Deriving *Ta‘ama de-Kra*: 3. “Gezerat ha-Katuv” (Column 716)

With God’s help

Disclaimer: This post was translated from Hebrew using AI (ChatGPT 5 Thinking), so there may be inaccuracies or nuances lost. If something seems unclear, please refer to the Hebrew original or contact us for clarification.

In two previous columns we discussed the Tannaitic dispute about deriving *ta‘ama de-kra* (the reason for a verse) and its implications. Here I wish to continue with a related concept: “gezerat ha-katuv.” This term already appears in the Sages’ literature and is used to this day. In its common meaning it refers to a law whose rationale is not understood, and, as we shall see, it even contains within it methods for handling such laws. There are different approaches to gezerat ha-katuv and its implications, chiefly regarding how frequently it appears in Halakhah.

The basic motivation is the sense that indeed there are certain laws defined as gezerat ha-katuv, but that very fact shows us that the body of law as a whole is not like that. If so, we do understand the rest. That itself sounds illogical, for across large swaths of Halakhah it seems we lack any real understanding, beyond the particular laws defined as gezerat ha-katuv. Moreover, once again the question arises: if we do in fact understand them, why not derive their reasons?

To address this we must clarify the very concept of “gezerat ha-katuv.” As we shall see, this will itself return us to ideas from the previous two columns, after which we can revisit knock-on implications for the issue of *ta‘ama de-kra*. I note that in my article, “What Is Gezerat ha-Katuv: A Study of the Law of Edim Zomemim,” I treated the matter at much greater length. To proceed in a more orderly and systematic way here, I will first address gezerat ha-katuv methodically and only afterward return to the issue of deriving reasons and its connections.

Point of departure: two principled approaches to reasons for commandments and laws

As a starting point we must examine our attitude toward explanations offered for commandments. In column 457 I discussed the Platonic Euthyphro dilemma, which originally dealt with Greek deities but is applied today to a monotheistic God as well. The dilemma is whether the good is good because God commanded it, or God commanded it because it is good. These are two approaches to defining “the good”: Is an action good because God commanded it—yet He could have decided otherwise and then the content of the good would change? Or is the good defined objectively, and God commands it because it is good—such that even God could not have defined the concept of “good” otherwise.

I noted there Avi Sagi’s claim that the overwhelming majority of Jewish thinkers adopted the second view. One must understand that this means the good constrains God and does not depend on Him (just like logic). The main argument I see for this approach is that if we adopt the first position (that God determines what is good), the result is that the statement “God is good” becomes a mere definition. There is no praise in it; it is almost tautological (that God behaves as He wills, or wills what He wills). By contrast, our intuition is that the claim “God is good” amounts to praise. It is an assertion, not just a definition. If so, it follows that “the good” must exist independently. There is a yardstick of goodness by which we are to “measure” God and find that He conforms to it (and therefore we call Him good). I will not enter here into the question of whether this is borne out in the world. For purposes of this discussion I assume the conventional religious view.

More broadly, this means that the values we are meant to pursue are not the product of God’s arbitrary decision but an objective truth. In that column I extended this also to religious values, not only moral ones. For our purposes, the reasons for the commandments are a kind of objective truth—ethical or religious facts—from which the commands are derived. Hence, the attempt to understand the commandments is akin to a scientific activity: we try to uncover these abstract facts via their practical consequences (the commandments), similar to how a scientist seeks to uncover abstract general laws of nature through the specific events he observes in the world.

Another expression of this dilemma appears in Guide of the Perplexed III:31, where Maimonides writes:

There are those for whom giving a reason for a commandment is burdensome; for them it is best that no meaning be grasped for any commandment or prohibition. What leads them to this is an illness they find in their souls which they cannot articulate, namely: they think that if these laws are beneficial for existence and for that reason we were commanded in them, then they would have come from deliberation and reflection by an intellect. But when a matter has no discernible meaning and yields no benefit, it must surely be from God, for no human thought would produce such a thing. As if—heaven forfend—in their foolishness, man would be more perfect than his Maker: man says and does only what brings about some end, while God does not do so but commands us to do what does not benefit us and forbids what would not harm us. Far be it, far be it! The matter is the reverse: the entire intention is to benefit us, as Scripture says, “for our good always, that He might preserve us alive, as it is this day,” and “which shall cause all these statutes to be heard, and they shall say, Surely this great nation is a wise and understanding people.” He has already explained that even all the statutes will indicate to all the nations that they are in wisdom and understanding. If there were something for which no reason is known and which yields no benefit and averts no harm, why would one who believes or practices it be called wise and understanding, and why would the nations marvel? But the matter is as we stated without doubt: every one of the 613 commandments is either to impart true beliefs or to remove false beliefs, or to establish a just order, or to remove injustice, or to train us in good traits, or to warn us against bad traits. Everything is attached to three things: beliefs, traits, and the conduct of political society. (…)

As emerges from Maimonides’ words, some wish to see all the Torah’s laws as gezerot ha-katuv, i.e., without reason; in their eyes this is the Torah-giver’s glory—His ways are higher than ours. Others strive to base all of the Torah’s laws, as far as possible, on reason and human rectitude; for them, the Torah-giver’s glory lies precisely here—that His laws are just and upright. Maimonides himself holds that all commandments have sound rationales—whether in the realm of beliefs, of character, or of social-political order. He himself offers reasons there, and I must say those reasons are far from persuasive (to put it mildly). It may be, however, that while the commandments indeed have reasons (i.e., God acts rationally and toward goals), we cannot grasp them. We saw this in the previous columns on deriving *ta‘ama de-kra*.

Back to gezerat ha-katuv

The two approaches described by Maimonides have further ramifications that can touch the very meaning of “gezerat ha-katuv.” If everything God does has reasons, then even for commandments we define as gezerot ha-katuv there are reasons—we simply do not understand them. If it is possible that God commands without reason, then perhaps some commandments truly have no reasons at all—and maybe those are the ones called gezerot ha-katuv.

According to the approach that believes we can understand the commandments’ reasons, one may wonder why the Torah had to write them if we could have reached them by reason alone. Evidently, reason by itself would not suffice to establish the law (either because it is not logically necessary, or because the quantitative threshold is not sharp, or because there are competing rationales, etc.). Our understanding of the commandment receives confirmation from the Torah’s writing it: now it is clear that such an obligation indeed exists. We will return to this below.

Let us leave ideology and general proclamations and examine the concept of gezerat ha-katuv from the sugya in which it appears. Our point of departure is the law of edim zomemim.

Contradiction and hazama

As is known, Halakhah recognizes two kinds of clashes between sets of witnesses: hakhasha (contradiction)—two sets testify to opposite accounts of some event; and hazama (refutation)—one set testifies to an event, and the other testifies that the first set could not have been at the event at all.

These are the verses (Deut. 19:18–20):

“The judges shall inquire diligently, and behold, if the witness be a false witness, and has testified falsely against his brother, then you shall do to him as he schemed to do to his brother; so shall you purge the evil from your midst. And those who remain shall hear and fear, and shall no more commit any such evil among you.”

The Torah addresses a case where witnesses are found to be lying in court. In such a case, the court is commanded to do to the lying witnesses exactly what they schemed to do to the litigant: if they schemed to make him pay money, they themselves must pay; if they schemed to have him flogged or executed, they themselves are punished in like manner. The problem is that it would seem a set of two witnesses can never be established as false. If another set appears against them, we have a contradiction of “two versus two” (trei u-trei), and there is no way to determine that one of the sets is lying. Moreover, even if the second set has a hundred witnesses, Halakhah accepts that their force is equal, for “two are like a hundred.” If so, how and where—if at all—can we apply the law of edim zomemim? Is there a case in which witnesses can be certainly identified as liars, such that they can be punished under “as he schemed”?

The Mishnah (Makkot 5a) addresses this question:

“Witnesses do not become ‘zomemim’ until they make themselves zomemim. How so? They said: ‘We testify that so-and-so killed a person.’ [The second set] said to them: ‘How can you testify? For the murdered man is alive,’ or ‘the killer was with us that day in such-and-such a place’—these are not zomemim. But if they said to them: ‘How can you testify, for you were with us that day in such-and-such a place’—these are zomemim and are executed on their [own] testimony.”

The Mishnah distinguishes hazama from contradiction: when the dispute concerns the subject of the testimony, that is contradiction; but when the second set testifies about the first set—saying they could not have seen the events—this is not a “two vs. two” contradiction but hazama. In such a case the second set is believed and the first is deemed false. This is the case to which the Torah refers when it says to punish the liars “as he schemed.”

The Gemara there brings a scriptural source for this distinction:

“From where are these matters [derived]? Rav Adda said: for the verse states (Deut. 19): ‘and behold, a false witness—he has testified falsely,’ meaning until the falsehood pertains to the body of the witness [himself]. The school of R. Yishmael taught: ‘to testify against him falsehood,’ meaning until the falsehood pertains to the body of the testimony.”

The derashah rests on the phrase “a false witness” indicating that in hazama the falsehood attaches to the witness himself, not to the content of the testimony as in contradiction,[1] and hence the above definition of hazama. Note that no logical rationale is brought here for the distinction; rather, a verse is cited—raising the question whether any rationale exists at all, or whether this is a gezerat ha-katuv.

A note on the relationship between source and sevara (reason)

I have already cited (see, e.g., column 411) what is written in the Encyclopaedia Talmudit under “One does not punish on the basis of an a fortiori” (its source in R. Yosef Engel): three understandings of that rule: (1) We do not punish lest there be a refutation we have not considered; (2) We do not punish lest the punishment given for the lighter case is insufficient for the more severe; (3) It is a gezerat ha-katuv from “and to his sister.” I noted there that this reflects a misunderstanding. The first two are rationales underlying the rule; the third is a source for the rule. Thus it is not a “third approach.” There are only two approaches—and in addition, a verse. (Incidentally, I brought further possibilities there to explain the rule.)

What causes this common mistake? People assume that if there is a verse it must be a gezerat ha-katuv; and if there is a rationale, then a verse is unnecessary (“why do I need a verse—reason suffices!”). The third “approach” above assumes the rule lacks a rationale. But this is nonsense: if the Torah legislates something, there is surely a rationale. So why is a verse needed? I will come to this below.

From here follows the conclusion I raised in that column: every verse teaches the opposite of what it says. The example I brought there is a quarrel between a yeshiva student and his wife. She says to him: after all, Scripture says, “Whatever Sarah tells you, heed her voice”—that is, a husband must obey his wife. The student snickers and replies: ignoramus! It’s precisely the opposite. Why would a verse need to tell Abraham to obey Sarah? Because the general rule is that one does not obey one’s wife, and only in his case was there a special scriptural decree to the contrary. Again, the assumption is that when a verse states something, it must be a *gezerat ha-katuv*.

This mode of thinking is not a new invention of yeshiva students or the later authorities; there are examples in the Talmud itself. For example, Shabbat 106b derives that “one who damages by wounding” is liable on Shabbat (i.e., wounding is not considered mere destructive work that is usually exempt), because the Torah needed to permit circumcision on Shabbat. If circumcision was permitted, it implies that wounding in general is prohibited. This is not a perfect example of the broader idea, since here one can indeed infer that wounding is prohibited—otherwise, why write the verse? Moreover, we do not derive from there that all destructive acts are liable, only wounding; i.e., the inference is local, not a global reversal. There are better examples, and elsewhere I laid out a rule (see that column) for when a verse serves as a paradigm and when it hints that the general principle is the opposite of what is written.

In any case, for our purposes, the fact that a verse distinguishes hazama from contradiction does not obviate the search for a rationale. Even if there are cases of gezerat ha-katuv (see below), it is not true that the existence of a verse excludes a rationale. So let us return to the law of hazama and ask whether there is a rationale at its root. I note in advance that I chose this law specifically because the Gemara itself presents it as a gezerat ha-katuv, while at the same time the Rishonim offer reason-based explanations.

Is the law of edim zomemim a “gezerat ha-katuv”?

In Sanhedrin 27a it is explained that beyond the punishment of “as he schemed,” edim zomemim are also disqualified as witnesses as wicked men. There Abaye and Rava dispute whether edim zomemim are disqualified from the time of their testimony or only from the time of the hazama. According to Abaye, they are disqualified from the time they testified, because it has now been clarified that they lied then and were therefore already wicked at that time. According to Rava, they are disqualified only from the time they were made zomemim. The straightforward rationale follows Abaye: since they lied then, they were wicked then; the moment of hazama seems irrelevant. The Gemara brings two explanations for Rava’s (innovative) view: (1) Edim zomemim are a novelty (chidush); take no more than what the novelty provides. (2) Because of loss to purchasers (who did not know of their disqualification and would rely on them). For our discussion, we will focus on the first explanation. I note that the halakhah follows Abaye here (this is the “ayin” of the mnemonic “YaL KaGaM”), so edim zomemim are disqualified retroactively—from when they testified. It would seem, then, that the Abaye–Rava dispute is whether the disqualification of edim zomemim is a novelty or not. For Abaye it appears not to be a novelty but straightforward law.

What is the “novelty” in the law of edim zomemim? Clearly not the special punishment; that is novel to everyone, but it is not the sugya’s topic. The sugya here concerns their disqualification as witnesses, not their punishment. About this Rava (in the first version) claims it is a novelty. The reason given is:

“An ed zomem is a novelty. What did you see that led you to rely on these? Rely on those!”

That is, the novelty is that hazama is not like a “two vs. two” contradiction; rather, the latter set is more credible than the former.[2] Rashi ties this concept of “novelty” to “gezerat ha-katuv”:

“It is a novelty—that two are disqualified on account of two who say, ‘you were with us’; for what did you see to rely on these—rely on those!—rather, it is a gezerat ha-katuv. Therefore take no more from it than from the time of the novelty onward—i.e., from when they were made zomemim.”

All this is Rava (first version). In practice we rule like Abaye; for him, it would seem there is no novelty here at all.

Indeed, we find in several Rishonim rationales that explain the law of edim zomemim and why hazama differs from contradiction.

Rationales for hazama

The first to suggest such a direction was Rav Nissim Gaon, who wrote in his Sefer ha-Mafte’ach (see Teshuvot ha-Geonim, Asaf, no. 392, pp. 195–196):

“The difference between contradiction and hazama is that in contradiction the clash between the two sets of witnesses falls upon the body of the testimony… and the law is to nullify both testimonies, for we do not know which is true. In hazama… the law is to execute… and the denial by Reuven and Shimon will accomplish nothing, for the testimony is directed at them; they become like litigants, and when one testifies about them, their denial is of no avail against the witnesses who testify about them. By contrast, in contradiction the clash falls only upon the body of the testimony, where they are in the category of witnesses; therefore we nullify both testimonies.”

Similarly the Ramban, in his commentary to the above verses (Deut. 19:18), writes:

“‘And the judges shall inquire diligently, and behold, a false witness’—the verse did not explain how one knows he is a false witness, for when there are two witnesses testifying to a matter, even if a hundred come and contradict them it will not be clarified that they spoke falsely. Nor can we say that the murdered man came walking into court, for then it would not say ‘and the judges shall inquire diligently.’ Therefore came the faithful received tradition and explained that hazama is when they say, ‘but on that day you were with us’ (Makkot 5a). And the reason is that this testimony addresses the witnesses’ own bodies, and they are not believed about themselves to say, ‘we did not do so,’ for these others can say about them that they killed a person or desecrated Shabbat.”

Ramban explains the difference between contradiction and hazama as follows: in hazama the second set testifies about the first set’s persons; the first set, however, now testifies about themselves. Such a situation is like the second set disqualifying the first as robbers, a case in which (as explained in Sanhedrin 27a) there is no novelty even according to Rava; the second set is believed, because the first set cannot testify about themselves—for they are parties to the case.

So too the Tur, Choshen Mishpat 38:

“What is the difference between contradiction and hazama? Contradiction does not address the body of the witnesses, but contradicts them: these say ‘so-and-so borrowed from so-and-so,’ and those say ‘we know that he did not borrow, for we were with him all day and saw he did not borrow.’ Hazama concerns the body of the witnesses: ‘at the time you say he borrowed, you were with us.’ Therefore the latter are believed, for they testify about the witnesses’ persons; it is as though they testified that the witnesses killed a person or desecrated Shabbat, and [the first set] are not believed about themselves to say, ‘we did not do such-and-such.’ Even if the first were a hundred, these are believed against them to have them executed—whether the hundred testified at once or in successive pairs and the refuters made each pair zomemim in turn.”

The Sefer ha-Chinukh (mitzvah 524) also cites, “as a bit of rationale,” this very reason.

Beyond this one can suggest further rationales for the second set’s credibility. For example, the P’nei Yehoshua (Makkot 5a) argues we should believe the second set because if they wished to save the defendant by lying, they should have disqualified the first set as robbers or at least contradicted them.[3] Although the Rishonim did not bring this rationale, it may well join with theirs. Because of the second set’s credibility, the focus of the discussion shifts to the first set’s qualification; thus the two witnesses in that first set become litigants testifying about themselves.

Furthermore, Ramban himself mentions an additional, more metaphysical rationale (see the passage cited in the next section). It is not entirely clear how this relates to the previous rationale in his view.

Also in Derashot ha-Ran, sermon 11 (p. 197 in Feldman ed.; cited as well by Abravanel to Deut. 19:14, who adds further reasons for believing the latter set), a rationale is offered regarding edim zomemim: the first witnesses relied on the assumption that no one saw them, hence their credibility is lower; by contrast, the refuters must fear that if they lie someone will see that they were not in the place they claim—or that the first set was not with them—since the first witnesses themselves will exert themselves to prove it. Therefore the latter enjoy greater credibility.

And yet a gezerat ha-katuv: Maimonides’ view

Despite all this, Maimonides writes in Hilkhot Edut 18:3:

“That the Torah gave credence to the latter testimony over the first regarding witnesses is a gezerat ha-katuv, even if the first witnesses were a hundred and two came and made them zomemim, saying to them, ‘we testify that you hundred were all with us on that day in such-and-such a place’—they are punished on their word, for two are as a hundred and a hundred as two. Likewise, when there are two sets of witnesses that contradict each other, we do not follow the majority; rather, we set aside both.”

Recall that Maimonides here describes Abaye’s view, which is the halakhah. According to him, Abaye agrees with Rava that the latter set’s credibility is a gezerat ha-katuv and not based on rationale. In his understanding, the Abaye–Rava dispute concerns only whether, as a result, the disqualification takes effect only from now on or retroactively.

The Lechem Mishneh to this law cites the Tur above who disputes Maimonides, and explains that they follow their respective principles regarding a clash that includes both hazama and contradiction. There are further indications that Rambam and the Tur each follows his own principle (e.g., see Rambam there 18:5 and R. Shmuel Rozovsky’s Shiurim to Makkot, §228), but this is not the place to elaborate.

A note from the Acharonim

Several Acharonim (see, e.g., R. Shmuel’s Shiurim there and his “Inyanei Edut,” §3; and Zecher Shmuel, §40) noted that the rationale of Ramban and the Tur is very difficult. Those Rishonim assume that the first set’s testimony about themselves turns them into litigants. But as long as the first set’s testimony does not concern themselves, it is not testimony of litigants. The mere fact that it has implications for them does not render them “parties.” This assumption can lead to apparent absurdities. For example: Reuven and Shimon testify that Levi borrowed money from Yehudah. According to those Rishonim we should invalidate their testimony, for their testimony also entails that they (the witnesses) were present where the loan occurred—and this would be testimony of litigants about themselves.[4] Note that such invalidation would not require a second set to perform hazama; the testimony would be invalid on its face, since it is testimony by parties. This argument seems to show the untenability of that rationale.

Accordingly, R. Shmuel Rozovsky writes that it is clear Ramban and the Tur did not mean to offer a genuine rationale. Everyone agrees this is a scriptural decree with no rationale. Their point was only that once Scripture innovated that the second set is believed, the legal characterization (as distinct from the reason—see the previous column) is that the first set’s testimony counts as though they were interested parties, and therefore it is disqualified.

Indeed, this can be seen in Ramban’s comment to the next verse (Deut. 19:19): although he gives an explanation for hazama, he still views the law of edim zomemim as “the decree of the Ruler”:

“‘As he schemed’—and not ‘as he did.’ From here they said: if they had him executed, [the zomemim] are not executed (Makkot 5b). The reason is that the judgment of the edim zomemim is by decree of the Ruler: two against two. Thus, when two come and testify that Reuven killed a person and two others come and make them zomemim, Scripture commanded that they be executed. For by Reuven’s merit—who was innocent and righteous—this matter came about; had he been a guilty man worthy of death, God would not have saved him from the court’s hand, as it says (Ex. 23:7), ‘I will not acquit the wicked.’ But if Reuven was executed, we should think that the first witnesses spoke truth, for he died for his sin; were he righteous, God would not have left him to their hand, as it says (Ps. 37:33), ‘The Lord will not leave him in his hand, nor condemn him when he is judged.’ Moreover, God will not cause the righteous judges standing before Him to shed innocent blood, for ‘judgment is God’s,’ and among the judges He judges. All this is a great elevation of Israel’s judges, a promise that the Holy One agrees with them and is with them in the matter of judgment. This is the meaning of ‘and the two men who have the dispute shall stand before the Lord’ (Deut. 19:17), for they stand before the Lord when they come before the priests and judges, and He will guide them in truth. I have already mentioned this in Parashat Mishpatim.”

Ramban continues to explain the law of edim zomemim on a metaphysical plane and uses it to illuminate the rule of “as he schemed, not as he did.” Again, he seems to offer no (human) rationale for believing the second set.

How does this square with his own explanation on the previous verse? According to R. Shmuel Rozovsky, they are one and the same: the law is a gezerat ha-katuv, and its legal characterization is that the first set’s credibility is dismissed as if they were litigants. There is no actual conflict of interest; the Torah decreed that we treat them as if they were.

Is there truly no rationale?

R. Shmuel Rozovsky’s assertion is not necessary. It is true that every testimony implicitly claims the witnesses were present at the event; but it is obvious that with respect to that point they are not “interested” in their testimony—so long as the subject is the event itself, the fact that a witness speaks about himself neither helps nor harms. When a second set comes to make them zomemim, however, that very testimony takes on a different significance: now the subject of the discussion is not only the event, but also the qualification of the first set of witnesses. In such a case we cannot accept the first set’s testimony about themselves, since they themselves are now the subject of the case.[5] For this reason we do not disqualify regular witnesses as litigants simply because they testify that they were present. But when another set comes to refute them, the subject shifts, and now the first witnesses themselves are the topic; only then—and only then—does their testimony become that of litigants.

Still, as we shall see in the next column, there is a deeper difficulty with Maimonides’ view that treats this as a gezerat ha-katuv.

[1] It may be that this is based on the extra word “witness.”

[2] From here it follows as well that the punishment imposed on them is a novelty—though a derivative one.

[3] The possibility that perhaps they seek to have the first witnesses executed (and therefore it would not help them to disqualify them as robbers or to contradict them) is irrelevant, since we are discussing their credibility before the Torah established the penalty of “as he schemed.” It is like migo: before we grant credibility to his present claim, we compare it to alternative, stronger claims—calculating under the assumption that we have not yet established the credibility conferred by migo.

One must discuss the effect of this very determination. In migo there is the well-known “genius of Meitshat’s” question: one who knows about migo will prefer to offer the “weaker” claim. But here, as there, we do not take this into account (see R. Yair Ettun, “Migo as an Umdena,” Meisharim I).

[4] It is strained to attribute this to the rule of palginan dibbura (“we split the statement”)—that we would believe them that they saw the event but not that they were present—since the latter is entailed by the former. Moreover, the testimony’s admissibility would then hinge on the dispute whether we split the statement or not (see Sanhedrin 9a).

[5] One can also frame this as a kind of “splitting credibility”: they are believed to say they were present for purposes of testifying about the event, but not believed to say so for purposes of opposing testimony that refutes them.