Unspoken Questions in the Rule 32.1 Debate:Precedent and Psychology in Judging

David E. Klein*

I. Introduction

The legal cOlnlllunity's debate over the seelllingly esoteric ProposedFederal Rule of Appellate Procedure 32.1 l11ight be expected to draw little l110rethan yawns frolll people who do not sit on a bench, practice law, or spend l11uchtil11e in law schools. But, as quickly beconles apparent frol11 the papers in thissylnposiul11 and the literature upon which they draw, the debate can be highlyinstructive, even fascinating, to non lawyers who want to understand judges andthe law they produce. Part of the fascination is in seeing how passionate l11anyl11el11bers of the legal cOlnlnunity are about the issue and how impressively eachside l11arshals nonnative argulnents for its position. But an equally valuableaspect of the debate, froln a social scientist's perspective, is how well ithighlights gaps in our knowledge about what judges do and how they think.

The purpose of this short essay is to direct attention to several empiricalquestions that are raised by but not answered in the debate. Naturally, there arel11any such questions. SOlne of those relating l110st closely to the proposedrule-such as how unpublished dispositions are produced and how they areused and regarded by lawyers and judges-are already receiving systelnaticscrutiny.l This Article focuses on questions that transcend the debate over Rule32.1, itllplicating fundalnental issues ofwhy judges act as they do and how thelaw el11erges, and that have not been the subject of a broad and sustainedresearch progral11 in law schools. These issues fall into two Inain areas:

* Associate Professor of Politics, University ofVirginia. I am grateful to David Caudillfor his invitation to contribute this essay, and to him, the other participants in the synlposiull1,and Greg Mitchell for helpful discussions.

1. In addition to the interview and survey evidence discussed by Professor Barnett in thissymposium, there is related evidence in the Federal Judicial Center's recently cOll1pleted report,TIM REAGAN ET AL. , CITATIONS TO UNPUBLISHED OPINIONS IN THE FEDERAL COURTS OF ApPEALS:PRELIMINARY REPORT (Apr. 14, 2005), available at http://nonpublication.com/t]c.pdf. Seegenerally Stephen L. Wasby, Unpublished Court ofAppeals Decisions: A Hard Look at theProcess, 14 S. CAL. INTERDISC. LJ. 67 (2004).

1709

1710 62 WASH. & LEE L. REV. 1709 (2005)

judges' use of precedent and the workings of the judicial mind. In discussingwhat we do and do not know about these issues and SOlne potentially fruitfulareas of inquiry, I will draw on work from political science and psychology. Asecondary but still itnportant aitn of this Article is to argue for the value ofgreater crossfertilization across disciplinary boundaries.

II. The Use a.!Precedent

A. How Much and Why?

Consider a pair of unconnected claiIns, one fron1 opponents of Rule 32.1and one from proponents. SOlne opponents clailn that allowing citation ofunpublished opinions would inevitably result in those opinions being givenundue weight by lawyers and lower court judges. In his letter opposing therule, Judge I(ozinski writes:

[L]ower court judges, whose rulings will be appealed to the circuit, areextrelnely reluctant to ignore fine nuances of wording that they believereflect the views of three court of appeals judges. Unlike law reviewarticles, opinions of district courts and other nonbinding authorities,unpublished dispositions of the circuit are seldom dismissed asinconsequential, yet they should be.2

According to another opponent, "Every judge and lawyer in Alnerica hasinternalized the hierarchical nature of our justice systeIn; the word of a federalCourt of Appeals, even unpublished, will not be treated the saIne as the word ofa legal scholar or newspaper coltllnnist. ,,3

The concern raised by the Rule's proponents is that disallowing citationrobs lawyers and judges of a valuable set of precedents. In the words ofajudgewriting in support of the Rule:

Judges rely on this Inaterial for one reason; it is helpful. For instance,unpublished orders often address recurring issues of adjective law rarelycovered in published opinions.... We have all encountered the situation inwhich there is no precedent in our own circuit, but research reveals thatcolleagues in other circuits have written on the issue, albeit in an

2. Letter froln Judge I(ozinski, U.S. Circuit Judge for the Ninth Circuit, to Judge Alito,U.S. Circuit Judge for the Third Circuit, (Jan. 16, 2004), available athttp://www.nonpublication.com/kozinskiletter.pdf.

3. Memorandum froln Patrick J. Schiltz, Reporter, to Advisory COInInittee on AppellateRules 45 (Mar. 18,2004) (quoting Letter from E. Vaughn Dunnigan to Peter G. McCabe,Secretary for COInmittee on Rules of Practice and Procedure 1 (Feb. 8, 2004), available athttp://www.secretjustice.org/pdf~files/con1n1ents/03-AP-322).

UNSPOI(EN QUESTIONS IN THE RULE 32.1 DEBATE

unpublished order. I see no reason why we ought not be allowed toconsider such material.4

1711

What the clailns have in COlnmon is an assulnption that judges takeprecedents very seriously even when they are not technically required to.Perhaps this assulnption could have been accepted without demur a fewgenerations ago. But today we lnust approach it with at least SOlne skepticisln.

It has been many years since legal realists like Frank, Llewellyn, andPound first argued that precedents are far less constraining than they aresOlnetitnes lnade out to be, in that "a court can usually find earlier decisionswhich can be made to appear to justify almost any conclusion"s and willfrequently be telnpted to do so. Applying empirical tests to realists' logic,political scientists have gathered a great deal of evidence suggesting thatjudges' individual beliefs about right and wrong and what is good or bad policyexert a substantial influence on their decisions. There is still a lively debateover the "attitudinallnodel,,6 in political science, but the debate focuses on howlnuch rOOln judges' personal attitudes leave for other influences to act~ fewwould contest the clailn that personal attitudes play an ilnportant role in theirdecisions. Indeed, the belief that judges' attitudes are important is so well andso long entrenched alnong political scientists that it would be a challenge tofind a single political science study ofjudges' decisionlnaking published in thelast thirty years, whether of suprelne, intennediate appellate, or trial courts, thatdid not include a variable intended to lneasure judges' ideology.7 It wouldprobably be itnpossible to get such a piece published today.

Of course, the legal comlnunity knows the realist/attitudinalist positionwell, and it is cOlnlnonly cited by legal acadenlics and judges. Sonle of thepublic COlnlnents on Rule 32.1, such as this one froln an opponent, have adistinctly real ist flavor:

Facts drive cases. It doesn't help to add authorities. It's a waste of timeand paper for a litigator in the Ninth Circuit to add a "persuasive" cite front

4. Schiltz, supra note 3, at 47-48 (quoting Letter from Judge Ripple, U.S. Circuit Judgefor the Seventh Circuit, to Judge Alito, U.S. Circuit Judge for the Third Circuit 1 (Feb. 12,2004), available at http://www.secretjustice.org/pdf_files/cOlnlnents/03-AP-335. pdf).

5. JEROME FRANK, LAW AND THE MODERN MIND 163 (Peter Smith Publ'g 1970) (1930).6. The tenn cOlnes froln the most prolninent proponents of the realist view in

contelnporary political science. See JEFFREY A. SEGAL & HAROLD SPAETH, THE SUPREME COURTAND THE ATTITUDINAL MODEL REVISITED (2002) (describing the "attitudinallnodel" as a theorystating that Supreme Court Justices decide cases by evaluating the facts and by considering theirown ideological values).

7. See generally LAWRENCE BAUM, THE PUZZLE OF ]UDICIAL BEHAVIOR (1997) for avaluable overview of the literature.

1712 62 WASH. & LEE L. REV 1709 (2005)

the Southern District of Florida or the Middle District of Pennsylvania.The same would be true of nonbinding cites from the circuit itself. 8

Nevertheless, the debate over Rule 32.1 illustrates what I believe to be ageneral truth-that the iIuplications of the realist position are not always takenas seriously as they Iuight be in thinking about the law and judging.

Let us aSSluue for the Iuoment that a strong version ofrealislu accuratelydescribes judging: Judges are powerfully swayed by their personal values andreactions to the facts of a case, and precedents can usually be evaded orIuanipulated to penuit desired outcolues. In this world, even bindingprecedents Iuight have only a weak effect. Precedent with less nonuative forceshould have very little influence. In the IUOSt exhaustive study to date of staredecisis at the United States Supreme Court, Spaeth and Segal found thatJustices who dissented frolu a decision were rarely willing to adhere to thatdecision as a precedent in later cases. 9 Unpublished dispositions are typicallythought to possess little, if any, nonuative force. It is easy to iluagine circuitjudges in the realist world citing unpublished opinions to help justify positionsalready reached, but hard to iIuagine thelu relying on unpublished opinions toreach a position that they otherwise would have rejected.

Lower court compliance with superior courts' precedents is less of ananolualy in the strong legal realist world: Lower court judges luay followprecedent siluply to avoid being reversed. However, it is not yet settled howluuch judges actually worry about reversa1. 10 Even if they are greatlyconcerned about reversal, judges are well aware that unpublished opinions arenot considered binding by circuit judges and usually were not even written bytheIu. Knowing these facts, only a highly risk-averse judge could feelluuchconstrained by an unpublished circuit court decision. Perhaps luany districtjudges are that risk-averse, but I do not see any reason to aSSluue so withoutproof.

8. Letter frolll Judge John Noonan, United States Circuit Judge for the Ninth Circuit, toPeter G. McCabe, Secretary for Con1mittee on Rules of Practice and Procedure (Dec. 16,2003),available at http://www.secretjustice.org/pdf_fi les/Comments/03-AP-052.pdf.

9. See generally HAROLD J. SPAETH & JEFFREY A. SEGAL, MAJORITY RULE OR MINORITYWILL: ADHERENCE TO PRECEDENT ON THE U.S. SUPREME COURT (1999).

10. For conflicting positions on the impoliance ofreversal, compare Charles M. Call1erOn,New Avenues for Modeling Judicial Politics, presented at the 2nd Annual Conference ofthe W.Allen Wallis Institute of Political EconolllY (Oct. 1993) and McNollgast, Politics and theCourts: A Positive Theory ofJudicial Doctrine and the Rule ofLaw, 68 S. CAL. L. REV. 1631(1995), with David E. K.lein & Robert J. HUllle, Fear ofReversal as An Explanation of'LowerCourt Con1pliance, 37 LAW & SOC'Y REV. 579 (2003), and Frank Cross, Appellate CourtAdherence to Precedent, 2 J. OF EMF. LEGAL STUD. 369 (2005). Unfoliunately, these pieces andothers focus largely on circuit court-Supreme Court interactions. There has been even lesssystelllatic attention to the importance of reversal for trial judges.

UNSPOI(EN QUESTIONS IN THE RULE 32.1 DEBATE 1713

Because it is difficult to reconcile the assumption that judges givesubstantial weight to nonbinding precedents with a strong realist view ofjudging, we should be reluctant to accept it unless we are confident that thestrong realist position is wrong. Like nlany legal scholars, I believe that thatposition is at least partly wrong. But elnpirical researchers have had a far easiertilne identifying attitudinal influences on judges' decisionlnaking than lnoretraditionally "legal" influences. I I Much lnore work Inust be done before thosewho doubt the strong realist position can justifiably disregard it.

Even assulning the realist picture of decisionnlaking is indeed incolnplete,what other itnportant influences are at work? This is another difficult eInpiricalquestion that elnerges froln but is not directly confronted in the Rule 32.1debate. Proponents' eInphasis on the iInportance of access to other judges'thinking suggests that they believe that judges look to the legal views of otherjudges who have confronted sitnilar cases in the past because they care aboutgetting the law right. Opponents seenl to have in Inind less laudable influences,such as unthinking adherence to nonns when they are not applicable or thedesire to reach a decision without too lnuch effort. I do not have space todefend this claiIn, but at the logical level it seelns clear that none of these threepossibilities is either ilnplausible or self-evidently Inore ilnportant than theothers. Nor does the eInpirical literature offer lnuch help in detennining whyand when each ofthese influences operates onjudges or which has the greatestitnpact.

Although difficult to address, these questions Inatter. Ifwe could answertheln confidently, we would have a better chance of agreeing about whether theadoption of Rule 32.1 would add to or detract froln the quality of judges'decisionlnaking. More broadly, we would be in a better position to understandhow and how well judges Inake law when they have the opportunity orresponsibility to do so-why an approach taken or rule announced in oneopinion is taken up, ignored, or rejected by other judges-and whether theresults for society are beneficial or not. Such an understanding would in turnhelp us to evaluate other possible judicial refonns.

B. How?

The debate over Rule 32.1 raises questions not just about how lnuch andwhy judges rely on nonbinding precedents, but also about how they use theln.

11. For interesting evidence of the influence of precedents and a helpful discussion ofother efforts to find it, see Mark J. Richards & Herbert M. I(ritzer, Jurisprudential Regin1es inSupren1e Court Decision Making, 96 AM. POL. SCI. REV. 305 (2002).

1714 62 WASH. & LEE L. REV 1709 (2005)

There are probably several that are worthy of discussion, but I will focus onthe question of how Inuch influence a single precedential case is allowed toexert over decisions in subsequent cases. Put another way, do judges tendto look to a body of precedent for help in deciding a case that necessarilydiffers froln each prior case, ifonly in a slnall way, or do they typically relyon a general rule announced in a single opinion to resolve any case thatfalls into a broadly related set?

A closely related question is whether "words chosen by judges fortheir opinion have sOlne controlling future force, just like a statute," or laterjudges "give force only to the ideas expressed in a judicial opinion, ideasthat can be expressed using a variety of words and phrases." 12 Undertraditional understandings of legal reasoning,judges are norlnally expectedto do the latter, piecing together a general rule froln a set of silnilar casesand disregarding as dicta overly broad statelnents in anyone opinion. Ifthey depart froln this approach and accept a previous court's language asbinding, they grant it Inore influence than it otherwise might have.

The consequences of adopting Rule 32.1 would depend partly onwhich type of practice is lnore COlnlnon. As Judge Kozinski points out,overloaded appellate judges "don't have the titne to consider how thelanguage of [an unpublished] disposition lnight be construed (ormisconstrued) when applied to future cases."l3 Hence, to the extent thatlawyers and district judges are guided by the language of circuit courtopinions, we Inust take seriously the argulnent of Rule 32.1 's opponentsthat the Rule could lead to the codification of language that would be betterleft in obscurity. On the other hand, insofar as lawyers andjudges focus onthe ideas underlying earlier decisions rather than the language expressingtheIn, Rule 32.1 proponents are right to elnphasize that even routineapplications of established rules have potential value for elucidating andtesting the logic of those rules.

Forceful scholarly objections to practices that give great weight tosingle precedents,14 such as deference to language, point to ilnplications

12. Richard B. Cappalli, The Conunon Law's Case Against Non-Precedential Opinions,76 S. CAL. L. REV. 755, 784 (2003).

13. K.ozinski, supra note 2, at 5.14. See, e.g., Cappalli, supra note 12, at 784 (arguing precedential value should not hang

on word choice); Penelope Pether, Inequitable Injunctions: The Scandal ofPrivate Judging inthe U.S. Courts, 56 STAN. L. REV. 1435 (2004); Joan M. Shaughnessy, COlnn1entary:Unpublication and the Judicial Concept ofAudience, 62 WASH. & LEE L. REV. 1597, 1603(2005) ("The concern with resolving every future permutation of any opinion before it has beencOlnpleted Inay have gone too far. "). Apparently, the issue is not entirely new. Jerome Frankrepol1ed concerns "that courts have been paying too much attention to the language of prior

UNSPOKEN QUESTIONS IN THE RULE 32.1 DEBATE 1715

that go well beyond Rule 32.1. Part of what seelns to concern critics is thevery fact that these practices appear to be inconsistent with traditionalunderstandings of what COlnnlon law judges do. But their objections godeeper than that. An idea-oriented, case-by-case approach to precedent islllore cautious and, at least from the perspective of the court deciding theinitial precedent, Inore hlllnble. An idea-oriented approach recognizes thatthe first court Inay be unable to foresee all possible circulnstances in whichits intended rule Inight apply, and it leaves later courts Inore freedoln toadapt the rule or limit its reach as necessary. On the other hand, there issOlnething to be said for a systeln in which the first court can announce ageneral rule and expect other courts to respect it--Inost ilnportantly, thatsuch a systeln lnight provide greater clarity and consistency to litigants,lawyers, and lower court judges.

So how do judges use precedents? I do not think we know yet.Scholars' ilnpressions that judges today expect the language in theirpublished opinions to bear a heavy burden is reinforced by SOlne statelnentsfrom judges thelnselves. 15 But I aln not aware of any systelnatic elnpiricalevidence delnonstrating that these iInpressions are correct. And even ifthey are, that does not Inean that judges in subsequent cases treat the earlieropinions as their authors intended. Nor do logic and assumptions aboutjudges' tnotivations yield a confident answer. On the one hand, we lnightexpect later courts to resist giving too nluch weight to earlier judges'statenlent of a rule, because doing so cedes SOlne of their own power toshape the law. On the other hand, being able to rely on a single source forthe statelnent of a rule can lllake judges' lives easier. In addition, judgeslllight appreciate the greater predictability offered by this approach toprecedent, and they lnight feel an obligation not to undennine thatpredictability by taking a different approach.

cases and ... should ... pay nlore attention to what judges in earlier cases have decided asagainst what they have said in their opinions." FRANK, supra note 5, at 162-63.

15. In her public COIUluent on the rule, Judge Cynthia Holcolub Hall writes:To the extent there is any possibility that an "unpublished disposition" luight serveas even persuasive authority in a future case, conscientious judges will refrain frolupreparing anything but the IUOst fully vetted opinion out of the concern thatiluprecise language might later be manipulated by a clever attorney to support aproposition that the judges never intended.

Letter from Judge Hall, United States Circuit Judge, to Peter G. McCabe, Secretary for theConlmittee on Rules of Practice and Procedure (Jan. 12, 2004), at 2, available athttp://www.secretjustice.org/pd(}1Ies/Columents/03-AP-133.pdf; see also Kozinski, supra note2, at 2 (noting published opinions "set the law of the circuit").

1716 62 WASH. & LEE L. REV 1709 (2005)

III. Psychology and Accountability

In its eInphasis up to now on Rule 32.1 's iInplications for the use ofprecedents, this paper reflects the larger debate on the Rule. But the Rule'seffects lnay not be litnited to what happens after an unpublished decision ishanded down; it also Inay affect the reasoning that goes into that decision. Theword "accountability" crops up fairly often in argulnents for the Rule, includingin the sylnposiuln comlnents of Professors Cravens and Pether. 16 SOlnetinlesthe tenn is used to refer to citizens' ability to lnaintain SOlne awareness of, andcontrol over, the actions of public officials in a delnocracy. But, it is alsosOlnetitnes used in a different context to suggest that a judge Inay beencouraged to exert Inore effort in order to write a better opinion, if the judgeknows the opinion will be associated with hitn or her and read by others. AsProfessor Carpenter puts it, "[W]ritten opinions encourage judges to producewell-reasoned, well-written decisions because they subject judges' conclusionsto public scrutiny. This leads to better, Inore consistent opinions because itholds judges accountable to the public which they serve.,,17

Pushing this notion just a bit further, we Inight suppose that when one isaccountable for a decision, not only will the opinion justifying it be written withgreater care, but the very reasoning that goes into the decision before anyatteInpt is Inade to justify it in writing Inight be of higher quality. The generalproposition that people who know that their decisions will be scrutinized willreason lnore soundly has received a good deal of systelnatic attention frolnsocial psychologists. In this section, I will give a brief overview of SOlnecentral theories and findings froln the psychology literature and raise questionsabout how well these apply to judges.

Ifwe were to ask meInbers of the legal cOl1llnunity to describe the kind ofreasoning that an ideal judge would elnploy to decide cases, the details woulddiffer froln one respondent to another. It seelns likely, however, that all wouldagree on sonle fundamental desiderata: for instance, that the judge take intoaccount all and only truly relevant evidence, weigh the evidence appropriately,be aware of the considerations that are playing a role in his or her decision, andnot allow elnotion to rule reason. Unfortunately for a would-be ideal judge,one ofthe Inost consistent findings in social and cognitive psychology [rOln thepast few decades is that human beings often do not succeed in thinking like

16. See also Petller, supra note 14, at 1483 (noting lack of accountability as a cost ofunpublished opinions).

17. Charles E. Carpenter, Jr., The No-Citation Rule for [}npublished Opinions: 1)0 theEnds ofExpediency for Overloaded Appellate Courts Justify the Means q[Secrecy?, 50 S.C. L.REV. 235,248 (1998).

UNSPOKEN QUESTIONS IN THE RULE 32.1 DEBATE 1717

this. Researchers have alnassed Inountains of evidence regarding variouscognitive shortcolnings. To name just a few, stereotypes cOlne to mindunbidden and sOlnetin1es affect our thinking whether we want theln to or not; 18

Inoods and emotions influence our thinking and behavior; 19 we often reachInoraljudglnents ahnost instantly through intuitions, with logic providing onlypost hoc rationalizations;20 we rely on hindsight to overestimate thepredictability of past events;21 and we exhibit "confirlnatory bias" by payinglnore attention to evidence that supports our favored hypotheses than evidencethat contradicts them. 22 Furtherlnore, lnany ofthese influences on our thinkingand behavior occur automatically and undetected. 23

Judges' Ininds n1ay not work the saIne way as those ofthe ordinary peopleor college students who are the subjects of most research in social psychology.Law school and experience in practice and on the bench provide judges withintensive training in reasoning, and training Inakes a difference. "Fonnaleducation, particularly in statistics and econolnics, and even short trainingsessions on the use ofnormative reasoning principles can itnprove performanceon a nUlnber ofjudglnent and decisionmaking tasks. ,,24 Still, judges are hun1an,and it is hard to itnagine that they escape all the cognitive pitfalls that otherpeople stumble into. bnportant support for this intuition COlnes froln a study byGuthrie et al.,25 which shows that judges are prone to several biases in theirreasoning, hindsight bias alnong them.Sitnilarly, psychologists have found

18. See Patricia G. Devine, Stereotypes and Prejudice: Their Autol11atic and ControlledCon1ponents, 56 J. PERSONALITY & SOC. PSYCHOL. 5, 5-6 (1989) (considering the "inevitabilityof prejudice" perspective on autolnatic activation of stereotypes) (citing E.R. Smith & N.R.Bransconlbe, Stereotype Traits can be Processed Auto/1'zatically (1985) (unpublishedInanuscript, Purdue University»).

19. See Joseph P. Forgas & Jennifer M. George, Affective Influences on Judgn1ents andBehavior in Organizations: An Infor/nation Processing Perspective, 86 ORG. BEHAV. & HUM.

DECISION PROCESSES 3, 4 (2001) (reviewing recent research demonstrating the influence ofmoods on judgment).

20. Jonathan Haidt, The En10tional Dog and Its Rational Tail: A Social IntuitionistApproach to Moral Judgn1ent, 108 PSYCHOL. REV. 814,814 (2001).

21. Chris Guthrie et aI., Inside the Judicial Mind, 86 CORNELL L. REV. 777,799 (2001).22. Angela M. Pfeiffer et aI., Decision-.!vlaking Bias in Psychotherapy: Effects of

Hypothesis Source and Accountability, 47 J. COUNS. PSYCHOL. 429, 429 (2000) (citing MarkSnyder et aI., Social Perception and Interpersonal Behavior: On the SelfFulfilling Nature 0.[Social Stereotypes, J. PERSONALITY & SOC. PSYCHOL. 656, 657-59 (1977».

23. See John A. Bargh et aI., Autol11aticity of'Social Behavior: Direct Effects of TraitConstruct and Stereotype Activation on Action, 71 J. PERSONALITY. & SOC. PSYCHOL. 230,230(1996) (discussing autolnaticity).

24. Gregory Mitchell, Why Law and Econo/11ics' Pe/fect Rationality Should Not BeTradedfor Behavioral Law and Econon1ics' Equal Incon1petence, 91 GEO. L.J. 67, 87 (2002).

25. Guthrie et aI., supra note 21, at 799.

1718 62 WASH. &LEEL. REV. 1709.(2005)

that trained therapists do not avoid all cognitive pitfalls; for instance, they"attend to hypothesis-consistent inforlnation about their clients more readilythan hypothesis-inconsistent infonnation. ,,26

So while it is an open question how susceptible judges are to shortcolningsin reasoning, it seelns reasonable to proceed on the assumption that they are nothnlnune to theln. And although they lnay engage in faulty reasoning less oftenthan other people, the costs when they do so lnay be much higher.Unrecognized stereotypes can infect factfinding, verdicts, and sentencing.Intuitive moral judglnents can inspire decisions that are inconsistent withprecedent or statutory or constitutional language. Confinnatory bias can causejudges to Inisweigh the evidence in individual cases or to overestiInate thewisdom or value of a rule they favor by tilting the salnple of hypothetical casesthey bring to Inind toward those where the rule would work well.

Because judicial reasoning errors can be so consequential, it behoovesthose of us who care about judging to ask how the risk of such errors can bereduced. Here psychology is not quite as helpful: psychologists know far lnoreabout the incidence of reasoning problems than about how they can be avoidedor overCOlne. But a large body of evidence points to the effectiveness of onealneliorative technique-accountability. It seelns clear that people often makefewer reasoning nlistakes when they know that their decision will have to bejustified to an audience. This holds for a wide range of reasoning problelns. 27

The ilnplications for citability rules are unInistakable. In an era whenInost "unpublished" opinions are readily available online, choosing not topublish in a reporter offers no real assurance that an opinion will not be readclosely by lawyers and by the judges receiving their briefs. Rules banningcitation offer much greater assurance. Because judges are aware of this, suchrules ahnost certainly reduce their feelings of accountability to a professionalaudience. Given the psychological evidence, it is reasonable to suspect thatreduced feelings of accountability lead to nlore errors in judges' reasoning.

Before awarding the point to proponents of Rule 32.1, it is important tonote that accountability is not a panacea; it has only been found to counter SOlnetypes of cognitive errors, and even for those, it does not work in allcirclunstances.28 Still, if we consider the conditions under which accountabilityhas the best chance of reducing cognitive errors, they appear to be satisfied bythe typical judging experience. According to Lerner and Tetlock,

26. Pfeiffer et aI., supra note 22, at 430.27. For a comprehensive overview, see generally Jennifer S. Lerner & Philip E. Tetlock,

Accounting for the Effects ofAccountability, 125 PSYCHOL. BULL. 255 (1999).28. Id. at 263~ see also Pfeiffer et aI., supra note 22 (noting that accountability has little

effect on the confirmatory bias of therapists).

UNSPOKEN QUESTIONS IN THE RULE 32.1 DEBATE 1719

accountability is most likely to work where "(a) suboptimal performanceresult[s] froln lack of self-critical attention to the judgment process and(b) in1provement require[s] no special training in forInal decision rules, onlygreater attention to the inforlnation provided. ,,29 The requisite

[s]elf-critical and effortful thinking is most likely to be activated whendecision Inakers learn prior to forming any opinions that they will beaccountable to an audience (a) whose views are unknown, (b) who isinterested in accuracy, (c) who is interested in processes rather than specificoutcomes, (d) who is reasonably wel1-infonned, and (e) who has alegitimate reason for inquiring into the reasons behind participants'judgments.3o

Judges already have special training in decisionlnaking, and the characteristicslisted above would seeln to describe the primary audiences for judges'opinions-bench and bar-quite well.

Ultilnately, then, it appears that we have good reason to suppose thatjudges will be less prone to cognitive errors if they know that their opinions canbe cited. And if accountability really enhances the quality of judicialdecisionInaking, then our interest in it should go well beyond Rule 32.1 to otherprocedures and rules that Inight affect levels of accountability. Doesnonpublication lessen accountability even where unpublished opinions arecitable? Does Inaking decisions in panels rather than alone (as I110St trial judgesdo), meaningfully increase feelings of accountability? What aboutrequirements that opinion drafts be circulated to all judges in a circuit? Arethere other policies or institutional arrangeInents that Inight enhanceaccountabi lity?

Still, we cannot be sure that greater accountability would affect judicialreasoning. Nor do we have a clear idea of how great the decrease in errorswould be, which specific types of suboptiInal thinking would occur lessfrequently, or even how much of a probleIll deficient judicial reasoning is in thefirst place. All of these are open eInpirical questions. And for anyone whocares about the quality ofjudging, they are quite iInportant.

IV Conclusion

I have tried in this Article to focus attention on SOIne questions aboutjudges' reasoning and use of precedent that scholars are a long way froInanswering satisfactorily. I do not insist that these questions Inust be answered

29. Lerner & Tetlock, supra note 27, at 263.

30. ld at 259.

1720 62 WASH. & LEE L. REV 1709 (2005)

for SOlneone to reach a position on Rule 32.1, though the position would beeasier to defend if we knew the answers to theln. What is more important isthat we can think ofn1any other questions, sin1i1ar and dissilnilar, that would beeasier to address if we had a fuller understanding of how judges actually thinkand act. These questions range froln narrower technical ones-aboutpublication, research practices, and the role ofclerks and staffattorneys-to theweightiest questions about why the law Inoves in certain directions and at whatpace, who benefits fron1 the directions the law takes, how well real judicialdecisionInaking measures up to the profession's and the public's ideals, and towhat degree policy Inaking by judges is legitin1ate in a den10cracy.

I do not ilnagine that the eInpirical questions highlighted here will be newto Inany readers, but because so Inany issues con1pete for our attention, it ishelpful at tin1es to relnind ourselves offundalnental issues that we still do notadequately understand. Nor do I Inean to ilnply that all of these questions havebeen neglected in eInpirical work. The cursory discussion in this essay does notbegin to do justice to the vast literature onjudicial behavior in political scienceor the vibrant and growing literature on the subject in law journals. I feelconfident, however, that few scholars active in the field would contest Iny basicprelnise that we do not know as much about these Inatters as we would like.And it seems iInpossible to deny that the study of what judges actually do (asopposed to what they should do) occupies only a sInall fraction ofthe acadelniclegal con1Inunity's attention.

To a social scientist studying courts, this last point is both puzzling anddisappointing. It is puzzling because, as the Rule 32.1 debate illustrates, legalacadelnics care a great deal about the practical itnplications ofjudges' behavior.It is disappointing because we would ahnost certainly Inake faster progresstoward understanding judging if Inore legal acadelnics studied the issue andscholars froln various disciplines were Inore falniliar with each other's work. Itis true that social scientists typically have the advantage of 1110re experiencedesigning and analyzing eInpirical studies and greater familiarity with theoriesof hUInan behavior and institutions froln fields outside of law. But legalacademics usually possess a deeper and lnore nuanced understanding of thejudicial phenolnena under study and ofthe contexts in which they occur, have aInore reliable sense of what theories are likely to work well when applied tojudging, and are able to identify important and intriguing real-worldquestions-such as how noncitation rules affect the process ofjudging-thatwill often escape scholars focused on broader theoretical debates.

The field of law and econolnics delnonstrates how fruitful collaborationacross disciplinary boundaries can be. If this kind of collaboration could beextended to the study of what judges do and how they think, the next tilne a

UNSPOKEN QUESTIONS IN THE RULE 32.1 DEBATE 1721

procedural change like Rule 32.1 is under discussion, argulnents for and againstInight be able to rely less on speculation. More importantly, it would help usbetter understand the law itself.