Instances of Self-Reference in Halakhah (Column 366)

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 recent columns I have dealt with self-reference, and I linked it to a distinction between two aspects of the human being: as subject and as object. I mentioned Ron Aharoni’s claim that one must not identify these two, and that such identification is the fallacy underlying all philosophical issues. His claim is that when we formulate a statement about person A who acts or relates in some manner to person B, we may not substitute A for B. Even if it is the same person, we must continue to relate to the statement as though these were two different people (A is the person as subject and B is the person as object). In Halakhah there are several sugyot in which such identification arises, and I thought to devote a column to a brief discussion of several of these sugyot (see also in the lessons 27–28 on the first chapter of Shabbat).

I will begin with two sugyot that discuss two people jointly causing damage, and then consider what happens when the same person plays both roles. The first sugya is the law of one who incites (sics) his fellow’s dog against another fellow, and the second is the law of one who throws a vessel from the roof. One could speak of additional sugyot, such as an ox that pushed another ox into a pit that belongs to the owner of the pushing ox, and more.

Joint Liability for Damage: The Law of “Inciting”

In the Gemara, Bava Kamma 24b, there is a query regarding one who incites his fellow’s dog against another fellow:

They raised a dilemma: One who incites his fellow’s dog against his fellow—what is the law? As for one who certainly incites [his own dog], he is exempt; but what of the owner of the dog? Do we say he can tell him, “What did I do to it?” Or perhaps we say to him: Since you know that your dog, when incited, is incited, you should not have left it [where it could be used to cause damage]…

The halakhah is decided that the inciter is exempt and the dog’s owner must pay half the damage. And later there it states that if one incited his fellow’s dog against his own property, the dog’s owner is, of course, completely exempt. The later authorities discuss the case where Reuben incites his own dog against Simon’s property. The Pnei Yehoshua (Pn”Y) writes there:

In the Gemara—what, is it not that the inciter is exempt and the dog’s owner liable…? It seems the initial assumption is drawn from the wording “he incited the dog against him,” and not “he incited his dog against him,” which would imply a greater novelty—that even though the dog is his, the inciter is nevertheless exempt. Rather, certainly in such a case he would be liable, and the answerer rejects this reasoning; and this is clear.

According to his view, the Gemara’s conclusion is that a person who incites his own dog against his fellow’s property is exempt. So wrote other later authorities as well (such as Ayelet HaShachar, the Chazon Ish, and others).

This is a first example of the substitution mentioned above. The basic claim is as follows: “If person A incites B’s dog against C’s property, the inciter is exempt.” Now we substitute person A himself in place of B, and we obtain the statement: “If person A incites his own dog against C’s property, the inciter is exempt.” The Pn”Y’s claim is that this substitution does not differ from the case where A and B are two different people, and the ruling is the same.

Formally, this is a reasonable and sensible substitution, for the identity of the people in question should not matter, and this substitution is a special case of the general principle. But specifically in this case, the conclusion emerging from this substitution is quite absurd: If a person is negligent and does not guard his dog properly and it bites another’s property, he is certainly liable (like anyone whose property caused damage). How, then, can it be that when he not only failed to guard but even actively incited the dog against another’s property with his own hands, he is exempt?!

Put differently: The inciter’s exemption stems from the fact that he is not the dog’s owner (a person is liable only for damages caused by his own property), and the dog owner’s exemption stems from the fact that he is not at fault (since the other incited it). But if these two people are one and the same, then there is a person who is both the dog’s owner and the one at fault for the damage, and it is unclear how he can be exempt. The exemption in the joint-damage case arises precisely from the split between A and B; therefore, such a substitution of A in place of B in this situation should not preserve the truth-value of the statement.

Sidebar: A Resolution

Our interest here is not the detailed laws of joint damage, but rather the substitution of the kind that involves self-reference. Therefore, explaining the difficulty is not our main concern; nevertheless, I will present a possible explanation (I do not know of another possible explanation for this paradox. See my article here).

As is known, later authorities debated the basis that obligates a person when his property causes damage: Is it negligence in guarding (of course he must also be the owner, but that is only a condition for his duty to guard), or is it that ownership itself imposes responsibility for the damages caused by his property (except that if he was not negligent he is exempt)? For the view that the obligating factor is ownership itself, it is not immediately clear why a person who guarded properly would be exempt. Seemingly even then we should impose liability for the actions of his property.

I explained that to understand the odd ruling about the inciter, one must apparently assume that liability when the dog causes damage first falls upon the dog itself[1] (this is especially plausible regarding “tam” horn damage, where the injured party collects from the body of the damaging animal). Only after the liability arises is it transferred to the owner because he is the monetary “pocket.” (Remember that seizing the dog itself is also an action that impinges on the owner’s assets, for the dog is his property.) Now we can understand the above claim of the Pn”Y: When the dog was incited, there is no reason to impose liability upon it, for it did not act culpably (we do not demand of an animal to resist a human who incites it). Consequently, no monetary liability arises on the dog, and therefore there is nothing to transfer to the owner. This is why the owner is exempt, even though he himself incited the dog. By contrast, when a person is negligent and his dog causes damage, in such a case a liability does arise upon the dog, and, as stated, it is transferred to the owner. The explanation for the “inciter” paradox is that the owner-inciter’s exemption is not related to his measure of blame (which is certainly greater than that of one who was merely negligent) but to the absence of “blame” on the dog.

Joint Liability for Damage: Throwing a Vessel from the Roof

The Gemara in Bava Kamma discusses one who threw a vessel from the roof, and another person came and smashed it just before it hit the ground. There is an opinion (see for example the Shach, Choshen Mishpat §386:27, and the Ramban in Milchamot, end of Bava Kamma, and others) that in such a case both participants are exempt: the first because he did not break it, and the second because he broke an already-broken vessel (since it was about to break in a moment). I once saw a discussion of a case in which the same person who threw the vessel from the roof ran below and smashed it—i.e., the same person plays both roles. It was suggested there that even in such a case he would be exempt. Of course, difficulties arise here similar to those we saw in the previous example.

In Bava Kamma 26b this case appears together with another case:

And Rabbah said: If one threw a vessel from the roof and another came and broke it with a staff, the latter is exempt. What is the reason? He broke a broken object. And Rabbah said: If one threw a vessel from the roof and there were pillows or cushions beneath it, and another came and removed them—or he himself came first and removed them—he is exempt. What is the reason? At the moment he threw it, his force was interrupted [by the cushions].

According to the views mentioned above, Rabbah’s novelty in the case of throwing the vessel is that the second is exempt, but it is clear to him that the first is also exempt.

In the second case (regarding the pillows and cushions) the Gemara also presents a scenario where the thrower himself ran down and “preceded and removed” the pillows, i.e., he acts in the place of the second person. But in that case we can understand why there is a view that exempts him: At the time he threw the vessel it was not destined to break, and therefore he cannot be liable for that; and for removing the cushions he is not liable because it is only grama (indirect causation) (though, incidentally, the halakhah generally rules that he is liable). By contrast, in the case of throwing a vessel from the roof, the one who breaks it is exempt only because of the thrower’s action (since in the breaking he does not act merely as grama, but with direct force). If the breaker and the thrower are the same person, it is difficult to understand how he can be exempt.

I would note that even here the situation is less extreme than in the “inciter” example above, for there it is clear that the inciter performed one complete damaging act. There are not truly two actors there but simply the ordinary case of “property that causes damage.” Here, however, even if the thrower also plays the role of the breaker, there are still two parts to the act, only that the same person plays both roles. Still, in the end it is difficult to accept a view that exempts him.[2]

Interim Summary

In the end, we saw two examples of substituting the same person into both places (variables) in the relevant halakhic proposition. In both examples, the initial intuition is that substituting A for B changes the ruling. That is, the proposition is not indifferent to the identity of the person placed in each variable. Yet the opinions I cited maintain that such a substitution indeed does not change the truth-value of the proposition. The proposition is indifferent to the identity of person B.

This appears to be precisely Aharoni’s claim. He argues that we must not identify the two roles of the same person, and the correct way to relate to such a situation is to view it as two distinct people. For him, substituting person A for B does not change the truth-value of the proposition, since for our purposes this is the case of two different people.

Two Forces in One Person

In Shabbat 5a we find another example of such a substitution:

R. Avin said in the name of R. Ila’i in the name of R. Yoḥanan: If [the receiver] stood in his place and caught it—he is liable; if he moved from his place and caught it—he is exempt. A baraita likewise teaches: “Others say: If he stood in his place and caught it—he is liable; if he moved from his place and caught it—he is exempt.”

If a person threw an object in the public domain toward another person four cubits away, and the latter caught it in his hand, the thrower is liable. The reason is that both the “lifting” and the “placing” are credited to the thrower’s force. But if the receiver moved toward the object and only then caught it, in that case the thrower is also exempt, for here the placing was done by the receiver’s force and not the thrower’s. Thus the thrower only lifted but did not place, and therefore did not violate a Torah prohibition.

Now the Gemara raises a dilemma employing our substitution:

R. Yoḥanan inquired: If one threw an object and then moved from his place and returned and caught it—what is the law? What is the question? R. Adda bar Ahava said: He was asking about “two forces in one person”: Are two forces in one person like one person, and he is liable; or perhaps they are like two persons, and he is exempt? The matter stands unresolved.

When the thrower himself moved and then himself caught the object at a distance of four cubits from where he lifted it, the Gemara is uncertain: Are two forces in one person like one person—thus he is liable—or like two persons—thus he is exempt (as in the previous case)? The Gemara’s formulation itself hints at the substitution problem I described: If we substitute the same person in the place of person B, does it change the truth-value of the proposition or not?

Yet here too it is not clear why there is a side that views this as two different people and therefore exempts. For when we are truly dealing with two distinct people (the Gemara’s first case), the exemption is precisely because they are two: one lifted and the other placed, and therefore no single person performed a complete labor. But if it is the same person, then he lifted and he placed. Why exempt him? Why even raise the possibility of treating this as two different people? This very much resembles the difficulties we encountered in the previous examples.

The Gemara itself speaks of “two forces in one person,” meaning it senses this difficulty and seeks to address it. It seems to mean that in the labor of throwing in the public domain it is required that the lifting and placing be done by one and the same force (and it is not enough that it be done by one and the same person). When a person throws an object in the public domain, the object’s placing is effected by the thrower’s force; that is the definition of the labor of throwing. Therefore, the question arises what happens when that same person runs and catches the object. It is indeed the same person, but the placing is not effected by the force of the throw, rather by a different force. Perhaps this is not the labor of throwing. Note that the sugya deals specifically with throwing four cubits in the public domain, not with carrying from a private domain to a public domain or vice versa. In transferring between domains it is obvious to the Gemara that there is no exemption when the same person lifts in a private domain and places in a public domain, for there the fundamental labor is not defined as an act by one continuous force. It is true that even in those cases, when one person lifts in a private domain and another places in a public domain (see the Mishnah at the beginning of tractate Shabbat), both are exempt; but there the exemption is because there are two people, not because there are two forces. If one lifted and the other placed, both are exempt because this is the case of “two who performed it.”[3] In ordinary carrying, the placing is not effected by the same force as the lifting, for a person lifts in the private domain, walks, and places in the public domain—meaning that even in the ordinary case, where there certainly is liability, it is done by two different forces. The entire discussion in our sugya arises in a situation where the exemption comes from the presence of two forces and not merely two people, and then the discussion emerges whether, when both forces are exerted by one person, this counts as one force (and thus the labor of throwing) or it still counts as two forces and is exempt (for it is not throwing).

In other words: The exemption in the case of carrying where one lifted and the other placed is the exemption of “two who performed it.” By contrast, the exemption in the case of throwing in the public domain—when two people are involved and the second moved and caught—is not because of “two who performed it,” but because there are two forces. Therefore here it is not necessary that, if it is one person, he must be liable. For further explanation, see my lesson 27 on the first chapter of Shabbat.[4]

Interim Summary

In the three examples presented thus far, Aharoni’s logic seems to describe well the halakhic positions cited. According to the opinions I described, the fact that the same person plays both roles does not change the ruling; it is exactly the same as the case where two different people play the two roles. So it emerges for the opinions that exempt the inciter of his own dog; so it emerges for the opinions that exempt the thrower who runs and himself breaks; and so it emerges for the side in the dilemma that “two forces in one person” are considered two distinct forces.

However, the last discussion (the case of two forces) sharpens why these examples are not analogous to Aharoni’s claim. Aharoni speaks of identifying the person as object and as subject, i.e., a situation of a person observing himself or relating to himself, and the question is whether to view this as a case where he relates to another person, or whether substituting himself changes the situation. By contrast, in all the examples discussed thus far the same person is placed into two different roles (variables). There is no relation here between subject and object.

Put differently: In these three examples we are not dealing with a person acting upon himself, but with the combination of two half-acts by the same person upon a third object or situation. There are two half-acts, and the same person performs both. And indeed we saw that the question arises precisely where the exemption does not result from the fact that there are two people but from the fact that there are two forces. Where the exemption is because there are two different people, it is obvious that in the case of one person there should be no exemption. Why? Because if liability requires two half-acts and that same person indeed performed both, there is no reason to exempt him. When two people do so, they are exempt only because they are two different people (the exemption of “two who performed it”).

Let us now turn to cases that appear closer to the situation of identifying object and subject—namely, cases where a person acts upon himself (and not two half-acts upon a third thing, as in the previous examples).

The Atvan De-oraita Inquiry Regarding an Act a Person Performs Upon Himself

R. Yosef Engel, in his book Atvan De-oraita, rule 20, addresses the following principle:

It will be explained that in every place where the Torah forbade performing something upon another, and also forbade receiving that action from another, then if a person performs that act upon himself he transgresses twice: once as the doer and once as the one acted upon.

R. Yosef Engel addresses precisely this question: When Halakhah treats an act of one person upon another, and rules that each of them transgresses, does a person who performs that act upon himself transgress two prohibitions or not?

The example with which he opens is the prohibition of rounding the head (hakafat harosh). In Makkot 20b it is stated that in such a case one incurs eighty lashes:

In the Talmud, tractate Makkot 20b, it is explained that one who rounds the corners of his own head receives eighty [lashes]: for rounding and for being rounded. From here emerges the foregoing principle: when the Torah imposes liability upon the actor and the recipient, then when the action is performed upon the person himself, he is considered both actor and recipient and becomes liable for both.

The discussion assumes there is a prohibition upon the “rounder” (one who rounds the corners of another’s head) and upon the “rounded” (one whose head was rounded). The question is whether, when a person rounds his own head, he transgresses two prohibitions and receives eighty lashes. This already more closely resembles Aharoni’s discussion, for here the same person appears on both sides of the very same act: he is the actor and he is the recipient. He is the subject (who performs the act) and he is also the object (upon whom the act is performed—its patient). His claim is that the substitution truly does not change the truth-value of the statement: if there are two prohibitions, then so it will be when the same person occupies both sides of the equation.

Immediately thereafter he brings additional examples:

This reasoning is further clarified from what is said in Sanhedrin 54a regarding male intercourse, that the Torah made the penetrator and the penetrated liable, and it is explained there that one who performs the initial penetration upon himself becomes liable twice: as the penetrator and as the penetrated. See there. You will find the same logic in Rambam, Laws of Idolatry 3:9: “One who makes an idol for himself—even if he did not make it with his own hands (but others made it for him)—is lashed, as it is said: ‘You shall not make for yourself a graven image…’; and one who makes an idol with his own hands for others is lashed, as it is said: ‘You shall not make gods of molten metal for yourselves’; therefore, one who makes an idol with his own hands for himself is lashed twice.”

These examples indeed evoke Aharoni’s claim, for in all of them the situation is one of a person performing an act upon another person, and we then substitute the actor also into the place of the other. This is unlike the tort cases above, where the person performs two parts of an act upon something else, and the question is whether they may be combined.

The Difficulty

However, the discussion conducted by Atvan De-oraita here is not clear. If there is a prohibition upon the rounder and a prohibition upon the rounded, then clearly if the same person violates both he should receive eighty lashes. How is this different from one who committed two different prohibitions (e.g., ate pork and rounded his fellow’s head), where he clearly receives eighty lashes? What could possibly have led us to think that such a person would receive only forty? We can ask this differently, in a way that clarifies the relevance to our topic: How does this discussion touch upon self-reference at all? In this case there happens to be a “rounder” prohibition and a “rounded” prohibition—two sides of one act—but they are still two distinct prohibitions. Therefore, if the same person violates both, it is obvious that he violates two prohibitions. There is no aspect here of self-reference, but simply a case of a person who committed two prohibitions with one act.[5]

Consider now the sugya in Makkot 20b:

A tanna taught before Rav Ḥisda: “Both the rounder and the rounded incur lashes.” He said to him: “Does one who eats dates in Arbil incur lashes? I tell you: Whose view is this? It is R. Yehudah, who says that for a prohibition without an action one receives lashes.” Rava says: [The baraita deals] with one who rounds his own head, and all agree [he is lashed]. Rav Ashi says: [It deals] with one who assists, and all agree [he is lashed].

Simply put, the discussion is only about lashes, since the rounded person does not perform an action (and for a prohibition without an action one does not incur lashes). But there is certainly a transgression for both the rounder and the rounded. If so, “rounder” and “rounded” are two separate prohibitions, and it is unclear what troubled Atvan De-oraita, and why he even sees here a question of self-reference.

Seemingly, however, in Rambam, Avodah Zarah 12:1, there appears to be a dispute between him and the Ra’avad on this matter:

In what case are these matters stated? In a man who shaves [rounds] another; but a man who is shaved [rounded] is not lashed unless he assisted the shaver; and one who shaves a minor is lashed.

/Ra’avad’s gloss/: “But a man who is shaved is not lashed unless he assisted the shaver”—Even though he is not lashed, since it was with his consent, he still violates a prohibition.

But there is not necessarily a dispute here. It is quite reasonable that the Ra’avad is merely clarifying the Rambam’s own position. And indeed this is how it is ruled in the Shulchan Aruch, Yoreh De’ah §181:4:

The rounded one is also liable if he assisted, such as by inclining his head toward the one rounding him; but there is a prohibition even if he did not assist. Therefore, it is prohibited to be rounded even by a gentile.

If so, the conclusion is that there are two distinct prohibitions. The question thus returns: What was Atvan De-oraita’s doubt? What is the side that, when the same person violates two prohibitions, he would not receive two sets of lashes? And why does he view this as a case of self-reference?

Explaining the Situation

From this discussion it is clear that R. Yosef Engel understood that these are not two distinct prohibitions but two sides of one prohibition. One might define it thus: When a person rounds another’s head he becomes liable because he bears responsibility for the outcome (the rounding that was done). The one whose head was rounded also becomes liable because of his responsibility for the rounding that occurred. Indeed, the enumerators of the commandments count only one prohibition, “You shall not round the corners of your heads,” though that prohibition may be transgressed—and lashes incurred—by the rounder and/or the rounded.

The Gemara states that when the same person is both the rounder and the rounded, he does not need to “assist” in order to receive two sets of lashes (also as the rounded). The act of rounding that he performs as the rounder is considered assistance for the act of the rounded. This is already a significant identification of the rounder with the rounded. Here there is a substitution similar to what we described above, and we have reached self-reference. Although Reuben acts here both as rounder and as rounded, his act as rounder is also considered an act of the rounded. It emerges that we do take into account that the rounder and the rounded are one and the same. In the ordinary case, the rule is: “When A rounds B and B does not assist, A is lashed and B is not.” Now we substitute A in place of B and obtain: “When A rounds himself and he himself does not assist in the rounding, he is lashed once.” R. Yosef Engel notes that we see in the Gemara that this statement is not correct (for he is lashed twice). That is, the statement is not indifferent to the identity of the rounded. This is a counterexample to Aharoni’s claim.

The Rambam’s Novel Ruling

At the end of the Laws of Kilayim (10:31) the Rambam presents a novel ruling:

One who clothes his fellow in kilayim: If the wearer was intentional, the wearer is lashed and the dresser transgresses “Do not put a stumbling block before the blind.” But if the wearer did not know the garment was kilayim and the dresser was intentional, the dresser is lashed and the wearer is exempt.

He rules that if a person clothes his fellow in kilayim, then if the wearer was inadvertent and the dresser intentional, the dresser is lashed. Since lashes are imposed, it is clear the dresser violated the kilayim prohibition itself (for there are no lashes for “before the blind,” and moreover that prohibition applies even if the wearer is intentional). This is despite the fact that the dresser himself did not wear kilayim. It seems he is lashed for the wearer’s transgression, because he caused it and is responsible for it. The Rambam repeats this novelty in two other places—regarding one who defiles a kohen and one who defiles a nazir (where a verse teaches that the defiler is not lashed)—and therefore it is clear that this is a general halakhic principle, not a special law of kilayim (as some of his commentators suggested).

Early and later authorities wondered about the Rambam’s source for this novelty. The glossators brought a source from a somewhat obscure midrash. I once thought there is room to derive this law from the Sifrei cited by Rashi at the beginning of Parashat Mattot (Numbers 30:16):

“After he has heard”—after he heard and upheld it, that he said “It is pleasing to me,” and he later annulled it, even on that same day;

“He shall bear her iniquity”—he enters in her stead. From here we learn that one who causes another to stumble enters in his place for all punishments.

This concerns a husband who told his wife that he annulled her vow when his annulment was ineffective (because he did so after he upheld it, and then it is no longer possible to annul), and she, thinking the vow was annulled, violated it. Rashi brings that in such a case the husband “enters in her stead” for punishments, and similarly for anyone who causes another to stumble. One might have interpreted this to mean punishment by Heaven, since the husband did not violate the vow and it is difficult to see how we could lash him for “He shall not profane his word.” But it is possible that the Rambam learned this literally: the husband will be punished by the court and lashed for “He shall not profane his word,” even though he himself did not violate the vow. The wife’s transgression is considered his transgression, and he is lashed for it.

But in light of what we saw above, there may be a source for the Rambam in the Makkot sugya regarding the rounder and the rounded. We saw there that the fundamental conception underlying the law of one who rounds himself is that the prohibitions of “rounder” and “rounded” are not two distinct prohibitions but one. Indeed, the enumerators of the commandments do not count two prohibitions, and there is no source that there is a separate prohibition for the rounder. Seemingly, this is precisely the Rambam’s ruling: when a person causes another to sin, he himself is considered the sinner in that very prohibition.

Back to the Rounder and the Rounded

Now the novelty of Atvan De-oraita in that sugya is sharper: When a person rounds his own head, only one rounding occurred. Sometimes the rounder bears responsibility for it, and sometimes the rounded does. But it is not plausible that for the very same violation two different people would be lashed; therefore, it is similarly implausible that when a person rounds himself he would be lashed twice for the same violation. The Gemara’s novelty, according to R. Yosef Engel, is based on the fact that if there were two different people and the rounded was also active, then indeed both would be lashed—even though there is only one prohibition, not two. This is a situation where two different people are lashed for one prohibition that was violated. Therefore, here—and only here—there is room to discuss the meaning of self-reference: When a person rounds his own head, even though as the rounded he is passive, will he nevertheless be lashed twice for the same prohibition—once for his responsibility as the rounder and once for his responsibility as the rounded? The Gemara’s conclusion is yes, and it is proven that substituting A for B does, in fact, change the ruling.

A Necessary Clarification: Back to Self-Reference

The example of the rounder and the rounded (as well as other examples raised in the same rule in Atvan De-oraita) already contains self-reference. Unlike all the previous examples, here the case is one in which a person acts upon himself, and the fact that the actor and the recipient are the same person has consequences.

It is important to note that I am not claiming one can derive from these sugyot the conclusion that, according to the Gemara, Aharoni is wrong in his philosophical claim. My goal here is only to illustrate the logic of substitution, and the question whether that substitution preserves the truth-value of the statement or not. In effect I am using Aharoni’s argument and logic to sharpen the connection and the difference between the various Talmudic sugyot I cited, and the role of self-reference in each of them.

In the next column I will address a case even closer to Aharoni’s paradoxical self-reference.

[1] On this matter see my article about liability for payment in “property that causes damage” (in particular one should consider this in light of the sugya of a grazing ox that caused damage).

[2] For a slightly different application, see also here.

[3] Although, as commentators there already noted, this is “two who performed it” in a longitudinal (sequential) sense and not a lateral (simultaneous) one (two who lift and place together), nevertheless even in such a case both are exempt. (Moreover, in the longitudinal division there seems to be even more reason to exempt, for neither performed a complete labor of carrying.)

[4] For now there is only a recording in the video lessons linked above. Written summaries of the lessons will be posted at the end of the year.

[5] See Pesachim 47a and Makkot 21b: “There is one who plows a single furrow and becomes liable for eight prohibitions.”