When the justices of the U.S. Supreme Court conclude that some law, or some decision of a lower court, violates the U.S. Constitution, no great difficulties of principle or sentiment need accompany their decision to abrogate the opinions of the earlier legislators or judges. The justices, and others, are expected to understand their decision either as correcting a mistake that had been introduced by fallible people who, with intentions good or bad, or unintentionally, had violated the system's basic rulebook, or as reflecting the fact that since the time those legislators or judges made their decisions something relevant (such as notions of “cruel and unusual punishment” or of what affects interstate commerce) had changed, so what was once constitutionally acceptable no longer is. Thus, however upsetting the substance of the justices’ decision may be, it need not imply a condemnation of their predecessors nor entail a disruption of the system's authority structures—as is seen in the fact that the justices, and American citizens, readily use such explicit verbs as “reverse,” “strike down,” or “overturned” for what the justices do
When, in contrast, the authorities of a religious community discover, or are forced to recognize, that some practice or doctrine to which they subscribe contradicts their sacred scripture, it might not be so simple for them to abrogate the practice or doctrine. For if, as is often the case, they believe that that scripture is as holy and immutable as the deity who inspired it, such abrogation would imply that previous adherents and teachers of the religion—to whom their heirs owe much more pietas than the justices of the U.S. Supreme Court owe legislators or other judges—had, willfully or not, been wrong, indeed sinful. That is a conclusion to be avoided at almost any cost.Footnote 1
When one adds to that consideration the likelihood that the offending practice or doctrine serves some need—for otherwise it would not have been adopted by the community in the first place—it is easy to understand that frequently such religious authorities will respond to the discovery of the contradiction by defending the practice or doctrine. That may be achieved by various routes, of which two standard options are elimination of the contradiction via creative interpretation of scripture or recourse to some overriding extra-scriptural authority, such as revelation or some special figure or council. Such defense of tradition is especially to be expected if the contradiction is uncovered and underscored in a polemical context, in which opponents of the tradition attempt to gain points by pointing out the contradiction.Footnote 2
In contrast, the present study will focus upon a case in which, in the absence of such a polemical context, and when dealing with a law that hardly served any practical need, ancient rabbis adopted another option: when they noticed that a traditional law violated the Bible, they abrogated the traditional law. However, since even in a non-polemical context, and even with regard to an issue of little practical importance, religious teachers will strive to avoid the conclusion that their venerated predecessors had been in sinful error, the rabbis had only two options. One would be explicit emendation of the traditional law, on the basis of the sincere or disingenuous assertion that the received text was in disorder (thereby shifting the blame from venerated teachers to second-rate figures: negligent tradents or scribes). However, ancient rabbis seem to have begun to apply this approach to the Mishnah (upon which we will focus) only in the fourth century, and in BabyloniaFootnote 3—and that might explain why in the case upon which we shall focus, where the problem was recognized already in the third century and in Palestine, another, more elegant solution was applied: the problem was resolved via reinterpretation of the received text.
Put in such a general way, this is not a rare procedure. The way it is usually done is known in rabbinic tradition as ʾuqimta (setting up): when two laws are found to contradict one another, the rabbis “set up” the terms of one or both of them in bounds narrower than those assumed by a more natural interpretation, thus eliminating the problematic collision.Footnote 4 Thus, for two simple examples: when rabbis who held that “the law of the kingdom is law” (i.e., Jewish subjects of a kingdom are obliged to obey its laws) was a basic principle confronted a mishnah that allows Jews to evade taxation by lying to tax-collectors (m. Ned. 3:4), they resolved the contradiction by asserting that the mishnah refers not to all tax-collectors, including those who work on behalf of the government (the only tax-collectors contemplated by natural interpretation!), but only to private ones, i.e., criminal extortionists of “protection” (b. Ned. 28a); similarly, when rabbis who held it to be a basic principle that one is not allowed to use any part of a human corpse confronted a mishnah that allows the use of the hair of a woman who has been executed (m. ‘Arak. 1:4), they resolved the problem by positing that the mishnah refers not to all of the woman's hair, including the hair that grew on her head (the only hair contemplated by natural interpretation!), but only to her wig (b. ‘Arak. 7b).Footnote 5
As these examples show, the typical ʾuqimta eliminates a contradiction by semantic legerdemain: it suggests we read the law as we did before but narrow the application of one or more of the words that it uses. That does not create any new law; it only restricts the application of a previously existing law, leaving the problematic case out of the purview of the offending law. In the case addressed in the present study, however, what had come to be recognized as biblical law was so completely contradicted by a mishnaic law that the problem could not be avoided merely by narrowing the scope of one or the other. Rather, to resolve the problem the rabbis had to do something more far-reaching, and what they chose to do was tinker with the syntax of the mishnaic text, construing the relations among its terms in quite an unnatural way. Given the fact that the text is a legal text, however, combining its words in a new way did not only limit the scope of the original law (as with a usual ʾuqimta); rather, it created a new law. That new law is reasonable enough, but given the way it originated it is not at all surprising that its formulation is anything but natural. It is, in fact, the unnatural character of its formulation, displayed in part 1 of this paper, that first suggested that it is the secondary result of some editorial process.
The two laws upon which this paper will focus deal, respectively, with animals and people. Part 1 will show that the law about animals is formulated very strangely; part 2 will show that the law about people changed over time; and part 3 will first argue that it was a swing toward the Bible that explains why the law about people changed and then show that it was the syntactical adjustment required by that swing that forced the unnatural application to animals of words that once quite naturally referred to people. I offer this case study in the hope that, beyond clarifying the particular case, one that amounts to the abrogation of an early rabbinic law and the genesis of quite a surprising formulation of a later rabbinic law, it will also contribute to the construction of a model explaining the circumstances under which a religious community will follow a particular option when such problems are discovered.
1. A Surprisingly Formulated Rabbinic Law
33 If a man opens a pit, or if a man digs a pit and does not cover it, and an ox or an ass falls into it, 34 the pit-opener shall render compensation: he shall return money to its owner, and the corpse shall be his. (Exod 21:33–34)
This biblical law, which imposes liability upon whoever left a pit open if an ox or ass falls into it and dies, is fairly straightforward and also in agreement, it seems, with our intuitive expectations: if an animal falls into a pit and dies, the one who created the dangerous situation should compensate the animal's owner.Footnote 6
However, some observers, and especially those who dig pits, might well rebel against the notion that they should be held liable for the death of such animals. First, they might argue, we should recognize that when the Bible referred to a “pit” it meant something quite useful: a well or a cistern. People who dig them are not doing anything that is ipso facto illegitimate; far from it, they should be encouraged, and one way of doing that is by limiting their responsibility when things go wrong.Footnote 7 Moreover, human experience shows that most animals successfully avoid falling into pits, so if one does not it is unfair to thrust all of the burden for such a fluke upon the pit-opener. And as for animals that are too young or too stupid to avoid pits—is it not the case that their owners should supervise them and therefore accept at least some of the responsibility for their accidents?Footnote 8
Thoughts such as these evidently lie behind a rabbinic law that limits the broad biblical imposition of liability upon pit-openers to cases in which the animal either could not see the pit (because it was blind or fell in at night) or was so young or so abnormal that it could not realize the danger posed by the pit. As Maimonides formulates this law, there is no liability for the death of a normal animal that could see the pit “for this is like a case of force majeure, because animals usually see impediments and avoid them.”Footnote 9 Indeed, our case is cited, in a modern analysis of rabbinic tort law, as a prime example of the usual rule that “a tortfeasor need pay only for such damage that he should have realized that his actions might cause because such a result is usual.”Footnote 10
But if the law itself is reasonable, insofar as it eliminates liability for normal animals but maintains it for subnormal animals,Footnote 11 the problem we will focus upon in this paper is that the formulation of this law in the central ancient source that states it, the Babylonian Talmud, sounds like a joke. The Talmud does not speak of “normal” and “subnormal” animals. Rather, the way it puts the law (at b. B. Qam. 54a–b, translated in the appendix) is that the mishnaic rule 
(m. B. Qam. 5:6) means “if an ox that is deaf-mute, imbecilic, or underage falls [into a pit and dies, the pit-opener] is liable [for damages].” In my experience, this law—codified that way in the standard compendia of Jewish lawFootnote 12—surprises virtually all who hear of it for the first time, even those with some familiarity with rabbinic literature and law. Often they respond with incredulity or amusement.Footnote 13 Upon reflection, it seems that there are eight reasons for such reactions:
1) This clause of the Mishnah is the sixth time m. B. Qam. 5:5–6 refers without any qualification to liability for an ox that fell into a pit. Readers, accordingly, should be surprised to be informed so belatedly that those references apply only to subnormal oxen, there being no liability if the ox is normal.
2) The distinction between animals that are “deaf-mute” and “imbecilic” and those that are not sounds ridiculous, just as does the talmudic discussion that focuses on whether or not animals are included in references to those who have daʿat (intelligence).Footnote 14 Normal animals cannot speak, and normal animals are devoid of intelligence, as we normally use that word—as is indeed taken for granted elsewhere in the Mishnah: “my ox or my ass, which have no daʿat” (Yad. 4:7). Similarly, § 6 of the Talmud's discussion of our law assumes that only one species—humans—has daʿat. The fact that to make our law acceptable a prominent medieval commentator was forced to assume that only exceptional animals lack daʿat shows how problematic this law is.Footnote 15 True, we could imagine distinguishing between daʿat in the sense of “intelligence” and daʿat in the sense of “instinct” (in our case: to avoid pits) and allowing normal animals the latter.Footnote 16 However, talmudic literature uses only the one term and expends no effort on making such a distinction. Similarly, although the Mishnah once refers to animals that are “deaf-mute” and “imbecilic” (m. Bek. 7:6), it invests no effort in defining these categories or in linking them to issues of legal responsibility.Footnote 17
3) In contrast, the triumvirate 
(“a deaf-mute,Footnote 18 an imbecile, or a minor”) appears very frequently in rabbinic literature, including fifteen times in the Mishnah alone (among them three other occurrences in our m. B. Qam.: 4:4; 6:4; 8:4), always as substantives denoting people of diminished mental capacity and, therefore, of inferior legal status.Footnote 19 This makes it all the more surprising that those three words are not allowed this meaning in the present case.
4) The talmudic law under discussion, that the pit-opener is liable for the death of an animal only if it was mentally incompetent, is, according to the Babylonian Talmud (§ 2), based on the interpretation of by R. Johanan b. Nappaha of Tiberias as adjectives (“deaf-mute, imbecilic, or underage”) describing the ox mentioned just before them.Footnote 20 Below (no. 8) we shall see that two other interpretations of these words are more natural. Here, however, I would emphasize that even if we were to accept R. Johanan's interpretation, we would still need to recognize that there is quite a leap from the assertion that the Mishnah refers to such animals to the Talmud's conclusion that the law is limited to such animals.Footnote 21 As the talmudic discussion reports in § 3, R. Jeremiah—who ought to have known what R. Johanan meantFootnote 22—explained that R. Johanan thought it goes without saying (perhaps: because the Bible assumes it) that the pit-opener is liable for the death of normal animals. According to R. Jeremiah, R. Johanan merely meant that the Mishnah had to spell out the pit-opener's liability for the death of mentally incompetent animals, for without such explicit clarification we might have imagined that in such cases the pit-opener is free of liability because the animal's owner was negligent insofar as he let the animal perambulate unsupervised (see above, n. 8). Given that obvious way of understanding the mishnah (on R. Johanan's interpretation),Footnote 23 the Talmud's decision to reject R. Jeremiah's explanation and limit liability to subnormal animals is a fourth surprise.
5) Moreover, note that according to R. Jeremiah's interpretation this mishnah is built very nicely, in a usual structure that juxtaposes two clauses headed by “even” to draw the border between two basic legal situations.Footnote 24 Namely, it states that, regarding animals, the pit-opener is liable even in the case of a pit-falling animal that was so subnormal that its owner was negligent in letting it perambulate, but regarding people the pit-opener is exempt from liability even in the case of pit-falling individuals so subnormal that the pit-opener could not have expected them to watch where they go.
The symmetry of that construction of the mishnah is very neat (although its second part is not at all without difficulty, as we shall see in the next paragraph), but the parallelism and its point are lost if the law in the first part in fact applies not “even” to subnormal animals but rather “only” to such animals. The fact that the Talmud, by rejecting R. Jeremiah's view, wrecks the mishnah's apparent structure is the fifth surprise.
6) Moreover, although that symmetric understanding of the mishnah (“even. . . even”) is the obvious one to expect, if the first words refer (as the Talmud concludes) to subnormal oxen, the second clause is quite problematic. For if the first clause means there is liability for all oxen, even for oxen without intelligence, the second clause would be expected to mean, correspondingly, that there is no liability for people, even for people without intelligence—but the people the second clause in fact mentions (“son or daughter, slave or slave-woman”) are not distinguished from other people by lack of intelligence. They are distinguished by their lack of full legal status, but as a rule they do possess, just as much as other people, the intelligence needed to understand the dangers of pits and avoid them.Footnote 25 For R. Jeremiah's symmetric explanation of the mishnah to work, the true parallel to substandard animals should have been people without intelligence—the people who, indeed, are usually denoted by the words . This intensifies our surprise that those words are not allowed that meaning here.
7) M. B. Qam. 5:5–6 is quite lawyerly, pedantically stating both sides of the rules it offers: the pit-opener is liable if the pit is ten handbreadths deep but is exempt if it is not ten handbreadths deep, and the pit-opener is exempt if he had properly covered the pit but nevertheless an ox somehow fell into it and died, while he is liable if he had not properly covered the pit. So if the Mishnah really meant to rule that “if an ox that is deaf-mute, imbecilic, or underage fell in and died the pit-opener is liable,” it should have also stated, especially given the surprising nature of that rule, that if the animal was normal the pit-opener is exempt. But it does not—a seventh surprise.
8) Finally, over against R. Johanan's suggestion, and the Talmud's conclusion, that the Mishnah's words 
mean “an ox that is deaf-mute, imbecilic, or underage,” there are two quite natural ways to read these four words as referring to an ox and—as is usual in rabbinic literature—three types of people lacking intelligence, and neither way makes informed listeners laugh. Indeed, they have both been followed by modern translators of the Mishnah. Namely:
a) If read as status constructus, the result is Jacob Neusner's translation: “[If] an ox belonging to a deaf-mute, an idiot, or a minor fell into it, [the owner] is liable.”Footnote 26 That is, in fact, the first interpretation considered in the talmudic discussion (§ 1), just as it is the way the Talmud (and everyone else) understands the exact same four words in another context (goring) earlier in this tractate (4:4).Footnote 27
True, § 1 of the talmudic discussion rejects this interpretation of our mishnah (5:6), the stated reason being that it would imply the unacceptable conclusion that pit-openers are exempt from liability if the ox that fell in and died had belonged to a normally intelligent person. However, the Talmud does not say why that implication is unacceptable. Rashi explains, in his commentary (ad loc.), that the reason is the lack of scriptural basis for such an exemption, but that is not very convincing, given that by the end of its discussion the Talmud settles for just as broad and just as un-scriptural an exemption (for the death of normal animals). Moreover, the Talmud could have read the Mishnah as status constructus and assumed, à la R. Jeremiah, that the Mishnah imposes liability even for animals owned by mentally incompetent people.
True, given the assumption, in biblical law and throughout the talmudic discussion, that even when an animal's owner is normal (and therefore could have been expected to supervise the animal) the pit-opener is nonetheless liable, it follows a fortiori that the pit-opener should be liable if the ox's owner lacks intelligence and so could not have been expected to care for the animal. Therefore, it would have been superfluous for the Mishnah to point out such liability–and that, as was pointed out to me by Robert Brody, might have been reason enough not to read the Mishnah as if it did point it out.
However, that too is hardly reason to reject the interpretation of 
as status constructus. The fact that it is logically superfluous to state something does not mean it will not be spelled out in a legal document. On the contrary, it is very usual for lawyers and legislators to cross all the t's and dot all the i's in their documents,Footnote 28 and the Mishnah, in the adjacent context, does just that, as we saw in no. 7. Such thoroughness of coverage is especially understandable when there is some reason for the special case to receive explicit attention—some reason why, if only briefly, we might have wondered if the law is otherwise. In our case there is such a reason: as noted, the previous chapter of the Mishnah, in its discussion of goring oxen, gives at 4:4 special attention to oxen owned by deaf-mutes, imbeciles, and minors, just as 4:5 refers separately—as does the Bible itself (Exod 21:31–32)—to damages incurred by a boy or a girl, a male slave or a female slave. This special attention to such cases in nearby contexts makes it quite natural for lawyerly draftsmen to address such cases with regard to pit-falling.Footnote 29 In sum: I believe it is easier and more natural to read this clause as status constructus and consider it pedantic and superfluous than to read it, as the Talmud does, as referring to mentally incompetent animals.
b) If read as a list of four beings that lack normal human intelligence, the result is Windfuhr's translation, which opens as follows: “If an ox or a deaf-mute, an imbecile, or a minor falls in.”Footnote 30 Although today we would have a problem with such a list, due to considerations of political correctness, these would not have bothered the ancient rabbis.Footnote 31 Whatever they felt for imbeciles, when it came to legal formulations it seems they would have been right at home in the school of the eleventh-century commentator who explained (in another halakhic context) that “the imbecile is the lowest type of person, for the only difference between people and animals is intelligence and understanding—and the imbecile, since he lacks intelligence, is considered like an animal.”Footnote 32
Thus, it is clear that the Talmud's decision to take the four words in question to mean “an ox that is deaf-mute, imbecilic, or underage” was not the only option nor the most natural one. This raises the question: What made the rabbis of the Talmud, the Amoraim, read this mishnaic text in such an unnatural way?
2. Biblical Law between Text and Logic
As was indicated in paragraphs 5 and 6 of the preceding section, an acceptable interpretation of the first clause of our mishnaic text, which deals with pit-falling animals, should make sense in light of the second clause, which deals with pit-falling people. Accordingly, any attempt to understand why the former was read so strangely might do well to study the history of the latter.
The biblical law of pit-falling makes no reference to human victims. Moreover, as ancient readers of Exodus 21–22 must have noticed, that list of laws distinguishes strictly between damage done to people and damage done to property, dealing first with people and only thereafter with animals, and the law of pit-falling comes in the second section.Footnote 33 Indeed, the law of pit-falling actually opens that second section, splitting the laws of goring oxen into two: it comes between oxen that gore people (21:28–32) and oxen that gore oxen (21:35–36), thus calling attention to the move from people to property. Furthermore, along with the fact that the biblical law of pit-falling refers only to animal victims, ancient readers will have noticed that v. 34 regulates the ownership of the pit-faller's corpse—and it is obvious that such a regulation cannot apply to people. Thus, from all points of view it is clear that Exod 21:33–34 imposes liability for the death of animals alone. Already Philo and Josephus recognized this point.Footnote 34
If ancient readers asked why the Bible makes the pit-opener liable for animals that fall in and die but not for people who do so, the obvious answer, for them as for us, was probably the following: as opposed to animals, people have intelligence and should watch where they are going. This goes together with the general recognition that, legally or not, there are all sorts of dangerous things in public thoroughfares, and so there is a reasonable expectation that people walking about will watch their step, certainly with at least the minimum of attention needed to avoid pits so deep that falling into them can be fatal.Footnote 35 People who do not take such proper care are at least guilty of contributory negligence, and that relieves the pit-opener of liability (see above, n. 8). This is a natural explanation, supplied by various commentators.Footnote 36
If, however, ancient readers concluded that by referring to animals but not to people the Torah expressed the principle that pit-openers are liable for the death of unintelligent beings but not that of intelligent beings, for the latter can and should watch their step, sooner or later questions will have arisen on both sides. For despite the general assumption that people have intelligence but animals do not, some people lack intelligence and, as experience shows, most normal animals avoid falling into pits. Anyone, therefore, who would follow the apparent logic of the Torah here and is willing to ignore the plain fact that Exod 21:33–34 refers only to animal pit-fallers might well make one or two moves that depart from the text of the biblical law: such an authority might on the one hand deny liability when a normal animal falls into a pit and dies, and/or on the other hand impose liability when a person lacking intelligence falls in and dies. Indeed, talmudic law did make the former move; that, and especially the strange way in which it is phrased, was the point of departure of our discussion. It seems, however, that that move, which abrogated liability for the pit-falling death of normal animals, was only a secondary development after a change in the law concerning people, to which we shall now turn.
2a. Tannaitic Law Imposed Liability for the Pit-Falling Death of People Who Lack Intelligence
It would not be surprising if the early rabbis indeed took such a position. After all, they had no trouble in generalizing “pit” into any stationary obstruction in the public domainFootnote 37 and “ox or ass” into any animal,Footnote 38 so why should they not have generalized animals into “beings without intelligence” and included such people too?Footnote 39
In fact, it seems that the Tannaim of the first two Christian centuries did impose such liability.Footnote 40 That this seems hitherto to have gone unnoticed is testimony to the skill of the interpreters and copyists who, as we shall now see, neutralized the original versions of all three baraitot (extra-mishnaic tannaitic texts) cited in the Babylonian Talmud's discussion of our mishnah, of which my translation is offered in the appendix.Footnote 41
According to the first baraita cited by the Talmud (§ 4 [
]), “if an intelligent being fell into it, he [the pit-opener] is exempt.” This clearly means the governing criterion for exemption is the intelligence of the pit-faller and so implies liability for the death of mentally incompetent people. The Talmud evades that implication by insisting, in §§ 5–6, that the baraita refers to the species that is intelligent, namely, the human species—including all of its members, even the exceptional ones who lack intelligence. However, the words “an intelligent being” do not indicate such an inclusive interpretation. Indeed, if all that is meant were “if a person fell in” the pit-opener is exempt, why did the baraita not say just that, using adam (“a person”), as is so usual?Footnote 42
The second baraita (§ 7) too was subjected to similarly heavy-handed interpretation, and to emendation as well. It states that the pit-opener is exempt if the one who fell in and died was 
, and the Talmud apparently puts a great deal of weight upon the conjunction, as if it means that the two terms are not equivalent; it takes the text to refer to “a person or someone else with intelligence.” That understanding of the text allows the Talmud to conclude that this baraita means that the pit-opener has no liability for two types of pit-fallers who died: all people, and those animals that have normal mental capacity. It is this text, on that interpretation, that clinches the debate of this issue in the Talmud's discussion. However, the baraita does not mention animals, and is a strangely vague way to say “a person, or an animal that has intelligence.” In fact, the conjunction need not bear all the weight the Talmud assigns to it, for these three Hebrew words could easily mean “a person who is one who has intelligence,” just as, for example, earlier in this tractate the Mishnah (1:3) rules that witnesses must be 
;Footnote 43 although literally that translates into “free people and (or: or) Jews,” it obviously means “free people who are Jews.”Footnote 44 When applied to our baraita, this reading, which is quite natural,Footnote 45 means that the pit-opener is not liable if an intelligent person falls into the pit and dies. That the Talmud's understanding of this text as referring to people and other intelligent beings is quite unnatural, already suggested by the fact that “nowhere else do we find an ox termed a ben daʿat,”Footnote 46 is shown by the stunning datum that although the text as cited above () is the reading found in all manuscripts, the printed versions all have, instead, 
.Footnote 47 That is, according to the printed versions the baraita says the pit-opener is not liable for the death of “an intelligent ox.” Evidently, someone along the line, no later than the late fifteenth century, realized that the talmudic interpretation of was unconvincing, so if this baraita is supposed to refer to intelligent oxen its text had better be emended.
Finally, after clinching its point in § 7 on the basis of that forced interpretation of the second baraita and seconding it in § 8 with a statement by Rava (a Babylonian sage of the fourth century) that makes the conclusion explicit by stating both parts of the law (the pit-opener is liable for the death of a mentally deficient ox but not for the death of a mentally normal ox), the Talmud now confirms its conclusion by citing a third baraita, which has the same two clauses as Rava's statement. However, as Judith Hauptman has noted, this baraita seems actually to be the same one quoted by the Tosefta (B. Qam. 6:14, discussed in our next paragraph)—but in the Tosefta's version, which presumably is more pristine, the second clause contrasts the mentally deficient animal of the first clause (if indeed that is what that clause means!) not, as in the Talmud's version, with a mentally normal animal, but, rather, as in our mishnah (5:6), with people of limited legal status (sons, daughters, and slaves). That is: Even if (!) this baraita meant to underscore that there is liability for mentally deficient animals, it did not, originally, deny such liability for normal animals; there was, originally, nothing in it that excluded R. Jeremiah's assumption that liability for the pit-falling death of normal animals was taken for granted. If the baraita's current text nevertheless excludes such liability, it seems, as Hauptman noted, that “in the course of time, part of Rava's memra [statement—DRS], ‘an ox of normal intelligence—the owner [of the pit] is exempt’ was tacked on to the baraita as a second clause.”Footnote 48 The very fact that whoever was responsible for the current version of the baraita in § 8 felt the need to emend its text, in order to make explicit the denial of liability for the death of normal oxen, implies that that person (who certainly knew rabbinic Hebrew!) thought the original text of the baraita did not imply such a denial.
We can, it seems, go a step further and show that this third baraita underwent yet another stage of revision. Hauptman's analysis of this baraita was based upon the text of the passage in Zuckermandel's edition of the Tosefta, which follows the Erfurt manuscript: 
; following R. Johanan's interpretation of the Mishnah, that means: “If an ox that is deaf-mute, imbecilic, or underage, or a blind ox [or]Footnote 49 one that walks about at night fell into it [and died, the pit-opener is] liable, [but if the one who fell in was] a son or daughter, a slave or slave-woman—he is exempt.”Footnote 50 However, the same year that Hauptman's volume was published, Lieberman's edition of the Tosefta appeared, based upon the Vienna manuscript, and it mentions “ox” only in the middle of the list: 
(“if a deaf-mute, an imbecile, or a minor, or a blind ox fell into it . . .”).Footnote 51 The natural interpretation of this text is that the first three mentally deficient pit-fallers mentioned are, as usual in rabbinic literature, types of people, and the pit-opener is liable for their death. True, in his notes on this passage Lieberman states his preference for reading with the opening 
. However, Lieberman's only argument in support of that Erfurt reading here is the fact that it agrees with talmudic law.Footnote 52 That is quite a surprising argument given the fact that, as Lieberman himself noted in several other contexts, the Erfurt manuscript was frequently emended to bring it into line with the Babylonian Talmud.Footnote 53 Lieberman's position in this case thus appears strangely harmonistic and conservative. Rather, from the point of view of historical philology the fact that the Vienna manuscript disagrees with accepted talmudic law is a point in its favor, as the lectio difficilior—just as, in general, the Vienna manuscript is usually preferred. So just as Hauptman pointed out that the second clause of this third baraita underwent revision on the basis of Rava's statement, so too, it seems, its opening clause was revised in order to exclude liability for the pit-falling death of people who lack intelligence. That liability, however, was—as we saw—also imposed, originally, by the other two baraitot cited in b. B. Qam. 54b.
2b. So Did the MishnahFootnote 54
Since the three tannaitic texts discussed in the preceding section imposed liability for the pit-falling death of people who lack intelligence, it is likely that the Mishnah did too. This brings us back to the two interpretations set forth above in part 1, no. 8: that the Mishnah refers either to pit-falling animals owned by three types of people who lack intelligence or to four types of pit-fallers that lack intelligence—an ox and three types of people.
According to the latter interpretation, our mishnah explicitly imposes liability for the pit-falling death of people who lack intelligence.Footnote 55 As for the former explanation—which has our mishnah clarify that one is liable even for animals whose owners lack intelligence but not for people even if they are of less than full legal status—it does not address the issue. But given the fact that the most liminal people it mentions in its second clause are those who lack full legal status (children and slaves), saying nothing about people who lack intelligence,Footnote 56 it allows, and might even take for granted, the imposition of liability that we found in the tannaitic texts discussed in part 2a.
3. What Changed in the Talmud? Back to the Biblical Text
Why did things change in the Talmud? Why did the Amoraim abandon the rule that a pit-opener incurs liability if a mentally deficient person falls in and dies? What was it that generated such thorough manhandling as we saw in part 2: Amoraim manhandling the Mishnah and also the baraitot in § 4 and § 7 and copyists helping to neutralize the baraitot in §§ 7–8?
I see no alternative to accepting what the Talmud itself adduces as its fundamental motivation here: the need to avoid contradicting what was taken to be the fiat of Scripture. Namely, in § 6 the Talmud rejects the possibility that the words “If an intelligent being [
] fell into it the pit-opener is exempt” refer to intelligent people, for that reading would imply that the pit-opener is liable if a person lacking intelligence fell in and died, an implication that contradicts what it cites as a scriptural rule: 
(“it is written [in Exod 21:33–34]: ‘an ox,’ [meaning:] ‘and not a person’”). This is not presented as lemma-cum-explanation, as a midrash, but as if it were a plain citation of a biblical verse. Note the failure to use any exegetical terminology or otherwise to argue on behalf of the interpretation; the rule is presented as something of the same finality and givenness as the Bible itself. Were it not for our familiarity with the Bible we would probably translate and punctuate the four Hebrew words as “it is written: ‘An ox and not a person,’” and then fruitlessly search our concordances for such a biblical verse. It is, however, obvious that the rabbis who quoted this “verse,” as other similar “verses,”Footnote 57 knew that they were not all biblical, so what their manner of citation really means is an insistence that we adhere strictly to the letter of the biblical law. It was this stance that forced the Talmud, so as to avoid going beyond Scripture, to claim that when the first baraita denies liability for an “intelligent being” it means “(member of) the intelligent species,” that is, any person, even one without intelligence.
But if it was literal adherence to the biblical text that led the Amoraim to deny liability for the pit-falling death of any person, from a historical point of view it is important to note with Abraham Goldberg that, although the rule “‘an ox’ and not a person” appears three more times in b. B. Qam. (28b; 53b; 54a), it is not found anywhere in tannaitic literature.Footnote 58 This finding, that literal adherence to this biblical text first appears in the post-tannaitic period, dovetails perfectly with my demonstration above that the earlier, tannaitic, law indeed did impose liability in the case of some human pit-fallers: the ones who lack intelligence.
Moreover, it also dovetails perfectly with the work of other scholars who have pointed to a rabbinic turn to the Bible in other ways in the same general period—the days of the late Tannaim and the Amoraim (ca. third–fifth centuries c.e.). I will cite three complexes of relevant evidence.
First, Benjamin de Vries and Yitzhak D. Gilat have pointed out that although the categories “Torah law” and “rabbinic law” existed in the tannaitic period, they became much more popular and strictly defined later on: laws that had previously been thought to be of biblical authority were recognized, late in the tannaitic period and in the amoraic period, as being of merely rabbinic origin.Footnote 59 That indicates the intensification of a scholastic process: the concern to distinguish between sources of law. It is obvious that clearer recognition of the fact that some laws are of biblical authority while others are only of rabbinic authority would intensify the pressure to make sure the latter did not contradict the former.
Correspondingly, Gilat and, in his wake, Christine Hayes, have noted that tannaitic laws and edicts (taqqanot) that had been tolerated although they contradicted the Pentateuch were now brought into line with it, especially in the Babylonian Talmud, either with the aid of imaginative biblical exegesis that demonstrated that they were, in fact, themselves of biblical origin or by imaginative exegesis of the tannaitic law in order to eliminate the contradiction.Footnote 60 The development we are attempting to understand amounts to an impressive instance of the latter.
Finally, the rule that mere logic, even if as cogent as an argument a fortiori, is not a sufficient basis for the imposition of punishment if no biblical verse imposes it explicitly, became accepted, as Chanoch Albeck observed, “in the days of the last Tannaim.”Footnote 61 And if then it was born only as an exegetical rule, explaining why the Bible said things explicitly that we could have inferred logically,Footnote 62 it had the potential to block the adoption or maintenance of laws proposed only on the basis of logic, and in the Talmud it indeed functions that way.Footnote 63 Accordingly, a law that depended upon logic alone to justify the demand of compensation for damages (which, like punishments, entails invasion of an individual's rights and property), beyond what the Bible ordained, would be hard put to maintain itself in the Talmud.
Such moves back toward the Bible, in the late tannaitic period or amoraic period, reflect a more basic change in the attitude toward law: from a position that built on logic together with a relatively loose and “purposive” attitude toward biblical law to one that demanded more adherence to the details of the biblical text.Footnote 64 This was part of a broader development among the rabbis, as recent scholars have shown. Thus, Paul Mandel has shown that prior to the Tannaim the Pharisees dealt primarily with law, not with text, the move to text coming in the second century.Footnote 65 Similarly, Aharon Shemesh has delineated, in the transition from early to late tannaitic literature, a move from a view of “commandments” (mitzvoth) as actions that one is to do or to abstain from doing to a view of them as texts to be obeyed.Footnote 66 Indeed, this general tendency to move from scriptural law as law to scriptural law as text, which is fundamentally an aspect of the professionalization of law and lawyers,Footnote 67 was to have a life of its own in the talmudic period, when regarding the mishnaic text as well the same textual approach could be applied.Footnote 68
Taken together, these two trends—return to the Bible and interpretation of the text of biblical law, not only of its sense and purpose—engendered, as the rabbis moved into the late tannaitic period and the amoraic period, a need to ground received laws in the letter of the Bible.Footnote 69 That was, of course, especially needed in cases in which the received laws seemed to contradict the letter of the Bible. As noted above, often this could happen via imaginative exegesis of the biblical text. At times, however, that could not deliver the goods—and in those cases it would become imperative to revise or even to abandon the received law.
In our case, which is of the latter type, the decision to back away from imposing liability in cases in which the Bible did not mandate such liability created a fascinating problem, because the previous approach had already been ensconced in texts—tannaitic texts. Those texts, as we see in the Mishnah and in the other texts discussed in part 2, had introduced, into the laws of pit-falling, the distinction between normal people and people without intelligence, termed “deaf-mutes, imbeciles, and minors.” What were the rabbis to do with these words when they decided to adhere to the now notionally biblical principle of “‘an ox’ and not a person” that denied the relevance, in the context of pit-falling, of such distinctions among people? Since, unlike justices of the Supreme Court or polemical sectarians, they could not simply rule that their predecessors had erred, they had to reinterpret either the Bible or the words of their predecessors. Since the biblical text (as opposed to its logic) gives no opening for the view that some people are to be treated as oxen, the other route was chosen: the words 
, which had originally functioned, as usual in rabbinic literature, as substantives denoting types of people, were forced to become, instead, adjectives describing certain oxen.Footnote 70
This, therefore, appears to be the genesis of the surprising formulation of the law with which we began. The need to abrogate a law pertaining to people made words pertaining to people superfluous, but, since they could not be erased or ignored, they were put to work in the law pertaining to animals. The fact that the ensuing formulation is so unnatural is a result of, and witness to, the fact that originally the words functioned very differently.
However, while this explains how it happened that the Talmud came to apply, to oxen, adjectives that normally apply to people and make little sense in connection with animals, it does not explain why the Babylonian Talmud decided to limit the law to such subnormal animals. Why not follow R. Jeremiah's position, that the law of course applied to all animals but the Mishnah specifies the abnormal animals for otherwise we might think the law about them is different?Footnote 71 Here, it seems, some attention to the identity of the participants in the talmudic debate will be useful.
As noted in part 1 (no. 4), the interpretation of the Mishnah as referring to mentally deficient animals is ascribed to R. Johanan or his contemporary, R. Eleazar; both of these Palestinian sages are said to have died in 279 c.e. Their view, however, need mean only that the Mishnah specified this special case, not that the law is limited to it, and such a minimalist position was in fact defended by R. Jeremiah, who lived in Tiberias a few decades later, early in the fourth century. In contrast, the talmudic discussion shows that, in Babylonia, Rava taught that the law should be limited to the special case of mentally deficient animals, and R. Aḥa of Diphti pushed to have that view accepted. Rava was more or less a contemporary of R. Jeremiah, but R. Aḥa of Diphti lived a century later, in the fifth century.Footnote 72 This suggests that we should reconstruct a two-step process.
First, third-century sages, in Palestine, in line with the late-tannaitic biblicizing trend described above, made the basic move: in order to avoid overstepping the Bible, they adopted an outlandish and even amusing reading that turned 
from substantives denoting people into adjectives (“deaf-mute, imbecilic, or underage”) applied to oxen. But in this first stage no more was meant than the clarification that there is liability even for such oxen. Once that move had been accomplished, however, over the next few generations Babylonian sages reaped the fruit of it by building upon it to revise the law according to their experience and notions of fairness. Namely, human experience indicates that normal animals usually avoid pits, and therefore human notions of fairness indicate that pit-openers should not be held liable for damages for the fluke occurrence when some normal animal nevertheless falls in. Hitherto, however, biblical law had imposed such liability. Now, however, the new distinction between normal and subnormal oxen, created as a by-product of the elimination of the law about people, allowed the Amoraim to revise the law regarding animals too, and to assert, coming closer to what fairness demands (see n. 11), that the biblical law applies only to subnormal animals.
3a. The Low Cost of the Change
Apart from explaining how and why a text that imposed liability for the pit-falling death of some people turned first into a text that clarified that there is liability even for the pit-falling death of subnormal animals and, eventually, into one that clarifies that there is liability only for the pit-falling death of subnormal animals, it remains to be noted that adoption of these changes was facilitated by the fact that neither cost much. The families of pit-falling deaf-mutes, imbeciles, or minors would not have received much compensation for such economically worthless individuals, and, as for normal animals—since they only rarely fall into pits, abrogation of compensation for them would not matter much. It is likely that, had the value of adhering to the Bible, or to what human experience and fairness indicated about animals, needed to compete with economic interests, the results would not have been so clean-cut or so readily accepted.
Conclusion
In summary, I have suggested that the surprisingly formulated rabbinic law that limits liability for the death of pit-falling oxen to animals that were mentally deficient is the by-product of the following process:
1) Biblical law imposed liability for the pit-falling death of animals alone, making no distinctions among them.
2) Ancient Jews inferred that the reason for the restriction was that, as opposed to people, animals are not intelligent and therefore cannot be expected to understand the danger posed by the pit. This, by analogy, implied pit-openers should also be liable for the death of people without intelligence—as is declared by the original text and sense of the four tannaitic texts studied in part 2.
3) When, in the context of a later biblicizing swing of the pendulum (observable in other late tannaitic and amoraic contexts as well), the Amoraim decided they were not allowed to impose liability for any pit-falling people since the Bible so clearly abstains from referring to people in this connection, they had to deal with the texts that had previously been formulated. However, unlike biblicizing Protestants of the sixteenth century who could simply reject traditional Catholic departures from “sola Scriptura” as sinful, and unlike Supreme Court justices who can simply reject departures from the Constitution by legislatures or lower courts as erroneous, the Amoraim could not simply reject the laws of their revered predecessors, the Tannaim. All they could do, when they noticed that the Torah contradicted tannaitic statements, was reinterpret the former or the latter or edit the text of the latter. In our case, the Amoraim radically reinterpreted the Mishnah and, with the help of tradents and copyists, brought some other tannaitic sources into line. The result was that words that originally distinguished among types of people, contrary to the Bible, were made to refer to animals.
4) The introduction of these distinctions among animals allowed the rabbis, in a secondary step, to limit the liability for the pit-falling death of animals to cases in which the animal had been subnormal or unable to see the pit—thus relieving pit-openers of liability for the death of normal animals, which, as Maimonides noted (n. 9), they cannot, in all fairness, be expected to foresee.
As for why they chose to resolve the problem the way they did, reinterpreting the Mishnah rather than the Bible, we made two suggestions. First, the fact that biblical law unambiguously left people out of its purview meant there was no way the problem could be solved by ʾuqimta, by narrowing any terms used in the Bible or in the Mishnah. Therefore, if the Amoraim were to avoid outright heavy-handed emendation, they had to rework the syntax of one or the other, and given the authority of the Bible, and its lack of openings for ambiguity in this case, it was the syntax of the Mishnah that was reworked. Second, it is important to note that the new law that was produced, that denied liability for the pit-falling death of normal animals, was readily acceptable because it entailed little practical cost.
I hope that recognition of this instance may encourage others to identify more cases and thus help build up a dossier and articulate a model that may further our understanding of the interplay of tradition and human logic, on the one hand, and biblical text and authority, on the other, in Jewish tradition, or, more generally, in the history of scriptural religions.
Appendix: End of m. Baba Qamma 5:6, and discussion of it in b. B. Qam. 54a–bFootnote 73
Mishnah: If a 
fell into a pit—the pit-opener is liable. But if a son or daughter, slave or slave-woman fell in—he is exempt.
Talmud: What does mean?
1) If we were to say it means “an ox belonging to a deaf-mute person, an ox belonging to an imbecile, or an ox belonging to a minor”—that would imply that if an ox belonging to a normal person fell in the pit-opener is exempt! [That would obviously be unacceptable. Rather:]
2) R. Johanan said: The Mishnah means “an ox that is deaf-mute, an ox that is imbecilic, or an ox that is underage.” But that would imply that if a normal ox fell in the pit-opener is exempt! [And that too is unacceptable.]
3) [In order to rescue R. Johanan's suggestion,] R. Jeremiah said: [R. Johanan meant that the Mishnah is to be read] as if it said “it goes without saying,” namely: It goes without saying that the pit-opener is liable if the ox that fell in and died was normal [for that is the normal case, to which the biblical law relates]. But if an ox that is deaf-mute, imbecilic, or underage fell into the pit and died, I might say that it was its deafness or its imbecility or its being underage that caused it to fall, so the pit-opener is exempt—which is why the Mishnah points out that in fact the pit-opener is liable even for such animals.
4) R. Aḥa of DiphtiFootnote 74 said to Rabina: But have we not learned in a baraita, “If an intelligent being [
] fell into it the pit-opener is exempt?” Does not that formulation [which does not simply say “person”] refer to an intelligent ox [which would mean that R. Jeremiah is wrong]?
5) Rabina responded: No, it means an intelligent person. [So nothing contradicts R. Jeremiah's statement that the pit-opener is liable when any animal falls in.]
6) But [R. Aḥa—or an anonymous voice on his behalf—rejoins]: If the point of the baraita were that the pit-opener is exempt if an intelligent person falls into the pit, would that not imply that the pit-opener is liable if the person who fell in and died was not intelligent? [But that would be an unacceptable conclusion, for] it is written: “‘an ox’ [Exod 21:33] and not a person” [thus excluding liability for the death of any person]! Rather [Rabina's interpretation of the baraita is unacceptable, and must be reformulated, as follows]: What does “an intelligent being” [in the baraita cited in § 4] mean? A member of the intelligent species. [So all that baraita means is that the pit-opener is exempt if any person falls in and dies, because all people are members of the species that is intelligent. If, however, any animal falls in, the pit-opener is liable, for all animals are members of a species that is not intelligent. The discussion would have ended here, ratifying R. Jeremiah's position, were it not for another text:]
7) R. Aḥa said to Rabina: But another baraita says: “If a person or someone else with intelligence [
Footnote 75] fell into it the pit-opener is exempt.” [That means that the pit-opener is exempt not only for any person, but also for intelligent animals. So R. Jeremiah's view cannot be maintained. Rather, in conclusion:]
8) Rava accordingly said [diametrically contradicting R. Jeremiah: The Mishnah refers] only to an ox that is deaf-mute, an ox that is imbecilic, or an ox that is underage. If, however, a normal ox falls into the pit and dies the pit-opener is exempt. Why? It should have watched where it was going. And there is also a baraita that says the same: “If an ox that is deaf-mute, imbecilic, or underage, or blind or walking about at night falls into it the pit-opener is liable; but if it is normal and walking about during the day the pit-opener is exempt.”
