Two Types of Derash Rules: A Look at Gezerah Shavah (Column 355)
As is well known, there are two principal ways to read the biblical text: peshat and derash. Reading on these two planes can yield different results, and each is valid in its own right (and in rare cases there is even a contradiction; there are different ways to address it, and this is not the place). The world of derash is built on a different mode of reading the text, and over the generations the basic rules of derash were systematized; they are called the “middot of derash.” Here I will not deal with the middot of aggadic derash—whose status and origins, to my mind, are quite doubtful—but only with the middot of halakhic derash.
All the Rishonim agree that the middot of halakhic derash were transmitted by tradition and accepted as a law to Moses from Sinai (though this tradition is flexible and has been interpreted in different ways)[1]. One indication of the development of this system is the existence of several different lists of middot, the best known of which is Rabbi Yishmael’s list of the thirteen middot at the beginning of the Sifra. Within these lists one can distinguish two types of middot: formal and logical. Gezerah shavah is the most prominent member of the formal group, and kal va-chomer of the logical one.[2] I wish to touch briefly on this fundamental yet neglected topic and clarify a few basic aspects.
Two categories of middot of derash
As noted, Rabbi Yishmael’s middot can be divided into two categories: the logical middot and the formal middot. The logical middot are kal va-chomer and the two forms of binyan av, and perhaps also “two verses that contradict each other” (there are disagreements and various interpretations of this middah). The remaining middot, such as gezerah shavah, general-and-particular, and the like, belong to the formal category. There are two main differences between these two categories:
- The logical middot are used in all domains of interpretation and thought, not only in expounding Scripture. We employ analogy and induction (the two binyanei av) and kal va-chomer across all fields, and one could even say that the very inclusion of these middot in Rabbi Yishmael’s list requires explanation (why we needed “permission” to use such basic tools of reasoning, and why they belong to the world of derash rather than peshat).
- With respect to inferences by the logical middot, Hazal’s literature occasionally raises refutations (pirchot) that challenge them and can even overturn them. But regarding the formal middot of derash there are no refutations—and there cannot be.[3] For example, when the Gemara makes a kal va-chomer or a binyan av, counterexamples or considerations sometimes arise that show the assumed hierarchy (in a kal va-chomer) or similarity (in a binyan av) is not necessary. But if one derives something from general-and-particular or from a gezerah shavah, no refutation is ever raised against such a derashah, since derivation by these middot is not based on logic but on a textual trigger (the same word in two contexts, or a biblical formulation that moves from particulars to a general and vice versa). So long as the trigger exists, it must be expounded; a refutation can do nothing about that.
Thus, for example, if one makes a gezerah shavah from “slave” (eved) to “woman” (ishah), this is not an explicit assumption about some similarity between them, but an instruction to equate their laws (though it is reasonable that some similarity underlies it). Someone may then raise a refutation that one should not compare a slave to a woman because a woman is freeborn, or because some law applies to her and not to a slave. Such a refutation will change nothing, for the comparison is made on the basis of gezerah shavah and not on an assumed similarity between them. This is unlike a binyan av or kal va-chomer, where the inferences rest on an assumption about similarity or hierarchy between the source and the target, and that assumption can of course be challenged.
There is an obvious connection between these two aspects, since they are rooted in the difference in nature between the two categories (formal vs. logical).
The origin of the middot and the exegete’s autonomy
I already mentioned that the middot of derash are halakhah le-Moshe mi-Sinai, but the specific derashot themselves certainly are not. Exegetes throughout the generations made use of the middot to extract, by means of them, novel halakhot from Scripture. Such a derashah is usually the work of the darshan and did not come to him from Sinai.[4]
The Gemara in Pesahim 66a discusses the offering of the Paschal sacrifice on the eve of Passover that falls on Shabbat. There are derashot that Hillel the Elder presented to the sons of Beteira (from whom this law had been forgotten), by which he proved that the pesach is to be offered even on Shabbat. At first he brings a derivation from gezerah shavah, and the Gemara challenges this from another source where it appears that he had derived it by kal va-chomer:
The master said: Furthermore, an a fortiori inference: If the daily (tamid) offering, for which there is no liability of karet, overrides Shabbat, then the Passover offering, for which there is liability of karet—how much more so should it override Shabbat. One can refute: What about the tamid, which is frequent (tadir) and entirely consumed (kolil).
And it answers:
He first said to them a kal va-chomer, they refuted it, and then he said to them a gezerah shavah.
He first derived it by kal va-chomer, and when they refuted him he presented a gezerah shavah. Note that the gezerah shavah previously cited was also from the tamid to the pesach (in both it says “in its appointed time” – be-mo’ado). Seemingly one could raise against it the same refutation (that the tamid is frequent and entirely consumed, so one cannot learn from it to the paschal offering). But as we saw above, a refutation cannot attack a gezerah shavah, since a gezerah shavah is a formal and not a logical middah.
The Gemara now asks why Hillel raised the kal va-chomer at all if he had a gezerah shavah on the matter:
Since he had learned a gezerah shavah, what need was there for a kal va-chomer? Rather, he said it to them for their sake: Granted, a gezerah shavah you had not learned, for a person does not derive a gezerah shavah on his own; but a kal va-chomer, which a person does derive on his own—you ought to have derived. They said to him: A kal va-chomer is subject to refutation.
Hillel wanted to rebuke the sons of Beteira for not mastering the middot of derash, so he presented them with a kal va-chomer—which anyone can derive on his own. A gezerah shavah one does not derive unless he received it from his teacher. Rashi there explains that a kal va-chomer is the darshan’s own reasoning; hence one can both derive it on his own and refute it. But a gezerah shavah must come to a person from his teacher and its source is a tradition from Sinai; otherwise you cannot know that the verse did not come to teach something else. This, broadly, is how the Rishonim and Aharonim explain this rule.
Thus there is a contrast between kal va-chomer, which a person may derive autonomously, and gezerah shavah, which must come by tradition. And what about the other middot? The Rishonim disagree. In Sukkah 4a there is a discussion about whether to analogize lulav to etrog. Rashi there writes:
“We do not analogize” — they were not equated to learn from one another, and one may not derive on his own from any of the thirteen middot by which the Torah is expounded, except for kal va-chomer.
He understands that kal va-chomer is the exception; all the other middot are like gezerah shavah, and one does not derive them on his own unless he received them from his teacher.
By contrast, Tosafot there, s.v. “ve-R”Y”, write:
Rather, a person may derive all the middot on his own, except for a gezerah shavah, which one does not derive unless he learned it from his teacher.
In their view it is specifically gezerah shavah that is the exception; it alone one does not derive on one’s own unless he received it from his teacher. But the other middot are like kal va-chomer, and a person derives them on his own.
Do we derive the other formal middot on our own?
What is special about gezerah shavah? Is it tied to its formal nature? According to Tosafot, apparently not. For them, gezerah shavah is the exception, while a person derives the other middot on his own—and this likely includes all the rest, both logical and formal. Hence in their view there is no place for our distinction between logical and formal middot on this issue. But according to Rashi it is precisely kal va-chomer that is the exception, and a person does not derive the other middot on his own. It is very plausible that Rashi’s intent is only to the formal middot (like gezerah shavah), whereas the logical middot are presumably derived on one’s own like kal va-chomer.
Even if we accept that the binyanei av were given to Moses at Sinai, it is still reasonable that an interpreter of Scripture can make an analogy or a generalization on his own, just as we do in every other domain of thought. Principles of logic that serve in every domain certainly serve in Torah as well. The formal principles are an addition and may be suited only to interpreting (or expounding) Scripture.[5] According to Rashi, from the formal nature of these middot there follows the conclusion that a person cannot derive them on his own (since we lack the tools to do so without tradition; there is no “ordinary” logic here). This conclusion, however, is not necessary: even if the middot are formal, there is no reason to deny the possibility that a person who received the list of middot by tradition from Sinai could apply them on his own and expound Scripture by means of them. Hence Tosafot, who hold that a person does derive them on his own, do not necessarily disagree with the very distinction between the two kinds of middot.
What is the tradition from Sinai?
Up to this point we assumed that all the midrashic inferences that use gezerah shavah, and according to Rashi also all the other formal middot, were received by tradition from Sinai. On this view, not only are those derashot not creative, they seem not even to be supportive (somkhot). A supportive derashah is one that the darshan creates in order to support a halakhah received by tradition. But Rashi here seems to imply that both the derashah and its halakhic upshot were received by tradition from Sinai. This, of course, raises the question of why any of it is relevant: why wouldn’t God simply give us the law as a bottom line? That is very implausible.[6]
Indeed, the Ramban (in his strictures to the Second Root) and other Rishonim of his school have made this point, arguing that the Gemara cannot be understood literally.[7] They contend that these derashot were not fully received from Sinai; otherwise, we would not find disputes about gezerah shavah (see Hullin 85a and Sanhedrin 83b),[8] nor would we find the Gemara praising someone who found a gezerah shavah as the basis for a law (Shevuot 7a), nor would we see questions in the Gemara as to why one learns a law from this word rather than another (see Ketubbot 38b). The Ramban adds that if everything had been handed down from Sinai, there would be no point in counting gezerah shavah among the middot by which the Torah is expounded, since we would have no ability to use it in expounding Scripture.
Therefore, argues the Ramban, it is reasonable that what was received by tradition was partial information regarding these derashot. For example: a certain halakhah was received by tradition, and the darshan fashioned the derashah to ground it (e.g., he determined that it rests on a gezerah shavah and/or located the two words and made a gezerah shavah between them); or the tradition stated that a gezerah shavah is to be made between two words, and the darshan fashioned the halakhah that follows; or a tradition stated that a certain halakhah is based on a gezerah shavah, and the darshan identified the specific two words in question; and so on.[9] The Talmudic rule merely says that a person cannot create a gezerah shavah without some support that comes by tradition; otherwise, one could manufacture almost anything by means of this middah (and per Rashi, the same would apparently apply to all the formal middot).
Hefna’ah (redundancy)[10]
So far we have discussed the gezerah shavah but have not mentioned a very important and distinctive detail. The Talmudic rule states that one does not derive by gezerah shavah unless at least one of the words is mufneh, i.e., extra/redundant. When the words are not redundant, one cannot expound them, for they may have been written for the peshat of the verse. In order to expound a word by gezerah shavah, we must be persuaded that it is redundant from the standpoint of peshat, and only then can we infer that it was written for the sake of the gezerah shavah.
The Talmud (see Shabbat 64a and many other places) distinguishes between mufneh on one side (i.e., only one of the two words is redundant) and mufneh on both sides (both words are redundant). When a gezerah shavah is mufneh on both sides, the derashah cannot be refuted. Rashi in that sugya (Shabbat 64a, s.v. “mufneh”) explains that it is like something explicitly stated in Scripture, and therefore no objection can be raised. When the gezerah shavah is mufneh on one side only, the Tannaim disagree (see Niddah 22–23) whether it can be refuted or not. And for a gezerah shavah that is not mufneh at all, the Amoraim disagree (in Niddah there) whether one derives it (but it can be refuted) or one does not derive it at all in such a case (even if there is no refutation).
What does this mean? In light of what we saw above, there should seemingly be no room to both derive and refute. If one may derive a gezerah shavah that is mufneh on one side, then we ought to derive it and not object to it (for one does not refute a formal middah); and if it can be refuted, we should not have derived it at all. Even more difficult is the case of a gezerah shavah that is not mufneh at all: how does one derive it, and if one does derive it, why can it be refuted? And how is it possible that in a case where one word is mufneh we both derive and (perhaps) refute—if we refute the derashah, we no longer have an account for the redundant word in the text. As noted, from the standpoint of peshat it is redundant, and in derash we are not expounding it. One could say that if we refute the gezerah shavah we will have to find another derashah to account for the hefna’ah.
From this it seems that there are cases where the derashah looks like a gezerah shavah but is actually a binyan av (an analogy). When there is a word mufneh on one side, this is only a hint that we should make an analogy between the passages. Therefore, in principle, we may derive; but if there is a refutation, the derashah falls and we must seek another derashah (to explain the hefna’ah). In such a case we are left needing to account for the redundant word, so the refutation must be strong. By contrast, when neither word is mufneh, then it is a regular analogy: in such a case there is no textual hint that we ought to make a derashah. If nevertheless the Sages decided to make a gezerah shavah, it is presumably based on a similarity they saw, by reasoning, between the two passages—that is, it is a binyan av. The use of identical words in such a case is probably just a rhetorical flourish or support. It is a regular binyan av; hence there is no impediment to refuting it.
Indeed, some Rishonim noted that a number of gezerot shavot in the Talmud are “giluy milta” and not a true gezerah shavah. Thus the Ritva on Kiddushin 2a, regarding the derashah “kikhah–kikhah” from the field of Ephron to betrothal of a woman, writes:
It is written here “when a man takes (yikah)” and it is written there “I have given the money of the field—take (kach) it from me.” Just as there [Ephron’s field] it is with money, so here [betrothal] it is with money. And it is not a true gezerah shavah, for the “taking” there refers to money, while the “taking” here refers to a woman; rather it is merely a giluy milta—akin to “let the obscure be learned from the explicit”—that the term “taking…taking” written by the Merciful One comes to reveal that just as there it is with money, so too here it is with money; and there are similar cases in the Talmud.
He brings an indication that this is not a gezerah shavah and explains that it is only a giluy milta, i.e., a disclosure of the meaning of a biblical word. The field of Ephron teaches us that “taking” (kikhah) in Scripture denotes acquisition by money, and from there we interpret “when a man takes a woman” as betrothal by money. There is no halakhic equation of purchasing a field with betrothing a woman; it is merely an elucidation of a term.[11] Note that giluy milta is a logical operation, at least in the sense that we apply it to any text, not only to Scripture: in any text, if we understand the meaning of a word in one context, it is reasonable to learn from it about its meaning in another.
In which kind of gezerah shavah were the rules of hefna’ah said?
The rules of hefna’ah require explanation in light of what we saw above—that a person does not derive a gezerah shavah on his own. Seemingly, if we received a tradition about a given gezerah shavah, there is nothing to refute; and if we did not receive a tradition, we do not derive it even if there is no refutation. What is the point of the rules of hefna’ah, and where and how are we to apply them?
Rashi on Gittin 41a, s.v. “dekulei alma,” writes:
Everyone agrees that a gezerah shavah is preferable: since it is mufneh for derash, it is as if it were written explicitly in the verse, for the verse is extra; and even if it is not mufneh, one cannot object to it since we learn it from Sinai—for a person does not derive a gezerah shavah on his own—and it is preferable to a hekesh (analogy).
It seems he holds that when the gezerah shavah is mufneh, it is as if explicitly written in the verses; but even if it is not mufneh, if we have a tradition regarding it, no objection can be raised. It is unclear, then, what the requirement of hefna’ah adds if the status of the gezerah shavah does not depend on it. It seems that in his view, when the gezerah shavah is mufneh, no tradition from Sinai is needed for it—for it is as if stated explicitly in Scripture (and apparently one may derive it on one’s own); only when it is not mufneh does one require a tradition from Sinai.[12] And the Ritva on Bava Metzia 41b, s.v. “ve-idach,” presents the other side of the coin:
“And the other [Tanna]—like that which was taught: ‘and the homeowner shall approach…’” And if you will say: since you expounded one verse for the purpose of saying “dayo” (that the conclusion of an a fortiori is like the source), how do we then derive it for a gezerah shavah, for we require mufneh? And further, for Rava who expounded both as above, how does this gezerah shavah work for him? One must say that this gezerah shavah is a received tradition here, and in any such case we do not require mufneh at all, as we have explained in several places.
For gezerot shavot that were received from Sinai, there is indeed no need for hefna’ah. The rules of hefna’ah were said only for gezerot shavot that were not received from Sinai.[13]
From Rashi and the Ritva it appears that there is a kind of gezerah shavah that lacks even support from the Sinaitic tradition yet is nevertheless derived—and only for these is hefna’ah required. But for gezerot shavot that were received from Sinai, the rules of hefna’ah do not apply. It follows that my earlier suggestion—that a gezerah shavah mufneh on one side is really a binyan av—was said only about gezerot shavot not received from Sinai. If it is mufneh on both sides, it is a full-fledged gezerah shavah; if it is mufneh on one side, it is essentially a hint to make a binyan av; and if it is not mufneh at all, then according to the view that one nevertheless derives it (albeit subject to refutation), it is simply a binyan av by reasoning (which does not require even a textual hint).
The sugya in Yoma about the warning (azharah) for affliction on Yom Kippur
What prompted me to write this column is a particularly unusual and difficult sugya in Yoma that deals with the warning for affliction (inuy) on Yom Kippur. Here I will try to learn the sugya and comment on several thorny aspects of its progression.[14]
It is accepted among the poskim that on Yom Kippur there is a prohibition (lav) and a positive command (aseh) regarding cessation from work (as on Shabbat), and a prohibition and positive command regarding affliction. However, the Torah itself does not state, regarding Yom Kippur, a warning (azharah) about affliction (only a positive command and a penalty of karet), nor does it state a positive command regarding work (only a prohibition). The Gemara in Yoma 81a discusses the source for the warning regarding affliction, and in the course of its discussion it brings the following teaching:
But a warning for the affliction of the day itself we have not learned. From where [do we derive it]? Let the punishment regarding work not be stated, for it can be learned from affliction: If affliction, which does not apply on Shabbatot and festivals, carries the punishment of karet, then work, which does apply on Shabbatot and festivals—how much more so [should it carry karet]. Why, then, is [the punishment regarding work] stated? It is mufneh—to compare and to judge from it by gezerah shavah: “Punishment” is stated with respect to affliction and “punishment” is stated with respect to work. Just as for work, He did not punish unless He had warned, so too for affliction He did not punish unless He had warned.
One can refute: what about affliction, which was never permitted in any instance—will you say the same about work, which has instances in which it is permitted?
Rather, let the punishment for affliction not be stated, for it can be learned from work: If work, which has instances in which it is permitted, carries the punishment of karet, then affliction, which was never permitted in any instance—how much more so [should it carry karet]. Why, then, is [the punishment regarding affliction] stated? It is mufneh—to compare and to judge from it by gezerah shavah: “Punishment” is stated with respect to affliction and “punishment” is stated with respect to work. Just as for work, He punished and warned, so too for affliction He punished and warned.
One can refute: what about work, which applies on Shabbatot and festivals—will you say the same about affliction, which does not apply on Shabbatot and festivals?
Two opposite lines of reasoning are presented, both based on the fact that between affliction and work there are relative stringencies in both directions: affliction is more stringent in that it applies to every person at all times (of Yom Kippur), whereas work has instances in which it is permitted (in the Temple). Conversely, work is more stringent in that it applies on Shabbatot and festivals, whereas affliction is only on Yom Kippur.[15] For our purposes, in both lines one first learns, by kal va-chomer, the punishment of karet either for work or for affliction; therefore the punishment explicitly written regarding work or affliction is rendered extra, from which it follows that it is mufneh to enable a gezerah shavah.
We must note that here the gezerah shavah is mufneh on one side only. Seemingly the Gemara should have suspended the refutation on the dispute over whether a gezerah shavah that is mufneh on one side can be refuted or not. According to the view that it can be refuted, there is no problem; but according to the view that it cannot—how are we to understand the refutation of this gezerah shavah?
To analyze the logic here, note that each line has a two-step structure: first one learns the punishment by kal va-chomer (from affliction to work, or in the other direction from work to affliction), and then one learns the warning by gezerah shavah from work to affliction (in both lines the gezerah shavah runs in the same direction). The refutation relies on a stringency present in affliction (and, in the second direction, in work). We can now see that in the first direction the refutation (based on a stringency of affliction) attacks only the kal va-chomer, not the gezerah shavah. Hence there the refutation was directed at the kal va-chomer and not at the gezerah shavah. In the second direction, however, the refutation (based on a stringency of work) can attack both the kal va-chomer and the gezerah shavah. True, even in the first direction one could have refuted the gezerah shavah by the opposite refutation—i.e., by presenting the stringency of affliction. The very reasoning that supports the kal va-chomer of that line itself undermines the gezerah shavah at its end.
We thus learn that in the first line the Gemara raises a refutation of the kal va-chomer that does not attack the gezerah shavah, but it ignores the alternative of attacking the gezerah shavah. That line probably assumes that one does not refute a gezerah shavah that is mufneh on one side; hence it attacks the kal va-chomer and thereby shows that there is no hefna’ah. There still remains a non-mufneh gezerah shavah, but that is only an analogy and likely falls because of the opposite refutation (recall: a gezerah shavah that is a regular analogy can indeed be refuted).
In the second direction, the refutation attacks both considerations (the kal va-chomer and the gezerah shavah), and that direction can be reconciled either according to the view that one refutes a gezerah shavah that is mufneh on one side (in which case it attacks the gezerah shavah) or according to the view that one does not (in which case it attacks only the kal va-chomer). The two directions may thus hinge on the dispute about refuting a one-sided mufneh gezerah shavah.
[Parenthetically, one should examine the very structure of the second line: if there is a kal va-chomer from work to affliction, it is not clear why we need the gezerah shavah at all; why not suffice with the kal va-chomer itself? One could learn the warning for affliction by kal va-chomer from the warning for work (and of course refute it by the opposite stringency). Perhaps the Gemara assumes that we do not derive a warning by kal va-chomer, for “we do not warn from a logical derivation.” The Rambam, however, in his introduction to the Sefer ha-Mitzvot, writes that there is no difference between kal va-chomer and gezerah shavah in this regard (we do not warn from a derivation in any middah), but he also holds that one may warn from a derivation when the punishment is stated explicitly.]
Later in the sugya, another gezerah shavah is brought that is also based on hefna’ah:
Ravina said: This Tanna derives “etzem–etzem” (“on that very day”) by a gezerah shavah that is mufneh; for if it were not mufneh, one could refute it just as we refuted above. How is it mufneh? Five verses are written about work: one for the warning by day, one for the warning by night, one for the punishment by day, one for the punishment by night, and one extra for making the gezerah shavah—to learn affliction from work, whether by day or by night.
Here the derivation is from the word “etzem” (“on that very day”) written with respect to affliction and with respect to work. But here the Gemara does not dismiss this gezerah shavah by the earlier refutation (i.e., on account of the stringency that work also applies on Shabbatot and festivals). Apparently one does not refute a gezerah shavah that is mufneh on one side. This bolsters my claim that the refutations in the two earlier lines were directed only at the kal va-chomer, not at the gezerah shavah.
On the other hand, the very fact that these gezerot shavot are not agreed upon (and perhaps refutations arise against them, at least in the second direction) indicates that there is likely no tradition from Sinai here. This explains why hefna’ah is required, and as we saw, when hefna’ah exists only on one side there may be room to refute. Yet in this sugya it seems the gezerah shavah was in fact not refuted. Perhaps the reason is that when there is no hefna’ah at all we explained that we are dealing with a binyan av and therefore refutation is possible; but when it is mufneh on one side, it is difficult to refute simply, for that would leave us with an unaccounted-for redundant word.
If so, it remains unclear why a refutation of the kal va-chomer helps to collapse the hefna’ah, for even if what fell was the kal va-chomer and not the gezerah shavah, that itself undermines the gezerah shavah (since without mufneh one does not derive), and we are once again left with an unexplained hefna’ah. It would seem, then, that even when a refutation of the kal va-chomer brings down the gezerah shavah, it is still preferable to refute the kal va-chomer rather than the gezerah shavah. As for the hefna’ah, we remain with a question; perhaps the assumption is that there is some other explanation for the redundant word (or one might even say: something that could have been derived by reasoning or a binyan av was nevertheless explicitly written by Scripture).
Final note
If indeed there are inferences that are really binyan av wrapped in the literary garb of a gezerah shavah, this could provide a different explanation for the phenomenon of disputes regarding gezerah shavah—which, recall, is what led the Ramban to conclude that gezerah shavah was not fully given from Sinai.
It is possible that the original gezerot shavot did indeed come from Sinai (as Rashi implies in several places), but beyond them the Talmudic literature contains gezerot shavot that were coined over the generations based on redundant words (and perhaps even without hefna’ah at all). Some of these are essentially binyan av, and regarding them disputes may arise among the Sages—as we find with all the other middot. And of course there is a hierarchy among them: if it is mufneh on both sides, it is a compelling gezerah shavah not subject to refutation (as if explicitly written in Scripture); when it is mufneh on one side, it is also hard to refute, for one must still account for the hefna’ah; and a gezerah shavah that is not mufneh at all is merely a regular binyan av, which can be refuted simply and ordinarily.
In closing, let me note that Michael Chernick, in his book Gezerah Shavah, points out that there are earlier gezerot shavot in the Talmud that are all based on words that appear uniquely in Scripture—i.e., they appear only in the two places between which the gezerah shavah is made. In such cases it is very reasonable to make a gezerah shavah, for the Torah itself hints to us in unequivocal fashion. By contrast, gezerot shavot based on non-unique words are a later phenomenon and, of course, less compelling. It may be that only for them is hefna’ah required to ensure that a gezerah shavah is permissible.
[1] See on this in the second volume of Talmudic Logic, which treats the rules of general and particular. In my book Ru’ah ha-Mishpat, in the essay to the Second Root in the book Yishlah Shoreshav, and in the aforementioned volume, you will find general introductions to the middot of derash from various aspects: halakhic status, historical development, disputes, logical basis and meaning, and more.
[2] In several places in Hazal one finds the phrase “the lenient and stringent cases and the gezerot shavot” as a description of the totality of the middot (see, for example, Temurah 16a and elsewhere). It seems to me that the reason is that these two are the most representative examples of the two types of middot: the formal and the logical.
[3] In Hullin 65b, R. Ahai raises a refutation against a derivation of general-particular-general, which seemingly contradicts what I say here. But the Ba’alei ha-Tosafot wrote that R. Aha in our Talmud is R. Ahai Gaon, the author of She’iltot, and in particular here, where the printed text itself reads “R. Ahai,” about whom the Rishonim wrote that he was a Savora sage—meaning this is clearly a later addition not belonging to the original Talmud. In any case, in our aforementioned book on general and particular we discussed this sugya in detail and showed that it is not truly a refutation of the derash of particular-and-general.
[4] There is debate among Rishonim and among scholars whether there exist creative derashot (that is, derashot that generate new halakhot) or whether all are supportive (or preservative, i.e., grounding halakhot already received by tradition). In my view this debate should not exist: there is no doubt that there are creative derashot. This is clear from explicit examples in the Talmud itself, and also from reason (there would be no point in creating the world of derash if it added nothing). I laid this out in the introductions cited above, and here I will simply assume it.
[5] Indeed, in our aforementioned book on general and particular we brought numerous examples of derashot of general and particular used in the interpretation of the Rishonim’s writings. But these are very general derashot, and they do not really require the formal details of the rules of general and particular discussed there.
[6] In Kin’at Soferim to the Rambam’s Second Root, it was suggested that the Rambam has a unique view: that “a person does not derive a gezerah shavah on his own” does not mean that its source is Sinai, but that it is generated only by the Great Court (Beit Din ha-Gadol) who know how to interpret the words of Scripture. The rule comes to exclude a gezerah shavah by a private scholar. This is certainly not Rashi’s view nor that of most Rishonim, and I do not think it is the Rambam’s either. Moreover, the Gemara in Pesahim 66a that I cited above proves otherwise, for there Hillel did not expect the sons of Beteira—and all the people of Eretz Yisrael (who staffed the Sanhedrin)—to be able to derive a gezerah shavah, only a kal va-chomer. I therefore set this possibility aside.
[7] See Encyclopaedia Talmudit, s.v. “Gezerah Shavah,” around note 58 and on.
[8] This is especially true according to the Rambam, who holds that something subject to dispute cannot be a halakhah le-Moshe mi-Sinai. See responsum Havot Ya’ir §265, which discusses the Rambam’s words at length.
[9] See also Hilchot Olam, Sha’ar 4; the Shnei Luhot ha-Brit (Shlah), section “The Thirteen Middot”; and Sefer ha-Keritut, Bayit 2, among others.
[10] See Encyclopaedia Talmudit, s.v. “Gezerah Shavah,” section “Mufneh and not mufneh.”
[11] See also Rashi, Bava Kamma 2b, regarding the iron horns of Tzidkiyahu ben Kenaanah.
[12] In the Encyclopaedia Talmudit, s.v. “Gezerah Shavah,” note 105, it is stated that in the Shnei Luhot ha-Brit, section “Torah Shebe’al Peh — on Gezerah Shavah,” it is proven from here that Rashi too, like the Ramban, holds that the intent is not that the gezerot shavot were entirely received by tradition from Sinai, but that a tradition is needed to lend some support to the derashah. But see there that he understood Rashi differently, and his words are unclear. See the next note.
[13] The Encyclopaedia Talmudit there writes that the Ritva disagrees with Rashi; this is incorrect (as noted, they misread Rashi).
[14] From the outset this sugya shakes many assumptions: about “warning from a derivation,” about the need for a warning at all, about deriving a warning (and other things) from the middot of derash, and of course about the difference between kal va-chomer and gezerah shavah. Here I wish only to illustrate some of the points raised in this column thus far.
[15] Seemingly this is actually a leniency, not a stringency, for a law that applies on more days is presumably easier to apply. But one can still infer from this that it is also easier to impose a penalty for it.