עדים זוממים and publication bias

The mitzvah of עדים זוממים is both intuitive and logical. False witnesses who pervert the legal system for their own ends are punished in direct proportion to the intended effect of their actions. If their false testimony was liable to cause someone to incur a financial penalty, they get a financial penalty; if lashes, they get lashes; if the death penalty, it’s curtains. However, there is one aspect of עדים זוממים that is not intuitive at all, to say the very least. Devarim 19:19 commands us וַעֲשִׂיתֶם לוֹ כַּאֲשֶׁר זָמַם לַעֲשׂוֹת לְאָחִיו and upon this Rashi writes:

כאשר זמם. ולא כאשר עשה, מכאן אמרו הרגו אין נהרגין

According to this (with which, as a far as I know, no subsequent authority disagrees) a perjurer is punished only if his testimony is discovered to be false before his victim is executed. If, however, the unfortunate victim has already been stoned, strangled or decapitated, then the false witness gets off scot free. What’s done is done.

This is obviously highly odd. Every legal system in the world distinguishes between crimes and attempted crimes, and between crimes that actually resulted in harm versus those that were intended to but never did. The degree to which one should differentiate between crimes attempted and those successfully carried out is a question that admits of no easy solution, but no-one would ever suggest that a completed crime should carry a lesser sentence, let alone no sentence at all! This problem has not gone unnoticed, of course, and two explanations for this surprising halacha can be found among the commentators.

The first and most famous of these dates back to Ramban who argues that if false testimony results in the death penalty being administered by a Torah court, this is ipso facto evidence that the victim had done something to deserve death, albeit not what he was actually executed for. The עדים זוממים were therefore unwitting executors of G-d’s will and, as such, should not be punished. The second explanation, by contrast, argues that the guilt of false witnesses who use the judicial system to murder someone is so severe that it cannot be expiated by receiving the death penalty by human hands. The Torah denies them the mercy of a death sentence so that they receive their full punishment in due course.

In the most obvious sense, these two apologetic explanations are opposites, but what they have in common is that they rest on the assumption that Torah laws are not designed to regulate the life and affairs of the Jewish nation, but to effect metaphysical justice. In the one case, a Torah court is one of the many means which G-d has at His disposal to visit judgment upon the wicked and the actual legal proceedings of the court have no more intrinsic relationship to this process than the instincts of a snake who bites an adulterer’s heel. In the other, the function of the court is not so much to punish sinners as to protect them from punishments in the world to come. The Torah’s judicial punishments are really expiatory sacrifices.

But this is not how the Torah itself presents the issue:

וַעֲשִׂיתֶם לוֹ כַּאֲשֶׁר זָמַם לַעֲשׂוֹת לְאָחִיו וּבִעַרְתָּ הָרָע מִקִּרְבֶּךָ׃ וְהַנִּשְׁאָרִים יִשְׁמְעוּ וְיִרָאוּ וְלֹא־יֹסִפוּ לַעֲשׂוֹת עוֹד כַּדָּבָר הָרָע הַזֶּה בְּקִרְבֶּךָ׃

The punishment of the false witnesses is intended not just to purge the evil from our midst, but also, as with many other punishments proscribed in Devarim, to act as a deterrent. Perjury is a fundamental attack on the basis of a nation’s legal system and, if left unchecked, the use of the courts by citizens to vicariously attack their enemies is the path to anarchy. It is absolutely necessary for perjury to be punished harshly and to be punished in a way that makes an impression on the general public. This is precisely what the Torah demands.¹ This goal, however, is completely undermined if the false witnesses who engineer the judicial murder of the fellow Jew are rendered, by their very success, immune from prosecution.

Most Jews are not particularly peturbed by an issue of this sort. For Hareidim, the Torah, or most of it, is in essence a theoretical concept whose purpose is to be studied not implemented. True, Mashiah will bring the return of the Torah to its place, but this will hardly be of any practical relevance since a world without sin, without want, and without politics will hardly require an actual legal system. For the Modern Orthodox, it does not matter greatly that this or that law seems illogical, for the entire Torah is an antiquated document that one, for reasons of delicacy and communal cohesion, should not interrogate too closely, but which would certainly not be expected to survive such an interrogation intact. It goes without saying that both world views militate strongly against any rational efforts to implement the Torah as the governing system of a state today or forever.

It had been my experience, however, that whenever a law or practice of the Torah appears contrary to reason or evidence, further investigation will revealFbas that the contradiction is a result either of a misconception about reality or a misconception about the Torah itself. In the latter case, the way to determine this is what I have dubbed for polemic reasons the ‘Reactionary Method’. This consists of setting aside the conclusions we already know, collecting all the sources relevant to a particular topic, arranging them chronologically, and applying ordinary rules of interpretation and analysis, resisting, in particular, the urge to interpret sources against their plain sense in order to assimilate them to later documents. Let us now see where such a method leads.

*

The primary source for this topic is Mishnah Makkot 1:6:

אין העדים זוממין נהרגין, עד שיגמר הדין, שהרי הצדוקין אומרים, עד שיהרג, שנאמר נפש תחת נפש. אמרו להם חכמים, והלא כבר נאמר (דברים יט) ועשיתם לו כאשר זמם לעשות לאחיו, והרי אחיו קים. ואם כן למה נאמר נפש תחת נפש, יכול משעה שקבלו עדותן יהרגו, תלמוד לומר, נפש תחת נפש, הא אינן נהרגין עד שיגמר הדין:

The Mishnah here informs of the opinion of Sadducees, which is an inversion of what we saw earlier from Rashi. According to them, the false witnesses can only be punished if their testimony resulted in the defendant being punished by the court. This is based on the concluding verse of the passage which states ‘life for life, eye for eye, tooth for tooth’. The Sadducee understanding, therefore, is that the concept of עדים זוממים is really no more than a specific application of the general principle of retributive justice. Just as one who uses a bow and arrow to murder his fellow is punished with death, so is one who (ab)uses the justice system to perform the same deed. The position of the hachamim is more severe and expansive. It is not necessary to actually engineer someone’s death to be punished as a false witness, it is enough that your testimony secured an erroneous sentence even if that sentence was not carried out.

The next source to look at is Sifrei:²

ועשיתם לו מיכן אמרו אין עדים זוממים נהרגין עד שיגמר הדין שהרי הצדוקים אומרים עד שיהרג שנאמר נפש תחת נפש

We see here, in briefer form, that the dispute between the Sadducees and the sages was about the earliest stage at which עדים זוממים are considered liable. According to the Sadducee interpretation, this only happens at the latest possible point, after the defendant has been executed on the basis of their false testimony. For the sages, however, the witnesses are liable from the moment that they succeed in their immediate goal of perverting the legal process. What I hope has become clear is that there is absolutely no hint in either source of the view according to which the עדים זוממים become exempt from punishment if the victim of their defamation is executed. The argument as presented refers only to the earliest stage at which one can be classified as an עד זומם; absent what we ‘know’ from elsewhere, there is no reason to imagine that the status is temporary. It should be noted that, despite what you may have heard about tzedoqim ‘interpreting the Torah literally’, the interpretation of the hachamim is in much better accord with a reasonable reading of the entire passage, while the Sadducee view is based on taking three words outside of their natural context and over-reading them.

A fuller picture of the situation emerges when we look at Tosefta Sanhedrin 6:6:

 אין העדים נעשים זוממין עד שיגמר הדין לא לוקין ולא משלמין ולא נהרגין עד שיגמר הדין לעולם אין אחד מן העדים נעשה זומם עד שיהיו שניהם זוממין ואין לוקה עד שיהיו שניהם לוקין ואין נהרג עד שיהיו שניהם נהרגים ולא משלם עד שיהיו שניהם משלמין אמר ר’ יהודה בן טבאי אראה בנחמה אם לא הרגתי עד זומם בשביל לעקור מלבן של בייתוסין שהיו אומרים עד שיהרג הנדון אמר לו שמעון בן שטח אראה בנחמה אם לא שפכת דם נקי שהרי אמרה תורה (דברים י״ז:ו׳) על פי שנים עדים או שלשה עדים יומת המת בעדים ב’ ובזוממין ב’ מה עדים שנים אף זוממין ב’ באותה שעה קבל עליו יהודה בן טבאי שלא יהיה מורה הלכה אלא ע”פ שמעון בן שטח.

The Tosefta connects two important laws laws about false witnesses. The first is the one we have already seen, namely that they cannot be punished until the victim of their perjury has been convicted. The second is that they can only be punished in pairs or more, despite the fact that the Torah talks in the singular about an עד זומם. A little reflection reveals that the second law is a consequence of the first: since a conviction requires two or more witnesses, limiting the category of עד זומם to witnesses who caused a conviction necessarily implies that there must be at least two of them.

The Tosefta then tells us that there that there were not two views in the second temple period about the boundaries of an עד זומם, but three. Yehuda ben Tabai presided over the execution of an individual עד זומם, implying that he believed that a witness found to be giving false testimony could be tried and convicted as an עד זומם from the moment he perjured himself, even if this never resulted in a conviction.³ The view of his partner Shimon ben Shetach, and of all the hachamim in subsequent generations, was more restrictive: only once false testimony had resulted in a conviction could a perjurer be convicted as an עד זומם. For Shimon Ben Shetach, the offense consists of perverting the course of justice; giving false witness is a means to this end and a perjurer whose testimony is not accepted as definitive is only an attempted עד זומם. Seen in this light, the view of the Sadducees can be seen as an extreme version of the limitations placed on the category of עד זומם by the hachamim themselves. The Sadducees agree that an עד זומם is one who succeeds in perverting justice, but go further in defining success as the actual execution of the court’s penalty. Put another way, the authoritative view is a midway point between the view (of Yehuda ben Tabai) that a false witness is liable to execution from the moment he perjures himself in a capital case and the (Sadducee) view that he only becomes so if his perjury results in the actual death of an innocent man.

The Yerushalmi has nothing further to say on this subject and so, having almost completed our survey of the sources, our preliminary conclusion is as follows. A false witness can only be convicted and punished as an עד זומם if his testimony results in the conviction of the man whom he perjures himself against. This means that an individual can never be tried on his own as an עד זומם. This is in contradistinction to a more stringent view which would make any person giving false testimony in a court immediately liable for whatever punishment the crime about which he perjures himself carries. It is also opposed to a more extreme sectarian view according to which the perjurers can only be punished if they cause an innocent person to undergo actual punishment. Based on these sources, there is absolutely no reason to imagine that the halacha should follow an inverted form of the Sadducee view, according to which you can be tried as an עד זומם only for a short intermediate period in between your testimony resulting in a sentence and that sentence being carried out. At this stage, however, enters the Bavli (Makkot 5:b):

תנא בריבי אומר לא הרגו נהרגין הרגו אין נהרגין אמר אביו בני לאו קל וחומר הוא אמר לו לימדתנו רבינו שאין עונשין מן הדין דתניא (ויקרא כ, יז) איש אשר יקח [את] אחותו בת אביו או בת אמו אין לי אלא בת אביו שלא בת אמו ובת אמו שלא בת אביו בת אמו ובת אביו מנין ת”ל ערות אחותו גילה עד שלא יאמר יש לי בדין אם ענש על בת אביו שלא בת אמו ובת אמו שלא בת אביו בת אביו ובת אמו לא כל שכן הא למדת שאין עונשין מן הדין

Before reading further, make sure you have understood what it says here fully (click here for an English translation, if necessary, but also pay attention to the wording in Hebrew). It’s necessary, I think, to read it over a couple of times in order to appreciate just how very odd this section of the Bavli really is. Once you have done so, hopefully, you will have noticed the following:

1. The opinion is given in the name of בריבי. This looks like a variant spelling of ‘son of Rabbi _____’, but that doesn’t make much sense. Whether it’s supposed to be a name or a title is unclear and, whatever the case, we have literally no idea who the author of this opinion is.⁴
2. The opinion of בריבי is introduced using the formula תנא, apparently indicating that he is from the era of the tanaim, but it later becomes clear that this is not the case, since בריבי himself cites a baraita using the introductory formula תניא, a term used by the amoraim when quoting a source from the earlier era.⁵
3. בריבי addresses his father using the formula לימדתנו רבינו, which is used in tanaitic texts when a group of students are trying to remind their Rav of something he has previously taught. Here, though, where an individual son is talking to his father, it makes no sense for him to refer to himself in the plural.

All this suggests either that the original source for this view has been substantially, and not very successfully, tampered with or that some of information has been lost or distorted in transmission. The full oddity of this source, though, comes to light if we compare it with the codification of the stated halacha by Rambam (הלכות עדות כ:ב):

נהרג זה שהעידו עליו ואחר כך הוזמו אינן נהרגין מן הדין. שנאמר “כאשר זמם לעשות” ועדין לא עשה. ודבר זה מפי הקבלה. אבל אם לקה זה שהעידו עליו לוקין. וכן אם יצא הממון מיד זה ליד זה בעדותן חוזר לבעליו ומשלמין לו:

Rambam cites a verse to prove the point, in a very similar fashion to Rashi. This exegesis does not come from the gemara itself, which instead argues negatively from the lack of proof in the Torah to the contrary. However, Rambam, evidently aware of the weakness of his proof, then adds that this interpretation is מפי קבלה, meaning that it is an oral tradition stemming from Sinai. But look at what the Bavli says. Even leaving aside the fact that this tradition is not mentioned in any of the earlier sources, the gemara makes abundantly clear that whatever בריבי is doing, he is certainly not quoting an accepted tradition, because when he first states his view, his own father is incredulous!

Let us look further at how בריבי supports his hiddush. The argument is based on the principle that we cannot learn punishments through a fortiori logic. Unless the Torah itself tells us, or unless we can derive it through exegetical methods, we cannot punish someone for a given offense, even if logic would seem to dictate that we must do so. Therefore, argues בריבי, even though we know that a false witness is punished from the moment his testimony has secured a conviction, we cannot infer from this that he should be punished once his false testimony has resulted in an execution.

This argument, however, is full of holes. First of all, the principle of אין [לומדים] עונשים מן הדין applies to distinct but similar offences. What it means is that if the Torah tells us that offence a carries a given punishment, we cannot infer from that fact that offense b does too, even if logic would indicate as much. The principle has nothing to do with the time-frame within which a sinner can be tried and punished for an offence. Nor can it, since there is no other mitzvah in the Torah where any such concept applies. Even more fundamentally, בריבי’s argument only makes sense if we assume that the Torah tells us positively that עדים זוממים are punished within a certain time frame, but is silent about what happens beyond it. However, this is not an accurate description of the situation at all! All the Torah tells us is that עדים זוממים are punished in accordance with the intended consequence of their false testimony. Hazal learn through exegesis that this punishment is limited only to cases in which a defendant was convicted. There is no need to learn by inference what happens to עדים זוממים if the defendant is subsequently executed because none of the earlier sources have given even the slightest hint that this might be any different from when the defendant is still on death row.

To say that בריבי (the identity of whom, to reiterate, remains a mystery) makes a weak argument, then, is actually to somewhat understate the case. The truth is that he does not make any argument at all. What he does is state a novel claim which has no basis in any reading of the biblical text, no support in a Rabbinic source, and is obviously contrary to ordinary logic and legal norms. Instead of offering an argument for there being a time limit on punishment of עדים זוממים, he simply assumes one into existence, then places the burden of proof on his bewildered father to demonstrate otherwise. Finally, he sets the bar for doing so infinitely high by excluding considerations of logic or sense through the citation of an entirely irrelevant legal tradition. We may justly admire the rhetorical skill involved here, but it is no exaggeration to say that standards of argument such as these could just as easily used to to justify any and every outlandish claim that might enter one’s head.

*

To sum up, after reviewing the sources we have uncovered two distinct opinions. The first, presented as the view of the sages in the Mishnah, Tosefta and Sifrei, is that עדים זוממים are only punished after their false testimony has resulted in a conviction. The second, presented as the view of an anonymous Rabbi in the Bavli, is that they are only punished for an intermediary period after the defendant has been convicted, but before he has actually been punished. Clearly, the weight of authority is behind the first view and though a sufficiently strong argument can always triumph over consensus opinion, in this case the Bavli’s opinion is illogical and no positive argument for it is even offered.

This brings us, however, to the nub of the issue. Why does the Bavli ignore the commonsense view found in the earlier sources and state only a counter-intuitive opinion of an anonymous Rabbi? The answer is that the Bavli frequently demonstrates the phenomenon of ‘publication bias’, a preference for recording unusual, exciting, or otherwise eye-catching ideas and opinions, and remaining silent about more modest and mundane views, even if these views are those of the majority.

This in itself is not a problem. The Bavli is hardly excessively short and it was perfectly reasonable for its editors to prioritize recording unusual views that would likely otherwise have been forgotten. After all, you can learn about the default view perfectly well by studying the Mishnah, Tosefta and halachic midrash. Nothing is gained by recording something a fourth time. The problem only comes when people do not learn the other sources and instead approach the Bavli as a fully representative, if not actually comprehensive, account of the Torah in its entirety. One reason Jews did this, historically, is simply that they didn’t actually have access to the Tosefta, Sifrei or Yerushalmi and were forced to rely on the Bavli alone and make the best of it. However, the main reason they did so, and the only reason anyone would do so today, is an act of will based on a particular ideological claim. Take the following:

Other works, written prior or contemporary to the Babylonian Talmud are likewise very important for the understanding of laws, beliefs and history. However, since they were all known to the compilers of the Talmud, it is assumed that when the Talmud disputes these works, it does so for a reason. Therefore, whenever they disagree with the Talmud, decisions found in the Jerusalem Talmud, Midrash and Tosefta are ignored. There are, however, certain special cases, where, because of long established custom, the opinions of other early works are accepted, even when they disagree with the Talmud.

It would be foolish to deny that this is sometimes true. There are, indeed, many sugyot in the Bavli which discuss an issue from many different perspectives, considering, and sometimes rejecting, views found in earlier sources. However, it is clear that it is not always true, as it is not in the case we are looking at. The compilers of the Bavli plainly did not set out to provide an exhaustive discussion of this mishnah, but rather recorded an outlier opinion that would otherwise have been entirely forgotten. A halachic theory in which we ‘ignore’ the vast majority of Hazal and instead follow a lone opinion from the Bavli which necessitates extensive apologetic may be called many things, but an ‘exacting process of metaphysical science, handed down from Sinai’ is not what most immediately comes to mind.

Accepting the fact that the Bavli demonstrates publication bias obviously comes with its own set of stresses and strains, but these are hardly comparable to the apologetic burdens placed on all those who, in the name of defending ‘Hazal’, are forced to defend outlandish and implausible minority opinions. There is no impiety in recognizing the Bavli for what it is and not what later authorities claim it to be, and there is no piety in ignoring what the authors of the Tosefta, Mishnah, Yerushalmi and halachic midrash teach us about the Torah, based on a false premise about the nature of the Bavli.

You should not, therefore, be too surprised to learn that neither Rif nor Behag mention בריבי’s opinion at all. They knew it was there, but they also knew how to read the Mishnah and distinguish the view of the majority and the minority. Rambam, as we have seen, does codify it, but he was clearly uncomfortable doing so and limited its application to capital cases. The reason why בריבי’s opinion became universally accepted is because it included by Rashi in his peirush on the Chumash and thus adopted as part of the official explanation of the Torah. This topic is a microcosm of a broader process in which a particular, extreme view of the Bavli, pioneered by the Rabbis of medieval France spread and radicalized until it came to be seen as synonymous with Torah study itself. But it doesn’t have to be this way and, really, there’s no good reason why it should.

1. Ralbag elucidates this point well
2. The text here is taken from the Finkelstein edition. The version used by Sefaria is full of additions that make it basically identical to the Mishnah.
3. The Tosefta does not spell this out and it is possible to interpret Yehuda ben Tabai’s opinion otherwise. He could have agreed with the view that גמר דין is a necessary precondition for trying someone as an עד זומם, but nevertheless thought it possible to convict only one of the witnesses whose testimony brought about the conviction. This is technically possible, because while two false witnesses are necessary to secure the conviction of an innocent man, not all false witnesses meet the definition of an עד זומם. I believe that the the most reasonable reading of the Tosefta, however, is that Yehuda ben Tabai did not require גמר דין, which is, in any case, the default position that one would arrive at from reading the Torah.
4. See Rashi here and also on Chullin 11b where it is spelled ברבי
5. Even this ‘secondary baraita‘ does not appear to be actually tanaitic. It consists of a scriptural proof of the principle of אין עונשין מן הדין, but the existence of this principle is itself a mahloqet between Abaye and Rava (B Sanhedrin 54:a)