1. The Equal-Weight View and Reasonable Doubt Instructions

First, as discussed above, whether a disagreement should lead to a vote of not guilty turns on whether the disagreement is among epistemic peers. One question, then, is whether our jury system has the resources to instruct jurors to adjust their votes by considering whether the disagreement that they are facing, if any, is among epistemic peers. It seems to me that we may already be telling jurors to try to decide whether they are facing epistemic peers or others when there is disagreement.

Consider the following, which are some standard reasonable doubt instructions:

• If . . . you think there is a real possibility that he is not guilty, you must give him the benefit of the doubt and find him not guilty.Footnote ⁷⁷

• Proof beyond a reasonable doubt is proof that precludes every reasonable hypothesis except guilt and is inconsistent with any other reasonable conclusion.Footnote ⁷⁸

• [I]t is a doubt for which a reason can be assigned.Footnote ⁷⁹

• It is a doubt for which you can . . . conscientiously give a reason.Footnote ⁸⁰

• It is a doubt that a reasonable person hearing the same evidence would have.Footnote ⁸¹

• [I]f the jury views the evidence in the case as reasonably permitting either of two conclusions, one of innocence, the other of guilt, the jury should . . . adopt the conclusion of innocence.Footnote ⁸²

These are not phrased in terms of epistemic peerhood, but these instructions could be interpreted as a way of telling the jurors how to assess whether the disagreeing juror's viewpoint is “reasonable,” in which case one ought to treat it as a disagreement among epistemic peers.

How would these instructions then apply to a case the first Etan Patz trial? In the first Etan Patz trial, eleven jurors wanted to convict, and one person wanted to acquit. Now imagine that the eleven jurors consulted the definition of “proof beyond a reasonable doubt” to decide whether this is the kind of disagreement that ought to produce an acquittal. One definition says that proof beyond a reasonable doubt is “proof that precludes every reasonable hypothesis except guilt.” Another instruction tells the juror that if she “views the evidence in the case as reasonably permitting either of two conclusions, one of innocence, the other of guilt,” then she must vote to acquit. Yet another says that reasonable doubt is “a doubt that a reasonable person hearing the same evidence would have.” Each one of these definitions may be thought of as a way of telling the jurors to determine whether the disagreement they are facing is one among epistemic peers, and phrases like “reasonable hypothesis,” “reasonably permitting,” and “reasonable person” are doing that work. If they think the dissenting juror is an epistemic peer, they ought to acquit; otherwise, they need not acquit.

2. The Equal-Weight View and Postconviction Review

It is also worth discussing whether the analysis presented here can illuminate the way in which convictions are reviewed for insufficiency of evidence. The Supreme Court, in Jackson v. Virginia, addressed the question of whether the constitutional requirement of proof beyond a reasonable doubt had constitutional implications for the standard of review for guilty verdicts and held that the challenge to a guilty verdict on the grounds of the insufficiency of the evidence is successful, “if it is found that upon the record evidence adduced at the trial no rational trier of fact could have found proof of guilt beyond a reasonable doubt.”Footnote ⁸³ This means that even if the reviewing judge disagrees with the jury and thinks that that there is reasonable doubt as to the defendant's guilt, the judge cannot set aside the jury's verdict on the basis of that alone. The judge also must find that no rational jury would have convicted.

The Jackson standard is a highly deferential standard, and one may see it as inconsistent with the view presented here. To see this, consider Table 9.

TABLE 9Footnote ⁸⁴

To simplify, let us assume a single reviewing judge and a unanimous verdict. If a defendant has been found guilty, there are twelve people who assess the probability of guilt to be at 0.9. If the reviewing judge is sympathetic to the jury's verdict yet is unconvinced beyond a reasonable doubt of the defendant's guilt and puts the probability of guilt at, say, 0.8, then, as “the thirteenth juror,” the judge's misgivings bring the average probability of guilt down to 0.89, below the 0.9 threshold for beyond a reasonable doubt, which would then call for setting aside the verdict and even acquitting the defendant.

Does the equal-weight view, then, imply that Jackson was wrongly decided? That conclusion would be too quick. The question of the appropriate standard of review involves several considerations, one of which is the assumption by our legal system that “it is the responsibility of the jury—not the [reviewing] court—to decide what conclusions should be drawn from evidence admitted at trial.”Footnote ⁸⁵ The original Jackson Court showed its sensitivity to the issue in stating that “the standard announced today does not permit a court to make its own subjective determination of guilt or innocence,”Footnote ⁸⁶ that under the standard, “the factfinder's role as weigher of the evidence is preserved,”Footnote ⁸⁷ and that the “criterion . . . impinges upon ‘jury’ discretion only to the extent necessary to guarantee the fundamental protection of due process of law.”Footnote ⁸⁸ The Court more recently observed that “[b]ecause rational people can sometimes disagree, the inevitable consequence of this settled law is that judges will sometimes encounter convictions that they believe to be mistaken, but that they must nonetheless uphold.”Footnote ⁸⁹

The Jackson Court, of course, had a few other options as it considered the proper level of deference to the initial verdict. The first option is the approach whereby the reviewing court would “ask itself whether it believes that the evidence at the trial established guilt beyond a reasonable doubt.”Footnote ⁹⁰ The second approach is the approach considered here, which is just like the first approach but with the equal-weight view applied. (In other words, the first approach is just like the second approach but applies the steadfastness view.) The third approach is the deferential standard articulated in Jackson, under which the conviction is upheld if the reviewing court finds that a rational jury could have found in favor of a guilty verdict. The fourth approach is an approach that the Jackson Court considered but rejected, according to which, “a conviction based upon a record wholly devoid of any relevant evidence of a crucial element of the offense charged is constitutionally infirm.”Footnote ⁹¹

So these are at least four approaches (and these approaches can be made even more complex) at varying levels of deference to the jury, but none of them are required by the equal-weight view. We are interested, of course, in the question of guilt and innocence, and the equal-weight view is relevant to that interest, but there may be other structural considerations, such as institutional reasons to shield the jury's considered judgments from second-guessing, given the important role of the criminal jury in the system, that are also relevant in deciding which level of deference is appropriate.Footnote ⁹² Though a complete analysis of the proper standard of review in this context is beyond the scope of this article, we can at least give an account of the Jackson standard in terms of epistemic peers.

As we see in Table 9, if the equal-weight view applies and the reviewing court disagrees with the guilty verdict, it appears, assuming that the jurors are the judge's epistemic peers, that the overall average probability of guilt has to be adjusted. The Jackson rule can be interpreted as a directive to the reviewing courts to first ask if the jurors are epistemic peers and not to disturb the verdict if the answer to that question is yes. However, considering all the evidence in the record, if the jury's decision is so far off base that they qualify as epistemic others, perhaps akin to a group of Apparently Mistaken Jurors, then the reviewing court may set aside the jury's verdict at that point. Perhaps this level of deference—where decisions made by epistemic peers are respected even if one believes them to be wrong—is appropriate in order to maintain the multitiered system of courts and reviews, or alternatively, perhaps it is underprotective of the due-process requirement of proof beyond a reasonable doubt. The equal-weight view does not suggest an answer one way or another.

1. The Equal-Weight View and Moral Disagreements

One feature of the criminal jury's decision-making process that the discussion so far has obscured is that crime definitions typically contain different types of elements. There are several types of questions that fact finders are tasked with in criminal trials, such as “What crime was committed?” “Did the defendant commit the crime?” “What did the defendant do?” And “Did what the defendant do constitute a crime?” These questions overlap, but the reason they are phrased slightly differently is because sometimes it is clear that a crime was committed (a person was killed), and the question is whether the defendant can be linked to the crime (Was O.J. Simpson the killer?), whereas at other times there is little dispute about whether the government has identified the correct defendant, and the important questions that remain are exactly what the defendant did (Did Martha Stewart tamper with potential evidence?) and whether what the defendant did constituted a crime and which crime (Did Dharun Ravi invade his roommate's privacy, and if so, was there bias intimidation?)

Deciding these cases requires reconstructing historical facts. Far more than that, however, is often involved in determining whether a person's conduct meets the definition of a crime. Take one of the definitions of “murder in the second degree” in New York law, for instance: “A person is guilty of murder in the second degree when under circumstances evincing a depraved indifference to human life, he recklessly engages in conduct which creates a grave risk of death to another person, and thereby causes the death of another person.”Footnote ⁹⁷ Under this definition, the prosecution has to prove that (1) “under circumstances evincing a depraved indifference to human life,” (2) an individual “recklessly engage[d] in conduct,” (3) “which create[d] a grave risk of death to another person,” and (4) thereby “cause[d] the death of another person.” In order to decide whether these elements have been proven, jurors must determine what happened, such as what the defendant exactly did and how it led to a person's death. But the jurors’ determinations of “recklessly,” “grave risk,” “depraved indifference to human life,” and “cause” involve more than simply determining what bodies occupied what spaces at what time, what kinds of movements such bodies engaged in, and what differences such movements made in the world. Jurors also have to determine whether terms like “recklessly,” “grave risk,” “depraved indifference,” and “cause” are appropriate descriptors of relevant features of the world, and such judgments involve making moral evaluations.Footnote ⁹⁸

Therefore, when applying crime definitions to particular cases, different kinds of determinations must be made about historical facts—or questions about what happened—and about norms—or about the evaluative significance of what happened. And the definition of second-degree murder in New York is not an outlier. Moral or normative terms such as “reckless,” “unjustifiable,” “depraved,” “grave,” “cruel,” “wanton,” “heinous,” “debased,” “perversion,” and “impair or debauch the morals” are routine in criminal law.Footnote ⁹⁹ Moral or normative elements generally have the following characteristics. Like the terms “unjustifiable,” “reasonable,” “depraved,” and “cruel,” they are vague and evaluative. Such terms are rarely defined precisely and instead invite the fact finder to make moral judgments in the process of determining whether the terms apply to the conduct in question. And they are typically applied to individual cases by trial judges and juries without detailed guidance from appellate decisions and legislation.Footnote ¹⁰⁰

Such moral or normative elements constitute a subset of the category of what is commonly called “mixed questions of fact and law” by lawyers.Footnote ¹⁰¹ Some moral philosophers, following Bernard Williams, refer to such terms as “thick concepts.”Footnote ¹⁰² Williams's own examples are terms such as “coward,” “lie,” brutality,” and “gratitude.”Footnote ¹⁰³ As Williams explains, thick concepts are “action-guiding” in the sense that they are “characteristically related to reasons for action,” as “[i]f a concept of this kind applies, this often provides someone with a reason for action.”Footnote ¹⁰⁴ They are, however, called “thick” concepts because they are “thick” with descriptive attributes, which make the concepts “world-guided” as well as “action-guiding.”Footnote ¹⁰⁵ Others have explained “thick concepts” in similar terms. Jonathan Dancy says that “thick concepts somehow hold together a property and an attitude . . . or, as it is also sometimes put, description and evaluation.”Footnote ¹⁰⁶ Allan Gibbard says “[a] term stands for a thick concept if it praises or condemns an action as having a certain property.”Footnote ¹⁰⁷

Williams uses the term “thick concepts” in order to contrast them with “thin” ethical concepts such as “good,” “right,” and “ought,” which are “thin” in the sense of having little descriptive content.Footnote ¹⁰⁸ The distinction between “thick” and “thin” concepts is controversial, especially because Williams has used the distinction to criticize contemporary ethical theorizing for relying mostly on thin concepts and neglecting thick concepts. Because so much of theoretical import seems to turn on the argument that many important ethical concepts for contemporary ethical theorists are “thin” as opposed to “thick” and hence indicative of defects in contemporary ethical theorizing, Williams's account invites the objection that the distinction is not always easy to draw.Footnote ¹⁰⁹ For instance, Samuel Scheffler, who mentions “privacy,” interests,” and “consent” as examples of concepts of some importance to contemporary ethical theories, observes that “every single one of the[se] concepts . . . appears to be determined to at least some extent by what the world is like,” and argues that “it is impossible confidently to classify various of the concepts on the list as either thick or thin.”Footnote ¹¹⁰ T.M. Scanlon, too, has pointed out that “ethical theories typically deal with a variety of moral concepts including well-being, consent, and specific rights, such as rights to privacy, all of which seem to have considerable empirical content,” and concludes that “the contrast in empirical content between thick and think concepts is at most one of degree.”Footnote ¹¹¹

For our purposes, this particular philosophical controversy is of little moment, because the distinction we are interested in is the distinction between factual and moral terms, whereas the debate over the distinction between thin and thick concepts presupposes that both types of concepts have some descriptive and prescriptive aspects, with the main disagreement being over how much description is necessary for a concept to go from thin to thick. What is relevant for our purposes is the similarity between such moral terms, whether they are properly classified as “thin” or “thick,” and moral terms in crime definitions. Because of the ways in which the factual and moral aspects intertwine in these terms, disagreements about whether a concept applies to a person or to a person's conduct tend to assume a particular shape.

For largely factual questions like “Did the man stab his ex-wife to death?,” there can be a disagreement among fact finders, and the nature of this disagreement would generally be about what factual inferences to draw from available evidence. A leather glove was found at the scene of the crime and it was the same type of glove worn by the defendant. Does the presence of the glove point toward guilt? What if the glove does not fit the defendant? Perhaps the glove has fit in the past but no longer fits because of a medical condition. Or maybe it is because the glove shrank after having absorbed blood. Or, even if the glove actually belonged to the defendant, it was planted at the scene of the crime by a cop in order to frame the defendant. If there is a disagreement about the significance of the glove, it would have to do with the ways in which the glove is or is not related to the crime. In most criminal trials, we are trying to piece together what happened after the fact, given the available evidence. The role of the fact finders is to use various types of evidence (witness testimony, documents, and physical evidence) to draw correct inferences and develop answers to these questions, and the field of evidence law is full of rules about what types of evidence may be considered and what inferences may be drawn from them.

Disagreements about moral elements, however, do not always look this way. Say jurors are having a disagreement about whether a particular form of corporal punishment was “cruel.” Let us assume that two jurors agree on exactly what happened in terms of how many blows were delivered, how much pain was likely felt, whether the blows were delivered intentionally or accidentally, and so on. Yet there can remain some disagreement between the two as to whether there was cruelty in the defendant's behavior. Disagreements about factual matters, then, may be resolved by more fact-finding. Disagreements about moral matters, however, can persist, as moral disputes—as to whether something is cruel or not—cannot always be settled through more fact-finding. It is true that many moral disagreements are actually factual disputes, and many moral disagreements disappear once people have the facts straight, but disagreements as to whether some act is cruel or not can remain even after all the facts have been put on the table.

In this sense, moral elements resemble thick concepts. Like moral elements in crime definitions, when two people disagree on whether a moral concept applies, the disagreement can be purely evaluative, in the sense that one's attitude toward the thing can determine whether the concept in question applies to the thing.Footnote ¹¹² Or, as others have put it, “with the thick, evaluation partly determines description,”Footnote ¹¹³ because “[i]n order to trace the contours of the ethical concept's applicability we have to understand its evaluative point,”Footnote ¹¹⁴ and this is because “[t]he uses of words or terms expressing thick concepts . . . involve a commitment to an underlying normative or evaluative judgment.”Footnote ¹¹⁵

It is true that it is not always easy to draw a distinction between moral and factual terms, but the fact that there are line-drawing problems does not by itself erase the distinction between the question of whether a person was occupying a particular space at a particular moment and the question of whether the person engaged in a behavior appropriately described as “cruel.” Given that there is such a distinction, and given that moral terms, as described so far, are common in crime definitions, the question of how the presence of moral terms raises distinctive issues for law can be addressed—at least for now—without being bogged down by the question of exactly how to draw the distinction between elements that are moral and those that are not. So the relevant question for our purposes now is how the equal-weight view handles disagreements that are morally inflected.

To explore this issue, let us focus on the idea of “consent,” which as we saw above, is often described as a moral term in discussions of thick concepts.Footnote ¹¹⁶ More specifically, consider the following facts from Commonwealth v. Berkowitz, a controversial 1994 case from Pennsylvania.Footnote ¹¹⁷ Berkowitz, the defendant in the case, and the complainant were college students and had mutual friends and acquaintances. Before the day of the incident at issue in this case, Berkowitz and the complainant talked on a few occasions, once shortly after the complainant attended a school seminar on the question, “Does ‘no’ sometimes mean ‘yes’?,” and twice afterward when she stopped by Berkowitz's room while intoxicated, once lying down on his bed. During that first conversation after the seminar, the complainant asked Berkowitz how large his penis was. One day, the complainant went to visit a friend in his dormitory room and found not her friend but only Berkowitz, who was the friend's roommate. Berkowitz asked her to stay, and she stayed and talked for a while with him. The two parties professed different understandings of what happened next. It is uncontroverted that Berkowitz initiated physical contact and that the complainant said “no” multiple times to him throughout sexual acts that culminated in intercourse. The complainant also stated that Berkowitz “put” her on the bed with something that was “kind of like a push” and that he then got on top of her, which meant that she “couldn't . . . go anywhere.” Berkowitz did not deny that she said “no” repeatedly but said that she “warmly responded to his advances by passionately returning his kisses” and was “whispering . . . no's” while moaning “amorously . . . passionately.”Footnote ¹¹⁸

At issue in the case was whether there was sufficient evidence to convict the defendant in the case of the crime of rape. So let us now imagine a jury considering the rape charge. The jury in the original case was working with a particular definition of rape that was in force at the time, but here are two possible definitions:

Using these definitions, people may disagree about whether someone like Berkowitz should be convicted of the crime of rape.

Fortunately, we do not have to work with such guesses, as there is an actual study on this case done by Dan Kahan.Footnote ¹¹⁹ Kahan's study took a national sample of 1,500 American adults who were representative of the general population in terms of race, gender, income, and education level. The study presented summaries of testimonies from the two individuals in the Berkowitz case and asked the participants to answer various questions concerning the incident described by the testimonies.Footnote ¹²⁰

One of the questions the study asked was whether the participants agreed with the statement that the man in the case “should be found guilty of rape.”Footnote ¹²¹ Different groups were asked this question with different definitions of rape, the two I presented above and two additional definitions.Footnote ¹²² Across all the groups (that is, aggregating responses from groups with different definitions of rape), 57 percent agreed with the statement and 43 percent disagreed with the statement.Footnote ¹²³ Depending on the definition of rape, the percentage of those who answered that the man should be convicted of rape ranged between 53 percent and 65 percent of the respondents.Footnote ¹²⁴ And in the aggregate again, 24 percent stated that they strongly agreed with the statement, and 15 percent stated that they strongly disagreed with the statement.Footnote ¹²⁵ So it appears that we see a robust disagreement within the population on this issue, and that there is a significant number of people who hold strong views.

This is an example of a case where the disagreements at issue do not seem to be simple factual disagreements. The facts may be somewhat murky, and additional details and background information may be useful, but there are also normative disagreements about what constitutes rape and, when applying the legal definitions, what constitutes “force or threat of force,” “without consent,” “can reasonably be expected to know the woman does not consent,” or “words or overt actions indicating a freely given agreement to have sexual intercourse.” All of these questions are like thick concepts discussed above in that one's evaluative perspective can determine the applicability of such terms.Footnote ¹²⁶ The question for our purposes is what ought to happen in such situations of disagreement according to the equal-weight view.

Take a look again at Table 4. This table assumes that there is an 11-to-1 vote on the question of guilt, since that is the strongest possible scenario to convict while allowing for a strong disagreement. Under the equal-weight view, when the jurors face the sort of disagreement over whether there someone like Berkowitz should be convicted of rape and the disagreement is among epistemic peers, those who feel strongly that he ought to be convicted need to lower the confidence in their belief and those who feel the opposite need to do the same with regard to their beliefs. The Epistemic Peers row of the table indicates that the jurors ought to vote to acquit, since their confidence in their belief in the defendant's guilt would not be high enough.

But the other possibility of course is that the disagreement is not among epistemic peers. How so? Here are some possibilities shown above.

If the jurors who think Berkowitz is a rapist believe that those who disagree with them are obviously insane, then they can ignore the one person who would acquit him. It seems to me, though, as a general matter, that those who would acquit Berkowitz cannot be dismissed so easily, given that among the 1,500 Kahan's study surveyed, 42.5 percent thought he was not guilty. Those who would acquit him might still be mistaken, but their belief is not as obviously wrong as the belief “that one is the messiah”Footnote ¹²⁷ or the prediction that “it will rain eggplants tomorrow.”Footnote ¹²⁸

If the Juror from Outer Space is not an option, what about this one?:

Is this a possibility? It is not clear. The example we consider above for this possibility is Christensen's example of splitting a check at a restaurant, where your friend comes up with a number for each person's share that is higher than the bill itself. This would be the kind of mistake that Elga might call “utterly insane” or “obviously out of the ballpark.”Footnote ¹²⁹ One could imagine a person thinking as follows: “Berkowitz is clearly a rapist. The complainant said ‘no’ repeatedly, and he ignored it. ‘No’ means ‘no.’ Anyone who thinks that Berkowitz did not rape is clearly in error in thinking that a person who says ‘no’ is consenting.” Following this logic, then, the eleven jurors may dismiss the one dissenting juror in the case as simply having gotten it wrong.

One problem with this answer is that it seems to require dismissing a large chunk of the population as holding an “insane” view. This may not be a problem as a general matter, but it may be a problem for the equal-weight view for the reason already discussed above. That is, if a person can ignore the opinion of anyone whose view he or she finds “utterly insane,”Footnote ¹³⁰ then it is unclear why one would want to subscribe to the equal-weight view in the first place, given its initial appeal as a philosophy that eschews “unfounded confidence” and offers “a valuable strategy for coming with our known infirmities.”Footnote ¹³¹ The strategy of ignoring those one disagrees with because they are “utterly insane” then seems to reintroduce the problem of “unfounded confidence.” As Jennifer Lackey asks, “For isn't it the case that the most impassioned sexists typically find arguments on behalf of women's rights laughable, the most devout racists find the view of proponents of equality crazy and the most committed homophobes find the idea of granting homosexuals even the most basic indignities perverse”?Footnote ¹³²

There is another option for proponents of the equal-weight view. Elga argues that in many controversial cases (say, implicating issues like the ones Lackey mentions), “the disputed issues are tangled in clusters of controversy.”Footnote ¹³³ And those with whom you disagree on one controversial issue tend to be the ones who also have different views from yours on closely related issues. What would such closely related issues be in the case of rape? Maybe the juror who thinks Berkowitz ought to be acquitted also thinks that women often falsely accuse men of rape because of shame or guilt or regret over sex acts that they consented to; that women often say “no” when they mean “yes” because they do not want to appear “easy” and want to be “talked into it”; that men are expected by the society to ignore what women say and be aggressive and forge ahead in order to be respected; and so on. The equal-weight view, as mentioned above, is motivated by the idea that in many cases of disagreement, “the explanation in terms of my friend's mistake is no more reasonable than the explanation in terms of my mistake.”Footnote ¹³⁴ So when considering a controversial rape case, if your fellow juror likely is someone who would come out the wrong way on a cluster of related issues, he is not an epistemic peer on this issue because he is more likely than you are to be wrong on this issue as well.Footnote ¹³⁵

The idea that people's views on particular rape cases often go hand in hand with their attitudes toward closely related issues is supported in Kahan's study, where he finds that those who subscribe to “a hierarchical world view, as opposed to an egalitarian one,” were more likely to find the defendant to be not guilty, and vice versa,Footnote ¹³⁶ and that “the individuals most inclined to form pro-defendant perceptions are hierarchical women.”Footnote ¹³⁷ By “hierarchical” or “egalitarian,” Kahan means “how favorably or unfavorably disposed individuals are toward a social order that features differentiation and stratification of social roles based on observable and largely fixed characteristics (including race, gender, sexual orientation, and class).”Footnote ¹³⁸ And Kahan elaborates that those who have a “hierarchical” outlook tend to believe that “[m]en . . . are entitled to exercise authority over women, whose status depends on successfully discharging domestic responsibilities,” and tend to view with “suspicion and contempt” “a woman who engages in sex outside of marriage.”Footnote ¹³⁹ So the fellow juror who wants to acquit Berkowitz may make statements along these lines during deliberation. Or, even if you do not know the juror's attitudinal orientation, you may be able to guess that he or she is someone with whom you do not see eye to eye on most related issues. Either scenario may give a juror a reason not to treat the dissenter as an epistemic peer.

As a number of philosophers point out, it is not clear whether this response works.Footnote ¹⁴⁰ First, is it not the case that we have here just another case of peer disagreement where the disagreement is not just about the issue of whether a crime of rape occurred but also about a cluster of related issues? If so, perhaps the equal-weight view would counsel a person holding a cluster of beliefs about related issues to reduce his confidence in his entire outlook. In Kahan's terms, perhaps one ought to question one's own hierarchical or egalitarian outlook instead of dismissing those whose general outlook differs from one's own as epistemic others. If this objection goes through, then we are back to the conclusion that the jurors in the case we are discussing ought to vote to acquit in the end.

Second, dismissing those with a different outlook from yours as epistemic others seems too quick. It may be the case that jurors who disagree on the question of whether someone like Berkowitz ought to be convicted of rape still share enough in common to qualify as epistemic peers. As Sarah McGrath points out in a similar context: “We would expect them to agree, for example, that slavery is morally abhorrent, that is wrong to cause others pain for the sake of one's own amusement, that lying is prima facie wrong, and about countless other issues.”Footnote ¹⁴¹

Third, on persistently controversial issues, there are often reasonable positions on both sides, and intelligent and well-informed people can often be found on both sides.Footnote ¹⁴² I present the case for acquittal of Berkowitz in a particularly unappealing way above in order to motivate the idea that only epistemic others would hold a belief like that. But of course we can present the contrarian view more positively in terms of the importance of refraining from punishing people for a crime as serious as rape when there may be an honest and reasonable misunderstanding as to the existence of consent. In the Kahan study itself, “40% of the subjects . . . indicated that they agreed that ‘[d]espite what she said or might have felt after, [the woman in the case] really did consent to sexual intercourse.”Footnote ¹⁴³ Even if one is inclined to find the purported misunderstanding on the part of Berkowitz to be neither credible nor reasonable, one would have to be in denial to fail to recognize what Stephen Schulhofer has called an “inconvenient truth”:

Kimberly Ferzan has similarly observed:

Given the possibility of such misunderstandings, one might reasonably believe that interpreting an ambiguous signal, in good faith, as an implicit consent may not rise to the level of culpability that warrants punishment and blame attached to a rape conviction. So the strategy of dismissing those one disagrees with on such controversial issues as epistemic others seems, again, to be in considerable tension with the motivations behind the equal-weight view.

Therefore, it appears that criminal cases that involve making decisions about controversial moral questions would frequently have to end in acquittals if we were to follow the equal-weight view. The only way to avoid this implication is for jurors to dismiss those who disagree with them on moral questions as epistemic others, which may be warranted in some cases, but not all. For some, then, the fact that the equal-weight view suggests that someone like Berkowitz may have to be acquitted may be sufficient to reject it as a jury deliberation ideal. To come to the conclusion, though, we would need to establish as a normative matter that someone like Berkowitz ought to be convicted, and that is not an obviously correct position to take. The following section, though, raises more concerns about the equal-weight view that do not rely on taking this position.

2. The Equal-Weight View for Factual Questions but Not for Moral Questions

To address the question of whether the equal-weight view is an attractive ideal for jury deliberation, we need to separate factual and moral elements of crime definitions. Should the equal-weight view apply to both types of issues? The answer is no. The equal-weight view should guide the jurors in determining factual issues but not moral issues.

As discussed above, there are several types of questions that jurors are tasked with in criminal trials, such as “What crime was committed?” “Did the defendant commit the crime?” “What did the defendant do?” And “Did what the defendant do constitute a crime?” Many of these are factual questions as to what exactly happened. And for factual questions like “Did the man stab his ex-wife to death?” there can be disagreements among jurors, and the equal-weight view implies that in such situations of disagreement it would be rational for jurors to revise their beliefs and become less confident about their conclusions, and, given the high standard of proof that the beyond a reasonable doubt requirement demands, a rational juror would vote to acquit, assuming that the disagreeing person is an epistemic peer. As noted above, this conclusion seems correct given the purposes of the beyond a reasonable doubt requirement, but this impression remains only as long as we are discussing factual elements and disagreements about facts.

Things look different when we are dealing with moral elements. The jurors who revise their beliefs and end up voting differently because of disagreement about the kinds of factual matters that criminal jurors generally deal with can say, “I believe x but I could be wrong about it,” and go on with the rest of their lives. This is not to say that all factual issues are like that. As discussed below, on controversial public-policy issues that are endlessly debated, such as climate change, childhood vaccinations, and gun control, relevant empirical evidence may be interpreted differently by different people depending on the various moral, cultural, and ideological commitments they have. Factual decisions that jurors make are generally not like that, because the evidence jurors consider tends to have implications only for the cases they themselves are deciding. Sometimes factual revelations can turn a person's world upside down, and the world of literature is full of such moments. But jurors do not decide such questions, either. As a general matter, jurors considering criminal cases do not have a personal stake in, say, the question of the factual significance of the glove found at the scene of the crime. Moral disagreements—at least the types jurors generally have—are not like that.

Say jurors are having a disagreement about whether a particular form of corporal punishment was “cruel” or a sexual encounter was “consensual.” If the jurors are told at the end that they are not to make the decision on the basis of moral beliefs they in fact subscribe to, the cost of such a decision is high. Jurors may feel confident in their moral judgments, but the equal-weight view seems to imply that it would be irrational for them to act according to such judgments if they encounter disagreement. Why is this a problem?

In order to see why this is a problem, we need to reexamine a core justification for the jury system. The case for delegating particularized moral judgments to juries has been made by many scholars. Juries in criminal cases are tasked not only with fact-finding but also with “the power to elaborate the governing norms underlying criminal laws from the perspective of the community and its sense of moral blameworthiness.”Footnote ¹⁴⁶ It is thus often argued that the jury serves as the community's moral conscience in bringing the community's ordinary moral sentiments to bear on specific cases.Footnote ¹⁴⁷

With that institutional setting as the backdrop, now consider what we do when jurors make moral judgments. First, as a general matter, when making moral judgments, we think and talk about fundamental, weighty issues. People take morality seriously. As T.M. Scanlon puts it, when we make moral judgments, we refer to standards and, “each of us regards the standards to which we refer as authoritative and binding—as standards which we have good reason to regard as outweighing all others, and whose violation is proper grounds for feelings of guilt and shame.”Footnote ¹⁴⁸ Therefore, when jurors make moral judgments, they are considering matters of some importance, as they are committing themselves to a set of implications about how they think people are required to behave and different ways in which people are prohibited from behaving.

Now this may appear to be an overstatement, and it may indeed be. Whether one should leave a tip for a person who shines one's shoes, for instance, may implicate moral questions, but it may also be relatively costless for people to go with one answer or another. And this is not the right place to defend sweeping generalizations about the fact and value distinction. For the purposes of this article, though, it is sufficient to note that that the kinds of moral questions criminal jurors consider are important and to be taken seriously.

Second, as alluded to above, a person's moral judgment on a particular question does not stand isolated from her beliefs on other issues. Many controversial cases, as mentioned above, “are tangled in clusters of controversy,”Footnote ¹⁴⁹ in Elga's phrasing, and how people evaluate such issues corresponds to different, broader worldviews that people generally subscribe to, as Kahan documents.Footnote ¹⁵⁰ Whether a given criminal defendant was at the crime scene as accused of by the prosecutor is an issue on which a person can change her mind with minimal damage to everything else she believes in. Revising her views on a particular moral issue, by contrast, often requires rethinking other, related things that are embedded in her web of beliefs, attitudes, and emotions. As Philip Pettit puts it, “[y]ou can come to think that the car went through on the green without revising any of your other beliefs . . . [b]ut you cannot come to think that [, say,] . . . abortion is not grievously wrong, without a range of adjustments in other matters of belief.”Footnote ¹⁵¹ In other words, unlike factual questions that arise in criminal cases, which tend to be about that particular case and do not generally have any deeper implications, how a moral question is decided in a criminal case has the potential to reverberate through one's life because of the way the question is embedded in one's web of belief.

Third, when we make moral judgments, we engage those parts of ourselves that are sensitive to aspects of the world that provoke moral emotions and thinking, and in the case of jurors, they are dealing with real, not hypothetical, situations. An important aspect of what it means to be a person with a good moral character is to be sufficiently sensitive and responsive to features of the world that are morally relevant.Footnote ¹⁵² A person who witnesses an act of cruelty or unkindness or injustice but appears indifferent or even delighted or amused is morally defective. So, when the legal system asks jurors to make a moral judgment, we are putting them to a test in a sense and asking them to bare their souls or make themselves vulnerable to others’ moral evaluative stance. They are in a vulnerable position because, depending on how they come out, there is a risk that they would be judged by others to be morally defective. What sorts of factual conclusions a person arrives at can be revealing about a person's general perspective on the world as well, but there is much less risk in the factual context because one can occlude one's moral views by burying them in a jumble of factual observations.

In short, the types of moral questions jurors address tend to be important, embedded, and reflective of the person holding them, and these three features give gravitas to the task of making moral judgments in the context of a criminal trial. Given the seriousness of the task, the legal system should treat each juror's moral judgment with appropriate respect. The equal-weight view counsels that the jurors revise their views and “split the difference” in such situations when facing disagreement, and a directive that tells jurors to vote on the basis of adjusted views as if they are their own fails to treat each view with such respect. Instead, it sends a confusing, if not offensive, message that the views that these jurors hold on such weighty matters are somehow optional and can be ignored at will, as they may be asked to vote against them, which can have a highly disorientating and deflationary impact on the jurors and in turn undermine the basic justification for the jury as an articulator and enforcer of morality.

Of course, it should go without saying that jurors ought to conduct their moral deliberations carefully and responsibly. We want them to slow down, to consider the other side, to deliberate carefully, and to reach thoughtful conclusions, as opposed to make decisions on the basis of, say, blanket prejudice against minority groups. And there is nothing wrong with the practice of reflecting on one's own moral beliefs and questioning them when facing disagreements.Footnote ¹⁵³ The point is that requiring that the jurors follow the equal-weight view is far more demanding than that.

It is also true that the justification for the criminal jury is often presented in terms of the jury as a stand-in for “the community”—that what matters ultimately is what the jury as a group concludes and not what any particular individual juror believes.Footnote ¹⁵⁴ Given the group setting of the jury, would these institutional costs outlined in fact materialize when jurors clearly understand that they are there as a group and that their individual voices have to be combined and translated as a group decision in the end? Why can the equal-weight view not simply be such a procedure to aggregate different individual votes to produce a group decision without undermining the logic of the jury system?

The answer is that instructing to a juror to follow the equal-weight view is equivalent to instructing her to revise her view and vote on the basis of that new, revised view when facing disagreements as if the new view is her own view. The kinds of problems I outline here come about only when each juror is told to vote differently from his or her conscience due to disagreement. There is no problem like that if a juror votes differently from the others and the ultimate outcome is a hung jury, even though the group decision is different from the individual decision, because each juror votes according to his or her belief. In sum, the equal-weight view may imply a correct course of action when handling factual disagreements but it fares poorly when it comes to handling moral disagreements.

3. Perceptions of Facts and Moral Commitments

Some may object that moral questions are not the only questions that are important, embedded, and reflective in the ways that I described, and that factual judgments share the same features in that, roughly speaking, people may be as committed to seeing the facts in a particular way in the same way they are committed to seeing moral dimensions of the world in a particular way. On its face, this objection can seem silly. What possible commitments might I have that could lead me to want to believe that the desk I am sitting at is about a foot away from me as opposed to two feet away or that the light in my office is on as opposed to off? Recently, though, there have been more sophisticated articulations of the view that one's factual perceptions can be heavily influenced by values one holds, and the objection that my concerns about moral questions cut broadly and impact factual, as well as moral, judgments may rise from such views. I consider two such views, which we may call the “cultural cognition” view and the “epistemic injustice” view. The “cultural cognition” view is the one espoused by Dan Kahan, whose study is discussed above, and his sometime co-authors. “Epistemic injustice,” on the other hand, is taken from the title of a recent book by Miranda Fricker.Footnote ¹⁵⁵

In the same article about the Berkowitz case described above, Kahan describes his general view about the relationship between factual perceptions and moral commitments this way:

Kahan calls this phenomenon “identity-protective cognition.”Footnote ¹⁵⁷

If Kahan is right about the ways in which people tend to form factual beliefs that are consistent with their values, then whatever arguments I am making against instructing jurors to adjust their moral judgments consistent with the equal-weight view may seem to apply to factual judgments as well. However, this conclusion does not follow from Kahan's arguments. Although Kahan and his numerous co-authors argue that people's factual perceptions are often shaped by their values or identities, the kinds of “facts” they focus on in order to make their case are special kinds of facts. The “facts” that Kahan discusses are not like the fact of the matter about the best explanation of, say, the presence of a glove at the scene of a crime. Kahan discusses two kinds of facts.

The first kind is the kind of facts on which people have heavy ideological stakes, such as, as mentioned above, childhood vaccinations, climate change, and gun control.Footnote ¹⁵⁸ Because jurors are as a general matter not tasked with assessing empirical evidence bearing such large public-policy implications, Kahan's argument that “identity-protective cognition” leads people to interpret these facts one way or another for the most part does not apply to the topic of this article. It is possible that certain factual issues implicate broader ideological commitments that jurors have. Consider beliefs like “Women do not lie about being raped,” and “Cops always lie about shooting in self-defense after killing young black men.” It is possible that some jurors would have trouble reaching a conclusion contrary to these beliefs without doing violence to them, but, for the most part, jurors should be able to see that these beliefs are generalizations to which there can be exceptions. They do not have the comparable option of seeing exceptions when their moral beliefs seem to dictate an answer in one way.

The second kind of facts that Kahan and his co-authors consider are, however, closer to the topic of this article, as they are routinely considered by jurors. And those are issues, as we saw above, such as whether a man “reasonably understood” a woman as “consenting to sex,”Footnote ¹⁵⁹ whether a person “killed in self-defense,”Footnote ¹⁶⁰ whether a person fleeing from cops “drove in a manner that put” members of public or the police at “great” or “serious” “risk of death,”Footnote ¹⁶¹ whether it was “worth the danger to the public for the police to engage in a high-speed chase” of a person fleeing the cops and whether they “should have tried to find and arrest him later,”Footnote ¹⁶² how much “the parties were at fault for the risk posed to the public by the chase,”Footnote ¹⁶³ and whether the danger the person's fleeing “posted to the police and the public justified” the police officer's decision to “end the chase in a way that put [the driver's] life in danger.”Footnote ¹⁶⁴

Even though Kahan and his co-authors repeatedly and dramatically describe these as factual questions that are answered differently depending on one's values, they are more accurately classified as moral questions (often featuring “thick concepts” as discussed above), given that they all have heavily evaluative components. And, of course, the objection that factual perceptions are heavily influenced by one's values in the same way one's moral perceptions reflect one's values loses much of its force if the label “factual” in factual judgments that are supposedly value-dependent turns out to mean something other than “factual” and actually means “moral.” The objection based on the cultural-cognition theory thus misses the mark.

Another might object that factual judgments, like moral judgments, can be embedded in a person's belief system by relying on Miranda Fricker's work on epistemic injustice. Fricker's central arguments concern what she calls “testimonial injustice,” which results from “prejudicial credibility deficit.”Footnote ¹⁶⁵ Prejudicial credibility deficit arises when one assigns less credibility to a speaker due to prejudices one holds, and when such a prejudice corresponds to the speaker's social identity as a woman or a person of color, Fricker classifies the credibility deficit that is assigned to the speaker as “identity-prejudicial credibility deficit.”Footnote ¹⁶⁶ So if a juror is not inclined to believe a female speaker testifying in a rape case because the juror believes that women often lie about being raped, the speaker would be a victim of epistemic injustice inflicted on her due to the “identity-prejudicial credibility deficit” assigned to her by the juror. Since much of what takes place in a trial is jurors listening to testimonies about historical facts and making credibility judgments about such testimonies, one might say that this shows the ways in which people's factual judgments can be motivated by a set of values they hold,Footnote ¹⁶⁷ and such motivated factual judgments would be an instance of factual judgments being “embedded” like moral judgments.

The answer to this objection is that there is a difference between a situation where one's perception of facts in the world is influenced by other beliefs one holds and a situation where revising one's perception of facts would require revising a number of other beliefs one holds. Fricker's argument as described so far speaks to the former situation. As to the latter situation, Fricker speaks to it in her discussion of how one ought to cope with one's own prejudice. She explains that:

The relevant aspect of this prescription for our purposes is that the act of “revising the credibility upwards to compensate” need not threaten other views that one may hold about the world, given that the credibility judgment that matters is about people testifying in that particular case, and one may still hold whatever generalizations one might be motivated to retain about ways of the world. Unless one encounters evidence in a trial that the earth is flat or that the laws of physics cease to exist on certain stoves in an ordinary home, implications of particular factual judgments one makes in a typical trial are limited to that particular trial. Therefore, even if one may be motivated to believe or disbelieve particular witnesses due to one's prior emotional commitments, ignoring such initial inclinations need not be costly. As noted above, it is not the same in moral judgments.