkhot Nizkei Mamon.

All this follows according to Rashi, the great authorities cited by the Rosh, and perhaps also Maimonides and Rabbi Yosef Migash above. But as we have seen, the Rosh (Bava Kamma 60a) disagrees with Rashi in the laws of the Sabbath, and holds that one cannot perform conceptual construction-from the primary labors of the Sabbath, at least not the sort suggested by Rashi. In addition, we find that he disagrees with the great authorities in Bava Kamma 6a, and there too he holds that conceptual construction should not be made from the primary categories of damage.

A possible generalization of the Rosh's view

It may be that this is a general approach of the Rosh, namely, that he does not accept this logical process at all. In his view, concepts cannot be built out of other concepts, and all that remains to us is the derivational process of 'mah ha-tzad,' which is a hermeneutic rule by which the Torah is interpreted.

We should note that the Rosh in Shevuot does not address this sugya, but in his Tosafist commentary there (42b, s.v. 've-halleh') he explicitly writes that the grapes were harvested, for otherwise the case would be one of heilekh. This proves that he understood the case there as involving the oath of admission of part of the claim and not the bailee's oath, contrary to Maimonides and Rabbi Yosef Migash. Perhaps there too he follows his general approach, in line with the majority of the medieval authorities.[29]

Of course, according to the Rosh, the question returns: why do we find 'mah ha-tzad' among the primary categories of damage and not such derivatives in the laws of the Sabbath?[30]

G. Additional examples: the significance of conceptual construction

It seems that we find conceptual construction in Jewish law in additional contexts as well. We shall now consider two such examples.[31]

Maimonides' view concerning the additional offerings of the Sabbath on Yom Kippur

Maimonides (laws of the Yom Kippur Service 4:1) rules that the additional offerings of Yom Kippur that falls on the Sabbath are offered only by the High Priest. The author of Or Sameach there explains that Maimonides means that the sacrifices offered on Yom Kippur that falls on the Sabbath are regarded like the sacrifices of Yom Kippur itself, and therefore they too must be performed by the High Priest.

Or Sameach there draws from this a practical conclusion: a sick person who must eat on Yom Kippur that falls on the Sabbath is not obligated in the sanctification of the day-as is well known, later authorities disagreed on this point.

The explanation is that Maimonides understands Yom Kippur that falls on the Sabbath not as a combination of two days that happened accidentally to coincide. This combination defines a day whose essence is different from that of either of them, a third essence. Therefore the laws applying to it are not a mere aggregate of the laws of the two days that compose it. Hence on Yom Kippur that falls on the Sabbath there is no law of kiddush at all, even not from the side of the Sabbath laws within it, for this day is not simply a Sabbath. Of course it is also not an ordinary Yom Kippur. It is Yom Kippur that falls on the Sabbath-that is, a combination that defines a day of a third type. Naturally, we learn the laws that apply to it from the known types that compose this day, but this is not a simple aggregation of those laws.[32]

If so, we have here another example of conceptual construction. We build a new concept of a festival day of a new kind out of two existing concepts of sacred days.

Here one can discern the nature and significance of such a composition, perhaps more sharply than in the earlier examples. As we see here, in a concept produced by conceptual construction there is indeed a component from each of the source primary concepts, but it is certainly not their simple sum.

If it were a simple sum, we would not say that a new concept has been created, but, in Rabbi Chaim's language, that there are 'two laws.' Yet here we see that these are not merely two laws, but one law of a third type. If we had Rabbi Chaim's 'two laws' here, then the Sabbath sacrifices would be offered by an ordinary priest, whereas the Yom Kippur sacrifices would be offered by the High Priest. Maimonides' main innovation is that this is not the correct way to view the matter. Even the Sabbath sacrifices take on the character of Yom Kippur, as explained.

A festival that falls on the Sabbath

Something similar is found in the responsa Tzafnat Paneach (sec. 2), where he discusses the question whether one who performed labors permitted for food preparation on a festival that falls on the Sabbath-such as one who cooked on such a day-transgressed a prohibition also on account of the festival aspect, or only on account of the Sabbath aspect.

His conclusion there is that he transgresses also on account of the Sabbath aspect of the day. See there, however, that his formulation is not like that of his fellow townsman, the author of Or Sameach, and his explanation is different. For our purposes, however, we would say here too that the combination 'a festival that falls on the Sabbath' is not an accidental juxtaposition of two types of day, but the creation of a day of a third type out of the two known types.

If so, it appears that here too we encounter conceptual construction. See there in the responsum of Tzafnat Paneach for further examples of this matter.[33] In any event, we learn that conceptual construction creates a new concept that is not a simple sum of its components. The construction is not mere juxtaposition, but a more complex logical building of a new concept out of two known concepts. In the language of the Rogatchover, one may say that there is here an 'amalgamative composition' between the two concepts, and not a merely 'juxtapositional composition'.[34]

A possible generalization in Maimonides' approach

Let us conclude the section with a general remark. If above we saw a consistent approach on the part of the Rosh against conceptual constructions, perhaps here we are seeing that Maimonides too is following a consistent line in the opposite direction. We saw that in the laws of oaths he relies on conceptual construction, and here too it appears that he does so-unlike the Rosh mentioned above.

H. A brief logical discussion

As we have seen thus far, there are at least two ways to learn a new derivative from two-or several-primary categories:

1. The derivation of 'mah ha-tzad' is based on generalizing from the two primary categories, and the learning with respect to the derivative is made directly from that generalization to the derivative. In fact, nothing new is created here; the learned case-the derivative-is simply another example of the general principle produced by the generalization of the two primary categories. More than that: the derivation from 'the common factor' is essentially a derivation about a general principle that is relevant to a general class of cases-all the cases included in the generalization, that is, in the common factor, of the two primary categories.
2. Conceptual construction, which is not based on generalization at all, and therefore is not based on the common factor of the two primary categories. Rather, it is a specific composition of the two primary categories, one that takes account of their distinctive natures and not only of their common side. This composition creates a specific concept composed of those two particular primary categories. Here there is no creation of a rule regarding a general class of concepts or situations, but the creation of one new concept.

The process of 'mah ha-tzad' is an expansive one. It involves generalization from two known cases and the creation of a broad rule (see the Sifra example above concerning 'immediate and for future generations,' where a principle is learned that applies to all commandments). To do that, we must identify the common factor, which is the element essential to the rule being learned, and remove the elements that are not essential. This is the logical process of generalization, which in logic is called induction.[35]

Conceptual construction, by contrast, is almost mathematical. It does not involve expanding the primary categories into a broader class, but rather specifying them-or at most learning on the basis of similarity to them. One can view the process of construction as a binyan av, and this is how the great authorities cited by the Rosh in Bava Kamma understood it, since they placed it under the heading of binyan av. According to this view, what we have here is an inference on the basis of resemblance-what in logic is called analogy.

As stated, however, one can also view the process of construction as an entirely analytic one. According to this view, we are simply analyzing the two primary concepts and necessarily deriving their derivative from them. Here the relevant mode of reasoning is what in logic is called deduction.[36]

The Rosh certainly did not understand it in this way, for otherwise he could not have disputed the validity of the process of conceptual construction-deduction is a necessary logical process. But it seems that even the great authorities, although they did accept the process of construction, did not conceive of it as deduction. Had they understood it as deduction, they would not have classified it among the hermeneutic rules by which the Torah is interpreted ('mah ha-tzad'), for such rules are analogical in nature.[37]

In the words of Maimonides and Rabbi Yosef Migash it is not explained how they understood this process, and their view still requires further study.[38]

Appendix: The meaning of the intermingling of passages

The significance of our explanation of Maimonides' view

As is well known, most medieval authorities hold that a bailee is required to swear only when he claims that the property was lost through unavoidable circumstances, and his claim is meant to exempt him from liability for it. But, as we saw above (note 7), Maimonides' view is different: according to him, a bailee must swear whenever he advances an exempting claim. For example, when a bailee claims 'I returned it,' according to Maimonides he is required to swear.

Here we may add that this also appears to be Rashi's view (Bava Kamma 107a, s.v. 'ma'iz u-me'iz'), who holds that in the case of a deposit, unlike a loan, there is an oath obligation even on complete denial. Rashi there explains that the verse 'for this is it' refers only to a loan and not to a deposit, even though it appears in the Torah within the passage of bailees-thus there is an intermingling of passages-because in a loan the defendant is exempt from an oath when he denies everything, since there is a presumption in his favor that a person is not brazen before his creditor.

Tosafot, in s.v. 'eiruv,' and other medieval authorities, raised many objections to him and proved from several places that one who completely denies a deposit by saying 'I never received it' is exempt from an oath.

According to our approach, one may say that Rashi's intention is not that every complete denial concerning a deposit incurs an oath. It may be that Rashi intends to require an oath only of one who completely denies a deposit by claiming 'I returned it'; in that case, even if he says so concerning the whole amount, he is liable to an oath. The reason is, as in Maimonides' view, that this is an exempting claim and not a true denial. Rashi's point is that with a deposit, every exempting claim-even where it is made concerning the whole claim and not merely part of it-incurs an oath. This resolves the objections of those medieval authorities, since their proofs concern only a bailee who says 'I never received it,' and this is not the case under discussion here.

If so, it may be that Rashi, like Maimonides, holds that a bailee is liable to an oath on a claim of return, and indeed on any exempting claim (as distinct from 'I never received it,' which is a denial and not an exempting claim). Earlier in the body of the article we saw that Rashi also engages in conceptual construction in the labors of the Sabbath, and therefore it may be that he accepts such construction in the context of bailees as well, in line with Maimonides and Rabbi Yosef Migash.

It should further be noted that according to our proposal it is highly plausible that the two oaths-the oath of admission of part of the claim and the bailee's oath-are different applications of one and the same oath. One might call it the oath of defendants on the defensive. Every defendant who stands in a defensive posture must swear. A bailee who advances a defensive claim (such as unavoidable loss or return), and one who admits part of the claim-these are both examples of a defendant on the defensive, and therefore both incur an oath obligation.

This explanation clarifies the intermingling of passages described by the Gemara in Bava Kamma 107a. The Gemara there concludes that the verse 'for this is it' refers to the oath of admission of part of the claim in a loan, even though it appears in the Torah within the context of a deposit. At first glance, it is difficult to understand why the Torah interweaves these two passages with one another.

According to our proposal, one may say that these are two examples of the same oath, and therefore the passages are appropriately interwoven. The entire section deals with the oath of a defendant who is on the defensive, and it appears within the framework of the passage of bailees. The bailee is a clear example of a defendant on the defensive, and within that framework the Torah brings another example of such a defendant-who is also liable to that same oath-namely, a defendant who admits part of the claim.

This also explains why we find, among some of the Amoraim (see the sugya in Bava Kamma there, Bava Metzia 98, and the parallels), and also among some decisors (see Tosafot Yom Tov on the Mishnah in Shevuot cited in the body of the article), that they require admission of part of the claim even for the bailee's oath. The source of their view is apparently the very fact that the Torah intermingles these two passages, and therefore they understand that a bailee too, like a borrower, requires admission of part of the claim in order to incur an oath.

Maimonides, however, does not so rule (see Tosafot Yom Tov there), and he obligates a bailee-who is on the defensive-to swear even without admission of part of the claim. It is therefore clear that Maimonides must explain the intermingling of passages differently. His understanding of it is as we proposed above: every bailee on the defensive is like one who admits part of the claim. Admission of part of the claim is only one example of a defendant on the defensive, as explained. In other words, the passage deals with defendants on the defensive, and it all appears within the passage of bailees, which is the clearest example of such a defendant. The oath of admission of part of the claim and the bailee's oath are two examples of one and the same oath, and therefore it is no wonder that the Torah interweaves them with one another.

The significance of the intermingling of passages: another possible angle on the case of the vines

We have seen that the oath obligation regarding the five vines that the bailee denies arises from the bailee's oath in light of the logic of admission of part of the claim. Yet in light of our discussion of Maimonides' view, it follows that in every exempting claim made by a bailee he will be liable to swear. As we saw here, it may be that every defendant on the defensive-who is not simply denying-must swear, and a bailee on the defensive, as well as one who admits part of the claim, are only two examples of this.

If this is indeed the underlying conception of these oath obligations, then the bailee's oath and admission of part of the claim are two examples appearing in the Torah of one and the same oath: the oath of the defendant on the defensive.

According to this, it may be that the construction of the bailee's oath in the case of the vines is not conceptual construction, but simply another example of the oath of the defendant on the defensive. This is another case of a defendant in a defensive posture, and it may be proved from the two known examples-bailees and admission of part of the claim-that it too constitutes an example of this oath.

But if so, this case is to a considerable extent parallel to the 'mah ha-tzad' of the primary categories of damage. There too, the two different primary cases may be understood as teaching their common factor-that both are his property and that their safeguarding is upon him-so that every other damager sharing this feature, namely that it is his property and its safeguarding is incumbent upon him, will obligate him to pay. There too we saw that the process can be understood both as 'mah ha-tzad' and as conceptual construction. It is therefore not surprising that in the case of the vines too one may perhaps see a derivational process of 'mah ha-tzad,' of the kind involved in the case of the stone, knife, and burden that fell from the roof. Here too, as there, it may be understood in both ways: both as conceptual construction and as 'mah ha-tzad'.

Accordingly, one may perhaps formulate the oath obligation in the case of the vines as follows: from admission of part of the claim and the bailee's oath one can identify a common factor-that in both there is a defendant on the defensive. Therefore any other case of a defendant on the defensive, as in the case of the vines, will likewise be liable to an oath. Once we complete that derivational process, either one of the two teaching source cases alone will again suffice, and we will no longer need both. If so, here too the process is derivational, exactly like the ordinary 'mah ha-tzad'.

Perhaps all this will depend on the dispute between the Rosh and the great authorities-whether such a process is 'mah ha-tzad' or conceptual construction. This still requires further study.

[1] See, for example, at the end of section D (before the summary, and in the summary itself).

[2] See on this matter Encyclopedia Talmudit, entry 'binyan av'.

[3] There are, however, places where the reading is that there are things that are 'in the ground,' with a bet, as in our Mishnah, and places that read 'as the ground,' with a kaf. There are also places where the statement of Rabbi Yosei bar Hanina appears as an ukimta, and the dispute is interpreted more broadly, and places where it seems that he establishes the dispute specifically in the case of grapes ready to be harvested itself. It may be that these two differences depend on one another, but this is not the place.

[4] According to this, in the other contexts in the Talmud that are presented as depending on the dispute between Rabbi Meir and the Sages, we will have to say that there are circumstances similar to the case of bailees, and therefore there too the grapes are not deemed detached for those reasons. This is not the place to elaborate.

[5] In my humble opinion, Rabbi Yosef Migash may understand that the Sages disagreed only in the context of a bailee, whereas Maimonides understands the dispute as sweeping. According to Maimonides, the split between the laws of safekeeping and the rest of halakhic law exists only on the level of practical ruling. There he rules like the Sages only in the laws of bailees. But according to his view-unlike what one might have inferred from Rabbi Yosef Migash-the Sages themselves held that the grapes are not deemed detached in all halakhic contexts.

This also seems to follow from his wording in the Commentary on the Mishnah quoted above, where the split is in the practical ruling and not in the interpretation of the Sages' position. In Rabbi Yosef Migash the matter is not clear. It seems to me that this would also resolve all the contradictions in Maimonides' rulings regarding grapes ready to be harvested and hair ready to be shorn-contradictions that the later authorities have already discussed at length-but this is not the place.

The later authorities, however, did not interpret Maimonides in the way I have suggested here. Most of them explained that he understands the dispute between Rabbi Meir and the Sages only with respect to the laws of bailees. In truth, according to my proposal it would follow that the continuation of the Mishnah, which deals with a claim arising from safekeeping, is not practical law, and that is forced.

Therefore, in what follows, we shall assume that according to Maimonides and Rabbi Yosef Migash the dispute between Rabbi Meir and the Sages in our Mishnah pertains only to the laws of bailees.

[6] Tosafot Yom Tov apparently understood that Maimonides would read the Mishnah as dealing with a claim that the grapes were damaged-that is, that the bailee was negligent in safeguarding them. The defendant admits that he was negligent with respect to some of the vines, and in that case it would be an ordinary bailee's oath. Tosafot Yom Tov then asks why a case of admission of part of the claim is needed. Yet in terms of the wording of the Mishnah this is very forced, and therefore Rabbi Akiva Eiger and the Shakh did not accept it. Below I shall propose an interpretation of Maimonides that fits the wording of the Mishnah, and according to that there is no need to enter into the forced explanation of Tosafot Yom Tov.

[7] At first glance, one might object to what we have said, for it is commonly accepted that a bailee who claims 'I returned it' is exempt from an oath, and there too he is a bailee seeking to free himself from an obligation of payment. 'I returned it' is a quintessential exempting claim, and therefore according to our proposal the bailee ought to swear in order to be exempt. This would seem to prove that not every time a bailee seeks exemption is he liable to an oath, but only when he claims unavoidable loss. Hence in our case too one might ask why to liken the defendant-who denies the five vines-to one who claims unavoidable loss, rather than to one who claims 'I returned it'. Is the claim 'I returned it' of one known to be a bailee weaker than the situation of admission of part of the claim in a safekeeping claim?

Now if we examine Maimonides' opinion on this point, it appears that the Maggid Mishneh and Kesef Mishneh disagreed about it on Hiring 2:11. Kesef Mishneh brought decisive proof for his view from Maimonides, Borrowing and Deposit 6:4 (see there also the Raavad, who disagrees), where Maimonides explicitly writes that on a claim of 'I returned it' the bailee swears while holding a sacred object. The Maggid Mishneh there was indeed greatly puzzled by Maimonides' words. See also Kesef Mishneh there, who repeats his dispute with the Maggid Mishneh on this basis.

And Kesef Mishneh writes in explanation of Maimonides' view in both places that his reasoning is that anyone known to be a bailee who seeks to exempt himself from payment is liable to the Torah-imposed bailee's oath. This is true whatever claim he makes.

If so, our remarks above fit Maimonides' view very well, and it is clear that he is indeed following his own consistent position. Here too we see that the obligation of the bailee's oath applies to anyone known to be a bailee who seeks to be exempt from payment, whatever his claim may be.

It should be emphasized that according to most medieval authorities, the bailee's oath applies only to a bailee who claims unavoidable loss, whereas one who claims 'I returned it' is not obligated in the bailee's oath. According to them, one cannot truly say that every bailee who advances an exempting claim (as distinct from a bailee who simply denies) is liable to an oath. But we have seen here that Maimonides' consistent position is that a bailee cannot be exempt-whether by claiming unavoidable loss or by claiming 'I returned it'-unless he takes the bailee's oath. It is clear that even according to Maimonides, a bailee who denies the entire safekeeping is exempt without a Torah oath, aside from a rabbinic oath, as explained above.

According to this, what we wrote above is well understood: from the logic of admission of part of the claim it follows that in our case the bailee is considered one who advances an exempting claim, and therefore according to Maimonides he becomes liable to the bailee's oath.

See below in the appendix for an additional explanation of the significance of this conception of Maimonides.

If we return to our present dispute, it follows that according to this the Raavad and the majority of medieval authorities who disagree with Maimonides here are also proceeding consistently in line with their method. As will be seen below, the Raavad holds like the other medieval authorities who do not accept Maimonides' innovation, and they maintain that our Mishnah deals with the oath of admission of part of the claim and not with the bailee's oath.

It should be noted that according to this school there is no plausible way at all to understand the Mishnah as dealing with the bailee's oath, even if one accepts the logical process that will here be called 'conceptual construction.' The reason is that the bailee's oath-one of the source concepts-does not teach us that every bailee who advances an exempting claim is liable to an oath, as is clear from the case of one who says 'I returned it'. Only Maimonides' understanding of the bailee's oath makes possible the creation of the new oath obligation described above.

If so, even the medieval authorities who explain the Mishnah in Shevuot differently from Maimonides and Rabbi Yosef Migash do not necessarily reject the logic of conceptual construction. Below we shall see that from the Rosh's view it emerges that he rejects it altogether.

[8] One may discuss whether we have here a new oath, which is neither the bailee's oath nor the oath of admission of part of the claim, or whether this is literally the bailee's oath, except that the conclusion that he is liable to it is based also on the logic of admission of part of the claim. See below in the dispute between the Rosh and the great authorities regarding a similar issue.

[9] This too, however, depends on the different understandings of the rule of heilekh, and this is not the place.

[10] For a possible explanation of this, see Maggid Mishneh on To'en Ve-Nit'an 3:15.

[11] See the Shakh, Hoshen Mishpat sec. 72, subsec. 50, who wrote this in the name of the Rosh on chapter 6 of Shevuot: he understood Maimonides as speaking of one who admitted less than a perutah, and he rejected his words. See also Ramban, Ritva, Rashba, and Ran in the sugya of a pledge in Shevuot (43a), who discuss this with respect to the view of Rabbi Yosef Migash.

[12] See, for example, the above-mentioned medieval authorities in the sugya of a pledge (Shevuot 43a). In various ways they explain Maimonides as meaning that there was an admission of a perutah, and the Shakh then objects, as I noted above. They also object to their own proposed reading of Maimonides on the basis of the rule of an item subject to measure, number, and weight.

[13] The medieval and later authorities discussed all this at great length in Maimonides' view. See, for example, Hiddushei Rabbi Yosef Dov Soloveitchik to Shevuot 43a (pp. 209-212 in his pagination), and Pri Moshe, Inyanei Shevuot Mamon, sec. 8, letters 2-3.

[14] Below we shall see a dispute among the medieval authorities-the Rosh and the great authorities in Bava Kamma 6a-as to whether features of the teaching source cases pass over to the derivative learned from them.

[15] When I discussed this with Rabbi Asher Deutsch, one of the heads of the Ponevezh Yeshiva, he suggested two explanations:

1. 'Mah ha-tzad' is a rule in interpreting the Torah. The primary categories of damage are written explicitly in the Torah, and therefore they can be expanded by means of the hermeneutic rules by which the Torah is interpreted. The labors of the Sabbath, however, are not written explicitly in the Torah; rather, the Sages defined them on the basis of the comparison of the Sabbath to the Tabernacle (and perhaps on the basis of a derivation that there are thirty-nine primary categories, from the fact that the verb 'melakhah' appears in the Torah thirty-nine times; see Babylonian Talmud, Sabbath 49b). Therefore they cannot be expanded by means of any hermeneutic rule.
2. The Sages define thirty-nine distinct primary categories of labor on the basis of the comparison of the Sabbath to the Tabernacle. If there were another labor learned by 'mah ha-tzad' from two primary categories, it should have to count as an additional primary category-the fortieth. But it is known that there are only thirty-nine primary categories of labor, and therefore it is clear that they cover all the fundamental possibilities of labor.

In my humble opinion, however, these explanations are insufficient. As for the first explanation, it is established that throughout the Torah one may learn from that which is itself learned, except in the area of sacrificial law (see Zevahim 50 and surrounding passages). That is, everywhere in the Torah except sacrificial law, one may use the hermeneutic rules even with respect to laws that are themselves the result of derivation. If so, why should we not be able to use 'mah ha-tzad' regarding labors learned from a derashah?

And as for the second explanation, one could say that here we are dealing with a derivative and not with a primary category. The labor that would be learned by 'mah ha-tzad' from two different primary categories of labor would be a derivative of both, not a new primary category. There is no reason to define it as a primary category, for it can be learned from the existing primary categories-true, two are needed, and one alone does not suffice. In damages too there are derivatives of two primary categories, and they are classified as derivatives. The reason they are so classified is precisely that they can be learned from the existing primary categories. Primary categories are fundamental modes of labor, or of damage, which cannot be learned from the other primary categories-not even from a combination of several of them-unless the Torah first teaches them directly. See below, section F, where we shall expand on the definitions of 'primary category' and 'derivative'.

If so, I still do not see a sufficient reason why such a case should not exist in the labors of the Sabbath as well.

In a note below, at the end of section F, we shall return to these two explanations and see that perhaps the first of them underlies the Rosh's view.

[16] It still requires examination why he should not be liable because of grinding. According to the view that grinding applies only to produce of the soil, one might say that a human being is like land; several later authorities explained similarly even with respect to spitting under the heading of winnowing. Or perhaps he is not at all interested in grinding, whereas in winnowing he is interested, since his desire is to distance the saliva from himself. See the note following the quotation.

[17] The medieval authorities, however, disagreed as to whether carrying and throwing are primary categories of labor or derivatives. See Nachmanides to Sabbath 73a, who holds that carrying four cubits is a derivative. Ritva there was uncertain about this. Maimonides, Sabbath 12:8, wrote that carrying four cubits is like taking out from the private domain to the public domain, and in 12:10 wrote that extending and throwing are derivatives.

If carrying is itself only a derivative, then spitting would be a derivative of a derivative. Yet according to most medieval authorities it seems that on the Sabbath one is liable even for a derivative of a derivative, apart from the view of Rabbenu Hananel on Sabbath 73b, and this is not the place. In any event, for our purposes there is no practical difference here, since Rabbi Menasheh explicitly learns from carrying and throwing with respect to spitting.

It should be noted that it is not clear whether his intention is that the relation between spitting and the two teaching source cases is the relation of a derivative to its primary categories; see below, section F.

[18] It should be noted that such a derivation depends upon a particular understanding of the sugya in Bava Kamma 60a, where Rav Ashi says regarding liability for winnowing with the assistance of the wind that on the Sabbath one is liable, whereas in damages one is exempt, because on the Sabbath the Torah prohibited intentional labor. The medieval authorities disagreed regarding the meaning of his words. Rashi there explains that his intended result was achieved, and he was pleased with the assistance of the wind. This implies that the statement is a general principle: if the intended result is achieved, then even when one performs labor with the wind one is liable. The Rosh there, by contrast, explains that although this is labor performed indirectly by means of the wind, nevertheless in the labor of winnowing its essential performance is by means of the wind, and therefore one is liable. The implication of the Rosh is that this is a principle true only of the labor of winnowing, and no more.

The derivation of Rabbi Menasheh of Ilya seems reasonable, and even compelling, according to Rashi's understanding-that there is a general principle in all the labors of the Sabbath that when one is pleased with the wind's assistance he is liable. But according to the Rosh this is a rule specific to the labor of winnowing, and one cannot learn from it to our case.

It should be noted, however, that Rabbi Menasheh's understanding is not forced by Rashi's words alone. Rashi does not necessarily imply that spitting is a derivative of winnowing and carrying; it may simply be a derivative of carrying. Winnowing itself is only one of the examples that teach that the wind does not exempt from liability in the labors of the Sabbath, but it is not the parent category of the other examples. If so, according to that formulation, even in Rashi's view there is no derivative of two primary categories here. This still requires study.

In any event, Rabbi Menasheh's words themselves require examination: why is this considered a case in which one is pleased with the wind, given that his purpose is not to bring the saliva elsewhere but to remove it from himself? One could analyze this in terms of labor not needed for its own sake and from additional angles, but this is not the place.

[19] See Encyclopedia Talmudit, entry 'binyan av', columns 3-5.

[20] See my book Two Carts and a Flying Ball, pp. 404-406, where it is explained that underlying the rule of 'mah ha-tzad' is the assumption that every law in the Torah has only one reason.

[21] At the beginning of Bava Kamma, the Gemara discusses whether to compare the derivatives in damages to the derivatives of the Sabbath, where their derivatives are like them, or to the derivatives of impurity, where they are not like them. Nahalat David explains there that the Gemara's uncertainty is whether the derivatives of damages are derivational derivatives, as on the Sabbath, or practical derivatives, as in impurity. In impurity, the derivative is created by contact with the primary category; that is not a way of learning the derivative but a way of creating or generating it. By contrast, on the Sabbath the derivative is derivational, that is, one learns it from its primary category. According to this, one may explain why in the course of the Gemara there is a view that the derivatives of the category of foot, for example, are things kicked up by the animal's foot that caused damage-that is, practical derivatives and not derivational ones. In the conclusion of the Gemara, however, even in damages the derivatives are derivational and not practical (perhaps with the exception of half-damages for pebbles, where perhaps that is exactly the conclusion of the Gemara there, 3a, that they are not like them because they are practical rather than derivational, but this is not the place).

In light of our discussion above, this distinction and its relation to our own distinction must be clarified further. The two types of derivatives discussed by the Gemara in Bava Kamma according to Nahalat David differ from the examples of conceptual construction. Here we are creating a concept out of two other concepts. That is a theoretical construction. Here the derivative is not actually produced by its primary category, nor is it merely learned from it as a principial law. Rather, a new concept is created as a generalization, or as a combination, out of two earlier concepts.

[22] In the Gemara, 6a, each of the derivations concludes with the words 'and the rule returns,' and not with 'the common factor.' The Mishnah there, however, does conclude with those words, and the Gemara is only elaborating what the Mishnah says-the sugya begins with the question 'what does this come to include?' See below.

[23] On this point, see Hiddushei Ha-Griz at the beginning of Hilkhot Nizkei Mamon, in what he wrote there about this dispute and in his broader discussion. It may be that his intention is like what will be suggested below.

[24] Usually, the later authorities understood that there are only two views in the Rosh: that of the great authorities, and his own. They do not address the fact that he also cites those who are uncertain. The assumption is that the uncertainty is between those two positions, and therefore no essentially distinct third view is present.

In my humble opinion, however, the Rosh presents three distinct positions. More than that: the uncertain position is not wavering between the two other positions, but between the first position-that all exemptions apply-and a completely opposite position-that no exemptions apply at all. The Rosh's own position has still not been presented at that stage, and therefore it is not likely that it constitutes one side of the uncertainty of that middle position.

[25] And in the previous note we saw that it is indeed plausible to understand them in this way.

[26] Perhaps with the exception of the exemption of tooth and foot in the public domain, which according to the Rif stems from each person's right to walk in the public domain, and is therefore not a scriptural decree but a substantive exemption. The medieval authorities already objected to him on this point, since it amounts to offering the reason for the verse. See the Rosh and Yam Shel Shlomo, but this is not the place.

[27] See also the books of rules by the Pri Megadim: Shoshanat Ha-Amakim, rule 1, and Ginat Veradim, rule 18.

[28] In this respect, we have returned somewhat to the formulations of Rabbi Asher Deutsch cited above.

[29] See, however, the earlier note in which we observed that those who disagree with Maimonides' interpretation of the Mishnah in Shevuot, and refuse to explain it as the bailee's oath, do not necessarily do so because they reject the logic of conceptual construction. The reason may simply be a conceptual view-which most medieval authorities adopted-that a bailee is not always liable to an oath when he seeks exemption. For example, on a claim of 'I returned it' he is exempt from a Torah oath. If so, then even if one accepts conceptual construction as a logical process, one still cannot infer that our Mishnah deals with the bailee's oath.

In the Rosh's own view, however, we see a consistent disagreement on this matter, and therefore it is entirely possible that he does not accept the very process of conceptual construction, and that the difficulty begins for him already on the logical plane.

[30] Perhaps the Rosh's position accords with Rabbi Asher Deutsch's explanation cited above. This is possible, however, only according to his first suggestion-that 'mah ha-tzad' is a hermeneutic rule regarding what is explicitly written in the Torah, and not regarding the labors of the Sabbath, which emerge from derivation by comparison to the Tabernacle. According to the second suggestion-that if there were such derivatives they would have to be counted as a separate primary category, which would contradict the count of thirty-nine primary categories of labor-this cannot be said in the Rosh's view. The reason is that the Rosh in Bava Kamma 6a understands 'mah ha-tzad' as a derivation only from one of the teaching primary cases, while the second primary case serves only to remove a refutation. If so, in the end it is a derivative of one primary category-of the pit, in that case-and not of both. Hence even if such a derivative existed, there would be no reason at all to treat it as an additional primary category. Above we argued similarly even according to the great authorities, who hold that the derivation is made from both teaching primary cases. But in the Rosh's own view this is entirely clear, and it seems very difficult to dispute.

[31] For a fuller principial discussion, see my book Two Carts and a Flying Ball, Beit El, 5762, at the beginning of chapter 4 of the second gateway, and especially note 10 there.

A considerable part of the discussion in the second gateway deals with the meaning and nature of concepts in general, and that discussion too is indirectly related to our present topic, but this is not the place.

[32] See my article in Tzohar 2, 'What Is a Halakhic Effect?', which deals with a similar example of layering laws-on top of one another, and in the case discussed there, even conflicting laws.

[33] These two examples of conceptual construction are not presented explicitly in the Talmud as a derivation from two primary categories. The Talmud grasps them intuitively as a combination of those two primary categories. The logical process is implicit. Only Or Sameach and Tzafnat Paneach point out that there is here a construction, and not merely a simple conceptual juxtaposition.

[34] See Rabbi M. M. Kasher's book Mefane'ah Tzefunot, Jerusalem 1976, chapter 8.

[35] See my article that is to be published in Tzohar 15.

[36] These distinctions are connected to what I called in the above-mentioned book an analytical mode of thought-that is, one that analyzes-as opposed there to a synthetic mode of thought-that is, one that expands.

[37] See my articles in Tzohar 12 and in Higayon 2, and the above-mentioned note 2 and notes 22ff. in my book.

[38] It should be noted that Maimonides presents a difficulty here: if he indeed understood the Mishnah in Shevuot as we have suggested, it is unclear why he did not record the halakhic innovation of conceptual construction anywhere in his code. Perhaps if he understood the process of construction as necessary deduction, the matter was self-evident to him, and therefore he saw no need to record it.

Above, at the end of section B, we saw that it may be that he does bring this in chapter 5 of Borrowing and Deposit; see there.