The Law of a Pursuer in Mutual Pursuit (Column 437)
This past Shabbat I spoke in the synagogue about the story of Moses and the two quarreling Hebrews (Dathan and Abiram), and I examined a complex situation of mutual pursuit. Beyond the very topical question of what to do when someone witnesses a duel in the Wild West or in eighteenth-century France, I argued that this discussion also has more practical ramifications (see on this in an article in Midah Tovah for Parashat Shemot, 5767).
The Midrash: “Gilui Milta” (Clarifying a Term)
The discussion begins with Bereishit Rabbah on the parashah (Shemot Rabbah, Parashah 1, sec. 29):
“Alternatively: ‘Nitzim’—they intended to kill one another, as it says (Deut. 25), ‘When men fight together,’ and R. Elazar said: Scripture speaks of a case involving the death penalty. ‘And he said to the wicked one: Why will you strike your fellow?’—it does not say ‘you struck’ but ‘you will strike.’ From here we learn that from the moment a person raises his hand to strike his fellow, even if he has not struck him, he is called wicked. ‘Your fellow’—who is wicked like you—teaches that both are wicked.”
The Midrash states that Dathan and Abiram intended to kill one another—that is, there was mutual pursuit. It compares this to the verses in Deuteronomy (25:11–12):
“When men fight together, a man and his brother, and the wife of the one draws near to rescue her husband from the hand of him who is striking him, and she puts forth her hand and seizes him by his private parts—then you shall cut off her hand; your eye shall not pity.”
At first glance it seems the Midrash learns by gezerah shavah (verbal analogy) that “fighting” implies an attempt to kill, i.e., a lethal altercation. That is how the commentators on this Midrash usually understand it. One must understand that this is not a full-blown gezerah shavah, but what is called a “gilui milta,” namely a lexical inference (see Encyclopedia Talmudit, s.v. “Gilui Milta,” and at the beginning of the entry “Gezerah Shavah”—on the two types of gezerah shavah). Early authorities note that some inferences look like a gezerah shavah but are in fact a gilui milta—deriving the meaning of a term from a biblical parallel. Here the Midrash derives the meaning of the word “nitzim” (“fighting”), and from the parallel it infers that we are dealing with a lethal altercation.
The Midrash: An Interpretive Gezerah Shavah
However, from the Midrash’s own wording one can see that this is not a gilui milta but a gezerah shavah. The Midrash implies that the term “nitzim” does not necessarily denote a lethal outcome, for it says “Scripture speaks of a case involving the death penalty,” which implies that there are also fights of a different sort—mere strife (whose aim is not to kill). Thus, from the Midrash itself it emerges that the meaning of “merivah” here is conflict in general, not specifically a lethal quarrel. The inference is that in this instance it was a special type of quarrel—one with lethal intent. Hence, this is not a gilui milta but an interpretive gezerah shavah, i.e., a verbal analogy that does not legislate a new law but interprets the events and verses.
This can also be seen in Scripture itself, Exodus 21:22–25:
“When men fight and they strike a pregnant woman so that her children come out, but there is no fatality—he shall surely be fined as the woman’s husband shall impose upon him, and he shall pay as the judges determine. But if there is a fatality, then you shall give life for life, eye for eye, tooth for tooth, hand for hand, foot for foot, burn for burn, wound for wound, bruise for bruise.”
Here it seems we are not speaking specifically of a lethal fight, but of an ordinary quarrel between two men that devolved into injury or death for a third party—the woman.[1] Thus it is not correct that the word “they fight” (yinatzû) must mean a fight to the death. If so, whence do we know that with Moses it was specifically a lethal fight? Therefore the Midrash brings a gezerah shavah from the verses in Deuteronomy to prove it. As noted, a gezerah shavah usually teaches law, but at times there is an interpretive gezerah shavah, and it would seem that this is the case here. I will now argue, however, that this is in fact a halakhic gezerah shavah (or at least an interpretive one with halakhic implications).
The Midrash: A Halakhic Gezerah Shavah
The Midrash formulates its conclusion thus: “Nitzim—they intended to kill one another.” Above I explained that the novelty is that it was a lethal fight. In my view, the emphasis is not on “to kill” but on “one another.” That is, nitzim means a quarrel (lethal or not), and a quarrel is by nature mutual. When one person pursues another to kill him, that is not a quarrel and we would not describe them as nitzim. A quarrel is when the two combatants are pursuing one another—a mutual pursuit. From the Deuteronomy verses, the Midrash learns that the fight in question was mutual. Beyond that, it was a lethal fight—but the point is that it was a mutual fight in which both parties sought to kill one another (and not merely to brawl).
Note the interpretive novelty: Moses addresses one of them and says, “Wicked one, why will you strike your fellow!”[2]—which might suggest that only one of them was wicked and pursuing his fellow. The gezerah shavah therefore teaches that there was mutual pursuit. Indeed, the continuation of the Midrash states explicitly: “‘Your fellow’—who is wicked like you—teaches that both are wicked.” In other words, both were pursuing each other and both were wicked. When Moses said, “Wicked one, why will you strike your fellow,” he apparently addressed both, telling each that he was wicked for striking his fellow—even if that fellow was himself wicked. In other words, the Midrash teaches that even in a situation of mutual pursuit there is a law of “pursuer” (rodef), and each party is called wicked. That is already a halakhic innovation. Thus, the gezerah shavah that teaches us it was a mutual quarrel is indeed interpretive, but it has a halakhic consequence: even in mutual pursuit the law of rodef applies to both sides.
How is this learned from the gezerah shavah? For this we must turn to Bava Kamma 28a:
“Come and hear: ‘…and she shall have her hand cut off’—[yet] monetary liability [applies]. What—are we not speaking of a case where she could not save [her husband] in another way? No—[it is] where she could have saved him in another way; but if she could not save him in another way, she is exempt.”
The Gemara understands that the woman acted under the law of rodef (to save her husband). Therefore, if she had another option and nevertheless harmed the pursuer, she is liable for damages; but if she had no other option, she is exempt. A problem arises, however: as we saw, this was a case of mutual pursuit between the two fighters. Does the law of rodef apply in such a case? From the Gemara’s reading of the verses we see that it does; despite the mutual pursuit, the law of rodef applies and the woman may strike one side, even though the “pursuer” she strikes is himself being pursued by her husband.
This, apparently, is what our Midrash learns from there to our parashah. From there we see that even in mutual pursuit the law of rodef applies; therefore, in our case both are deemed wicked and both have the status of rodef. Thus, it is indeed a gezerah shavah and not a mere gilui milta (lexical point). Moreover, it is not only interpretive—it carries a halakhic implication: it teaches the law of mutual pursuit.[3]
Yet as we shall now see, at least two sources indicate that in such a case there is no law of rodef.
Mutual Pursuit: The Jerusalem Talmud
The well-known Mishnah in Ohalot (7:6) discusses a fetus endangering its mother:
“A woman who is having difficulty giving birth—they cut up the fetus in her womb and remove it limb by limb, because her life takes precedence over his. Once its head has emerged, one may not touch it, for one life is not pushed aside on account of another.”
The Mishnah distinguishes between a fetus endangering its mother, whom we kill because her life takes precedence, and a case in which its head has already emerged, in which we leave matters as they are. The Bavli (Sanhedrin 72b) explains the latter on the grounds that there is no law of rodef because “from Heaven they are pursuing him” (mishmaya kardefu lei). So why, when it is still a fetus, do we kill it? Because its life does not have the status of a full human life, and therefore the mother’s life takes precedence (the Rambam gives a slightly different explanation; not our topic here). The Yerushalmi (end of Shabbat ch. 18; Sanhedrin 8:9), however, explains the prohibition to kill the newborn after the head has emerged as follows: “You do not know who is killing whom [who is the pursuer].” That is, there is mutual pursuit, and in such a case the rule is that “refraining is preferable” (shev ve’al ta‘aseh)—one must not intervene.
True, the Bavli, which explains the law differently, might be disagreeing with the Yerushalmi and, in its view, even in mutual pursuit the law of rodef applies. Recall that we saw this also in the Midrash Rabbah cited at the beginning.
Mutual Pursuit: R. Shlomo Eiger
R. Shlomo Eiger (cited in R. Akiva Eiger’s Hiddushim to Ketubot 33b) reaches a similar conclusion from a different sugya. He begins by posing the question:
“I am in doubt: where two people are pursuing one another to kill, and they both began simultaneously (akin to what is in the Rosh and Shulchan Arukh, later §421:13, that where two people injured one another, each pays the excess to the other, but that is only where they began simultaneously), may an onlooker kill either one of them or not? Perhaps in such a case neither has the status of rodef to permit saving [the other] by taking his life, since just as he is pursuing his fellow, so is he being pursued by his fellow.”
Up to this point he is uncertain what the law is in mutual pursuit: do both have the status of rodef, such that a bystander may kill either one (but of course not both), or is there no rodef at all because rodef exists only in an asymmetric case?
He then seeks to resolve the doubt from the Sanhedrin sugya dealing with the Exodus passage we cited about two fighters who strike a pregnant woman. Recall that the verses teach that if the woman dies, the killer is liable to death, but if only the fetuses are lost, he must pay damages.
The Gemara in Sanhedrin 74a cites R. Yonatan ben Shaul:
“As it was taught: R. Yonatan ben Shaul says: A pursuer who was pursuing his fellow to kill him, and one could have saved [the victim] by [wounding] one of the pursuer’s limbs but did not—[the ‘rescuer’] is executed on his account.”
According to R. Yonatan ben Shaul, if one could have saved the pursued by injuring one of the pursuer’s limbs, and the rescuer nevertheless killed the pursuer, he is liable to execution. His proof is from the above verses:
“What is R. Yonatan ben Shaul’s reason? For it is written, ‘When men fight together…,’ and R. Elazar said: Scripture speaks of a case involving the death penalty, for it is written, ‘And if there is a fatality, then you shall give life for life’—and even so the Torah said, ‘But if there is no fatality, he shall surely be fined.’ If you say it is when one could have saved [the pursued] by one of [the pursuer’s] limbs—then he is not given over to be saved by taking his life; hence we find a case where he pays damages. But if you say that even where one could save by [injuring] a limb one may also save by killing, how can you ever find a case of monetary liability?”
The one who killed the woman was pursuing his fellow; in the course of that pursuit he either killed the woman or caused damage. If he merely caused damage, he must pay—despite being a pursuer who would be liable to death, which would ordinarily exempt him from monetary payments.[4] We are forced to say it was a case where the pursued could have been saved by injuring a limb, and in such a case the pursuer is not liable to death and may not be killed; hence he remains liable for the damages he caused. This is the proof for R. Yonatan ben Shaul’s ruling that one who kills a pursuer when he could have saved with a lesser injury is liable to execution.
R. Shlomo Eiger proceeds to resolve his doubt from here:
“Now, in the verse ‘But if there is no fatality’ it says, ‘and they strike a pregnant woman’—in the plural. This can be interpreted in two ways: either that both of them struck her, i.e., the two men mentioned in the verse who fought together; or that ‘they strike’ means ‘this one or that one,’ as Ibn Ezra discusses regarding this expression (at Exodus 21 in Parashat Mishpatim).”
If both fighters together struck the woman, a difficulty arises:
“If so, one can challenge R. Yonatan ben Shaul’s proof in Sanhedrin. For one could say the verse is thus: ‘When they fight… and they strike’—meaning both struck. ‘If there is no fatality, he shall surely be fined’—one of them, i.e., the pursued, for a pursued person who breaks vessels is liable. And don’t ask: in that case why should he pay the entire sum—let him pay only his half. That is not so, for one can say he pays all of it by the law of ‘R. Natan’ (derabbi natan), that wherever one cannot collect from this one, one collects all from the other (Bava Kamma 53 and Shulchan Arukh HM 410:35)… And the verse ‘If there is a fatality, you shall give life [for] life’ means any life—either that of the pursuer or that of the pursued who sought to save himself by the pursuer’s life. Since he intended to kill, if he ended up killing the woman he is executed, for one who intended to kill this person and killed that person is liable; and the fact that he was pursued does not exempt him—just as he is not exempt when he breaks vessels, so too he is not exempt if he kills a person.”
If both struck her, R. Yonatan ben Shaul’s proof can be deflected: perhaps the pursued killed the woman, and then clearly he must pay damages. He intended to kill the pursuer but mistakenly struck the woman; the law is that one who intends to kill one person and kills another is liable.
He therefore concludes that if R. Yonatan nonetheless brought a proof from here, he clearly was unwilling to interpret the verse as both having struck the woman:
“One must say that, according to R. Yonatan ben Shaul, it is impossible to read the verse that way, for then the latter part of the verse would be difficult: ‘If there is a fatality, [you shall give life for life].’ If ‘and they strike’ meant both struck, there would be no liability to death, for it would be like when ten people strike a man and kill him—none are executed (Sanhedrin 78a). Therefore, ‘and they strike’ must mean this one or that one. And then the difficulty stands: if the pursued struck, why should he pay? He had permitted himself to be killed [as a pursuer].”
The reason is that if both together killed her, one cannot understand why anyone is liable to death, for when two people jointly kill, both are exempt from execution.
But this, too, raises a difficulty:
“Even so, a problem remains: how can one establish the verse as a case where one could have saved by wounding a limb? For if the pursued is the one who struck, why should he be liable to death if there is a fatality? Though one who intended to kill nefalim (stillborns) and killed a viable person is liable (Sanhedrin 79), the Sages agree that if one intended to kill a non-viable and killed a viable person, he is exempt. Here, the pursued was permitted to kill the pursuer—even if he could have saved by wounding a limb. Thus, the pursuer, with respect to him, has the status of a non-viable being (nefel) for whom the pursued would not be liable; if so, even if he killed the woman he should not be liable to death…”
If it is not symmetric—there is a pursuer and a pursued—then the pursued cannot be liable, for he is permitted to kill his pursuer. If he tried to kill his pursuer and hit the woman, that is like one who intended a permitted killing (as in killing a non-viable being) and struck someone else, and he is exempt.
He therefore concludes that according to R. Yonatan ben Shaul the verse must be read as follows:
“We must say (for the view that the pursued is always permitted to kill him) that the verse ‘When men fight…’ means that both began simultaneously in a fight of lethal intent, this one to kill that one and that one to kill this one. Since neither had the status of ‘pursued’ vis-à-vis the other, and neither was permitted to kill the other, the other is like a fully viable person to him; therefore, if he killed the woman he is liable.”
We must conclude that, according to R. Yonatan ben Shaul, the verse speaks of a case where both pursued one another; in such a case, neither has the status of pursuer or pursued, and neither may kill the other. Consequently, if they kill the woman they are liable; if only damage results, they must pay (since they do not have the status of rodef). From here he proves that where two people pursue one another there is no law of rodef, as we saw in the Yerushalmi above.
I note that his proof can be deflected in several ways, but I will set that aside here.
Two Ways to Read the Conclusion
The conclusion from these two sources is that in mutual pursuit the law of rodef does not apply. This seems to contradict the Midrash, which indicates that it does apply in such a case. I noted that perhaps the Midrash follows the Bavli, which disputes the Yerushalmi. But it seems possible to reconcile the sources.
Let me preface by noting that the Yerushalmi and R. Shlomo Eiger’s conclusion can be read in two ways:
- The law of rodef exists only in an asymmetric case. In a symmetric case, neither party is a rodef.
- The law of rodef exists even in a symmetric case; however, where both pursue one another it is forbidden for anyone to intervene, because in any case one of them will die and intervention gains nothing. In a symmetric case, “refraining is preferable”; we must leave them to their fight and let whichever one dies, die. We are not to determine who dies, because neither has priority over the other.
It seems to me that on either reading—especially if we are speaking of Dathan and Abiram—in a symmetric case of two such pursuers, we can only wish both sides “good luck” and watch the bout with interest. Indeed, Moses rebukes one of them (or both, as we saw in the Midrash), but there is no hint in the verses that he actually intended to kill either of them (as they suspected him of intending).
Note that on the second possibility both are deemed rodef even in a symmetric case. Still, due to the symmetry one may not intervene and kill either of them. One might argue that perhaps there is, in principle, a license to kill, but we have no way to choose which one to kill, since both are in the same condition and equally culpable. The symmetry compels non-intervention. But that could be solved by a lottery: one could draw lots to decide which of the two to kill. My claim is that this is not the issue: in a symmetric case there is no principled license to kill at all. It is not merely that we cannot choose which one.[5]
Returning to the two readings, the first seems implausible. By way of preface: in halakhah, the law of rodef is treated as a kind of capital liability; killing the pursuer is not only for the sake of saving the pursued—it also has a punitive dimension. The clearest indication is that if the pursuer breaks vessels along the way he is exempt from payment, like one who is liable to capital punishment and monetary penalties and pays neither (the rule of kim lei bederabbah minei; see Rava in Sanhedrin 74a, ruled as halakhah). Indeed, in Afikei Yam, vol. 2, §40, it is demonstrated at length that the pursuer’s death has a punitive aspect. If killing the pursuer were solely to save the pursued, one might say that in a symmetric case there is no license to kill and no law of rodef. But if there is also punishment for being a pursuer, it is unlikely that mutual pursuit would exempt either party from the penalty he deserves; both are guilty, and it would be more reasonable that both are liable to death as pursuers (I speak of a case where both began and both are culpable, as R. Shlomo Eiger explains). Of course we would not actually kill both—there is no point—but the status of rodef would apply to both. Why, then, do the Yerushalmi and R. Shlomo Eiger rule that they are not killed? Because the pursuer is not executed in the usual punitive sense: if one can save the pursued by injuring a limb, there is no license to kill. Saving the pursued must be present alongside any license to kill. He may indeed be “liable to death” as a pursuer, but we do not kill him unless it saves the pursued. In the symmetric case, killing a pursuer does not achieve rescue; therefore, although both have the status of rodef, there is no license to kill either of them.[6]
The upshot is that the second reading is likely correct: in mutual pursuit both have the status of rodef, but there is no license to implement it. This also resolves the apparent contradiction with the Midrash cited at the outset. The Yerushalmi and the Sanhedrin sugya agree that in a symmetric case there is mutual rodef status, but there is no license to implement it in practice.
A Practical Difference for Mutual Rodef: Separating Conjoined Twins
At first glance, the analysis above would have no practical difference: even if both have rodef status, there is no license to implement it by killing one of them. So why was it important for the Midrash to teach, by gezerah shavah, that even in mutual pursuit the law of rodef applies? Is this only a hypothetical law (with implications such as whether a woman’s betrothal would be valid)? I think there are practical ramifications, and a possible example is the case of separating conjoined twins.
Conjoined twins are twins joined to each other; in some cases they share organs—sometimes even a single heart or a single brain. In certain situations we know that if we leave matters as they are, both will die within some months. The question is whether, in such a case, one may perform a separation surgery in which one twin receives the shared organ (the heart or the brain) and the other is left to die. If the surgery succeeds, one life is saved; if we do nothing, both will die.
About fifteen years ago I heard of a case of conjoined twins (I believe they shared a heart), and the halakhic authorities instructed the parents that there was no permission to perform a separation, as it would be considered murder; thus both were left to die. This shocked me; my immediate intuition was that in such a case one should be permitted—indeed obligated—to perform the separation, since it saves at least one life. What sense is there in upholding the prohibition of murder at the cost of losing both lives? The aim of the prohibition is to prevent loss of life; here, because of the prohibition, we lose more life. This seemed utterly irrational to me.
I therefore delved into the sugya, examined the sources and arguments on all sides, and surveyed the views of the halakhic authorities (the result is my article in Techumin on separating conjoined twins). To my astonishment, I discovered a complete consensus among leading decisors that there is no permission to perform a separation in such a case. I am speaking of a symmetric case—there is no reason to assign the organ to one more than to the other, and the chance to save each is identical. In short, there is no collateral consideration that would allow choosing whom to save. In such a case there is, halakhically, a consensus that the surgery is forbidden as an act of murder, and both are left to die. See my article cited above.[7]
One can view the conjoined-twins case as one of mutual pursuit (albeit without intent—but an unintentional pursuer is also a pursuer),[8] but its outcome differs from the cases discussed earlier. In those cases, without intervention one of the two would die; hence I explained that even if there is a law of rodef, there is no license to implement it. By contrast, in conjoined twins, without intervention both will die; here there is room to implement the law of rodef and intervene (i.e., perform the separation and kill one under the law of rodef).
A more clear-cut application would be in the case of two culpable adults (as with Moses’s case) engaged in mutual pursuit, where, absent intervention, both are likely to die. In such a case, if we assume the law of rodef applies, there is certainly justification to intervene to save one life. Thus, even apart from conjoined twins (a more problematic case), in mutual pursuit between adults there can be practical ramifications to recognizing that both have rodef status: there are circumstances that call for intervention.
I will now suggest two ways to permit separating conjoined twins in the case described above: either on the basis of mutual rodef or without invoking rodef.
Permitting Separation of Conjoined Twins Based on Mutual Rodef
Here I propose a permission premised on mutual rodef. True, one might argue that “from Heaven they are pursuing one another,” but where both parties have naturally been placed in this situation, it is possible that the law of rodef nevertheless applies. In the case of the fetus and its mother there is no rodef, because the mother’s danger stems from the fetus and not vice versa (although according to the Yerushalmi it seems to be mutual). In such a case there is no permission to kill the newborn whose head has emerged, because he is not a pursuer; Heaven placed him there. With conjoined twins, however, both have been placed in this situation, and killing one to save the other is warranted. To leave both to die because neither has rodef status is illogical. The absence of rodef status should lighten their situation (i.e., increase our obligation to save them), not make it worse.
Assuming mutual rodef, there is a license to kill either under the law of rodef. Obviously, there is no sense in killing both; the point is to save one life. The question, of course, is how to choose whom to save, given the symmetry. Several disputants argued to me that because of this symmetry one must prefer inaction. But, as noted, there is the option of a lottery; in my view, we should draw lots and save the one who is chosen (more on this in the next column).
Importantly, if there were no rodef status for either, a lottery would not help and would not be permissible: a lottery cannot permit murder. If there is no principled license to kill either, the lottery changes nothing. But if both have rodef status, then killing either is not murder; the only question is how to choose whom to save and whom to let die. In such a case, a lottery is a natural solution. My interlocutors objected that it is forbidden to cast lots over lives; I will address that in the next column.
Permitting Separation of Conjoined Twins Without Mutual Rodef
Now let us assume that conjoined twins do not have mutual rodef status. I contend that even so there is permission (indeed, an obligation) to draw lots and perform the separation to save the one chosen. To explain: imagine two adult twins who must decide for themselves. They would presumably decide to draw lots to determine which one will be saved. Recall that the alternative is that both die; thus, such a lottery is the preferable option for both, since it grants each a 50% chance of survival.
Could one conceivably forbid two competent adults from deciding to hold such a lottery for themselves? Surprisingly, this claim was raised repeatedly in debates I had about this topic, though in my view it is patently unreasonable. I explained that it is akin to a person standing on the roof of a burning building, contemplating a jump: suppose there is a 50% chance of dying in the jump but also a 50% chance of surviving, whereas remaining means certain death. Would anyone forbid him to jump because it is a “possible suicide”? Would anyone forbid a bystander to push him because of a “possible murder”? Any sensible person understands that in such a case one must jump (or push him) to give him a 50% chance of survival. Likewise, it is obvious that adult twins may decide to draw lots to gain a 50% chance to live.
What about twins who are minors and not legally competent to decide? In such a case, the court (beit din) is “the father of orphans,” and must decide for them what they themselves would decide were they competent. Therefore, I argued that the court should conduct a lottery on their behalf and send them for separation surgery—precisely what they would have done as competent adults.
Note that to permit a lottery in this way one need not posit rodef status at all. Even without their being liable to death, performing the surgery is permitted, because we are not “killing” anyone: the one who dies as a result of the surgery would in any case have died. On the contrary, the surgery saves the other who otherwise would also have died. In the idiom of the Chazon Ish, this is an act of rescue, not an act of killing/murder.
I mentioned that one interlocutor argued there is a prohibition against drawing lots over lives. I will address that claim in the next column.
[1] See discussion in Sanhedrin 74a (the sugya will be cited below). According to at least one opinion, that too concerns a lethal altercation.
[2] Only this Shabbat I read in Rabbi Ilai Ofran’s book, Torah Shel ha-Nefesh (in the essay beginning p. 201), that in Scripture the word “lamah” is not a question but an expression of protest or rebuke. The word “madua” conveys a question seeking an answer. Hence I was careful not to use a question mark here (putting theory into practice).
[3] One might object that the reading which applies the law of rodef is that of the Gemara; the verses themselves say to cut off her hand (i.e., monetary payment), and on the surface they say the very opposite: that there is no rodef here. If so, how can one make a gezerah shavah from this? Perhaps the Gemara has a textual hint (it may be learned from yet another gezerah shavah to the Exodus verses I cited, where there is clearly pursuit—though not pursuit of each other, but of the pregnant woman. This requires further study).
[4] The Gemara later there discusses the case as one of “money for this one and death for that one.”
[5] From Rashi in Sanhedrin it would appear that the license to kill the pursuer is based on saving him from sin. If so, even in mutual pursuit one might argue for killing one of them, for although no life would be saved, we would save whoever would otherwise prevail from the sin of murder. After all, even in an asymmetric pursuit we do not “gain” a life numerically: if we leave the pursuer, the pursued dies; if we kill the pursuer, he dies. In either case one dies; the choice to kill the pursuer could then be grounded in saving him from sin. Note that a lottery would in any case prevent murder, since either we kill the would-be killer (the one who would have prevailed) or the would-be victim.
But this reading of Rashi is very difficult for several reasons. For example, the Amoraim dispute the case of a minor pursuer (Sanhedrin 72b), and the halakhah is that he has the status of rodef and must be killed to save the pursued—even though a minor who kills is not liable for sin. It is therefore clear that even in Rashi the “saving from sin” component is coupled with the need to save the pursued; without the latter, we would not permit killing. Accordingly, even per Rashi, in mutual pursuit there is no rationale to permit killing either of them.
[6] This parallels the well-known words of the Brisker Rav on the Rambam, who distinguishes two aspects of the pursuer’s law: (1) an obligation to kill him; (2) that he has “no blood” (the prohibition of murder is suspended with respect to him). See there the proofs and ramifications. My claim here is that in mutual pursuit only (2) applies, not (1).
[7] I now noticed a discussion in Rabbi Shabtai Rappaport’s article here (see notes 1 and 3). He also cites Iggerot Moshe, YD II §60.
[8] This is evident from the case of a minor pursuer: as noted above, the halakhah is to kill him. In the Ohalot case of the fetus whose head has emerged, however, it would seem otherwise—the newborn is not deemed a pursuer because “from Heaven they are pursuing him,” i.e., nature put him in that situation; hence there is no pursuit. But in the Rambam it appears there is pursuit (a “quasi-rodef”), and elsewhere I explained that this is a weaker form of rodef, permitting killing only if he is still a fetus whose life has lesser halakhic weight. One must discuss the case of conjoined twins who have already been born: if that situation is “natural” and there is no full pursuit, then seemingly there is no license to kill either. Again, one must examine whether there is no rodef status at all, or whether only its implementation is barred in practice.
Discussion
C. Mutual pursuit.
C1. You explained the idea that in mutual pursuit it is forbidden to kill one of them (by lottery, or free choice, or according to precedence as in Horayot) by saying that although both are considered pursuers and thus are not protected by the law against bloodshed, there is no law here of saving the pursued party. But the statement of R. Chaim, who innovated a special law of saving the pursued beyond the ordinary law of saving life found throughout the Torah (and this applies even to a Noahide, though a Noahide does not have the general rule of life-saving override. And only in this special law, for some reason, is the life of a fetus less than the life of the mother), was already rejected by the Hazon Ish: in fact the gemara uses this as a paradigm to learn that definite life-saving overrides Shabbat from the law that one may kill a pursuer, and we see that there is nothing here but the life-saving of the pursued and no more. And such life-saving surely exists for each of them even in mutual pursuit—why should it not? So no halakhic rationale remains for distinguishing mutual pursuit from ordinary pursuit.
C2. R. Shimon Eiger does not raise the idea of a lottery, only that “one may kill any one of them,” and you added the lottery in order to break the symmetry. As usual, I do not understand the point of a lottery when the probabilities are equal. If we are indifferent to the outcome, then one can simply decide on some outcome according to whatever slight consideration there may be, just as a tiny straw can tip the scale between two heavy piles, even though the straw by itself is as nothing compared with the other pile. In a responsum you once explained to me that one must grant the chance of life equally. But like every non-consequentialist consideration, I find that hard to accept. (And there is also room to say that a personal decision based on some consideration is generally also egalitarian, because every decision-maker decides according to his own will, and the circumstance that it was precisely this decision-maker who happened to be present and he decided specifically to kill So-and-so is parallel to a case where So-and-so lost in a lottery.)
C3. R. Shimon Eiger says that one particular person struck, not both of them (because otherwise, if ten people struck him they would all be exempt), and in asymmetric pursuit, if the pursued one struck and killed the woman he would be exempt (because he intended to kill the pursuer and it is like one who intended to kill a stillborn fetus), and therefore he explains the case as one of symmetric pursuit, where neither is allowed to kill the other, and therefore if he killed the woman he is liable. You discussed whether he means that there is no law of pursuer here at all, or that there is a law of pursuer but one may not intervene, and you decided on logical grounds that there is a law of pursuer but no rescue here.
Something here escapes me in R. Shimon Eiger and in your explanation. Clearly there is pursuit here, because the gemara derives from here proof for Rabbi Yonatan ben Shaul that one must save the pursued specifically by injuring one of the pursuer’s limbs if possible. If no one is a pursuer at all, then of course he is liable for damages in the case where there is no fatality. R. Shimon Eiger apparently says that each of them has the status of a pursuer with regard to exemption from damages, but no one is allowed to kill either of them (and therefore, if he accidentally killed the woman he is liable and it is not like intending to kill a stillborn fetus). This, of course, seems on its face a strange idea—that there is a law of pursuer for purposes of exemption from damages, but it is forbidden to kill him. It is not comparable to kam lei bed-rabba minei in the case of inadvertent offenses punishable by death, because there he did perform the act that incurs the punishment. But if in the end, in this act, he incurs no death penalty in any way at all (because no one is allowed to kill him), then why is he exempt from damages? Presumably R. Shimon Eiger deals with this, but I do not have it in front of me at the moment, and from the Tur by itself it seems to me impossible to understand.
C4. You explained that there is a pursuer, but one may not intervene. The wording “intervene” is a bit misleading, because R. Shimon Eiger says that even each of them is forbidden to kill the other (otherwise, if he accidentally killed the woman he would be exempt, as one who intended to kill a stillborn fetus, according to his view). In my opinion this does not affect the reasoning itself, but it somewhat weakens its force for someone who thinks there is some kind of magical problem with “outside intervention” more than with what the person himself does.
[Since I came here after an absence of several days, the spirit of joy immediately came upon me, and I remembered the pilpul people say about the silly song “Chad Gadya”: if the cat is pursuing the kid, then the dog is righteous and the stick is wicked, etc., and the Angel of Death is proper—and if so, how could the Holy One, blessed be He, punish him? Rather, the dog is forbidden to intervene, and the stick too is forbidden to intervene, and so on, until the Holy One, blessed be He, comes and “the King establishes the land by justice,” and we see that forever and ever it is absolutely forbidden to intervene, Heaven forbid. Or perhaps the dog is forbidden to intervene in order to save the oppressed, but—as is the baffling custom of the world—it is permitted to intervene against those who intervene (and therefore, if two people are fighting and a third intervenes on behalf of one, all the bystanders who side with the other immediately gang up on him and intervene with the cry, “Don’t intervene!”—and this can be rejected, etc.); and so the cat is wicked because it pursued, and the dog is wicked because it intervened, but the stick is fit because it intervened against the intervener, and the fire is wicked, etc., and the Holy One acted justly.
And now it is explained even better: why is the dog forbidden to intervene if the cat is a pursuer? Obviously because the cat and the kid are pursuing one another, like Jerry the mouse pursuing the cat, and therefore the dog may not intervene, as explained in the Tur. Another interpretation is that the cat’s pursuit had already ended, as it says, “Then came the cat and ate the kid,” and therefore from then on all the dog could do was bring witnesses and accuse him in court, so it was forbidden for him to bite. But it is very difficult that the punishments are not equal: the cat ate the kid and killed it, while the dog only bit the cat, as it says, “and bit the cat,” whereas the slaughterer slaughtered the ox completely—why did he not merely wound the ox like the dog’s bite? But I cannot elaborate. ?.].
C5. By the way, you wrote that according to halakhah, if one intended to kill this person and killed that person, he is liable. I am not sure that is the halakhah (I saw a dispute between Rambam and Raavad), but R. Shimon Eiger said it according to Rabbi Elazar, and he certainly holds that way.
If someone sees Reuven pursuing Shimon and Levi pursuing Yehudah, and he can save only one of them by killing the pursuer, what should he do?
A1. I showed this from the midrash itself. The support I brought from the continuation of the midrash, that both are wicked, is only supporting evidence.
A3. Not necessarily. One of them answered him.
A4. Perhaps. I cited that in the gemara there is an opinion that there too this means intent to kill.
B1. Possibly.
B2. Perhaps, but it seems forced to me.
A1. Where in the midrash does it say that there is here a “law of pursuer”?
A4. ?? I do not understand what you mean, and I found nothing there. The gemara there is discussing a person taking the law into his own hands in a case of loss.
B2. Then why doesn’t the Yerushalmi simply say, “Both are pursuing each other”? [To explain the phrase “you do not know who is killing” as meaning “you know that both are killing” is not merely “forced”; it is simply not what is written.]
C1. There is much evidence for this, and I do not see how it can really be rejected from the paradigm. For example, we see that in the law of pursuer there is an element of punishment and not only saving the pursued. And in Rashi on Sanhedrin this is also saving the pursuer from sin. But I do not understand the connection between the claim you brought from R. Chaim and my remarks. R. Chaim says there is a law of saving the pursued beyond the life-saving rules of the whole Torah. I said something else (in the name of the Griz): that there is another law beyond saving the pursued (which is due to life-saving). In any case, regarding the Hazon Ish’s argument itself, it can be rejected. First, this source for life-saving overriding Shabbat does not remain as halakhah and is disputed there. So it is possible that even if this is the logic of that source, we do not rule that way in halakhah. Beyond that, the proof itself is also not compelling. First, the proof is from a burglar breaking in, not from a pursuer. Second, in truth this is a very problematic proof. Does the fact that one may kill the thief in order to save the homeowner from a possible threat to life show that life has a very high value? On the contrary: by killing the thief, one is taking a life. And third, even if the law of pursuer includes a death-penalty element beyond saving the pursued, one can still derive the same thing. But as I said, this is not relevant to my remarks.
C2. I did not bring the issue of a lottery from R. Shimon Eiger, only the point that in a mutual situation there is no law of pursuer. Regarding the difference between a lottery and arbitrary choice, I explained: you are supposed to give both an equal chance, because that is the fair allocation when it is impossible to divide the result itself. Of course, a truly arbitrary choice is also a kind of lottery. But a choice based on some other consideration is not a fair distribution of the chances and is therefore forbidden.
C3. Rabbi Yonatan ben Shaul brings proof for saving by injuring one of the pursuer’s limbs, but the interpretation of all this as mutual pursuit is that of R. Shimon Eiger. Once we set it up that way, you are right that there must be a law of pursuer in such a case, although from R. Shimon Eiger it appears that he holds there is not.
C4. Well said.
A lottery.
A1. It is the same nitzim as in the mutual pursuit there. And when Moses says that he is wicked because he strikes his fellow, that means he is a pursuer. I do not see the problem here.
A4. I mean what I wrote in note 1.
B2. That wording is easily interpreted as mutual pursuit. You do not know whom to regard as killing the other (to whom to assign the law of pursuer).
A1. This is just a side point, but still. You are attributing to the midrash the view that in mutual pursuit there is a “law of pursuer” on each of them (even if in practice it is forbidden to kill either of them). Whereas the midrash merely interpreted the case as mutual pursuit and added that both are wicked (perhaps with no connection at all). How can one fail to see the problem here?
A4. Please explain more. In Sanhedrin 84 it is explicitly discussing the verses in Exodus: “When men strive and they strike a pregnant woman” (and according to Rabbi Elazar this is intent to kill). In Bava Kamma 28 it is discussing the verses in Deuteronomy: “When men strive together and the wife of one approaches” (and there Rabbi Elazar did not say it means intent to kill, and the gemara there is not talking at all about a pursuer but about taking the law into one’s own hands in a case of loss, and the Rishonim there like Raavad and Rosh discuss one who strikes his fellow in order to injure him, and I suggested above in A4 that the word “together” in the midrash is a scribal error and that it too meant the verses in Exodus). Although none of this is practically significant, because of course I too agree that the midrash cited some verses in which there is a pursuer intending death.
That is, the symmetry between the pursuers here does not prevent the rescuer from intervening, because his intervention helps ensure that at least one righteous person will live. So why in mutual pursuit does this whole issue of symmetry suddenly arise? Just say that here no righteous person is being saved, and therefore one may not intervene.
C1. Indeed, I was mistaken. To my shame, I do not know the well-known words of the Griz, only those of R. Chaim in the first section of Hilkhot Rotzeach and the Hazon Ish’s gloss there, and I imagined they were the same. I now see that they really are not at all close.
[Your rebuttals to the Hazon Ish’s rejection of R. Chaim I do not understand. The gemara there says that for definite danger to life the sources are good, and only for possible danger to life they are not; and in any case we do not multiply disputes unnecessarily. What does “high value of life” have to do with it? The gemara there says that because of the homeowner’s possible danger to life, the burglar may be killed, and the gemara learns from this that because of danger to life one may also violate any other prohibition. If there is a special law of “saving the pursued” beyond the ordinary danger to life of the pursued, then there is no comparison to life-saving on Shabbat. Unless one says that the prohibition of desecrating Shabbat is also considered as “pursuing” the person, and some mysterious law of rescue was innovated there too.]
[By the way, parenthetically, it seems you incidentally made two corrections to my words, and I am honored. C2—as though I had written that you brought the lottery from R. Shimon Eiger (whereas on the contrary, I relied on him in support of my opinion that one may choose arbitrarily. Dividing the actual outcome produces a new result, and is not at all comparable to dividing the chance, where in every case one arrives at the same result in value terms). C3—it seems from your words that you came to correct something in what I wrote, though I did not know what it was.]
(Perhaps I need not say this, but in my opinion outside halakhah, in mutual pursuit one may kill whomever one wants, or sit idle, or hold an auction between them as to who will hire his services. Not only that—even if a rock is about to crush one of them while they are lying bound together and it is unclear which one, a bystander may in principle take one of them and place him against the rock so that it stops. And in the case of Siamese twins, if some madman of the kind who disagrees with you let both die in front of their grieving parents, he would be a murderer in the fullest and most scrupulous sense—filth, abomination—and of course the Sanhedrin to arise at the end of days would judge him so (it will have a lot of work to do revising the Talmud)).
This is a question about the reading of the word “together.” But for our purposes it is enough that it is speaking of intent to kill, wherever they brought it from, in order for the law of mutual pursuit to be novel in our case (an interpretive gezerah shavah with halakhic implications). One can put it this way: once it has been established that even in a case of mutual pursuit both parties are called wicked, the law of mutual pursuit follows as a consequence. One might have said that if both sides consent to mutual pursuit, it is like a duel or a mutual agreement to struggle for life, and there is no wickedness here and therefore no pursuit. This comes to teach us otherwise. And so we infer pursuit from wickedness. The claim is that mutuality is not an exemption that changes our attitude toward people striving against one another. According to this, it is not even essential that the case involve intent to kill. On the contrary, they brought that it is speaking of intent to kill in order to say that this is not merely mutual wickedness but also mutual pursuit.
Aside from the fact that this source is rejected in halakhah, it itself is disputed among the tannaim. The others brought different sources.
This is about saving the pursued, not saving the pursuer (from sin). So there is no point discussing rescue from violating Shabbat. But as I wrote, after all we kill the pursuer—that is, a life is being taken here. So what exactly do we learn from this? I would have learned that the value of life is marginal, since it is permitted to kill in order to save a person from a possible threat to life. If so, then all the more so one should be forbidden to desecrate Shabbat for a possible threat to life. In general, this begs the question, because if one may violate the prohibition of murder in order to save a possible life, what is the all-the-more-so argument that one may desecrate Shabbat? Only because one assumes that desecrating Shabbat is less important/less severe than the value of life. That is begging the question.
As for C3, I was only making a precision.
You save neither any person nor any righteous person. That is because of the symmetry, and that is exactly what I wrote.
I obviously do not agree. As I wrote, fairness requires dividing the chances of survival equally between them. If you did not do that, then you are in effect a kind of half-murderer, because you took away from someone his 50% chance of being saved. Of course this is not murder, only “do not stand idly by your fellow’s blood” (the transgression of failing to save), but you have committed a 50% failure-to-save. Half a failure-to-save, or a failure to save half a life—define it as you wish.
From this you will also understand that you too, who called him a murderer, were mistaken. At most you can say that he transgressed by failing to save. And once we concluded that this is a failure to save, my claim returns that there must be a fair lottery. And one who does not do so, but chooses between them as he wishes, a future Sanhedrin will judge him as is fitting.
According to this, however, I do have something to consider regarding two people walking in the desert and a third person has a flask of water. It is accepted among the decisors that he may give it to whomever he wants. But in my opinion he should conduct a fair lottery between them.
That is because for you the concept of “murder” in the value sense concerns the life of this particular person or that one, not the reduction of life in the world.
Presumably, regarding the prohibition of “do not destroy,” if we are forced to destroy one of two piles of barley, you would indeed say one may choose as one wishes; and in dividing a battery between two Siamese devices, everyone would agree that one should divide it and not sit and do nothing.
Indeed. Both and both. You once wrote that in your view failure to save and murder are the same thing. The argument continues.
But could you please sharpen the difference between “do not murder” and “do not destroy”? In bal tashchit, no one demands “sit and do nothing” (dividing a battery between two devices), nor does anyone deal with the rights of the object (therefore with two piles one will save whichever pile one wants, and one does not even need a lottery), whereas in murder one does. And in all this, for you halakhah and morality merge in complete unity. From where in halakhah do these differences emerge?
The difference is the subject. Killing a person is not eliminating a battery, and the difference is not merely quantitative. Killing a person is harming him, not only the world. Therefore in murder there are two aspects: the reduction of life in the world and the injury to the person who was murdered (an intrusion into his territory). There is here a subject who has rights and whom one may not harm.
You are right that in this context I identify the halakhic definitions with morality. I think I noted this in my article on harming innocents and targeted killing (if not, then at least I said it orally in lectures. In these areas there is overlap between halakhah and morality). The reason is very simple: where there is no reason compelling me to say that halakhah operates according to other categories, and it is possible to interpret it in line with morality—especially when dealing with very important moral principles, where violating them is very grave (human life)—I do so.
From the standpoint of the subject’s right, he can waive it. Meaning: if in mutual pursuit one of them stretches out his neck to the third party (while still maintaining that if the third does not kill him, he himself will kill the other pursuer), then the third party may intervene. Is that really so?
That is not necessary, because the status of the subject does not mean that he is the one who decides. Perhaps the Holy One, blessed be He, is the one who should decide, and a person is not the owner of his own body (as in the well-known words of the Radbaz).
But I personally tend to think that where one of the two volunteers to die, one may choose him and there is no need to conduct a lottery. Some hold that a person may volunteer to die in order to save others (at least in a case of doubt; see Hagahot Maimoniot ch. 1 of Hilkhot Rotzeach in the name of the Yerushalmi). So when he himself is in mortal danger, he is only volunteering to be the one sacrificed for the sake of the others (when in any event one must be sacrificed)—it seems obvious to me that this is permitted.
By the way, regarding the song Chad Gadya, lest a listener suspect me that the plain meaning was lost in the cloud of pilpul: presumably the poet did not think at all that the dog bit the cat because it ate the kid. Rather, the dog, for its own pleasure, acted according to the way of the world—that the great devours the small out of exploitation and wickedness, and the small too devours what is smaller than itself; and the stick too beat the dog for its own pleasure, etc. (or all of them are axes in the hand of Providence that hews with them). And in the end, against the greatest tyrant of all, above whom there is no other tyrant, the Holy One, blessed be He, will come and exact retribution, and all evildoers will be as straw, while the people of Israel live.
Now I have thought of something according to the view that in mutual pursuit it is forbidden even for each of them to kill the one pursuing him (this is R. Shimon Eiger’s opinion, while the Tur’s opinion here was not spelled out—perhaps only external intervention is forbidden). If so, this gives us excellent practical advice: know that every pursuer has a way to exempt himself from liability, for he can stop his pursuit and choose, for a brief moment, to decide not to kill the one pursuing him. Immediately he ceases to be a pursuer, and only the other is pursuing him. How manifold are Your works, O Lord! From that moment it is permitted for him to kill the one pursuing him, and even if he strikes the woman he will be exempt. If so, it is astonishing why he is liable for the woman: let us suspect that perhaps he entertained, for an instant, the thought of stopping his killing, and so became permitted. Are we dealing with fools, who would not scheme by deciding not to kill in order that they might then be allowed to kill? We should have said: leave the money in the possession of its current holder.
There is also here a piece of advice for permissible suicide. If a righteous man wishes to commit suicide permissibly and have another kill him, what should he do? He should go to his righteous friend and say: let us accept upon ourselves mutual pursuit, and arrange the distances so that it will take us at least five minutes to reach one another. During those five minutes, this righteous pursuer who wishes to commit suicide will stand and call to a third man, also righteous like them, and say to him: I hereby permit myself to you; hurry and kill me, and if not, it is decreed and determined upon me to kill my friend, my mutual pursuer, who has kindly agreed to assist me.
The stratagem, of course, is only in order to sift out the halakhic components in the prohibition of suicide where there is no moral prohibition, and all under the supervision of our great rabbis.
The verse is: “And he said to the wicked one, ‘Why would you strike your fellow?’”
Moses did not say “wicked”; rather, Scripture testifies to the wickedness of one of the men who were fighting, and it is specifically to him that Moses speaks.
Who does know?
This is not ordinary doubt, nor vagueness and the law of doubt.
That is an entirely reasonable way of expressing the idea of mutual pursuit, just as with vagueness many people speak as though there were doubt here.
A. The midrash interprets anashim ivrim nitzim (“Hebrew men fighting”) as a case involving intent to kill. You say this is not a merely interpretive clarification, since there are also cases of nitzim that are not about killing intent (perhaps like “when they strove against the Lord”), but rather a gezerah shavah. And this is not an interpretive gezerah shavah (I did not know that this too is called a “gezerah shavah”), but a halakhic one, teaching that in mutual pursuit the law of the pursuer applies and both are wicked.
A1. The midrash only says “wicked,” and you derive from it the law of a pursuer, without any hint in the midrash itself and on the basis of the sugya in Bava Kamma 28 that is not cited in the midrash (and below in A4 I wrote that even there they did not interpret it as a pursuer intending to kill). That does not seem plausible on the interpretive level. In addition, it is difficult to derive laws from a narrative in the Torah, especially one that took place before the giving of the Torah. [I saw in Sanhedrin 58b: “A gentile who struck a Jew is liable to death,” as it is written, “and he struck the Egyptian and hid him in the sand”; and Rambam, Kings 10:6, rules that he is liable to death but is not executed by a court. Perhaps his reasoning is that one does not derive halakhah from an incident, especially one before the giving of the Torah.]
A2. Perhaps the midrash can be explained more simply. The novelty of the midrash is that one who raises his hand is called wicked. Therefore, to interpret it as involving intent to kill minimizes the novelty—only one who raises his hand to kill is called wicked, but not one who raises his hand to hurt. So the mere possibility of interpreting it as intent to kill is enough to require us to interpret it that way. Therefore they quite properly bring Rabbi Elazar’s statement that we find nitzim in a context of intent to kill, and thus here too we should interpret it specifically that way in order to minimize the novelty. Also, the midrash might interpret it on its own as involving intent to kill for various reasons—for example, presumably Moses would not intervene in every ordinary scuffle between two Jews that he happened to see. Also, perhaps they simply interpret nitzim as intent to kill in order to magnify the wickedness of Dathan and Abiram, whom according to the midrash are the ones fighting (it is common among Hazal to attach every possible transgression to the wicked; for example, in Bava Batra 16b it says, “Esau came from the field and he was faint,” and they attributed to him five sins by means of an interpretive gezerah shavah that is not unequivocal—for example, that he violated a betrothed maiden, about whom it is written that he found her “in the field.” That is certainly not a halakhic gezerah shavah), and this has nothing to do with the continuation, that one who raises his hand is called wicked.
A3. By the way, you wrote that when Moses addressed “the wicked one” he apparently addressed both of them. From the continuation of the verses it seems quite clear that he addressed one of them, and that one answered him: “Who made you a ruler and a judge… Do you mean to kill me?” It is already forced to say that each of them said this. Presumably you interpreted it this way because if both are wicked, why did Moses address only one? But one can simply say that he addressed the one whose hand was uppermost at that moment, and nevertheless from the very fact of the quarrel Moses saw that both were trying to kill. And the midrash does not explain that the whole incident was before any blow had yet been struck (in which case both would be equal, and why would Moses address only one?); rather, they certainly had already struck and struck, but from Moses’ wording “will you strike” in the future tense we learn that even with regard to the future he is called wicked, although in the actual event he had also struck in the past.
A4. In Exodus 21 it says, “When men strive,” and in Deuteronomy 25 it says, “When men strive together.” In the midrash you quoted it says “together,” so you brought the verses from Deuteronomy. But in the gemara in Sanhedrin 84 that you cited later, they bring the continuation of Rabbi Elazar’s words: “The verse speaks of intent to kill, as it is written, ‘if there is harm, then you shall give life for life.’” That is, it is speaking about the verses in Exodus 21, and the word “together” in the midrash is apparently a scribal error (one should check newer editions, but it seems clear). And specifically in Exodus Rabbi Elazar interprets it as intent to kill, because that is what the continuation proves: “life for life” (according to his view, literally life), meaning that he intended to kill. In Bava Kamma 28 that you cited, I could not find in any way what you said—that the gemara understands the case there as a pursuer seeking to kill—and the Rishonim there explained it as an ordinary striking.
B. Regarding a fetus whose greater part has emerged, “we do not touch it”—the Yerushalmi explains: “You do not know who is killing whom.”
B1. You explained the Yerushalmi to mean that in mutual pursuit we adopt a passive stance. One could distinguish: in mutual pursuit without guilt (only rescue, not punishment) we adopt a passive stance, but when two murderers are pursuing one another it is permitted to punish one of the murderers and rescue the other murderer (even though on balance there is no net rescue here) through an active intervention. [A rationale for this, it seems to me, could perhaps be invented.]
B2. Seemingly, the Yerushalmi can be interpreted differently. “You do not know who is pursuing whom” does not mean that I know both are pursuers (how does that fit the wording of the Yerushalmi?), but rather that perhaps in truth only one is pursuing the other and we are in ordinary doubt as to who the pursuer is. For example, perhaps the mother could move in a certain way and then the fetus would come out smoothly, but the mother and the doctors do not know exactly how to move, and in such a case one could say that the fetus is not a pursuer at all, but only the mother is pursuing (the fetus and herself). Or perhaps the opposite—for example, the fetus is badly positioned, and if it were positioned properly it would come out, so it is the sole pursuer. And in a doubt as to who is pursuer and who is pursued, it seems obvious that halakhah takes a passive stance. The Bavli disagrees and says there is no doubt of pursuit here, but rather no pursuit at all, because “they are being pursued by Heaven.”