Printer Friendly


Judges sometimes disagree about the best way to resolve a case. But the conventional wisdom is that they should not be too swayed by such disagreement and should do their best to decide the case by their own lights. An emerging critique questions this view, arguing instead for widespread humility. In the face of disagreement, the argument goes, judges should generally concede ambiguity and uncertainty in almost all contested cases.

Both positions are wrong. Drawing on the philosophical concepts of "peer disagreement" and "epistemic peerhood," we argue for a different approach: A judge ought to give significant weight to the views of others, but only when those others share the judge's basic methodology or interpretive outlook--i.e., only when those others are methodological "friends." Thus textualists should hesitate before disagreeing with other textualists, andpragmatists should hesitate before disagreeing with like-minded pragmatists. Disagreement between the two camps is, by contrast, "old news" and so provides neither camp additional reason for pause.

We also suggest that judges should give weight to the views of all of their methodological friends, not just judges. And we suggest, even more tentatively, that our proposal may explain and, to some extent, justify the seemingly ideological clusters of justices on the Supreme Court. The most productive disagreements, we think, are ones that come from arguing with friends.



       A. Understanding Epistemic Peers
       B. Immediate Objections
          1. Second-Order Humility
          2. Overlap in Interpretive Methodologies
          3. Institutional Peers
       C. Do Judges Really Have Methodologies?
       D. Strategic Voting


In our legal system, judges decide appeals in groups. How should that affect their views about the law?

The conventional wisdom (unquestioned until recently) is that it should not. Each judge should consider the case by his or her own lights--looking to text, history, precedent, practicality, or whatever those lights dictate--and announce this vote to colleagues. If other judges disagree and vote differently, that should not shake one's confidence or change one's behavior. A little bit of compromising around the edges might be permissible to write an opinion that can command a majority, but in the main, each judge is entitled to stick to his or her guns in the face of disagreement. But this view, sometimes labeled "solipsistic," (1) has recently come under attack.

Several scholars now advocate a bracingly contrary position. Professors Eric Posner and Adrian Vermeule argue that all judges ought to become more humble in the face of just about any disagreement with other judges. Presumptively, they maintain, when judges learn that their colleagues disagree with them about interpreting a legal provision, they ought to conclude, at a minimum, that the provision's meaning is unclear. (2) Under existing doctrines conditional on legal clarity, they argue, this presumption should push judges toward deference to agencies, toward officer immunity from constitutional torts, and toward solicitude for criminal defendants under the rule of lenity. (3) Similarly, Professor Alex Stein argues that the solipsistic view is "fundamentally incompatible with the epistemological principles of rational fact-finding." (4) This emerging critique, which we will call "judicial conciliationism"--for the reason that it instructs judges to react in a conciliatory fashion almost whenever encountering judicial disagreement--makes a powerful point against the status quo.

The current debate, however, represents something of a false dichotomy. We propose a different approach, one that is supported both by common sense and by a more nuanced application of the "epistemological principles" invoked by one of these judicial conciliationists. In particular, we think it is a mistake to assume that all judges must treat one another alike, rather than recognizing consistent differences between some groups of judges. And it is a related mistake to assume that only other judges' views should be considered.

Instead, we argue, much turns on the question of legal methodology, rather than legal status. Judges ought to focus on the votes of those who share their interpretive methodology, approach, or outlook--who we call their methodological "friends." Moreover, we suggest that judges should look to all of their methodological friends who have studied the case and express a firm view about it--judges on other courts, lawyers, professors--and not just to their immediate colleagues. Rather than looking to the votes of "other judges," judges ought to look to the votes of their methodological friends.

This analysis is also grounded in formal epistemology and provides occasion to develop the important concepts of "peer disagreement" and "epistemic peers" in the context of legal interpretation. Philosophers' study of epistemology provides several important insights about when and how it is rational for one to change one's mind after learning that others disagree. We deploy those concepts here, arguing that they place great emphasis on disputes over methodology.

Finally, we note that these decisionmaking principles may have been hiding in plain sight. If judges do give substantial weight to the votes of their methodological friends, we would expect to see clusters of like-minded judges whose votes tend to align with one another, and with certain academics and interest groups who follow the cases. As it happens, many observers have documented exactly this phenomenon on the Supreme Court today. (5) But while those observers have generally attributed such clustering to politics, our account provides the potential for a more charitable, and more lawdriven, explanation.

The rest of this Essay proceeds through these three points. In Part I, we apply both common sense and formal epistemology to argue that judges ought to heavily weigh only the views of their methodological friends. In Part II, we extend those arguments to suggest, albeit more tentatively, that judges presumptively ought to consider all of their friends, regardless of whether or not they are judges. Finally, in Part III, we consider the implications of these conclusions and the judicial behavior they explain.


The intuition that judges ought to consider one another's views is a powerful and sensible one, and one with a strong philosophical pedigree. But an important question remains: How? In our view, the answer starts with a key distinction that the judicial conciliationists do not accept. That distinction is between other judges who share their own methodology, orientation, or approach, and those who do not. As a matter of both common sense and more rigorous epistemology, judges ought to give far more weight to the votes of other judges who share their approach, who we call their "friends." By contrast there is little reason to give much weight to judges with very different approaches, who we call their "foes."

A. Understanding Epistemic Peers

The argument for moderation in the face of disagreement starts with a powerful, motivating scenario. Imagine a group of justices sitting at conference table after oral argument, surprised to discover that they are deeply divided:
   [F]ive Justices say that the ordinary meaning of the statute is
   clearly X, and four say that it is clearly Y. Each camp is
   astonished to hear the other camp's view. Each is astonished to
   hear that the other camp not only fails to realize that (X or Y) is
   the clear meaning, but actually, and quite perversely, believes
   that instead (Y or X) is not only one possible reading, but is
   actually the clear meaning. (6)

In this scenario, Posner and Vermeule declare, "all nine Justices need a stiff dose of epistemic humility." (7) They argue, quite plausibly, that the justices here ought to moderate their views based upon their respective discoveries, becoming much less confident that they are right about what the statute means. "If other colleagues, who are presumptively reasonable, agree that the statute is clear, but believe that it is clear in precisely the opposite direction," they reason, "it would be indefensible epistemic practice to simply ignore their views." (8)

So far, so good. Indeed, this position fits into a more general philosophical literature on the phenomenon of "peer disagreement." (9) This philosophical literature turns out to provide a useful lens for examining disputes about legal interpretation. Because that literature is new to legal interpretation scholarship, we must first beg the reader's patience to introduce the key terms: peer disagreement and epistemic peers.

Peer disagreement is a dispute between "epistemic peers," (10) those who are equal in a certain sense. In particular, epistemic peers are individuals who are equally likely to get things right (or wrong) with respect to a given issue. (11) Generally speaking, two people are epistemic peers as to some issue when they are (1) equally rational and (2) have access to the same evidence. (12) For instance, two expert meteorologists issuing forecasts on the basis of the same meteorological data, or two expert linguists looking at the same text, are likely to be epistemic peers. By contrast, a lawyer explaining a complicated legal situation to his client, or an eyewitness describing a scene to somebody who was not there, are not. (13)

A natural instinct in response to peer disagreement is that the two peers ought to try to reconcile their beliefs. In particular, if one learns that an epistemic peer disagrees, one ought to have reduced confidence in one's own position. In the literature on peer disagreement this position is known as "conciliationism." (14)

The instinct for conciliationism is easiest to see in cases of an easily falsifiable disagreement. To take an example from the literature:
   Suppose that five of us go out to dinner. It's time to pay the
   check, so the question we're interested in is how much we each owe.
   We can all see the bill total clearly, we all agree to give a 20
   percent tip, and we further agree to split the whole cost evenly,
   not worrying over who asked for imported water, or skipped
   [dessert], or drank more of the wine. I do the math in my head and
   become highly confident that our shares are $43 each. Meanwhile, my
   friend does the math in her head and becomes highly confident that
   our shares are $45 each. (15)

Here, conciliationism is common sense. One of us is in the right, and if we are equally rational and proceeding on the same evidence, then it is equally likely that it could be you, or me. (16) Thus, goes conciliationism, it is equally likely that the correct calculation is either $43 or $45, so I should significantly reduce my confidence that $43 is right. (17)

The philosophy of tip calculation may seem trivial, as well as academic, because with a little bit of time and basic arithmetic we can check the results by hand and learn whose guess was right rather than assign them equal weight. (18) But conciliationism extends further, to differences in perceptual judgment as well. For instance, most philosophers agree that mutually reduced confidence is appropriate in the following scenario as well:
   You and I are traffic cops watching the cars pass on Main Street.
   We are equally good, equally attentive cops, with equally reliable
   eyesight. We see a truck pass through the intersection. I think
   that it ran the red light. You think it got through on yellow. (19)

Now to be sure, conciliationism is subject to various objections. One is that the position is self-defeating: Insofar as numerous intelligent, well-informed philosophers reject conciliationism, how can it be rational for conciliationists to hold steadfast in their conciliationism? (20) Another concern is that conciliationism, if true, would lead to rampant skepticism. As one philosopher puts it, "[t]he worry is that there is a lot of disagreement out there: if it is epistemically significant, then it will turn out that we aren't rational in believing much of anything." (21) This concern might seem especially warranted once we move beyond epistemically settled zones like math and eyesight to domains such as morality, religion, or politics. (22)

We think that conciliationism can withstand these objections if it is properly understood. (23) The key, however, is to understand that coming to regard someone as an epistemic peer is not as easy as it might seem. And this means that, in the legal context, not all judges should necessarily regard all other judges as epistemic peers. (24) In our view, the judicial conciliationist approach--that is, the position that calls for conciliation among all judges simply because they are judges--neglects a key distinction, namely the role of judicial outlook or methodology. Not all judges share the same outlook, and these disagreements in outlook are a foundation of divided votes. Understanding these disagreements is key to applying the concept of peer disagreement to legal interpretation.

In particular, we think that two judges ought to consider one another "epistemic peers" only to the extent that they share the same judicial outlook or methodology. This shared approach to judging is what marks the judges as "equally rational" from each other's point of view and committed to looking to the "same evidence." (25) The philosophical label may be slightly unfortunate, because it seems impolitic to label another judge "irrational" in the colloquial sense, but the substance is a square match.

In substance, interpretive disputes are disputes about what things to look at, and what to do with those things. (26) Judges who converge on those points are epistemic peers in the technical sense; judges who dispute them are not--or at least not so far as either can tell. (27) Even if one would not call a methodological foe "irrational" or "unreasonable" in the colloquial sense, one might candidly say that this otherwise capable, intelligent person is repeatedly and predictably mistaken when interpreting statutes.

With methodology more firmly in view, we can now see that something important was elided from the original motivating scenario for judicial conciliationism--the surprising vote at conference. Why were the judges surprised by their disagreement? Consider two variants on the original scenario:
   At the conference after oral argument, one judge, a conservative
   textualist, says that the meaning of the statute is clearly X. She
   reasons that X is the only interpretation that squares with
   ordinary usage of the language at issue. A second judge, also a
   conservative textualist, says that the meaning is clearly Y. She
   also claims to be moved primarily by ordinary language examples,
   albeit examples that come out the other way. Each judge is
   astonished to hear the other's view.

Or alternatively:
   At the conference after oral argument, one judge, a conservative
   textualist, says that the meaning of the statute is clearly X. She
   reasons that X is the only interpretation that squares with
   ordinary usage of the language at issue. A second judge, a liberal
   purposivist, says that it is clearly Y. She explains that Y is the
   only interpretation that is compatible with Congress's apparent
   policy aim. Each is dismayed but unsurprised to hear the other's

Whether as a matter of formal epistemology or common sense, there is an important difference between these two scenarios. In the first case, the judges have learned something important from their disagreement and should become much less confident in their respective readings of the statute. In the second case, by contrast, the judges have learned nothing new from one another and have little reason to update their views.

What should be important is not simply the presence of disagreement but what judges learn from it. As judicial conciliationists emphasize again and again, the whole rationale for judges updating their beliefs on the basis of the votes of other judges is that those votes contain "information" that those judges did not already have. (28) Indeed, this point is built into the original scenario's emphasis that the justices involved are "astonished" to learn of the imagined disagreement. (29)

A disagreement between those who share a methodology--between methodological "friends"--involves the discovery of something new. One learns that one's own methodology might lead to a different result in the case at hand. By contrast, a disagreement between those with opposed methodologies--methodological "foes"--pretty much comes as "old news" to the parties involved. (30)

Start with the case of methodological friends. In that scenario, both judges see eye to eye on how to interpret statutes. Their jurisprudential and interpretive methodological commitments are the same. To the extent that other attitudes inform legal interpretation, those attitudes are in alignment too. (31) (With the judges as longtime colleagues, we can assume that they know about this convergence as well.) Given these similarities, and given that each has access to the same briefs and legal materials, each judge would predict that the other would interpret the statute the same way. And each would be even firmer in this prediction once determining that its meaning is clear. (32)

Against the backdrop of these predictions, the two friends' diverging interpretations would be incredibly surprising to each. Previously, each thought their shared interpretive methodology yielded a clear answer in this case. Now, each judge learns that one of them is quite wrong. Each is left to wonder which of the two of them was the one who applied the methodology incorrectly. It is likely, and rational, for each judge to reconsider with diminished confidence because they should have reached the same result.

Now consider the case of methodological foes. There, the two judges are utterly at odds about how to interpret statutes, and, again, as longtime colleagues they presumably know about this longstanding dispute. For that reason, neither would have a great deal of confidence in the first place that the other would interpret the statute the same way. The other judge will be looking at different materials, or looking at them in a different way, so it will be little surprise if they come out with a different answer in the end. As each considers the arguments in isolation, each would start to predict that the other would disagree, recognizing that, for example, ordinary usage points one way and apparent purpose another. And, at conference, when their disagreement is revealed, each would respond with a sigh: "Yet another case where the wrong methodology produces the wrong result." (33)

To summarize: Because disagreements between methodological foes are unsurprising, (34) the rediscovery of such disagreements reveals little information to the parties involved and so provides little reason for those parties to change their beliefs. By contrast, when methodological friends disagree, they should be more surprised and so have more to learn. This gives them more reason to update their beliefs. (35) Disagreement between methodological foes is more like a tip disagreement between a skinflint and a spendthrift: neither is surprised because they were always bound to disagree.

We have stylized the disagreement between friends and foes by considering only two polar cases. In reality, judicial views may lie on a spectrum, with more similar judges being more "friendly" in the methodological sense, and more dissimilar judges being more adversarial. For that reason, the true implications of our argument are scalar rather than binary: the more similar two judges are in their interpretive methodologies, the more reason they have to reduce their confidence in the face of disagreement. (36) But the stylized description demonstrates the key point: the centrality of methodology.

B. Immediate Objections

Defenders of judicial conciliationism do anticipate that some readers will insist on the importance of methodology. The defenders deploy three preemptive responses: first, an argument that judges ought to display the same epistemic doubts at the level of their own methodology; second, an argument that different methodologies overlap so much that there is no real methodological disagreement; and, third, an argument that judges are formally required to treat one another as epistemic peers. (37) None of these arguments, we think, is really convincing.

1. Second-Order Humility

First, judicial conciliationists argue that the same principles of epistemic humility ought to apply to disagreements over methodology itself--that the fact of methodological judicial disagreement ought to weaken judges' confidence in their own methodologies. This is second-order epistemic humility. As Posner and Vermeule explain:
   On our view, epistemic humility should extend to the meta-level as
   well, at least presumptively. All nine Justices should recognize
   that reasonable minds can disagree about the proper approach to
   interpretation, at least within conventional boundaries that
   comfortably include self-identified texualists, self-identified
   purposivists, self-identified intentionalists, and various hybrids.

Just as judges should become less confident in specific interpretations upon learning that other judges interpret the same text differently, so too might judges be less confident in their preferred interpretive theory once acknowledging that other judges find different theories persuasive.

There is some intuitive force to this argument. (39) After all, why should there be anything magical about interpretive methodologies that exempt them from normal principles of epistemic humility? If nobody else is a pure-hearted originalist, should that not make the stout-hearted originalist worry that he is doing something wrong? (40) To the extent that we have methodological commitments or outlooks, many of us have formed them through ongoing discussion and give-and-take with others. So perhaps the continuing presence of others who disagree with us should shake our faith in our own methodological commitments.

Despite this intuitive force, we do not think second-order humility is appropriate when judging particular cases. At conference, when judges hear that some of their colleagues do not share their interpretive approach, they are hearing "old news" (at least after the first few times the judges sit together). With no new information, there is no reason for judges to update their views about the case.

This argument was implicit in our discussion above of methodological foes. Disagreement with one's methodological foes is a nonevent, epistemically speaking, because such disagreement is old news to both sides. The same is just as true for second-order humility. It is precisely because those parties are aware of each other's methodological positions that the specific disagreements between them are unsurprising.

This is not to deny that second-order humility could have a place in reaching one's own methodological views; this just denies that the place is the judicial conference table. By the time lawyers become judges, they are already well aware of different legal methodologies--they may even have become aware of them back in law school, before they were lawyers. Moreover, those judges have doubtless been reminded of those disagreements throughout their careers. Justices Scalia and Breyer relived their methodological disagreements every time they took their act on tour. (41) As such, judges have had ample opportunity to rationally update themselves on the basis of those fundamental disputes. Hearing, one more time, that their colleagues have a different approach tells them nothing new. Whatever degree of second-order humility is appropriate should already be baked into their views. (42)

There are also more fundamental philosophical arguments against second-order humility, namely that it pushes conciliationism beyond its conceptual limits. As philosopher Adam Elga has observed, conciliationism does not apply to instances of "deep" or fundamental disagreement. (43) The reason is that "if a disagreement goes deep enough ... there is little or no independent ground from which to evaluate our respective epistemic credentials," i.e., no way to determine whether the person with whom one disagrees is in fact one's epistemic peer. (44)

To see why, it is important to remember that epistemic peerhood requires us to step outside of the individual disagreement we are evaluating. A naive response to peer disagreement would be to automatically deny that anybody who disagrees with us can be a peer: "I thought that we were both relatively observant, but now that I hear you deny that the truck ran the red light, I see you cannot be trusted!" (45) But this response is obviously question begging. (46) For the concept of an epistemic peer to do any work, peer-hood must be "based on reasoning that is independent of the disputed issue." (47) Thus one's determination whether another person is equally rational must be made setting aside this particular observation--based on track record, vantage point, or whatever else. (48)

All of this means that independent peer evaluation becomes unavailable when disagreement runs too deep. Elga imagines two friends at "opposite ends of the political spectrum" considering whether abortion is morally permissible. (49) As Elga notes, even setting aside their disagreement about abortion, neither friend regards the other as her epistemic peer about politics or morality. (50) Each friend considers the other to have a terrible track record when it comes to moral and political questions generally--as demonstrated by their disagreements about affirmative action, gun control, etc.--and so to be her epistemic inferior. (51)

One might be tempted to say that the two friends ought to "attend to the larger disagreement," and set aside not only their reasoning about abortion, but about moral and political matters generally. (52) But having set aside all of politics and morality, neither friend has any basis left upon which to judge whether the other is an epistemic peer as to the sorts of questions at issue. (53) The question becomes: "Setting aside everything you think about morality and politics, do you think she is just as likely as you to get things right when it comes to morality and politics?" The reasonable response is: "I have no clue."

Elga's reasoning provides reason to doubt second-order humility in the legal domain as well. If two methodological foes are asked to set aside only their disagreement about the case at hand, each one would deny that the other is an epistemic peer. Each one would conclude that the other has a terrible track record when it comes to interpreting statutes generally. If each were instead asked to set aside not just their disagreement about the instant case, but about methodology in general, i.e., if the two were then to "attend to the larger disagreement," neither would have any basis left upon which to assess her respective epistemic credentials. Thus, once again, neither judge would have reason to be less confident in her views.

This inability to assess is starkest in situations where judges regard each other's methodology as incoherent or premised upon some fundamental misunderstanding. (54) To be sure, second-order humility, like first-order humility, could be scalar rather than binary. We could imagine, for instance, a world where all judges were committed consequentialists with the same social welfare function, and their methodological disputes would be disagreements about what methodology maximized social welfare. (55) Or a world where all judges were committed legal positivists and their methodological disputes would be disagreements about the application of positivism to various social facts. (56) But even this kind of shared philosophy is probably too thin to generate useful peer assessments, just as two people cannot surmount deep moral disagreement simply by sharing basic human empiricism and logical reasoning.

Finally, advancing one step further into epistemological theory, we acknowledge a further possible counterargument: one might think that even if one cannot tell whether another person is an epistemic peer, having set aside all of one's theoretical commitments, one also cannot rule it out, and that that alone should be enough to trigger epistemic humility. (57) In other words, maybe reduced confidence is the right response to disagreement absent evidence of another's epistemic inferiority, (58) But as philosopher Katia Vavova explains, this position is too demanding. It ignores the important difference between being "faced with [affirmative] evidence of error" and being in a situation in which we merely lack evidence of absence of error. (59) In part for that reason, it ultimately reduces to generic Cartesian skepticism. (60) Without a non-question-begging response to the moral skeptic, for example, one would be just as obligated to revise one's views in their directions as toward someone who is obviously a genuine epistemic peer on normative questions, such as a close friend with a similar evaluative outlook. Thus, the more defensible view is the more modest view: one should revise one's beliefs only if one has good reason to think that the other person is epistemically equal. (61)

Another way to think about these philosophical arguments is by analogy to "the principle of insufficient reason." (62) That principle instructs decision-makers to "assume[] that variables of unknowable importance, cutting in opposed directions, have equal values and thus systematically cancel out." (63) In other words, the principle of insufficient reason tells decision-makers to attend to considerations of known significance, to the exclusion of considerations of unknown significance. Here, the argument is that judges should attend to disagreements of known significance to the exclusion of disagreements of unknown significance, and, indeed, perhaps unknowable for the reasons Elga and Vavova articulate. (64)

In sum, second-order humility requires an incredibly demanding and somewhat implausible form of theoretical skepticism. The more plausible basis for peer disagreement is a principle requiring revision of one's beliefs on the basis of disagreement only if one has reason to think that the person with whom one disagrees is an epistemic peer. In the context of judging, this occurs, roughly speaking, only to the extent that the person with whom one disagrees is one's jurisprudential friend.

2. Overlap in Interpretive Methodologies

Judicial conciliationists also suggest that the difference between mainstream interpretive methodologies is not so great, making methodological disagreement relatively unimportant. Thus, according to Posner and Vermeule:
   Schematically, it is not the case that textualist judges consider
   sources or arguments {A, B, C} while purposivist judges consider
   sources or arguments {D, E, F}. Rather closer to the truth is a
   schema in which textualists consider (A, B, C} while purposivists
   consider {B, C, D}, or even (A, B, C, D}. There is a substantial
   overlap in the sources used by all the major camps of interpretive
   theory. (65)

And from this they infer:
   This overlap of sources implies that judges in both camps will
   often gain relevant information--relevant even on their own
   theories--from observing the votes of other judges, even judges in
   other camps, insofar as those other judges are considering the same
   sources. (66)

We agree with Posner and Vermeule that textualists and purposivists consider many of the same sources (e.g., text, structure) when investigating statutory meaning, and may sometimes even share the same methods. Thus, we grant that methodological agreement is scalar rather than binary. But we do not think that a mere overlap in sources and methods causes adherents of different interpretive methodologies to learn much from the votes of their methodological foes.

We are not denying that judges might usefully learn from arguments made by other judges during the course of an appeal. Those arguments can be taken on their own terms, run through one's own methodological filter, and accepted or discarded. But judicial conciliationists argue that epistemic humility ought to come from other judges' votes themselves. Yet these differing votes on the same interpretive sources are often caused precisely because of methodological differences. First, and most obvious, even if textualists and purposivists consider similar sources, the relative weight each assigns to those sources is quite different. (67)

Second, even if the evidence each considers is substantially similar, the object of inquiry for textualists and purposivists is potentially quite different. (68) If, for example, textualists are trying to determine "the import that a reasonable person conversant with applicable social and linguistic conventions would attach to the enacted words," (69) while purposivists are seeking the reading that best advances "the statute's goal," (70) the fact that similar sources are relevant to each inquiry is beside the point. Both astronomers and astrologists look at the stars, but neither has much to learn from the other.

Even if there are some cases where textualism and purposivism fully overlap--i.e., where textualists and purposivists use the same evidence in the same way--it is not at all clear that judges can determine when this is the case. And, even if they can, it is not clear it is worth the effort. Reconsider a variation of our earlier scenarios:
   At the conference after oral argument, one judge, a liberal
   purposivist, says that the meaning of the statute is clearly X. She
   reasons that X is the only interpretation that makes sense of the
   obvious purpose of the statute, and is also consistent with its
   ordinary meaning. A second judge, also a liberal purposivist,
   agrees. A third judge, a conservative textualist, says that the
   meaning is clearly Y. She argues that Y is simply what the ordinary
   meaning of the statute requires.

In this scenario, there seems to be some methodological overlap, but it is hard for either of the purposivist judges to give much credit to the third judge's vote and rationale.

Perhaps, if the third judge does go to the trouble to insist that she would vote the same way even considering Congress's apparent policy aim, that should spur the purposivists to give her vote some weight. Even so, this would require the third judge to be equally reliable at determining ordinary meaning, which there is some reason to doubt. The purposivists might reasonably worry that the judge has an excessively wooden, dictionary-bound conception of ordinary usage, and that she is insufficiently attentive to language's sensitivity to practical context. Moreover, as discussed above, the third judge's repeated insistence on interpreting statutes in the wrong way might cause them to doubt her statutory interpretation ability more generally. (71)

It is conceivable that through sufficient inquiry and extensive conversation, the purposivist judges could figure out which is afoot in each case. But then again, maybe not. And at this point, we must ask: Why go through the hassle of an extensive interrogation with limited expected utility when each already has confirmation from a true epistemic peer that her initial interpretation of the statute is correct?

As methodologies converge enough times, judges might come closer and closer to being methodological friends. But if their methodologies remain sufficiently distinct as to remain foes, then we doubt that they have much to learn from one another's votes--even if textualists and purposivists do use overlapping methodologies. And, even if those votes could be mined for some information, it is highly questionable whether it is enough to be worthwhile.

3. Institutional Peers

Finally, judicial conciliationists argue that judges must treat each other as epistemic peers because our legal system assigns them equal institutional authority. As Stein puts it:
   [Our] system entrusts the power of making decisions about people's
   rights, duties and liabilities and about the meanings of statutes
   and constitutions in the hands of equally informed and (more or
   less) equally competent decision-makers: judges and jurors. These
   decision-makers function as epistemic peers. (72)

As his argument implies, Stein (sensibly) does not appear to think that judges in fact regard all other judges as epistemic peers. (73) Rather, the claim is that, because judges enjoy the same legal status under our constitutional and statutory schemes, each is obligated to act as if the others are epistemic peers.

We think that Stein's argument conflates institutional authority with epistemic authority. More still, his position is at odds with the widely accepted judicial practice of criticizing the opinions of other judges on expressly methodological grounds.

Start with institutional, as opposed to epistemic, authority. Here, a helpful analogy is to citizen voting in public elections. Like judges, citizens enjoy equal democratic status under our constitutional and statutory schemes. (74) The way these equal rights manifest is through the assignment of equal legal significance to each citizen's vote. But is that enough? Does the principle of "one person, one vote" also require that citizens act as if all other citizens are epistemic peers on matters of politics? (75) While greater respect among citizens would benefit our politics greatly, it seems implausible to suggest that, as a legal (or perhaps ethical) matter, citizens are required to act--and, even more strongly, vote--as if the policy views of all other citizens are, in their view, equally plausible. So too, we argue, with judges. No one denies that judges must assign equal legal significance to the votes of other judges, adjusting for institutional hierarchy. (76) Be that as it may, why think that, in addition, judges must act as if the reasoning of other judges is equally persuasive?

That judges need not maintain this pretense, especially when it comes to methodological disputes, is further supported by the wide acceptance of open methodological disagreement by judges. No one thinks, for example, that it is somehow inconsistent with his judicial role for Judge Katzmann to criticize other judges for ignoring the intricacies of the legislative process. (77) Nor for Justice Breyer to insist that his originalist colleagues are interpreting the Constitution all wrong. (78) Interestingly, wide acceptance of open methodological disagreement by judges contrasts sharply with, for example, open political disagreement. Contrary to Stein's suggestion, this contrast between methodological and other forms of disagreement indicates that feigned methodological agnosticism is decidedly not compelled by a judge's legal (or professional ethical) obligations.

C. Do Judges Really Have Methodologies?

One might also object that our proposal rests on a false assumption-that judges indeed have well-defined methodologies to mark out their friends and foes. Indeed, Posner and Vermeule seem skeptical of this assumption, remarking that "the controversies over interpretive theory that preoccupy academics do not necessarily carry over to judicial decision making, which is often more pragmatic in spirit," (79) and that "many judges are not theoretical at all and just consider all sources and arguments in a sort of promiscuous jumble." (80) We are willing to defend the assumption in part, but our proposal does not really depend on it. It works just as well for judges who are pragmatic and impatient with interpretive theory.

So we note in passing that some judges do seem to have self-consciously methodological disagreements with one another. Indeed, a number of recent judges have written about their methodologies, and sometimes sharply distinguished themselves from one another. (81) And other judges nonetheless obviously fall into methodological camps. One doesn't need to be preoccupied with academic theory to recognize the difference between Justice Breyer and Justice Thomas. When Justice Breyer learns that Justice Thomas disagrees with him about the permissibility of an agency action or the viability of a constitutional claim, he is less likely to be surprised, and less likely to reconsider his views, than if Justice Ginsburg disagrees with him instead.

But even putting aside judges who have pronounced views on interpretation, methodology can just be a general shorthand for the way judges think about deciding cases. A judge who takes an atheoretical, all-things-considered approach to interpretation will learn little from the votes of a more single-track colleague (as we discuss above, that colleague's vote contains a little bit of information, but not much). (82) By contrast, the same atheoretical judge will learn a lot more from the vote of a similarly atheoretical, all-things-considered colleague. If the similarly open-minded judge comes to a very different conclusion, one would rightly wonder whether one is missing something dispositive.

Similarly, if two different camps of atheoretical judges have consistent differences between them--whether caused by political commitments, by considering different angles of the "promiscuous jumble," or by whatever else--those different camps can be shorthanded as methodological. Nothing rides on whether the judges in those camps can describe their differences in theoretical terms or on whether they could get an "A" in a statutory interpretation seminar.

Take, for instance, the very recent survey of appellate judges by Abbe Gluck and Richard Posner. (83) While Gluck and Posner conclude that "the labels 'textualist' and 'purposivist' [are] unhelpful in addressing the differences among the judges" they surveyed, their work still suggests consistent methodological disagreements. (84) They conclude that the judges they surveyed can be best divided into five categories--"Eclectic Textualists," "Legal Process Institutionalists," "Congressionalists," "Post-Scalia Canonists," and "D.C. Circuit Judges." (85) We do not necessarily endorse the authors' survey methodology or their definitions, but their work illustrates that one can have methodological camps without methodological labels or deep theory.

In other words, our point about friends and foes holds regardless of terminology. One could describe judges as having methodologies, or general outlooks, or simply consistent patterns of behavior that are shared between some judges more often than between others. The point remains that judges learn a lot more from the votes of judges who reason like them than judges who regularly reason differently.

D. Strategic Voting

Thus far, we have argued generally that, in areas beset by deep disagreement, one has reason to be less confident in one's beliefs upon discovering disagreement with one's friends, but not with one's foes. What about the possibility of strategic voting? Posner and Vermeule consider whether the risk of strategic voting by judges undercuts the thesis that judicial disagreement is presumptively epistemically significant. Concluding that the answer is no, they observe that strategic voting is a "more general" problem "hardly unique" to this setting. (86) In addition, they reason that reputation is a "major check" on strategic behavior by judges. (87)

We agree, but we would go further: The prospect of strategic voting provides another useful application for our approach, which assigns greater epistemic significance to disagreement with certain friends. We do not limit ourselves to the kind of strategic voting that Posner and Vermeule focus on--strategic voting made possible by a rule that judges should take the votes of other judges into account. (88) Instead we have in mind strategic voting of a much more banal variety: Voting one way to advance one's ideological interests even though the law, by one's own lights, compels or at least recommends an opposite vote. If such ideologically motivated behavior is afoot, then the differences between friends and foes become even more important--though in this application ideological friendship may be more central than methodological friendship, to the extent the two diverge.

Our theory, simply put, is that one is less likely to be misled by strategic behavior from one's ideological friends than one's ideological foes. Start with this: In a given case, an individual judge has special access to whether she is voting strategically, i.e., that judge knows whether she is casting her vote in bad faith. If a judge is voting strategically, she does not care much whether her ideological foe is as well, at least not for epistemic reasons. She is, after all, uninterested in "what the law is." (89) Limiting ourselves, then, to cases in which a judge is not voting strategically, i.e., is voting in good faith, if the judge is voting consistently with her ideology, she has less reason to believe that an ideological friend who disagrees with her is voting in bad faith than she does an ideological foe. The reason is that, as best she can tell, the ideological friend is voting against interest and so is unlikely to be voting strategically. By contrast, her ideological foe is voting with interest, making it at least plausible that her vote is strategic. (90)

Importantly, this argument does not rest upon the assumption that strategic voting is widespread. (91) Rather, it assumes only that strategic voting occurs with nonzero frequency. The argument has more significant epistemic implications, of course, the more frequent one estimates strategic voting to be. So long as strategic voting occurs some of the time, however, the argument entails that judges should take disagreements with their ideological friends at least somewhat more seriously.

To be sure, strategic voting suggests increased attention to one's ideological friends rather than one's methodological friends, to the extent those two sets of friends differ. It suggests that conservative textualists might prioritize the votes of other conservative textualists more strongly than liberal textualists, and that liberal purposivists might similarly flock together more closely than with conservative purposivists. To the extent that methodology correlates with ideology--or to the extent one is using methodology as general shorthand for decisionmaking patterns--the role of strategic voting reinforces our thesis in a straightforward sense. To the extent one thinks of methodology and ideology as distinct and uncorrected, it instead provides another axis on which to refine our theory.


We have so far suggested that it is important to narrow the concept of epistemic humility, so that it extends not to all judges but only to one's methodological friends. But it may be important to broaden the concept of epistemic humility in a different respect. In particular, we see no good reason for a judge's epistemic peers to be limited to other judges.

Might judges learn not only from the votes of other judges, but also from other professionals who have studied the legal question at issue? From lawyers who write amicus briefs in the case, from scholars who have written commentary, or even from journalists or bloggers who cover the case? Restricting ourselves, as we have argued, to one's circle of methodological friends, why not look to all of one's friends, and not only the ones who wear robes?

Posner and Vermeule briefly flag this point too, asking themselves: "Does the relevant information have to come from the votes of other judges, or will any decision maker do? What about a poll of law professors, practicing lawyers, or people on the street, about whether the statute is clear--should the Justices take such information into account?" (92) But they never return to answer their own question.

Throughout, however, Posner and Vermeule appear to assume that judges should look only to other judges, both from the title of their piece and from their repetition of "judge" and "colleague" to describe their epistemic humility principle. The logic of epistemic humility, we think, points in a different direction.

Posner and Vermeule repeatedly rely on the Condorcet Jury Theorem (93) and on the intuition that when "many other people who are presumptively reasonable" disagree with you, that should lower your confidence in your own views. (94) But if so, the insight generalizes off the bench. American judges come from long legal careers before they take the bench, as private lawyers or government officials or scholars. Their colleagues at their old jobs seem likely to be just as reasonable as their new ones--or at least some of them will be, and need not be ignored.

The point can be made more formally as well. Recall that the two basic proxies for being an "epistemic peer" are that one is "equally rational" and that one "have access to the same evidence." (95) There is no reason nonjudicial actors cannot meet both of these criteria.

As we have already suggested, the most important factor when considering whether one's peers are "equally rational" is whether they are applying a similar framework of reasoning to the problem--i.e., a similar interpretive methodology. One need not be a judge to have, or share, a methodology. Additionally, lawyers have the same kind of legal education as judges, and plenty of them have practical experience, scientific expertise, raw brilliance, or whatever virtues one thinks that judges possess. We see no reason to assume the rationality or ability of legal analysts outside the judicial system to be unequal to those inside the system. (96)

As for "access to the same evidence," one might think that a judge is justified in assuming that only her colleagues really have the same evidence about the case, having read the same briefs and listened to the same arguments by the lawyers. But many other people have access to the briefs and oral arguments as well, especially in federal court of appeals cases or Supreme Court cases. (97) In a controversial case, there is often no shortage of amici and bloggers making their views known on the same evidence that the judges are considering. Moreover, if the really important evidence in the case is not the filings but the relevant legal materials--the regulation or statute or constitutional provision at issue--access is even more widespread.

Another seemingly formal difference for judges is that they take an oath and undertake an obligation to apply the law fairly. (98) But even this difference does not necessarily support total exclusion of nonjudges. Other government officials, and even private lawyers, take an oath as well. (99) And in any event, the oath goes to one's obligation to apply the law; it does not magically increase one's ability to do so. Absent an unusually formalist turn by Posner and Vermeule, we doubt that they would argue that the judicial oath or the judicial power limits their proposal to judges.

To be sure, there are some reasons that courts are attributed additional persuasive authority, but none of those reasons are on point here. For instance, a leading treatment provides:
   When does the court of another jurisdiction have authority just by
   virtue of being another court? I isolate three instances: First,
   when circuit courts cite other circuit courts, not merely for their
   informational or persuasive value, but because they seek to avoid a
   circuit split; second, when state courts aim to harmonize their
   interpretation of state "uniform acts" with other states based on
   the fact that those other states have adopted the same uniform act;
   and third, in common law decisions, when states seek to harmonize
   their doctrines with the judicially crafted doctrines of other
   states. (100)

In each of these examples, the ruling of a whole court creates facts on the ground with which another court must reckon. But the votes of individual judges--before conference and before the court has even ruled--do not have the same effect.

Even if there are no formal reasons to sharply exclude nonjudges from epistemic humility, there might still be practical ones to give nonjudges less weight. One possibility is that nonjudges are more likely to engage in strategic behavior. One's judicial colleagues, the argument goes, have to be honest about how the case should come out, but amici and law professors and journalists might merely be "working the ref." (101) Again, we are not sure how real this difference is. Once the judicial conciliationist regime is in place, judges have plenty of incentive to engage in strategic behavior, as Posner and Vermeule recognize. (102) Instead, they suggest, judges might "be in a better position to gauge confidence levels" of their colleagues than of others. (103)

Perhaps. But it seems to us that the more important consideration is, as we have discussed, that strategic behavior is less of a problem from one's methodological and ideological friends. (104) Moreover, it is also easier to discern strategic behavior from one's friends because one has better access to their reasoning process and so a better ability to see if it is plausibly being followed. (105)

A different possibility is that judges and others have different tolerances for error because of their different professional goals. After denying that judges have any special "expertise" or legal "knowledge," (106) former professor, current judge Frank Easterbrook suggested that "good scholars are bolder than good judges, and accordingly are wrong more often." (107) We could imagine this being manifested in judicial reluctance to consider overly creative academic ideas. It would not surprise us if there were more enthusiasm for stare decisis among members of the bench than members of the academy.

On the other hand, we can imagine this effect to be diminished when we are specifically asking how scholars would vote to decide a certain case. When writing an amicus brief or commenting on a pending case, rather than trying to market a new idea, it is not obvious to us that law professors will be substantially more "apt to err" (108) than any other lawyer or judge of similar outlook. (This optimism is of course limited, though, to briefs and commentary that actually reflect scholarly or lawyerly consideration, not just a signature.) (109)

Nonetheless, Easterbrook's hypothesis might be a reason to discount the votes of law professors on how a case should come out. On the other hand, such votes could still provide valuable information to a methodologically friendly jurist. Suppose an originalist judge is trying to consider which of various dubious precedents he should devote some energy to overturning. The fact that one of the precedents has been singled out as wrong by an overwhelming majority of originalists might provide a good reason, all else equal, to push that one up in the queue.

In any event, our bottom line is this: we can see some reason to think that judges may find more true "friends" on the bench than off of it, but we see no good reason to categorically exclude reasonable nonjudges from the project of peer disagreement. Indeed, once one recognizes our basic point about the logic of looking to methodological friends, the most natural thing to do is to look to all of one's friends, however they are robed.


Having discussed the application of peer disagreement to legal interpretation, we now consider what implications our analysis might have if judges were to implement it. Though this is our most tentative claim in the Essay, we have at least some suspicion that peer disagreement and methodological friendship may already be hiding in plain sight.

In rare cases, judges do offer explicit acknowledgment that the votes of their colleagues have downgraded their confidence, or even changed their mind. For instance, in the 1804 case of Little v. Barreme, which affirmed liability for an illegal naval seizure, Chief Justice John Marshall wrote:
   I confess the first bias of my mind was very strong in favor of the
   opinion that though the instructions of the executive could not
   give a right, they might yet excuse from damages.... But I have
   been convinced that I was mistaken, and I have receded from this
   first opinion. I acquiesce in that of my brethren, which is, that
   the instructions cannot change the nature of the transaction, or
   legalize an act which without those instructions would have been a
   plain trespass. (110)

Eighty years later, we can find the United Kingdom's Lord Blackburn acknowledging in a contract case: "I had persuaded myself [of the other side of the case, and] had written my reasons for so thinking; but as they were not satisfactory to the other noble and learned Lords who heard the case, I do not now repeat them nor persist in them. I assent to the judgment proposed, though it is not that which I had originally thought proper." (111) And as Posner and Vermeule point out, in the more modern context we can occasionally find the Supreme Court acknowledging that lower courts' "differences of opinion from our own are substantial enough" to justify qualified immunity from damages because they "counsel doubt that we were sufficiently clear in the prior statement of law." (112)

But such instances may be most noteworthy for their rarity. Judges do not normally tell us how their initial inclinations changed over the course of deliberation, only what ended up as their final considered judgment. Nor do they specify which of their colleagues' votes counted the most, or whether all had equal weight. Still, we wonder if the overall pattern of judicial behavior might reflect the kind of methodological friendship we advocate here.

For instance, it is well-known that the Supreme Court decides many high-profile cases by a 5-4 vote, and that in many of those cases the 5-4 votes reflect the same general alignment of justices--Roberts and Alito and Thomas and Gorsuch (or their recent predecessors) on one side, and Ginsburg and Breyer and Sotomayor and Kagan (or their recent predecessors) on the other, with Justice Kennedy in between. (113) Indeed, empirical research has documented this ideological clustering on the Court. (114)

Moreover, some of the Court's most controversial 5-4 decisions sometimes give rise to the appearance of influence from outside the Court. On the Supreme Court's decision in District of Columbia v. Heller, (115) Reva Siegel argues that "[t]he correspondence between the law-and-order Second Amendment forged in culture wars of the New Right and the original public meaning of the Second Amendment that Heller vindicates is striking." (116) The Court's split decision on (and near-invalidation of) the Affordable Care Act (117) has been said to have been influenced by a small group of legal scholars and bloggers who "present[ed] legal arguments that professionals could take seriously" and were "difficult to dismiss" as ignorant or hacks. (118) And when invalidating state marriage laws in Obergefell, (119) the Court "invoked the authority of the many [lower] federal court decisions" that had already done so, and "portrayed [them] as having been informed directly or indirectly by the arguments of litigants, lawyers, and society." (120)

These results are generally studied as part of the phenomenon of political influence on judging, or of the phenomenon of popular constitutionalism. Our analysis, however, suggests a different way to think about them.

These patterns of convergence--among like-minded judges and among judges and certain allies off the bench--might well be the result of a rational approach to epistemic humility. As we have discussed, under the principle of epistemic humility, properly understood, we would expect judges to more closely match the votes of their methodological friends on the bench. And we would also expect them to match, perhaps slightly less closely, the votes of their methodological friends off the bench.

While Supreme Court voting blocs are often explained in terms of ideology, (121) they could well be the result of methodology, especially if methodology and ideology are correlated, as they seem to be on today's Court. Indeed, a methodological explanation might actually explain the Court's most recent voting patterns better than an ideological explanation. As Eric Posner has noted, in recent years the so-called conservative justices have tended to disagree among themselves more often than the so-called liberal justices. (122) And a methodological divide among some of the so-called conservatives would provide a convincing explanation for why. (123)

In essence our picture is this: Imagine the Supreme Court hears a case about whether some new statute exceeds Congress's enumerated powers under the Constitution. When the justices reveal their tentative votes to one another, Justice Thomas thinks the statute is unconstitutional. If the late Justice Scalia had been on the fence, he would be more likely to vote to strike down the statute. By contrast, Justice Breyer or Justice Ginsburg is unlikely to gain any new information from Justice Thomas's vote because his application of his methodology tells them little about theirs. In the very rare cases where they think a statute exceeds Congress's enumerated powers, it will be on a different methodological basis. If they take a cue from anybody, it will be from more similar justices. And in each of these scenarios, we would expect the justices to take a challenge to the statute more seriously when law professors and amici who share their jurisprudential outlook have endorsed the challenge.

That picture, we think, is a quite plausible representation of reality. (124) Indeed, to our knowledge the three times Justice Breyer has voted that a statute exceeded Congress's enumerated powers were in Eldred v. Ashcroft, where he agreed on the result only with Justice Stevens, (125) in NFIB v. Sebelius, where his vote was fully shared only with Justice Kagan, (126) and very recently in Murphy v. NCAA, where he joined an opinion invalidating a federal statute as beyond Congress's powers, but also a wrote a separate opinion dissenting in part and arguing that the challengers' victory should be "mostly Pyrrhic." (127) To our knowledge the only time Justice Ginsburg has voted that a statute exceeded Congress's enumerated powers was in City of Boerne v. Flores, (128) where she joined a six-justice majority that included Justice Stevens.

So judicial reasoning among epistemic peers may already be in action--at least in broad strokes. If so, this suggests that what is usually described as immediately political behavior by the Supreme Court may have a more benign, (129) or at least more complicated, explanation.


Two heads are better than one, even when they belong to judges. No individual judge can be sure that she is right about everything, and so she has something to learn from judicial disagreement. Judicial conciliationists are right to suggest that rational judges ought to take such disagreement into account, and not always shut out everybody else's decisions when making their own. But their proposal has a great flattening effect, pushing all legal materials toward ambiguity and all judging away from committed methodological views. For those with strong methodological commitments, it might seem as if the only alternative is to resist them in toto, falling into the solipsistic or heroic model of judging.

We take a position that is both more modest and more radical. Judges ought to attend to disagreement, but they should not treat all disagreements with equal weight. To give all judicial colleagues (and nobody else) equal weight is to confuse power with knowledge. In epistemic terms, what matters is not power but information--i.e., somebody else's views tell the judge something reliable about his own. Our methodological foes may have much to teach us, but not about how confident we should be about individual cases.

We suspect that many judges already know this, and may already do this, even if they would not say so in such exact terms. And if we are right, much of what cynics may be tempted to call extralegal decisionmaking really is consistent with a rational approach to legal reasoning, among friends.

William Baude* & Ryan D. Doerfler**

* Neubauer Family Assistant Professor of Law, The University of Chicago Law School.

** Assistant Professor of Law, University of Pennsylvania Law School. We have benefitted from friendly comments by Nicolas Cornell, Jonah Gelbach, Alon Harel, Daniel Hemel, Brian Hutler, David Plunkett, Eric Posner, David Pozen, Richard Re, Lawrence Solum, Asher Steinberg, Adrian Vermeule, Daniel Viehoff, and participants in workshops at the Duke University School of Law, the University of Chicago Law School, and the University of Pennsylvania Law School. Kevin Keating and Natalie Peelish provided useful research assistance, and the SNR Denton and Alumni Faculty Funds provided research support.

(1.) Eric A. Posner & Adrian Vermeule, The Votes of Other Judges, 105 Geo. L.J. 159, 159 (2016).

(2.) Id. at 162; see also Sharon B. Jacobs, Administrative Dissents, 59 WM. & MARY L. REV. 541, 600-01 (2017) (endorsing this argument); F.E. Guerra-Pujol, The Case for Bayesian Voting: A Response to Posner & Vermeule, SSRN (Jan. 8, 2018), 3096881 [] (emphasizing that judges should take into account their own and their peers' degrees of confidence).

(3.) Posner & Vermeule, supra note 1, at 159-60. Posner and Vermeule assume that the doctrines in question have to do with the clarity of the law, which we will accept for present purposes.

(4.) Alex Stein, Law and the Epistemology of Disagreement, 96 WASH. U. L. REV. (forthcoming 2018) (manuscript at 2), [].

(5.) See infra notes 113-114 and accompanying text.

(6.) Posner & Vermeule, supra note 1, at 163.

(7.) Id.

(8.) Id. at 164.

(9.) See generally David Christensen, Disagreement as Evidence: The Epistemology of Controversy, 4 PHIL. COMPASS 756 (2009) [hereinafter Christensen, Disagreement as Evidence] (surveying philosophical literature).

(10.) See, e.g., Hilary Kornblith, Belief in the Face of Controversy, in DISAGREEMENT 29, 31 (Richard Feldman & Ted A. Warfield eds., 2010) [hereinafter Kornblith, Belief in the Face of Controversy],

(11.) See Adam Elga, Reflection and Disagreement, 41 Nous 478, 487 (2007) [hereinafter Elga, Reflection and Disagreement] ("[Y]ou count your friend as an epistemic peer--you think that she is about as good as you at judging the claim."). And though a true peer is a true equal, peerhood can also be a matter of degree. See infra note 36 and accompanying text.

(12.) See Thomas Kelly, The Epistemic Significance of Disagreement, in 1 OXFORD STUDIES IN EPISTEMOLOGY 167, 168 n.2 (Tamar Szabo Gendler & John Hawthorne eds., 2005) [hereinafter Kelly, The Epistemic Significance of Disagreement]; Christensen, Disagreement as Evidence, supra note 9, at 756-57 (characterizing an epistemic peer as "one's (at least approximate) equal in terms of exposure to the evidence, intelligence, freedom from bias, etc."). But see Elga, Reflection and Disagreement, supra note 11, at 499 n.21 (explaining why such factors are best understood as rough proxies).

(13.) Though we focus on the well-developed concept of peer disagreement, similar points have been made occasionally in other areas of philosophy as well. See, e.g., Dagfinn Foellesdal, Indeterminacy of Translation and Under-Determination of the Theory of Nature, 27 Dialectica 289,298 (1973).

(14.) See Adam Elga, How to Disagree About How to Disagree, in DISAGREEMENT, supra note 10, at 175, 175 [hereinafter Elga, How to Disagree About How to Disagree]; Christensen, Disagreement as Evidence, supra note 9, at 757.

(15.) David Christensen, Epistemology of Disagreement: The Good News, 116 PHIL. REV. 187, 193 (2007) [hereinafter Christensen, Epistemology of Disagreement]; see also Stein, supra note 4, at 5 (adapting this example).

(16.) Christensen, Epistemology of Disagreement, supra note 15, at 193. For this reason, conciliationism is sometimes referred to as the "equal weight" view. E.g., Thomas Kelly, Peer Disagreement and Higher-Order Evidence, in Disagreement, supra note 10, at 111, 112 [hereinafter Kelly, Peer Disagreement] (describing the "equal weight" view as the view that "[i]n cases of peer disagreement, one should give equal weight to the opinion of a peer and to one's own opinion"). Much of our analysis would likely hold true on weaker assumptions, so long as one agrees that the views of a peer are owed significant weight.

(17.) Arguments for conciliationism are thus similar to, but distinct from, Robert Aumann's famous proof that two rational agents with shared priors but different conclusions as to some event will ultimately converge as their posteriors become "common knowledge." See Robert J. Aumann, Agreeing to Disagree, 4 ANNALS STAT. 1236 (1976). Aumann's agreement theorem assumes that both sides process information identically (and correctly), and on this assumption their disagreement must be because they have different information. As their differing views become common knowledge, each infers that the other has private information, in turn adjusts her view, and thus indirectly takes that private information into account. This process repeats again and again until both sides ultimately converge. Id. at 1238. Cases of peer disagreement fix a different assumption--it is assumed that the information available to both sides is the same, such that the disagreement can only be owed to a failure by one or both in information processing. See Kelly, The Epistemic Significance of Disagreement, supra note 12, at 176 ("Disagreement among epistemic peers then, is disagreement among those who disagree despite having been exposed to the same evidence. Thus, our question concerns a case which stands outside the range of cases for which Aumann's result holds."). Judicial disagreement generally involves judges reaching different conclusions "based on the same sources and arguments," Posner & Vermeule, supra note 1, at 159-60, so peer disagreement, rather than Aumann's theorem, is a better match.

(18.) See David Christensen, Disagreement, Question-Begging and Epistemic Self-Criticism, 11 PHILOSOPHERS' IMPRINT, no. 6, 2011, at 8-11, [] [hereinafter Christensen, Disagreement, Question-Begging and Epistemic Self-Criticism] (discussing cases in which the disputants each check their math carefully).

(19.) Katia Vavova, Moral Disagreement and Moral Skepticism, 28 PHIL. PERSP. 302, 307 (2014) [hereinafter Vavova, Moral Disagreement],

(20.) See, e.g., Brian Weatherson, Disagreements, Philosophical, and Otherwise, in THE EPISTEMOLOGY OF DISAGREEMENT 54, 55 (David Christensen & Jennifer Lackey eds., 2013).

(21.) Katia Vavova, Confidence, Evidence, and Disagreement, 79 ERKENNTNIS (SUPPLEMENT) 173, 182 (2014) [hereinafter Vavova, Confidence, Evidence, and Disagreement],

(22.) See, e.g., Christensen, Epistemology of Disagreement, supra note 15, at 189; Kornblith, Belief in the Face of Controversy, supra note 10, at 33.

(23.) Most of the arguments here go to the concern that accepting conciliationism would lead to rampant skepticism. As to the argument that conciliationism is self-defeating, see Elga, How to Disagree About How to Disagree, supra note 14, at 184-86.

(24.) See Lawrence B. Solum, How NFIB v. Sebelius Affects the Constitutional Gestalt, 91 WASH. U. L. REV. 1, 39 (2013) ("The relationship that creates epistemic authority might be viewed as dyadic--a relationship between a pair consisting of the possible epistemic authority and the individual who might defer to the epistemic authority. A given lower court judge, say Learned Hand, might not view a given Supreme Court Justice, say Tom Clark, as an epistemic authority or even as an epistemic peer or equal.").

(25.) See supra notes 10-12 and accompanying text.

(26.) William Baude & Stephen E. Sachs, The Law of Interpretation, 130 HARV. L. REV. 1079, 1112-18(2017).

(27.) See infra notes 33-35 and accompanying text.

(28.) Posner & Vermeule, supra note 1, at 159-60; Stein, supra note 4, at 33.

(29.) Posner 8i Vermuele, supra note 1, at 163.

(30.) Vavova, Confidence, Evidence, and Disagreement, supra note 21, at 181.

(31.) As we explain below, even if nonlegal normative attitudes are irrelevant to interpretation, ideological alignment is plausibly still relevant to the determination of whether another is an epistemic peer with respect to statutory interpretation. See infra Section I.B.

(32.) If, by contrast, each had a confidence level just higher than 0.5, each would presumably have only limited confidence that the other would come out the same way.

(33.) To be fair, judges might learn something from their foes' votes--such as what the opposing methodology requires in this case, or how committed one's foes are to following their inferior methodology. To the extent that judges engage in a sort of continual reflective-equilibrium process to evaluate interpretive methodologies, cf. Richard H. Fallon Jr., Arguing in Good Faith About the Constitution: Ideology, Methodology, and Reflective Equilibrium, 84 U. CHI. L. REV. 123, 123-25 (2017), this information might have a little bit of influence on the margin. But we doubt that any particular case provides an effect of real significance. Cf. Cass R. Sunstein & Adrian Vermeule, The Unbearable Rightness of Auer, 84 U. CHI. L. REV. 297, 299-300 (2017) (describing "the sign fallacy," viz: "to identify the likely sign of an effect and then to declare victory, without examining its magnitude--without asking whether it is realistic to think that the effect will be significant").

(34.) Unsurprising, that is, given the context. There is a wider frame in which judicial disagreement is always surprising because most applications of the law are easy and obvious. See Brian Leiter, Explaining Theoretical Disagreement, 76 U. CHI. L. REV. 1215, 1227 (2009).

(35.) In Bayesian terms, disagreement between methodological foes could best be described as a difference in priors. Cf. Aumann, supra note 17 (arguing only that rational agents with shared priors will converge in their posteriors) and accompanying discussion.

(36.) See, e.g., Elga, Reflection and Disagreement, supra note 11, at 478-79 ("There are experts and gurus, people to whom we should defer entirely. There are fakes and fools, who should be ignored. In between, there are friends and other advisors (including our own future and past selves), whose opinions should guide us in a less than fully authoritative way.").

(37.) Posner and Vermeule do allow that "it is possible that judges should disregard the votes of other judges based on those judges' reasons" but only under conditions that they "suspect ... are rarely satisfied." Posner & Vermeule, supra note 1, at 181-82.

(38.) Id. at 166.

(39.) See Jacobs, supra note 2, at 601 n.294 (endorsing second-order humility argument).

(40.) On pure-hearted originalism, see Antonin Scalia, Essay, Originalism: The Lesser Evil, 57 U. CIN. L. REV. 849, 864 (1989) [hereinafter Scalia, Originalism: The Lesser Evil], On the relative prevalence of commitment to more and less pure forms of originalism, see generally William Baude, Essay, Is Originalism Our Law?, 115 COLUM. L. REV. 2349 (2015).

(41.) See, e.g., Tony Mauro, Breyer and Scalia Take Their Road Show Inside, THE BLT (Mar. 24, 2010,10:51 AM), [].

(42.) Here it is worth noting that versions of conciliationism that call for dramatically reduced confidence in the face of persistent, deep disagreement, see supra notes 39-41 and accompanying text, are a great deal more revisionary than the one defended here.

(43.) Elga, Reflection and Disagreement, supra note 11, at 493-97.

(44.) Vavova, Moral Disagreement, supra note 19, at 319.

(45.) See supra note 19 and accompanying text for the red light example.

(46.) See Christensen, Disagreement, Question-Begging and Epistemic Self-Criticism, supra note 18, at 2.

(47.) Elga, Reflection and Disagreement, supra note 11, at 492 (emphasis omitted).

(48.) This "independence" principle--that one must assess epistemic peerhood on the basis of independent reasoning, Christensen, Disagreement, Question-Begging and Epistemic Self-Criticism, supra note 18, at 1-2--is inherently intertwined with conciliationism, see Christensen, Disagreement as Evidence, supra note 9, at 758-61; Thomas Kelly, Disagreement and the Burdens of Judgment, in THE EPISTEMOLOGY OF DISAGREEMENT, supra note 20, at 31, 3637. Some philosophers largely reject conciliationism by objecting to this independence principle. See Kelly, Disagreement and the Burdens of Judgment, supra, at 40-43; see also Kelly, Peer Disagreement, supra note 16, 122-35. Addressing those objections, however, goes beyond the scope of this Essay, which is focused on the implications of a broadly conciliationist approach for legal interpretation.

(49.) Elga, Reflection and Disagreement, supra note 11, at 492-93.

(50.) Id. at 493.

(51.) Id.

(52.) Id. at 495.

(53.) See id. at 496 ("To set aside [one's] reasoning about all of these issues is to set aside a large and central chunk of her ethical and political outlook. Once so much has been set aside, there is no determinate fact about what opinion of [the other] remains."); Vavova, Moral Disagreement, supra note 19, at 315 ("This suggests that there is an inverse relation between how confident I should be that you are my peer and how deep our disagreement goes. The deeper the disagreement, the less confident my evaluations of your epistemic credentials. And the less confident those evaluations, the less significant our disagreement can be.").

(54.) See, e.g., Antonin Scalia, Common-Law Courts in a Civil-Law System: The Role of United States Federal Courts in Interpreting the Constitution and Laws, in A MATTER OF INTERPRETATION 3, 22 (Amy Gutmann ed., 1997); Richard A. Posner, The Incoherence of Antonin Scalia, NEW REPUBLIC (Aug. 24, 2012), scalia-garnerreading-the-law-textual-originalism [],

(55.) Cf. Cass R. Sunstein, Must Formalism Be Defended Empirically?, 66 U. CHI. L. REV. 636,642 (1999).

(56.) Cf. Baude & Sachs, supra note 26, at 1140-47.

(57.) See, e.g., Nathan L. King, Disagreement: What's the Problem? or A Good Peer Is Hard to Find, 85 PHIL. & PHENOMENOLOGICAL RES. 249, 268 (2012) ("[I]t is hard to see how one's belief that P can be rational if one considers, but withholds [judgment] concerning the claim that one's total epistemic position renders one more likely to be correct than one's dissenter....").

(58.) Philosopher Katia Vavova calls this thought the "No Independent Reason Principle." Vavova, Moral Disagreement, supra note 19, at 316.

(59.) Id. at 318; see also id. at 314-15.

(60.) Id. at 321.

(61.) This one is the "Good Independent Reason Principle." Id. at 317.


(63.) VERMEULE, supra note 62, at 173.

(64.) Attending to methodological friends and ignoring foes also meets Vermeule's two proposed criteria that there be "a pronounced informational advantage" for the attended consideration, and that the attended consideration be, at least "in some rough sense, of the same order of importance as the discarded imponderables." Id. at 175.

(65.) Posner & Vermeule, supra note 1, at 167.

(66.) Id.

(67.) See, e.g., John F. Manning, What Divides Textualists from Purposivists?, 106 COLUM. L. REV. 70, 92-96 (2006).

(68.) Baude & Sachs, supra note 26, at 1112-18.

(69.) John F. Manning, Essay, Textualism and Legislative Intent, 91 VA. L. REV. 419, 424 (2005).

(70.) WILLIAM N. ESKRIDGE, JR. ET AL., LEGISLATION AND STATUTORY INTERPRETATION 229 (2d ed. 2006) ("Purposivism sets the originalist inquiry at a higher level of generality. It asks, 'What was the statute's goal?' rather than 'What did the drafters specifically intend?' ").

(71.) See supra notes 45-54 and accompanying text.

(72.) Stein, supra note 4, at 17 (emphasis added).

(73.) Hence the remark about judges being "(more or less) equally competent." Id.

(74.) See, e.g., Reynolds v. Sims, 377 U.S. 533, 565-66 (1964); Baker v. Carr, 369 U.S. 186, 207(1962).

(75.) Again, for the reasons above, we think citizens at opposite ends of the political spectrum have little reason to actually regard one another as epistemic peers. See supra notes 45-54 and accompanying text.

(76.) A conservative textualist judge on a court of appeals is, for example, no less bound by a Supreme Court decision supported by liberal purposivist majority than one supported by a conservative textualist majority.



(79.) Posner & Vermeule, supra note 1, at 181-82.

(80.) Id. at 167.

(81.) Compare, for instance, ANTONIN SCALIA & BRYAN A. GARNER, READING LAW (2012), with ROBERT A. KATZMANN, supra note 77, and Brett M. Kavanaugh, Fixing Statutory Interpretation, 129 HARV. L. REV. 2118 (2016) (reviewing KATZMANN, supra note 77). Or compare Richard A. Posner, Statutory Interpretation--In the Classroom and in the Courtroom, 50 U. CHI. L. REV. 800 (1983), with Frank H. Easterbrook, Statutes' Domains, 50 U. CHI. L. REV. 533, 534 (1983). Or compare Antonin Scalia, supra note 54, at 37-47, and Scalia, Originalism: The Lesser Evil, supra note 40, with Breyer, supra note 78, and David H. Souter, Address, Harvard University's 359th Commencement Address, 124 HARV. L. REV. 429 (2010).

(82.) See supra Section I.B.2.

(83.) See generally Abbe R. Gluck & Richard A. Posner, Statutory Interpretation on the Bench: A Survey of Forty-Two Judges on the Federal Courts of Appeals, 131 HARV. L. REV. 1298 (2018).

(84.) Id. at 1303.

(85.) Id. at 1303-05.

(86.) Posner & Vermeule, supra note 1, at 183.

(87.) Id.

(88.) See id. at 184 (discussing "rational epistemic free riding--acting in such a way that puts more of the burden of deciding the case on other, possibly more informed, judges" (footnote omitted)).

(89.) Marbury v. Madison, 5 U.S. (1 Cranch) 137, 177 (1803).

(90.) We suspect that it is the typical case where a judge takes herself to be voting in a way that is at least consistent with her own ideology. Our logic would apply in reverse to the opposite case, where a judge takes herself to be voting contrary to her own ideology (e.g., if her ideological friend disagrees with her in such a case, she would have some reason to suspect her friend was behaving willfully).

(91.) As Posner and Vermeule observe, any normative theory of judging must assume that "judges are occasionally, but not uniformly, strategic," the reason being that "[i]f all voting is strategic, then it is idle to argue about how judges 'should' vote." Posner & Vermeule, supra note 1, at 182.

(92.) Id. at 174.

(93.) See id. at 177 n.54.

(94.) See id. at 180.

(95.) See supra notes 10-12 and accompanying text.

(96.) Cf. Eric A. Posner & Adrian Vermeule, Inside or Outside the System?, 80 U. CHI. L. REV. 1743, 1753 (2013) ("From [the Madisonian] standpoint, the judiciary is just another of the branches struggling to encroach upon the others or to aggrandize itself at the expense of the others; judges are just part of the invisible-hand system, not some sort of external regulator of the system."); see also id. at 1757-58 ("Judges are inside the political system, not outside it. If the system is structured and pervaded by partisan competition ... then one cannot turn around and assume that the judges will be immune.").

(97.) See Pamela Samuelson, Possible Futures of Fair Use, 90 Wash. L. Rev. 815, 847-48 (2015).

(98.) See generally PHILIP HAMBURGER, LAW AND JUDICIAL DUTY (2008); Diane P. Wood, Reflections on the Judicial Oath, 8 Green Bag 2d 177 (2005).

(99.) Richard M. Re, Promising the Constitution, 110 NW. U. L. REV. 299, 301 n.1 (2016).

(100.) Chad Flanders, Toward a Theory of Persuasive Authority, 62 OKLA. L. REV 55, 7576 (2009).

(101.) Brett M. Kavanaugh, Essay, The Judge as Umpire: Ten Principles, 65 CATH. U. L. Rev. 683, 688-89 (2016) ("Sometimes, there is even 'working the ref before the game is played, with blog posts and opinion commentaries. Politicians sometimes do this, journalists do this, and professors do this. And you see this of course in sports. Coach Mike Krzyzewski, a legendary basketball coach, is pretty good at working the ref during the game. Nothing wrong with that for the coaches or advocates. But as judges, we have to tune out the Coach K's of the legal-political world who are trying to work the judges." (footnotes omitted)).

(102.) Posner & Vermeule, supra note 1, at 182-84.

(103.) Id. at 180.

(104.) See supra Section I.D.

(105.) Cf. Frank H. Easterbrook, What's So Special About Judges?, 61 U. Colo. L. Rev. 773, 782 (1990) ("Judges apply well only rules that they have internalized.").

(106.) Id. at 773-74.

(107.) Id. at 779; see also id. at 111 ("A free mind is apt to err--most mutations in thought, as well as in genes, are neutral or harmful--but because intellectual growth flows from the best of today standing on the shoulders of the tallest of yesterday, the failure of most scholars and their ideas is unimportant. High risk probably is an essential ingredient of high gain. Academic tenure is desired for reasons opposite to that of judicial tenure: scholars have freedom so that they may be creative, and in spite of the possibility that tenure may protect routineers who unblinkingly do today what was done yesterday."); id. at 779 ("A high proportion of all ideas is unsuccessful (most papers are never cited, even by their authors). A very few scholars, both in ages past and today, produce a high percentage of all the ideas we find useful.").

(108.) Id. at 777.

(109.) See generally Richard H. Fallon, Jr, Scholars' Briefs and the Vocation of a Law Professor, 4). LEGAL ANALYSIS 223 (2012).

(110.) 6 U.S. (2 Cranch) 170, 179 (1804).

(111.) Foakes v. Beer (1884) 9 App. Cas. 605 (HL) 622-23 (appeal taken from Eng.).

(112.) Safford Unified Sch. Dist. No. 1 v. Redding, 557 U.S. 364, 378-79 (2009). See generally Posner & Vermeule, supra note 1, at 168-73.

(113.) That is not to say that this is the only important voting pattern in 5-4 cases, see Joshua B. Fischman & Tonja Jacobi, The Second Dimension of the Supreme Court, 57 WM. & Mary L. Rev. 1671, 1673-74 (2016), but it is an overwhelmingly common one, id. at 1677-78.

(114.) Paul H. Edelman, The Dimension of the Supreme Court, 20 CONST. COMMENT. 557 (2004); Lawrence Sirovich, A Pattern Analysis of the Second Rehnquist U.S. Supreme Court, 100 Proc. NAT'L ACAD. SCI. 7432 (2003); see also Tom S. Clark, Measuring Ideological Polarization on the United States Supreme Court, 62 Pol. Res. Q. 146 (2009).

(115.) 554 U.S. 570 (2008).

(116.) Reva B. Siegel, Comment, Dead or Alive: Originalism as Popular Constitutionalism in Heller, 122 HARV. L. REV. 191, 239 (2008); see also id. at 240 ("The mobilization of living Americans around the text and history of the Second Amendment did more than tutor popular and professional intuitions about the amendment's core and peripheral purposes; it imbued the amendment with compelling contemporary social meaning by connecting the right to bear arms to some of the most divisive questions of late twentieth-century constitutional politics.").

(117.) Nat'l Fed'n of Indep. Bus. v. Sebelius, 567 U.S. 519 (2012) (plurality opinion in part).

(118.) Ilya Somin, Speech, William Brennan Lecture 2014: NFIB v. Sebelius and the Constitutional Debate over Federalism, 39 Okla. City U. L. Rev. 415, 435-36 (2014); see also Lawrence Baum, Supreme Court Justices as Human Decision Makers, 41 Ohio N.U. L. REV. 567, 579 n.81 (2015).

(119.) Obergefell v. Hodges, 135 S. Ct. 2584 (2015).

(120.) Neil S. Siegel, Reciprocal Legitimation in the Federal Courts System, 70 VAND. L. REV. 1183, 1194 (2017). To be sure, Siegel points out that in fact this legitimation was reciprocal, since the lower courts were responding to an apparent Supreme Court invitation in United States v. Windsor, 570 U.S. 744 (2013).

(121.) Clark, supra note 114, at 153 ("A primary finding in the judicial politics literature is that judges who have divergent policy preferences are less likely to vote together. These theoretical and empirical results highlight ideological disagreement as a primary determinant of vote splits among the justices." (citing JEFFREY A. SEGAL & HAROLD J. SPAETH, THE SUPREME COURT AND THE ATTITUDINAL MODEL REVISITED (2002))).

(122.) Eric Posner, A Fractious Majority: The Supreme Court's Conservatives Increasingly Disagree with One Another, Slate (June 30, 2015, 3:26 PM), news_and_politics/view_from_chicago/2015/06/supreme_court_conservatives_disagreements _make_the_court_more_liberal.html []; see also Kedar Bhatia, Final October Term 2015 Stat Pack, SCOTUSBLOG, at 28-30 (June 29, 2016), [].

(123.) See Posner, supra note 122 ("Another explanation is jurisprudential disagreement. Here the division is between formalists (Scalia and Thomas) and pragmatists (Roberts, Alito, and Kennedy).").

(124.) That said, we note that recent work has failed to document a more specific "group polarization" or "political polarization" effect, at least at the Supreme Court. SEE LEE EPSTEIN ET AL., THE BEHAVIOR OF FEDERAL JUDGES 144-49 (2013).

(125.) 537 U.S. 186,243 (2003) (Breyer, J., dissenting).

(126.) Nat'l Fed'n of Indep. Bus. v. Sebelius, 567 U.S. 519, 522-23 (2012).

(127.) Murphy v. Nat'l Collegiate Athletic Ass'n, 138 S. Ct. 1461, 1488 (2018) (Breyer, J., concurring in part and dissenting in part). We put aside a fourth case, Golan v. Holder, in which Justice Breyer and Justice Alito would have invalidated a statute under "the Copyright Clause, interpreted in the light of the First Amendment," 565 U.S. 302, 367 (2012) (Breyer, J., dissenting), since the reasoning did not rest exclusively on enumerated powers grounds.

(128.) 521 U.S. 507(1997).

(129.) To be sure, one might argue that these methodologies ill serve the justices' or the country's interests, cf. Posner, supra note 122 ("If this is true, however, conservatives might wonder whether they are being well served by their justices. Our society has assigned legislative power to the Supreme Court, authorizing it to settle the hardest political questions by fiat. Gay marriage and Obamacare are now unshakable political facts in America, and will remain so long after the jurisprudential debates among the conservatives have been forgotten"), but we leave those debates for another day.
COPYRIGHT 2018 Michigan Law Review Association
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2018 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Baude, William; Doerfler, Ryan D.
Publication:Michigan Law Review
Date:Nov 1, 2018

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |