The Logic Behind the Majority Criterion (Column 422)

In the previous column I discussed a comparison between two value-laden options. I noted there that people get entangled in problems and contradictions because they fail to notice that one must weigh the benefit and the harm (the “costs”) of the two options before them. I analyzed in that way the conflict between saving life and Shabbat. That reminded me of a discussion with Tirgitz following Column 420, where the same point arose. In this column I will try to spell out what I answered him there.

Tirgitz’s Question

Tirgitz shlit”a raised there the following question:

What, in your reasoned view, do you make of the proposal that only in a public matter (such as raising large livestock, or fostering an atmosphere of mourning over the Temple) does the decision whether to enact a decree depend on the majority of the public? For without a majority, the minority has no practical standing. But when making a fence around the Torah, it is certainly worthwhile to decree in order to save the minority. And this resembles the Rambam’s remark that if the truth benefits one excellent person but harms a thousand simpletons, one should still state the truth for the sake of the excellent person.

What is the logic in thinking that if the Sages assessed that the majority of the public would read by lamplight and tilt the lamp, they would therefore not enact a decree, thus letting the minority stumble? Elsewhere you wrote that “the majority of the public” is feedback from laypeople indicating the decree is mistaken—but that feedback is only one factor and not always correct. Sometimes the laypeople are simply lazy.

And he writes similarly in his continued comment there:

An elegant sevara, Rabbi Yifaor. It is indeed plausible that “better they remain inadvertent sinners” is relevant even if a majority of the public will comply with the decree and only a minority will act deliberately—if the breach of the fence by the minority is particularly severe. Yasher koach.

[I’ll add a little, though you probably won’t agree with this part. The Torah’s rule to follow the majority is very surprising. In a court you must follow the majority, for otherwise you would follow the minority, which is less reasonable. But in prohibitions—say, a piece separated from several pieces—what is so special about the arbitrary 50–50 line, such that the Torah said that if there is more than 50 percent permitted it is allowed, and otherwise not? To the same degree one could have chosen any arbitrary line, say 34.998%. Explanations like those attributed to R. Gedaliah Nadel—that such is the human way to evaluate an aggregate—obviously lack substance. Therefore, here, where there is a difference between the benefit the majority will receive and the harm created for the minority, and it’s not exactly a zero-sum, we revert to the original point that the arbitrary 50–50 line has no intrinsic meaning.]

He returned to this in the Q&A here, and this column is my response.

Initial (Linguistic) Remarks

As you will see there, at first glance his proposal sounded reasonable and logical to me. Moreover, the expression “rov” (majority) can indeed be understood as “many people,” and not necessarily as 51% and up. Thus, for example, at the end of the Book of Esther it says about Mordechai the Jew that he was “pleasing to the rov of his brethren,” and despite all the jokes, the plain meaning is not that he was pleasing to a numerical majority of his brethren, but that he was pleasing to many of his brethren (and perhaps to all of them), or to a substantial number of them—though not necessarily 51%. Likewise, when one speaks of someone who receives “compliments larov,” the intention is not that most people compliment him, but that he receives many compliments (perhaps from all present). And so too “in a multitude of people is the King’s glory,” where the phrase does not necessarily mean a numerical majority of the people, but a large public (see also I Chronicles 4:38; II Chronicles 31:10). This linguistic consideration allows for Tirgitz’s proposal on the interpretive level.

The Source: The Avodah Zarah Sugya

The source of the law is in Avodah Zarah 36a:

“And our Rabbis relied on the words of Rashbag and on the words of Rabbi Elazar ben Tzadok, who would say: One does not enact a decree upon the public unless the majority of the public can abide by it. For Rav Adda bar Ahava said: What is the verse? ‘You are cursed with a curse, yet you rob Me—the whole nation’ (Malachi 3:9). If there is ‘the whole nation’—yes; if not—no.”

Throughout the sugya two assumptions are taken for granted (for some reason the Gemara does not distinguish between them—see the Ritva below): 1) A decree must spread through the public in order to take effect; if it does not spread, it is annulled. 2) One does not enact a decree upon the public if the majority of the public cannot abide by it. The second principle addresses the criteria a court should consider before enacting the decree; the first addresses the need to revoke a decree after the fact if it did not take hold. In both principles the term “majority” appears (“spread among most of Israel,” “most of the public can abide by it”). According to our approach, however, one could explain this as speaking of a substantial segment and not necessarily a numerical majority, as per Tirgitz’s proposal.

Seemingly, from the verse in Malachi cited as the source one sees that “the entire nation” is required, which would apparently demand a majority, for “a majority is like the whole.” But the Malbim there explains: “Do this—the whole nation; not individuals alone,” which can certainly be understood as a substantial segment, not specifically a numerical majority.

The early authorities on that sugya cite Rabbeinu Yonah that when excommunication is declared in a synagogue it is valid only if most of the public is present. The simple reading is indeed “most,” but again one could interpret it as meaning a substantial segment, i.e., that it not be a mere handful.

The Ritva there notes the connection between these two principles (which the Gemara does not differentiate):

“And our Rabbis relied on Rashbag… For if the only reason were that the prohibition did not spread throughout Israel, a court greater than the first would still be required to annul it, as with other decrees that are not among the eighteen [enactments]. But now that most of the public cannot abide by it, even a lesser court can annul it. And if you will say: Since there is this reason—that most of the public cannot abide by it—behold, it is not a decree at all, as we learn from the verse ‘You are cursed with a curse,’ I would answer: That verse teaches both—that they cannot abide by it and that it did not spread; and whenever one of these is present, ‘the whole nation’ is present. Moreover, when it did spread it is more severe, such that even Elijah cannot annul it. And in Horayot 3b we cite that verse regarding the matter of spread and Israel’s acceptance of it. Therefore they needed to say that both were present there—that it did not spread and that they could not abide by it. But whenever it did spread and they accepted it, an important court is required to annul it at the very least. And in our entire sugya there is no distinction between ‘all Israel’ and ‘most of Israel,’ for a majority is as the whole.”

In any case, from the end of his words it seems he understood this as a literal majority, since he invokes the rule that “a majority is as the whole.” With some effort one could still say the phrase is used metaphorically and he too means a substantial segment.

My Reservation: A First Look

In my response there I qualified my initial agreement with his sevarot, and to that end I also invoked the rule of nullification by majority. Let me preface by noting that proposals of this kind typically proceed while ignoring one side of the problem. Life’s equations have two sides, and therefore the natural balance between them is 50%. If you ignore one side, the question indeed arises why not shift the line wherever you wish.

For example, in a democracy we follow the majority (51%). Why precisely a majority—why not 40% or 70%? Everyone understands that choosing some other figure would produce injustice. If we set the threshold at 70%, then if there were a 60% majority we would in effect follow the minority. Why prefer the minority over the majority? And similarly if we set from the outset that “majority” means 40%, the injustice would arise in every scenario. There is always a cost associated with where you place the threshold for following the majority (a cost borne by the minority whose view loses—the sitra achra). Balancing the costs leads us to the natural 50% line.

One may indeed object: there are issues where it matters far more to the minority to have the decision go their way than it matters to the majority. For example, with drafting Haredim to the army one might argue that while equality of burden is important to the majority, for the Haredim their entire world hinges on this decision (let us assume this here, at least for the sake of discussion). In such a case it makes sense to weight the Haredi view differently despite their being a minority. According to this proposal one should decide by a weighted majority—not headcount, but headcount multiplied by the weight (importance) each side assigns to having its position prevail.

But such a criterion is a recipe for trouble. First, it is very hard to measure such importances. Second, people can misrepresent and inflate how much they “care” in order to win. These two difficulties create two practical, societal problems: every decision would force us into deep polling and studies, and you cannot run a society like that (even referendums for every decision are impractical—so studies of this sort are all the more so). Beyond that, endless suspicion would arise between sides about manipulative self-reporting. It is therefore sensible to set a simple criterion—even if less fair—of following a headcount majority. If the majority voluntarily chooses to consider the minority because it recognizes how important the issue is to them, all the better—but it is not reasonable or practical to fix that as a hard rule.

I now contend that this is also the situation in the other contexts about which Tirgitz raised his questions.

Following the Majority

Tirgitz discussed there the rule of “better that they remain inadvertent sinners,” and at the end he also alluded to the laws of nullification and following a majority that I had mentioned:

Regarding the mysterious matter of nullification and following a majority (and not any other line) in prohibitions—as opposed to court decisions—I think I shall wait until the topic returns to center stage.

Well, here it is.

I argued there that his mathematical calculation is flawed because it ignores the cost. He claims regarding the rule “better that they remain inadvertent sinners” that even if harm will occur to a substantial segment of the public (who will violate the decree deliberately), that suffices to refrain from rebuke and leave them inadvertent. That is not correct only in a case where a majority would act deliberately. By the same token, one could argue concerning the principle in Avodah Zarah—not to enact a decree that most of the public cannot observe—that if a decree will not spread in a substantial segment of the public, i.e., they cannot abide by it, then one should not enact it even if it is not a literal majority.

But both claims ignore the other side of the equation. When weighing whether to enact—or whether to rebuke—there is a problem on both sides of these decisions. On the one hand, the starting point is that there is a need for the decree; therefore, if you do not enact it, you harm that need. Likewise with rebuke: there is a need to rebuke so that the public will fulfill its duty. When we decide not to rebuke, that harms those who would have complied. On the other hand, there is reason not to enact—to avoid burdening and causing people to stumble; and not to rebuke—to avoid turning people into deliberate sinners. The natural balance between these two sides is to follow an actual majority, not merely a substantial segment. In both cases we weigh the costs and benefits of the two alternatives against one another, as I will now detail.

Regarding rebuke (“better that they remain inadvertent sinners”), the two alternatives are: A) Rebuke, in which case part of the public will comply and gain proper behavior, while another part will not comply and will act more harmfully (deliberately rather than inadvertently). Because there are costs on both sides, we must set a criterion for deciding between the two options. It is sensible and logical to set the line at the majority, for the good of the majority overrides that of the minority. True, there can be cases where the harm from not rebuking is enormous, or the harm from rebuking is enormous, which can neutralize the numerical advantage—just as we saw in the democracy example above. But if we need a general, simple rule, it is very reasonable to set the line at the majority.

So too regarding whether to enact the decree in the first place. Again, the two alternatives are: A) Enact it, and then part of the public will not comply and we will have created transgressive behavior; whereas the other part will comply and gain proper behavior. B) Do not enact it, in which case the expected benefit from the decree (avoiding prohibition) will not be realized for that part of the public that would have observed it—but others will at least not be halachic offenders (though with respect to the underlying reason for the decree they will indeed be acting improperly). And again, if one must draw a general line, it is reasonable to set it at 51%. The good of the majority overrides the good of the minority.

Therefore, in the end I argued that it is logical to set a criterion of a majority rather than suffice with a substantial segment of the public. Because the other side also bears costs, we must weigh the costs of both paths before deciding between them. When an equation has two sides, the natural balance point is the midpoint—50%.

A Note: Is the Very Restriction Itself Also a Cost?

In my remarks there I also raised the sevara that the very enactment of a decree bears a cost, because I am imposing a restriction on the majority who will observe the decree—a restriction not required by strict law. Tirgitz argued that this is not a cost, for the halachic logic says there is a need to enact in such a case (the decree benefits that same majority who will observe it). Therefore, the restriction imposed upon the majority is not a “cost.”

I agree—but only partially—because that restriction is also a cost. After all, we oblige people to act in ways they are not obligated by the letter of the law to act. True, this is justified; still, one cannot say there is no cost. Therefore it makes sense to factor this cost into the overall assessment comparing the two alternatives.

An analogy can be found in R. Akiva Eiger’s sevara regarding a biblical doubt (discussed at length in the Radzin Rebbe’s book Shifuni Temunei Chol about tekhelet). As is known, the general rule is that in a biblical doubt one must be stringent. But R. Akiva Eiger argues that this applies only where, if we are stringent, we certainly avoid transgression. Where even by being stringent one may still transgress, there is no obligation to be stringent. For example, if before us is a doubt whether a piece is forbidden fat (chelev) or permitted fat (shuman): if we are stringent and do not eat, we certainly avoid prohibition—hence the duty to be stringent. By contrast, if we have tzitzit dyed with tekhelet and we are unsure whether this is the correct tekhelet, there is no duty to be stringent and attach it to the garment. One might say there is logic to be stringent—improving our odds of keeping the law—but, argues R. Akiva Eiger, in such a case even if we are stringent and attach it, if this is not the correct tekhelet, we still walked without a tekhelet thread and nullified a positive commandment. The stringency does not ensure we have avoided an offense. In such a case there is no obligation to be stringent.

Similarly here: there are certainly cases where it is justified to impose a restrictive decree. But if even imposing the decree will not necessarily yield the hoped-for benefit—or at least not its full benefit—perhaps there is no justification to restrict the entire public.[1]

Example: “One may not read by the light of a lamp lest he tilt it”

The Talmud in Shabbat rules that it is forbidden on Shabbat to read by lamplight, lest the flame dim and one tilt the lamp so that the oil reaches the wick—thus transgressing kindling or extinguishing. Consider the Sages as they debated whether to enact this decree. There is certainly a need for it, because people might come to tilt the lamp. On the other hand, this is a restriction not required by law, and more: it prevents people from reading and learning on Friday nights—i.e., it also has a spiritual cost. Furthermore, even if enacted there will be those who will not comply, and then we cause them to transgress a rabbinic prohibition.

The basic assumption is that if the entire public complied, the benefit would be worth the cost. It is worth preventing possible Shabbat desecration at the price of restricting reading. But it is clear that some segment will not comply. Let us denote that fraction by P (a number between 0 and 1 of the whole public). We must now compare the alternatives:

• Enact the decree. The gain is that (1–P) of the public will certainly not stumble into Shabbat desecration (of course, not all of them would have stumbled even otherwise). The harm is that those who do stumble (P) will transgress a rabbinic prohibition of reading and may also stumble into tilting (which is not deliberate but also not coerced or inadvertent action—rather, inadvertent—because there was a rabbinic decree).[2] In addition, the other part (1–P) suffers a restriction on their Shabbat reading and learning, even though this restriction does not yield the full benefit it is supposed to bring (since a segment does not comply).
• Do not enact the decree. The gain is that we impose no restriction on anyone, and no one transgresses the rabbinic prohibition of reading by lamplight. The loss is that some of the (1–P) who would have obeyed the decree may now stumble into a biblical Shabbat labor (tilting), which is a stoning-level offense.

Without doing the math with formulas—since it is hard to quantify the harms themselves—it is quite clear there are two sides here and the decision between them is neither simple nor self-evident. Therefore, as we saw above, if we seek a general, across-the-board criterion, it is very reasonable to set the line where P = 1/2.

This is the discussion regarding the rule “a decree that the public can abide by.” The same calculation can be made regarding whether to rebuke those who do not observe this decree and read by the lamplight (the “better they remain inadvertent sinners” sugya).

The Rule of Majority

I mentioned above that in my remarks I also invoked the rule of following the majority or nullification by majority. Tirgitz regarded this as an arbitrary, unjustified determination. In my view its logic is quite similar to what I have explained so far.

My starting point is the rule of a “thing that has permissible times” (davar sheyesh lo matirin). The Talmud in several places (see Beitzah 3b–4a and elsewhere) says that if there is a prohibition for a limited period—meaning a time arrives when the prohibition lapses—then it is called “a thing that has permissible times,” and various stringencies apply to it. For example, if it is a rabbinic prohibition then its doubt is treated stringently, unlike other rabbinic doubts; and if a biblical prohibition has been mixed into a majority of permitted matter, then by Torah law it is nullified—but where it has permissible times it is not nullified. Most Rishonim explain this by a sevara: “Rather than eating it in a prohibited manner, eat it in a permitted manner”—i.e., there is no reason to permit a problematic case if you can simply wait and eat it in a fully permitted state.

This introduces a novel assumption in the laws of doubt and nullification. We tend to think that when halacha permits, there is no value in stringency. For example, in a rabbinic doubt, or when eating a mixture where the permitted overwhelms, halacha allows leniency—leading many to think there is no prohibition at all, such that the one who is “exempt but does it anyway is called a fool.” But if it were truly a full, unqualified permission, why be stringent even with a thing that has permissible times? You would also now be eating it permissibly—so why wait?[3] How can the Rishonim say that eating it now is “in a prohibited manner,” whereas later it is “in a permitted manner”—when now too it is permitted?

From the law of a thing that has permissible times it would seem that even in cases where halacha permits there remains some degree of prohibition; therefore it may be appropriate to refrain, and the stringent person is praiseworthy. The halachic permission is aimed at those for whom it is difficult or in cases of no alternative. There is no obligation to be stringent in such cases, but it is certainly praiseworthy. This is “deferred” (dechuyah) rather than “wholly permitted” (hutrah). When there is no way to eat permissibly (a thing that has no permissible times), halacha does not obligate stringency, so one may be lenient in a rabbinic doubt or eat a nullified prohibition. But if you have a choice to eat permissibly, there is no license to eat in a prohibited manner. So too we see from the rule that one may not deliberately nullify a prohibition.[4]

But if it is prohibited, why does halacha ever permit it? Even where there is no alternative, “there is no wisdom or counsel against the Lord.” If it is prohibited, it is prohibited. It seems this can be explained as follows: if you have a mixture of prohibition and permission, assuming it were forbidden to eat the whole mixture, we would have to discard the entire mixture and lose the permitted food merely to avoid the prohibition. Again, there is another side to the equation. The side of avoiding the prohibition is clear, but there is also the side of sacrificing permitted property. When one sees this as a two-sided equation, one can understand that halacha has pity on us and says it does not wish to cause us a great loss of permitted food merely to avoid the prohibition.

Again we may ask: where is the line? Does any amount of permitted matter in the mixture allow me to eat all of it? If one permitted piece fell into a pot full of prohibited items, would we allow eating the prohibited items? Again we must decide between two alternatives: eating a prohibition and saving the permitted food, or avoiding the prohibition and losing the permitted food. It is reasonable that halacha set the line at 50%: if the prohibited portion exceeds the permitted, there is no justification to allow prohibition in order to save the permitted. But if the permitted portion is larger, the loss justifies permitting the whole. And again, one could wonder about severe prohibitions and cheap food—perhaps different thresholds would be warranted in different situations. But as noted above, halacha prefers simple, general criteria, and the natural line is 50%. Along with this, one who is more stringent is praiseworthy. If a person sees that the prohibition is severe and the food to be discarded is inexpensive, it is indeed fitting for him to set a different line.

This explanation need not stand alone; it can join other sevarot (e.g., that a mixture does not have the status of “prohibition” and you are not considered to be eating a prohibition, etc.). But if those were the only considerations, I would expect the status to be full permission; in that case, even with a thing that has permissible times we would have permitted. It seems, therefore, that this sevara at least plays a part in the leniencies for rabbinic doubts or nullification by majority.

A similar explanation can be offered with respect to doubts. A rabbinic doubt is not full permission, but halacha has pity on us and does not wish to burden us by prohibiting what is permitted merely to avoid a possible prohibition. Here we are dealing with probabilities rather than quantities, yet the situation is quite similar. For example, in a doubt of muktzeh, if I had to be stringent there is a chance I would be refraining from a necessary use of something permitted. That is an unnecessary burden. It is justified if there is a significant chance of transgression; but if the chance is not significant, there is no justification to prohibit. The line that defines when we allow leniency in a rabbinic doubt is again the midpoint.

So too with following the majority. Following the majority is different from nullification by majority: unlike nullification, it is part of the laws of doubt. In nullification there is no doubt: we know there is prohibition and permission in the mixture and we know the quantities. The ruling is lenient, but there is no uncertainty. By contrast, in following the majority we have a doubt. For example, I find a piece of meat in the street and do not know whether it is kosher. By strict law, if most of the shops in that place are kosher, I may eat it; if not, I may not. That is following the majority. Here there is no nullification but a decision in a case of doubt. In such a case, the explanation will be akin to what I said regarding doubts, not to what I said regarding nullification. We prefer not to prohibit a person from eating the piece if the chance it is prohibited is less than 50%, since there is a chance he will lose kosher food that is rightfully his.

I think the analogy to the sevara I presented earlier—about a decree that most of the public cannot abide by—is now clear. It is the same weighing of cost and benefit between two alternatives; if one wants a general rule, it is sensible to set it at the midpoint. This is what I meant in my reply to Tirgitz when I referenced, in this context, the sevara underlying nullification by majority.

“What is their alternative?”

In conclusion, in my view it is not true that it is equally logical to choose any fraction we wish, as Tirgitz wrote. His error, I believe, is that he ignores the other side of the equation—as many of us tend to do.

Shimon Peres once said of the right-wing opponents of his peace process: “What is their alternative?”—for the right had no alternative that would bring peace. The answer (had he listened) is that indeed there is no alternative—but sometimes one must not choose even the only alternative. Here, too, there was another side to the equation that he tended to ignore: the cost if the peace process fails (failure has a cost; it is not merely a non-success). In such a case—even with no alternative—it is not always right to choose the lone option before us. Peres ignored the other side of the equation and focused on the success of the process and the prospect of peace. He was right as far as that side of the equation goes—but it has another side.

A Closing Note

It still seems clear to me that there is a place for Tirgitz’s sevara: when the benefit or harm is very great, one does not necessarily make a quantitative-majority calculation. There are cases where the Sages will enact a decree even if the majority will not abide by it, or will instruct us to rebuke even when the majority will not heed us, or will forbid nullification by majority despite the absence of “permissible times” (or will permit even when there are “permissible times”).

I have often noted that halachic rules are at best starting points; alongside them one must always consider the specific situation at hand and weigh whether it is right there to deviate from the rules. My claim here, too, concerns only the starting point. Absent a clear consideration warranting departure from the rules, there is strong logic to choose the quantitative majority (51%) as the decision criterion in all these contexts. Where necessary, there is room to deviate from it—both for the individual and for the Sages weighing the principled halachic consideration.

[1] One should indeed distinguish between the analogy and the case at hand, for anyone who observes the decree certainly benefits (by avoiding a prohibition); therefore, imposing the restriction on him is justified. And one who does not observe it is not restricted by it; thus for him there is no “cost” to the decree (only that he becomes a transgressor). Still, it seems one must consider whether in such situations the perspective is communal or individual.

[2] This was the main discussion with Tirgitz and Avishai there in the comments. The Talmud and the Rashba in Shevuot 18 indicate that if a person violates a rabbinic decree and thereby comes to a biblical prohibition that the decree was intended to prevent, he is not an anus (coerced) but an inadvertent sinner (Avishai argued he is deliberate). See also Kli Chemdah at the end of Parashat Balak (sec. 4), citing R. Yehoshua’le of Kutna (author of Yeshuot Malcho).

[3] Later authorities discuss whether nullification by majority is a full permission (hutrah) or “deferred” (dechuyah), and some indeed take the former view. If it is “deferred,” my explanation follows naturally. For those who hold it is full permission, one might, with difficulty, attribute the rule of a thing with permissible times to moral education—to distance a person from prohibitions.

Note that the “hutrah vs. dechuyah” discussion does not necessarily track other questions discussed by later authorities regarding nullification by majority. Two examples: (a) Does the minority portion actually “turn into” permission like the majority or not? One could say it “turns into” permission in the sense that it is permitted to eat, but not that its essence becomes permitted. Even so, the stringent are praiseworthy. In the opposite direction, the linkage between the positions seems stronger. (b) Is the permission as a matter of certainty or of doubt? Again the linkage is not necessary, because the permission could be as a matter of certainty and still be “deferred” rather than full permission.

[4] Though here one could counter that a mixture created by our own hands is not deemed a “mixture,” and the permission never applied there at all. Some Rishonim hold this is a biblical rule. See, for example, here.