10 2307@40041740 PDF

Common Law Constitutionalism and the Limits of Reason
Author(s): Adrian Vermeule

Source: Columbia Law Review, Vol. 107, No. 6 (Oct., 2007), pp. 1482-1532
Published by: Columbia Law Review Association, Inc.
Stable URL: http://www.jstor.org/stable/40041740
Accessed: 30-10-2015 15:20 UTC
REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/40041740?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.
Columbia Law Review Association, Inc. is collaborating with JSTOR to digitize, preserve and extend access to Columbia Law
Review.
http://www.jstor.org
This content downloaded from 128.192.114.19 on Fri, 30 Oct 2015 15:20:21 UTC
All use subject to JSTOR Terms and Conditions
ESSAY
COMMON LAW CONSTITUTIONALISM AND

THE LIMITS OF REASON
Adrian Vermeule*
A centralclaim of commonlaw constitutionalism has beenthat

precedentand traditionembody someformof latentwisdom. On this
view,judgeswillgenerally tothewisdomembodied
do bestbydeferring in
precedentand tradition,ratherthan trustingtheirunaided reason.
ThisEssay offers a criticalanalysisofthemechanisms thatare said to
generatethislatentwisdom.Therelevantclaimsand mechanisms suffer
frominfirmities of internallogicand froma failureto makeinstitu-
tionalcomparisons between and amongprecedent and tradition, on the
onehand,and theoutputsoflegislatures, executive and consti-
officials,
tutionalframerson theother.Statutes,administrative decisions,and
constitutionaltextsalso embody informationand are also theproductof
manyminds.Arguments fortherationalityorefficiencyoftheordinary
commonlaw, or ofsocietaltraditions, do nottranslatesuccessfullyinto
arguments for thecomparative or
rationality efficiencyof the constitu-
tionalcommonlaw, as comparedto statutesand othersourcesoflaw.
Introduction
Commonlawconstitutionalism is a theory,or rathera familyoftheo-
ries,aboutconstitutionaladjudication. The family includes theidea that
courtsdo and shoulddevelopthemeaningofgeneralor ambiguouscon-
stitutional
textsbyreferenceto tradition and precedent,ratherthanorig-
inal understanding,and the relatedidea thatcourtsdo and shouldpro-
ceed in a Burkean,ratherthan ambitiouslyrationalistor innovative
fashion.In recentyears,thecentraland moststriking claimof common
law constitutionalism has been that precedentand traditionembody
someformoflatentwisdom.Judgeswillgenerally do bestbydeferring to
the wisdomembodiedin precedentand tradition, ratherthan trusting
theirunaided reason,or so the generalclaimruns.
In whatfollows,I offera criticalanalysisof themechanismsthatare
said to generatethis latentwisdom. Drawingthroughouton Jeremy
Bentham'scritiqueof the subconstitutional commonlaw, I attemptto
update Bentham by using the tools of modern social science and by
adapting his claimsto the settingof constitutional law. Myconclusions,
however,remainlargelyBenthamitein spirit:The constitutional com-
* Professor of Law, Harvard Law School. Thanks to Adam Cox,

Anuj Desai, Jack
Goldsmith,Daryl Levinson, Frank Michelman, Cass Sunstein,and Todd Zywickiforhelpful
comments and conversations, and to Dan Klaff and Andrea Paul for helpful research
assistance.
1482
2007] THE UMTS OF REASON 1483
mon law is not plausiblyseen as a repositoryof latentwisdom. The rele-

vant claims and mechanisms sufferfrominfirmitiesof internallogic and
froma failureto make institutionalcomparisonsbetween and among pre-
cedent and tradition,on the one hand, and the outputs of legislatures,
executive officials,and constitutionalframerson the other.
The idea most frequentlyinvoked to connect common law constitu-
tionalismand the limitsof reason is Burke's dictum:
We are afraid to put men to live and trade each on his own
private stock of reason; because we suspect that this stock in
each man is small, and that the individualswould do better to
avail themselvesof the general bank and capital of nations and
of ages. Many of our men of speculation, instead of exploding
general prejudices, employ their sagacityto discover the latent
wisdom which prevailsin them.1
Burke's dictum is a claim about ordinarycommon law, not constitu-
tional common law; Burke celebrated a common law systemthatdid not
include judicial review.2 But common law constitutionalists,invoking
Burkean themes,have transposedthese mechanismsto the settingof con-
stitutionaladjudication, emphasizing the immanentrationalityand evolu-
tionaryor adaptive fitnessof constitutionalprecedents and traditions.As
we shall see, these ideas can be interpretedeitherin informationalterms,
through the lens of the Condorcet JuryTheorem, or in evolutionary
terms,by referenceto models in which common law processes evolve to-
ward efficiency.
In either case, however, I will suggest a range of puzzles and
problems that limitthe mechanisms' operation to a narrowset of condi-
tions and that even more sharplylimit their relevance to constitutional
law and theory. The rash of scholarship thatportrayedArrow'sTheorem
as a revolutionarycontributionto legal theory (albeit not intended as
such) has abated, in part because it has become apparent that the
Theorem's conditions are highly restrictive.3I suggest that the same
skepticismis warrantedhere. Common law constitutionalismmayor may
not be justifiableor superior to its competitorson other grounds,but the
informationaland evolutionarymechanisms recentlyinvoked to depict it
as a repositoryof latent wisdom turn out to be intrinsicallyfragileand
ungrounded.
institutionally
1. EdmundBurke,Reflections on the Revolutionin France74 (FrankTurnered.,

Yale Univ.Press2003) (1790).
2. Atleastnotin anything likeitscurrentform.See generallyPhilipA. Hamburger,
Revolution andJudicialReview:ChiefJusticeHolt'sOpinionin CityofLondonv. Wood, 94
Colum.L. Rev.2091,2137-46 (1994) (describing1702opinionthatconcludedthatactsof
Parliamentwerenot subjecttojudicial reviewbut acts of corporatebodies thatdid not
exercisepowerof sovereigntycan be) .
3. See GerryMackie,DemocracyDefended 72-157 (Ian Shapiro ed., 2003). For
responses,see Don Herzog,Dragonslaying, 72 U. Chi. L. Rev. 757, 769-74 (2005); Saul
Levmore,PublicChoice Defended,72 U. Chi. L. Rev.777, 782 (2005).
1484 COLUMBIA LAW REVIEW [Vol. 107:1482
The central theme in the followingcritique is this: Argumentsfor

the rationalityor efficiencyof the ordinarycommon law, or of societal
traditions,do not translatesuccessfullyinto argumentsfor the rationality
or efficiencyof the constitutionalcommon law, especiallyas compared to
statutesand other sources of law. The institutionalcontext of constitu-
tional adjudication is decisivelydifferentthan that of ordinarycommon
law adjudication. Both the epistemic and evolutionarymechanisms at
most suggestthat a singlejudge can do betterby deferringto the collec-
tivewisdomembodied in the decisions of pastjudges, or in largersocietal
traditions,ratherthan by relyingon her unaided reason. Perhaps this is
sometimesthe situationfaced byjudges in the shrinkingdomain of ordi-
narycommon law cases where thereis no statuteor administrativeregula-
tion in the picture. Yet thisis almost never the situationthatjudges face
in constitutionaladjudication. The alternativeto relyingon precedent or
tradition,in constitutionallaw,is never reliance on the unaided reason of
the singlejudge; the alternativeis reliance on the latentwisdom of collec-
tive legislatures,or of the executive branch, or of a group of constitu-
tional framers.In differentcontexts,each of these sources has distinctive
costs and benefits. However, the relevantinstitutionalcomparisons are
much differentthan, and more complex than, the simple comparison of
unaided reason to collectivewisdom emphasized by Burke.
The discussion is structuredas follows. Part I discusses Burke as
Condorcet - the idea thatBurke's dictum can be
interpretedepistemically,
in lightof the information-aggregating models developed under the um-
brella of the Condorcet JuryTheorem. I suggest that theJuryTheorem
has no clear payoff for constitutional adjudication. The Theorem's
ratherstringentconditions will frequentlybe violated by the precedents
or societal traditions that common law constitutionalistswould draw
upon. Most importantly, justifyingcommon law constitutionalismon the
basis of theJuryTheorem is fatallynoncomparative. Statutes,administra-
tive decisions, and constitutionaltextsalso embody informationand are
also the product of many minds. In a range of cases, those sources will
oftenbe superior to precedents and societal traditions,according to the
JuryTheorem's own criteriaand logic. Indeed, the epistemic interpreta-
tion of Burke leads most directlytoJames BradleyThayer- to deference
to legislaturesin constitutionaladjudication.
Part II discusses Burkeas Darwin- the idea that Burke's dictum can
be interpretedin lightof evolutionarymodels in which the common law
converges to efficiency.I suggest that these models, like the JuryTheo-
rem, have no obvious payofffor constitutionaladjudication. Even if effi-
ciencyor adaptive fitnessis a desirable propertyof the ordinarycommon
law, it is a problematic normative goal for constitutionallaw. In any
event, the mechanisms of common law evolution have no clear counter-
part in constitutionallaw. The central evolutionarymechanism in com-
mon law theory- that the selection effectsof litigationcause inefficient
rules to be weeded out over time- is of dubious relevance to constitu-
2007] THE LIMITS OF REASON 1485
tional law,where litigationis infrequentand stakesare chronicallydistrib-

uted in an unequal wayacross organized and unorganized interests.Most
importantis thateven ifthe constitutionalcommon law is constantlycon-
verging to efficiency,the background environmentis changing as well,
which means thatconstitutionalprecedent may or may not be efficientat
any given time. If the political environmentchanges withsufficientrapid-
ity,constitutionalcommon law willface a movingtarget,and willnot have
enough time to converge to or even close to efficiency.
For that reason lawmakingthrough alternativeprocedures, such as
legislation,will oftenprove at least as efficientas precedent. Even if con-
stitutionallaw is constantlyconvergingtoward the optimum, legislation
may be closer to the optimum at any given point in time,given a chang-
ing environment. Here too, the crucial institutionalcomparison is be-
tween legislative constitutionalismand common law constitutionalism.
There is no general reason to thinkthatthe evolutionarycapacityof com-
mon law constitutionalismis systematically superior to that of legislative
constitutionalism.Even if the efficiencythesis is useful in the ordinary
subconstitutionalcontext of common law adjudication, its utilityis ob-
scure in the verydifferentsettingof constitutionallaw.
Throughout, I pursue a twofoldstrategy,both identifyinginternal
problems withthe optimisticaccounts of the common law that I canvass,
and then proceeding to examine whetherthose accounts, even ifcorrect,
can be transposed to the institutionalsettingof constitutionaladjudica-
tion. In each case, the internalcritique is a necessarypreliminaryto the
constitutionaldiscussion; the mechanisms of common law constitutional-
ism thatI discuss cannot be understood in the abstract,withoutreference
to the common law context fromwhich theyarose. However, I take no
position on whether these mechanisms offersuccessfulaccounts of the
ordinarycommon law. My central claim is that even if theydo, no trans-
position to the constitutionalsettingis possible.
All in all, the idea thatcommon law constitutionalismis a repository
of latent wisdom, and enables judges to cope with the limitsof human
reason, findslittlesupport in informationalor evolutionarymechanisms.
This does not touch argumentsthatjustifycommon law constitutionalism
on other grounds,4nor does it speak to other debates over constitutional
adjudication. It does, however,remove a main strutof recent arguments
for common law constitutionalism.
I. Burke as Condorcet
This Partexplains,and questions,theinformationalversionof com-

mon lawconstitutionalism.PartLA outlinessome basicsof the Burkean
4. For example, on the basis of the value of stabilityand protectingsettled

expectations.For an overviewof the arguments, see Thomas W. Merrill,Does Public
JudicialActivism
Choice TheoryJustify AfterAll?,21 Harv.J.L. & Pub. Poly 219, 225-26
Merrill,PublicChoice].
(1997) [hereinafter
approach to constitutionaladjudication and details a Condorcetian inter-

pretationof Burke's dictum as a mechanism foraggregatinginformation:
The decisions of a line of past judges embody aggregated information
thatis superior to the unaided reason of any singlejudge or small group
ofjudges sittingtoday. Part I.B analyzes a series of problems arisingfrom
the internal logic of this interpretation,suggestingthat the Condorcet
JuryTheorem's conditions are too restrictivefor the interpretationto be
convincing. Part I.C turns to comparative institutionalproblems. The
idea thatdecisions of a line of pastjudges are betterthan the decisions of
a single judge today has little relevance for constitutionaladjudication.
There the question is not whetherjudges should be guided by their un-
aided reason; it is whethertheyshould defer to pastjudges ratherthanthe
viewsof constitutionalframersor currentlegislativeand executiveinstitu-
tions. In that comparison, the Condorcetian logic suggeststhat in many
cases, precedent is not the best source of information.
A. BurkeanBasics
1. Burke,Tradition,and Precedent. - Common law constitutionalism
representsan explicitlyBurkean strainin constitutionaltheory.5The ba-
sic commitmentof Burkean theoristsis to tradition,somehow under-
stood. For some, traditionhas intrinsicworth,and adherence to tradi-
tion is intrinsically
admirable or at least inevitable. On thisview,tradition
is constitutiveof our veryidentities.6To break withtraditionthreatensa
kind of cultural suicide, and in any event is necessarilyineffective,be-
cause traditionwill continue to shape the veryattemptto break withit.
5. See ThomasW. Merrill,Borkv. Burke,19 Harv.J.L. & Pub. Pol'y 509, 511-13
(1996) [hereinafter Merrill,Borkv. Burke] (defending"conventionalism" and claiming
thatit has Burkeanroots);DavidA. Strauss,CommonLaw Constitutional Interpretation,
63 U. Chi. L. Rev. 877, 893 (1996) [hereinafter Strauss,CommonLaw Interpretation]
("[I]deas . . . commonlyassociatedwithBurke. . . are . . . the underpinnings of the
commonlawapproachto precedent.");Cass R. Sunstein,BurkeanMinimalism, 105 Mich.
L. Rev. 353 (2006) [hereinafter Sunstein,BurkeanMinimalism](describingBurkean
approach to constitutional ErnestYoung, Rediscovering
interpretation); Conservatism:
BurkeanPoliticalTheoryand Constitutional Interpretation, 72 N.C. L. Rev. 619 (1994)
(same). Commonlaw constitutionalism is also compatiblewithFriedrichHayek'sclaim
that the common law, analogouslyto markets,embodies dispersedor decentralized
information thatis superiorin theaggregateto theinformation ofanyofitsparticipants. 1
FriedrichA. Hayek,Law,Legislationand Liberty:Rulesand Order 15-16 (1973). That
claimfitsnaturally witha Condorcetianinterpretation: As I willexplainshortly, theJury
Theorem relies on dispersedinformation for its aggregativeeffects,and describes
conditionsunderwhichthe information held by the votingor statistical groupwillbe
superiorto thatheld byanyof the group'smembers.However,althoughthereis little
difference betweentheBurkeanand Hayekianstrainsas regardsanyoftheclaimsrelevant
here,in constitutional theorythe self-conception of the centraltheoristsis Burkean,not
Hayekian.Accordingly, I focuson Burke'sdictumratherthanitsHayekianrelatives.
6. AnthonyKronman,The Lost Lawyer:FailingIdeals of the Legal Profession217
(1993) ("The constrainedcreativity of the experiencedlawyer. . . comesonlywithtime
and experience,and hence withage, and thusconstitutes a speciesof wisdomthatthe
youngand inexperiencedcannotpossess.").
For other Burkeans, however,traditionhas instrumentalor deriva-

tivevalue.7 Tradition can be and has been defended on a varietyof in-
strumental grounds, including the value of small-scale incremental
change as opposed to sudden large-scale change, and the related idea
that institutionsevolving incrementallyover time are more likelyto be
optimal than designed institutions. I examine such claims in Part II.
This section addresses the epistemic argument for tradition: the argu-
ment thatfollowingtraditionsis the best response to the limitsof human
rationality,cognitive capacity,and information. In one sense, this is a
limited inquiry,but the epistemic argumentfor Burkeanismis enduring
and central;it unites manydisparate strandsof Burkeanismin a common
distaste for "theorizing," "unaided reason," "rationalism," and
"innovation."
Against this background, I focus on the epistemicjustificationfor
Burkean adjudication. Burkeans who apply their views to adjudication
tend to assimilateprecedent and tradition;theysee common law constitu-
tionalism as a kind of judicial traditionunfolding over time, one that
drawson largersocial and political traditionsas well. To be sure, in some
cases, or at some times,common law constitutionalistsdistinguishprece-
dent from tradition. As we will see, however, there is also a standing
temptationto let the distinctionerode.8
The slippage fromtraditionto precedent and back again is a mate-
rial ambiguity,indeed a crucial one. Argumentsthat find latent wisdom
in traditionmay not translatesuccessfullyto precedent even if precedent
is understood as judicial tradition. My focus is on precedent, not on tra-
dition per se, but one of mypoints is that common law constitutionalism
offerssimilarrationalesforincorporatingboth into constitutionaladjudi-
cation, and not infrequentlyequivocates between the two. Accordingly,I
will generallyrefer(withsome awkwardness)to "precedent or tradition,"
but will distinguishthe two when the analysiswarrants.
2. Burkeanismand Second-Order Rationality.- By a kind of mental
commitment,the Burkeanjudge suspends his or her own first-order rea-
son about the meritsof individual policies, in order to follow a simpler
second-order rule of thumb: Policies that comport with the stream of
precedent or tradition,somehow defined, are permissible,at least pre-
sumptively;policies that do not are not permissible,or at least are sus-
pect. The consequentialistargumentfor this decision procedure is that
over an arrayof futurecases, the Burkeanjudge believes she will do bet-
ter by followingprecedent or tradition than by followingher unaided
reason. The Burkean approach is thus a kind of second-order rational-
ism thatdistrustsfirst-order rationalism,in whichjudges deciding particu-
7. See, e.g., Merrill,Bork v. Burke,supra note 5, at 515 (proposingthatchoice

betweenoriginalism and Burkean"conventionalism" shouldbe evaluatedbyaskingwhich
approach is most likely to furthera series of "values that conservatives
commonly
embrace").
8. See infranotes24-29 and accompanying text.
lar cases attempt,on a blank slate, to figureout for themselveswhat is

best.9
The shape of this argument suggestsfour main questions. First,is
thissortof mental commitmentfeasible,or psychologicallypossible? Sec-
ond, what is the criterionfor "doing better"- what is the first-best that
the Burkean second-order approach hopes to track, over an array of
cases? Third, whyis precedent the rightsecond-order guidepost, as op-
posed to, say,the findingsof law and economics, the platformof the Re-
publican Party,or the teachingsof Zoroaster? Finally,what are the impli-
cations of the Burkean approach for constitutionaladjudication? I take
up the last question in detail in PartsI.B and I.C. Here I offersome brief
remarkson the firstthree questions.
The mental commitmentto a second-orderrule- in thiscase, follow-
ing precedent or traditionmay seem psychologicallymysterious.Sup-
pose a case where, to the decisionmaker's unaided reason, the correct
resultseems clearlyto be X, whereas precedent or traditionclearlysaysY;
will not the second-ordercommitmentbreak down? However, common
law constitutionalistsimplicitlyassume that such mental commitments
are stable, although the psychologicalmechanisms thatsupport them are
as yetpoorlyunderstood; I willfollowthatassumptionin order to engage
the internallogic of the common law constitutionalists'view.
Moreover,thereare twonuances thatdilute the forceof the problem
in any event. First,Burkeanjudges use precedent or traditionnot only to
avoid errorbut to conserve on the costs of decisionmaking.10If so, then
the Burkean judge will not (re) decide the first-order question and will
not forman independent view thatX is the case; no tensionwillarise. As
we will see, however,while this nuance helps to sidestep the problems
withsecond-orderdecisionmaking,it actuallyundercuts the Burkean as-
sumption that precedent or traditioncontains latent wisdom,because it
means that most precedents or traditionsdo not embody any indepen-
dentjudgment about the merits. I take up this crucial point in Part I.B.
Second, some theoristsenvisionBurkean adjudication not as an iron-
clad commitment,but as a presumption or a rule with exceptions, and
thus as defeasible in particularcases where the Burkean judge's first-or-
der viewsor judgments are particularlystrong. "[A] rationalisticaccount
of traditionalismjust establishesa requirementthat one give the benefit
of the doubt to past practices. If one is quite confidentthat a practice is
wrong . . . this conception of traditionalismpermits the practice to be
eroded or even discarded."11 Because beliefs have a dimension of inten-
sity,or the epistemicconfidence the decisionmakerattaches to them,the
Burkean can reconcile his first-order and second-order beliefs by saying
9. See, e.g.,Strauss,CommonLaw Interpretation,

supranote5, at 892-94 (discussing
"thecommon-sense notionthatone reasonforfollowing
precedentis thatit is simplytoo
timeconsumingand difficult to reexamineeverything
fromthegroundup").
10. See id. at 912.
11. Id. at 895.
that the latterwill prevail only if the formerdo not exceed some thresh-
old of intensity.
A more troublesomequestion about the second-orderinterpretation
of Burkeanismis: What exactlyis the first-order good thatfollowingpre-
cedent or traditionpromotes? If precedent or traditionis of derivative
value, as an epistemicaid, of whatvalue is it derivative?But it is wrongto
assume thatthere must be some singlefirst-level answer. If Burkeanismis
a second-orderepistemic strategy,it can be used by decisionmakerswith
differingfirst-ordercommitments.12Burkeans with various first-order
theories about what makes outcomes good, or valuable, or just, can con-
verge on the second-ordervalue of precedent or tradition,withoutset-
tlingtheirtheoreticaldifferences.It is perfectlycoherent,whetheror not
correct, for a judge who believes that maximizing "welfare" (somehow
understood) is the touchstoneof good decisions and a judge who believes
that respecting "justice" (somehow understood) is the touchstone of
good decisions to agree on the second-order epistemic value of prece-
dent or traditionas a guide to implementing their verydifferentfirst-
order values.
The most basic question in such cases, however,is whythe second-
order guidepost should be precedent or tradition,ratherthan something
else. For Burkeans, one prominent answer is that followingprecedent
and traditionbest comportswith the limitsof human reason; precedent
and traditionembody the contributionsof many minds13in the past, in
contrastto the unaided reason of an individual decisionmaker today.14
Again, the manyminds mightbe conceived as expressingviewsabout wel-
fare,orjustice, or anythingelse; the basic idea thatmore heads are better
than one or a fewwill remain unaffected. Crucially,however,the insight
about many minds can be given either an informationalor an evolution-
aryinterpretation.Here I will address only the former,leaving the latter
to Part II.
12. Cf. Cass R. Sunstein,Commentary, IncompletelyTheorizedAgreements,108

Harv.L. Rev. 1733, 1740-41 (1995) (arguingthatparticipants in legal controversymay
agree on resultsand low-levelexplanations,but need not agree on fundamental
principles).
13. The phrase"manyminds"is takenfromCass R. Sunstein,Infotopia:How Many
MindsProduceKnowledge(2006).
14. See Merrill,Borkv. Burke,supra note 5, at 519-21 (emphasizingthatjudicial
"conventionalism" comportswith"skeptic [ism] about the powersof human reason");
Strauss,Common Law Interpretation, supra note 5, at 891-92 (noting that "central
idea" is cautionaboutrejecting
traditionalist judgmentsof people "actingreflectively and
in good faith,especiallywhenthosejudgmentshavebeen reaffirmed or at leastaccepted
over time"and that"[jJudgments of thiskindembodynotjust seriousthoughtby one
group of people, or even one generation,but the accumulatedwisdom of many
generations"); Sunstein,BurkeanMinimalism, supranote 5, at 369 (articulatingBurkean
theoryof judicial decisionmakingand suggestingthat "Burke opposes theoriesand
abstractions,developedby individualminds,to traditions, builtup by manymindsover
long periods");Young,supra note 5, at 644-47 (offering Burkeanaccountof limitsof
humanreason).
1490 COLUMBIALAWREVIEW [Vol. 107:1482
3. Precedent,Tradition,and theJuryTheorem. - The CondorcetJury

Theoremprovidesthe mostobviousframework foran informational in-
terpretation ofBurke's dictum, and itis the framework
onlywell-specified
that appears in the literatureon common law constitutionalism. Al-
in
though principle informational approaches need not draw upon the
JuryTheorem,whichisjust one epistemicmodel,commonlaw constitu-
tionalistseitherinvokedie Theoremto putstructure intotheirclaimsor
else remainvague about the informational mechanismstheyhave in
mind.15I willthereforefocuson theJuryTheorem,withoutclaiming
that it is the only informational or epistemicaccount that could be
offered.
The JuryTheoremexplainswhylargegroupsmightdo betterthan
small groups at solvingproblemswithexogenouslydefinedrightan-
swers.16In itssimplestform,theJuryTheoremstatesthat,wherethereis
a binarychoiceand a rightanswerexists,and whereaveragecompetence
exceeds.5- thatis,theaveragememberof thegroupis morelikelythan
not to choose correctly- thenthe likelihoodthata majority voteof the
group will produce the rightanswer approachescertainty the group
as
becomeslargeror as averagecompetenceincreases.17Thus,thegroup's
averagecompetencecan quicklybecome higherthan the competence
even of an expertindividual.
The Theoremcan be extendedto covermore than twochoices,18
and evento covercaseswherethereis no rightanswerexogenousto the
preferencesof some definedgroup,such as the populationat large.19
The latterpointimpliesthatthe Theoremneed not be understoodas
addressingtheaggregation of dispersedinformation, butit certainly
can
be interpreted to do so, and usuallyis. In anyevent,the Condorcetian
interpretation of Burke restson this informational interpretation of
Condorcet.The crucialpoint,amid thesecomplexities, isjust thatthe
aggregatedecisionmaking competenceof manymindswillalmostcer-
15. FortheCondorcetian ofBurke,see Sunstein,BurkeanMinimalism,

interpretation
supra note 5, at 370-71. Somethinglike the Condorcetianinterpretation is plausibly
implicitin theotheraccountscitedin note5, supra,butthosediscussions are vagueabout
theirepistemicassumptions.
16. See generally JamesSurowiecki, The Wisdomof Crowds176 (2004) (notingthat
"small groups can make verybad decisions,because influenceis more direct and
immediateand small-group judgmentstendto be morevolatileand extreme").
17. See Marquisde Condorcet,An Essayon the Applicationof Mathematics to the
Theoryof DecisionMaking(1785), reprinted in Condorcet:SelectedWritings 33, 48-49
(K. Bakered., Bobbs-Merril Condorcet,Applicationof Mathematics]
1976) [hereinafter .
18. See ChristianList& RobertE. Goodin,EpistemicDemocracy:Generalizingthe
CondorcetJuryTheorem,9 J. Pol. Phil.277 (2001).
19. This is the "pollingmodel"interpretation
of theJuryTheorem. See NicholasR.
Miller,Information, and Democracy:Some Extensionsand Interpretations
Electorates, of
the CondorcetJuryTheorem,in Information Poolingand Group Decision Making173
(Bernard Grofman& GuillermoOwen eds., 1986); Paul H. Edelman, On Legal
Interpretations of the CondorcetJuryTheorem,31 J. Legal Stud.327, 332-33 (2002).
tainlyexceed the competence even of the group's wisestmembers,if the

average member is more likelythan not to vote correctly.
How exactly does the JuryTheorem relate to Burkean traditional-
ism? On this view, traditions - including judicial traditionsor lines of
-
precedents are seen as a series of "votes" that aggregate to a collective
view. That view contains a kind of latent wisdom, in that the collective
view is under certain conditions much more likelyto be correctthan the
view of any individual. "If countless people have committedthemselves
to certain practices [over time], then it is indeed possible, on
Condorcetian grounds, that 'latentwisdom' will 'prevail in them,' at least
if most of the relevantpeople are more likelyto be rightthan wrong."20
On this account, the people who participate in a traditionor the
judges who participatein developing a line of precedent lived at different
times,and thus never participatedin an actual collectivevote. The Jury
Theorem does not literallyrequire that the decisionmaking group ever
take a collective vote under majorityrule; exactly the same aggregative
propertiescan be obtained just by taking the statisticalmean of guesses
fromwithinsome population of guessers. As the number of participants
becomes large, the median vote (which prevailsunder majorityrule) will
converge towardthe statisticalmean of guesses withinthe group. Later I
willquestion whetherthisaccount succeeds, but fornow I take it as given.
I will also put aside the standard question whetherthere are indeed
exogenously defined rightanswersin law generallyor constitutionallaw
in particular. Jurisprudentialcontroversiesabound here, between those
who argue that there are rightanswers even in hard cases,21those who
deny this,22and those who think that the existence of rightanswers is
irrelevant,because of persistentdisagreementover what the rightanswers
are.23 The JuryTheorem (in its informationalinterpretation)requires
only an exogenous rightanswer,and thus gets purchase on any case or
legal problem where one of the followingconditions are met: (1) thereis
a factual component to the legal question; (2) there is a prescriptiveor
means-end judgment about which legal ruling will best conduce to
achieving an agreed-upon goal; (3) the legal question, although neither
factual nor prescriptive,otherwisehas a rightanswer somehow defined.
Few willdeny thatmanycases fitone of these categories. Even those who
deny thatall legal cases have rightanswersrarelydeny thatsome cases do.
Those who are skeptical of the right-answerthesis,in any of these
senses, put themselvesoutside the informationalframeworkaltogether.
By contrast,common law constitutionalists who interpretBurke in episte-
20. Sunstein,BurkeanMinimalism, supranote 5, at 371.

21. See RonaldDworkin,Hard Cases,88 Harv.L. Rev. 1057 (1976).
22. Foran overview ofdifferentversionsof thebasicthesisthatlawis indeterminate,
at least in some cases, see LawrenceB. Solum, Indeterminacy, in A Companion to
Philosophy of Law and Legal Theory 488 (Dennis Patterson
ed., 1999).
23. Cf.Jeremy Waldron,The IrrelevanceofMoralObjectivity, in NaturalLawTheory:
Contemporary Essays158 (RobertGeorge ed., 1992).
mic termsnecessarilyassume the existence of rightanswers,suitablyde-

fined, in at least some constitutionaldomain. In order to examine the
internallogic of the argument,I will follow that assumption.
B. Burkeas Condorcet:InternalProblems
Nonetheless, there remain serious obstacles to any Condorcetian in-
terpretationof Burke, both in general and as applied to judicial prece-
dent in particular. I begin with the internal logic of the interpretation,
before moving to comparativeinstitutionalproblems.
1. Precedent,
Tradition,and Numbers. - As mentioned above, a central
ambiguitywithincommon law constitutionalismis the slippage back and
forthbetweenjudicial traditions - lines of precedent or doctrine- on the
one hand, and broader societal traditionson the other. In principle, a
precedent need not draw upon tradition,and a traditionneed not be
embodied in a precedent. Sometimes common law constitutionalistsex-
plicitlydistinguishprecedent from tradition.24Some proponents of tra-
dition,especiallythose who see traditionas a kind of spontaneous order,
explicitlydeny that praise for tradition implies praise for precedent,
which theysee as the positivelawmakingof a centralizedlawmaker (such
as the Supreme Court).25
However, the lines frequentlyblur, both in theoryand in practice.
In theory,a main claim of common law constitutionalismis thata stream
of precedent that has stood the test of time is a kind of tradition,and
deserves respect for the same reasons as tradition.26In practice, it has
been argued, for example, that the Warren Court was a "common law
court,"basing its decisions on experience.27 In some cases, this claim is
supported by pointing to earlier precedents on which the Warren Court
drew.28 Where the Warren Court overruled or broke dramaticallyfrom
precedent, however,the claim is supported by pointing to largerpolitical
and social traditionssaid to be inconsistentwiththe discarded precedent.
Thus, forexample, the innovativedecisions requiringa "one person, one
24. See DavidA. Strauss,Tradition,PrecedentandJusticeScalia, 12 CardozoL. Rev.

1699,1706 (1991) ("Someprecedentsmaybe said tobe partofa tradition.Butnotall are.
Some are simplythedecisionsof a groupofjudges rendereda fewyearsago.").
25. See A.C. Pritchard& Todd J. Zywicki, Findingthe Constitution:An Economic
Analysisof Tradition'sRole in Constitutional 77 N.C. L. Rev.409, 489-93
Interpretation,
(1999).
26. See, e.g.,Strauss,CommonLaw Interpretation, supranote5, at 892 ("Because,in
thisviewof traditionalism,theage ofa practicealone does notwarrantitsvalue,relatively
newpracticesthathaveslowlyevolvedovertimefromearlierpracticesdeserveacceptance
morethanpracticesthatare olderbut thathave notbeen subjectto testingovertime.").
27. DavidA. Strauss,The CommonLaw Geniusof the WarrenCourt5-6 (Univ.of
Chi. Pub. Law & Legal TheoryResearchPaper Series,Paper No. 25, 2002), availableat
http://papers.ssrn. 5682 (on file withthe Columbia
com/sol3/papers.cfm?abstract_id=31
Law Review)[hereinafter Strauss,CommonLaw Genius] (suggesting that"[i]t is unwiseto
tryto resolvea problemwithoutdeferring, to some degree,to the collectedwisdom
reflectedin whatothershavedone whenfacedwitha similarproblemin the past").
28. See id. at 17-24.
vote" standard in reapportionmentcases are justified by pointing to a

broader historical trend toward expansion of the formal franchise.29
Common law constitutionalismthat ranges over both precedent and tra-
dition has multiple degrees of freedom.
This ambiguityis often material,because the Condorcetian creden-
tials of precedent may be much differentthan those of broader tradi-
tions. In the case of Supreme Court precedents, for example, only 110
Justiceshave ever sat on the Court,and of those onlya fractionhave been
participantsin any given line of precedent. For a Condorcetian analysis,
numbersare critical. When the individualreason of a lone judge or small
group ofjudges is compared withtraditionsembodying (let us suppose)
thousands or even millions of individualjudgments, the superiorityof
collective wisdom is palpable. When the judgments of a group of nine
Justices,today,are compared withthejudgments of (say) twenty Justices
yesterday,the margin of superioritythins. Moreover, by the Theorem's
terms,group competence is a functionnot only of numbers,but of aver-
age competence- the probabilityof selecting the correct answer. If to-
day's Justicesare more competent than yesterday's,the differencemay
swampa small deficitin numbers. Shortly,I willgive reasons to thinkthat
average competence indeed rises over time.
Of course, even if that is true, the relative numbers of yesterday's
Justicesmayswamp the higher competence of today'sJustices;everything
depends on the particularvalues of the variables,as is alwaystruewiththe
JuryTheorem. The overall point remains, however. Precedent, espe-
ciallyat the level of the Supreme Court, has a verydifferentCondorcetian
statusthan large-scalepolitical and social traditions.Assimilatingthe two
or vacillatingbetween them, as often happens in common law constitu-
tionalism,givesprecedent an unwarrantedsheen. To the extent that the
informationalinterpretationof Burke is focused on judicial traditions -
the precedentsdeveloped withincourts,as opposed to broader social and
political traditions- its Condorcetian credentials are especially suspect.
2. Timeand Information. - There is a systematicreason for thinking
that today'sJusticesare of higher average competence than yesterday's,
all else equal: Today's Justicesknow more,just because theylive today,
not yesterday.Bentham, followingPascal, offereda crucial objection to
Burkean traditionalism: Decisionmakers today have more information
than decisionmakersin the past, because theyenjoy the benefitof seeing
how thingsactually turned out afterthe firstround of decisions.30 The
ancients are not older than we are, and thereforemore wise; they are
younger than we are, and thereforeless wise, because theylack whatever
29. Id. at 31-32.

30. See Jeremy Bentham,Bentham'sHandbookof PoliticalFallacies54-56 (Harold
Bentham,Handbook];BlaisePascal,Prefaceto
A. Larrabeeed., 1952) (1824) [hereinafter
theTreatiseon Vacuum,inThoughts,Letters, and MinorWorks444,449 (CharlesW. Eliot
ed., M.L. Boothet al. trans.,1910).
knowledge we have gained in the succeeding years. Experience in this

sense favorsthe moderns over the ancients.
3. CommonQuestions'?- A fundamental assumption of the Jury
Theorem, where the aggregation of dispersed information is con-
cerned,31is that votersor group members must be addressing the same
question.32 IfA makes a judgment about question X and B makes a judg-
ment about question Y, there is no genuine pooling of collectivewisdom.
Suppose thatX and Y superficiallyresemble each other,but are critically
differenton closer inspection. Then it may be easy to mistakean ersatz
agreement among many minds,which looks as though it restson collec-
tivewisdom but reallydoes not, for a genuinely Condorcetian process.
This ersatz collectivewisdom is, plausibly,the ordinarystate ofjudi-
cial precedents extended over time.33Suppose thatat Time 1 theJustices
of the Supreme Court hold that a state may not constitutionallyban flag
burning34or partial-birth abortion.35 Then suppose thatCongress enacts
a similarstatute,36and the constitutionality of that statutearises at Time
2. Does the collectivejudicial judgment at Time 1 contain information
useful at Time 2, froma Condorcetian perspective? What questions ex-
actlyare the Justicesaddressing in the two cases? It is not obvious that
the considerations are the same. Holmes, among others, thought that
the problem ofjudicial reviewwas verydifferentforfederal and stategov-
ernments;37and as I shall emphasize shortly,the Time 2 decision poses a
differentquestionjust because it is not a case of firstimpression. None of
these points turn on whetherthe Time 1 decision was corrector not, or
on what the Time 2 decision should be. The point is thattheyare not the
same decisions.
31. This pointdoes not hold forthepollinginterpretation, whichwe haveputaside

(becauseitis nottheinterpretation on whichcommonlawconstitutionalists rely). In that
interpretation,votersare each expressing theirownpreferences. See Edelman,supranote
19, at 332-33 (notingthatin pollingmodel,"whatis consideredrightis the same as the
outcomeofa majority voteamongall of thevoters").In theinformational interpretation,
however, votersare each guessingabout a common,exogenously definedrightanswer.
32. See David M. Estlund et al., DemocraticTheory and the Public Interest:
Condorcetand Rousseau Revisited,83 Am. Pol. Sci. Rev. 1317, 1322 (1989) (among
requirements oftheJury Theoremis thatvotersmustbe "addressing a commonquestion")
(Jeremy Waldron).
33. ThankstoJackGoldsmith foremphasizingthispoint.
34. See Texas v.Johnson,491 U.S. 397, 420 (1989) (holdingthatconvictionunder
stateflagburninglawviolatesFirstAmendment) .
35. See Stenbergv. Carhart,530 U.S. 914, 945-46 (2000) (holdingstatepartial-birth
abortionlawunconstitutional).
36. CompareJohnson, 491 U.S. at 420, withUnitedStatesv. Eichman,496 U.S. 310,
319 (1990) (holdingthatfederalflagburningstatuteviolatesFirstAmendment);Stenberg,
530 U.S. at 945-46, withGonzalesv. Carhart,127 S. Ct. 1610, 1639 (2007) (upholding
federalpartial-birthabortionstatute).
37. See OliverWendellHolmes,Law and the Court (1913), reprintedin Collected
Legal Papers291, 295-96 (1920).
Similar problems arise even when the earlier and later decisions in-
volve the same level of government,on preciselythe same constitutional
issue. Suppose thatat Time 1, the Court decided thatcapital punishment
did not violate "evolvingstandardsof decency."38 Is thisan information-
ally useful precedent if, at Time 2, the Court faces the same question
again? Not necessarily,because standardsmayhave evolved even further.
In this example the problem of change over time is righton the surface
of the doctrinal test,but the same problem may arise in less transparent
form. If at Time 1 the Court held that a certain rule of criminalproce-
dure is or is not "implicitin the concept of ordered liberty,"39 does this
mean that a Court facing the same issue at Time 2 is answeringthe same
question? Even if,say,the twocases are a generation or more apart? Any
account of constitutionalinterpretationand adjudication thathas even a
dollop of sensitivityto changed circumstances- which is to say most such
accounts, except perhaps for the most stringentlybackward-looking
forms of originalism- will code the questions that the Justices ask in
these two cases as criticallydifferent.
Finally,and most generally,in a regime of precedent there will nec-
essarilybe a differencebetween the questions asked in cases of firstim-
pression and in all subsequent cases. At Time 1 the question is whatlegal
rule the court should adopt; at Time 2 (or 3 or N) the questions include
how much weightto give to the Time 1 precedent. Because the question
at Time 1 alwaysdiffersfromthe question(s) at Time 2 (or Time N), one
cannot straightforwardly aggregateinformationacross the divide between
examination and reexamination of legal questions. Adherence to a re-
gime of precedent itselfundermines the commonalityof questions thatis
necessaryfor the Condorcetian interpretationof precedent.
One must not overstate these points, because a slight shiftin the
relevantquestions does not deprive previous precedents of all informa-
tional value to current decisionmakers. Suppose that a decade ago,
judges decided that a particularpractice did not violate widespread stan-
dards of decency. Suppose also thatjudges today must decide whether
the same is true today,under an evolvingstandard. It is surelyrelevant
informationto know that as of a decade ago, decency did not condemn
the practice. But we cannot straightforwardly rely on the miracle of
Condorcetian aggregation in a case like this, because we cannot lump
together in a single notional voting group two differentsets of judges
deciding two differentquestions.
The Condorcetian interpretationof Burke trades on a kind of meta-
phor, one not dissimilarto Burke's own description of traditionas the
"bank and capital of nations and of ages."40 In the implicitmetaphor of
the common law constitutionalists, theJusticeswho have participatedin a
38. Trop v. Dulles,356 U.S. 86, 101 (1958).

39. Palkov. Connecticut,302 U.S. 319,325 (1937), overruledbyBentonv. Maryland,
395 U.S. 784 (1969).
40. Burke,supranote 1, at 74.
line of decisions are thought of as deciding, in common, on the same

question. But in manycases, no meeting of the manyminds ever occurs.
The metaphor here is affirmatively harmful. It obscures that the many
minds participatingin the line of precedent answer questions that shift
subtlyover time with changing circumstances,changed legal contexts,
and novel applications of preestablished legal rules.
4. Independence,Herding,and theBurkeanParadox. - Another critical
condition for theJuryTheorem to apply (in its strongestform;I will ex-
plain the qualifiershortly)is thatthe votes or guesses thatare aggregated
must be independent of each other.41 For brevityI will speak of the "re-
quirement" of independence, but here too one must be careful.
Nonindependence just reduces the number of effectivevotes or guesses
in the pool. If there are ten voters,and nine utterlydefer to a leader,
guessing exactlyas the leader guesses because the leader does so, then
only one vote counts.
This need not mean that the average competence of the group de-
creases. If the leader is of much higher average competence than the
followers,it mayeven increase, at least ifthe followerswould otherwisebe
more likelyto guess wrong than right. Group performance under the
JuryTheorem is a functionof both numbers and competence; deference
to relativelycompetent opinion leaders can improve performanceunder
some conditions.42 Nonetheless, under a broad range of conditions,the
more independence, the more separate judgments and dispersed infor-
mation are aggregated into the group judgment, the better the Con-
dorcetian credentials of the precedent or tradition.
This criterionshould not be applied too stringently.Independence
is not violated bymere deliberation,orjust because people influenceone
anothers' views. The independence required by the JuryTheorem is a
strictlymathematicalnotion.43 It means thatVoter 2 does not decide to
votejust as Voter 1 votes, however that is; it does not mean thatVoter 1
cannot influenceVoter 2's independentjudgment, throughdiscussion or
otherwise. Deliberation by itselfdoes not compromise independence.
That said, however, true failures of independence can arise from
many causes. One cause is power- the power of ruling regimes,interest
groups,or even court leaders. Suppose thatthe precedents of a given era
were generated, not by statisticallyindependent considerationby a panel
ofJustices,but by a panel ofJusticeswho were political cronies following
the lead of the White House, or the lead of outside groups, or who (like
the Justicesof the Marshall Court) systematically deferred to the Chief
Justice. The reduction of independence reduces the number of effective
votes. When we turn to comparative institutionalproblems later in the
discussion,the role of power will become important: Argumentsthatim-
41. See Estlundet al., supranote 32, at 1326 (Jeremy

Waldron).
42. See generallyDavid Estlund,Opinion Leaders,Independence,and Condorcet's
Theorem,36 Theory8cDecision 131 (1994).
Jury
43. See Estlundet al., supranote 32, at 1327 (Jeremy
Waldron).
peach the Condorcetiancredentialsof otherinstitutions, such as legisla-

on
tures, grounds of interestgrouppower or other processfailuresmust
considerthe same possibilities withrespecttojudicial institutions.
An important specialcase in whichindependenceis violatedinvolves
informational cascades.44Such cascades can arise even if all actorsare
rationaland are attempting to make the best possibleindividualguess.
Whereeach personhas a smallstockof privateinformation, therational
guesser willdo best to mimic the of
judgments guessers earlier in the
sequence, because their totalinformation swamps his own. The result
can oftenbe a bad equilibriumin whicha long sequence of people, de-
ferringto erroneousguessesearlierin the sequence, make erroneous
guessesthemselves.45 This outcomeis collectively bad but individually
rational;each personfollowedthe best strategy forthemselves, but the
outcomeforall is a bad cascade.
Cascadescan occurin courtsas wellas in societygenerally; a line of
precedentsmayrepresentlittlemore than a rationaldecisionby later
judges to ignoretheirprivateinformation in favorofwhatearliercourts
havesaid. Wherethisoccurs,laterdecisionsin theline ofprecedentare
notindependentcontributions thatadd to theinformational valueofthe
whole; althoughdecisionsaccumulate,the information implicitin the
line of precedentsdoes not increase. Althoughsome argue on a priori
groundsthatprecedentialcascadesare unlikely,46 theyhavein factbeen
documentedunderquite ordinaryconditions.47
In somesettings, cascadesare fragileto smallchangesin information
and motivations; althougheasilystarted,theyare easilyshatteredas
well.48In markets, cascadescan oftenbe dispelledbyreleasesofinforma-
tionfromgovernment or firms.Cascadesembodysurprisingly littleinfor-
mation; afterthe firstfew guessers,all othersare rationally
following the
herd. Hence a smalladditionofinformation can cause actorslaterin the
44. For an overview,see Sushil Bikhchandani et al., Learning from the Behavior of
Others: Conformity,Fads, and InformationalCascades, J. Econ. Persp., Summer 1998, at
151 [hereinafterBikhchandani et al., Behavior of Others].
45. For a simple example, see Sushil Bikhchandani et al., InformationalCascades and
Rational Herding (June 1996), at http://welch.econ.brown.edu/cascades(on filewiththe
ColumbiaLaw Review).
46. Eric Talley, Precedential Cascades: An Appraisal, 73 S. Cal. L. Rev. 87, 92 (1999)
(concluding that necessaryconditions forprecedential cascades are "implausible" and that
"it is extremelydifficult... to verifywhetherobservedjudicial conformityis the byproduct
of a cascade or of some omitted thirdfactor that commonly affectsall judges").
47. Andrew F. Daughety & Jennifer F. Reinganum, Stampede to Judgment:
Persuasive Influence and Herding Behavior by Courts, 1 Am. L. & Econ. Rev. 158, 161-65
(1999) (describing precedential cascade involving retroactive reach of Coal Industry
Retiree Health BenefitAct of 1992).
48. See generallyCristinaBicchieri & Yoshitaka Fukui, The Great Illusion: Ignorance,
InformationalCascades, and the Persistence of Unpopular Norms, 9 Bus. Ethics Q. 127,
147 (1999) (concluding that "it may take surprisinglylittle new public informationto
reverse the original cascade").
sequenceto changetheirguesses,breakingthecascade.49Ifthisfragility
to newinformation workedstraightforwardly in the case of precedential
cascades, one might have reason to think that mostlines of precedent
represent the informational contributions of manyminds,ratherthana
cascade.
However,precedentialcascadesare probablyrelatively robust.The
reason is that"the passivenatureof courts. . . means thaterroneous
decisionsmade byone levelin a systemof courtspotentially remainun-
corrected,even thoughall courtsare tryingto make principledjudg-
ments."50If the SupremeCourtfollowsitsusual practicesof case selec-
tion,it will not grantcertiorariin cases where all lower courtshave
cascaded to an agreement,and so willhave no occasionto overturn the
lowercourtconsensus.Litigants, seeingthe lowercourtconsensusand
knowingthatSupremeCourtreviewis unlikely, are unlikelyto challenge
theconsensusin thefirstplace. Byvirtueof thestructure of thejudicial
system and thepassivity of thecourts,precedentialcascadesare plausibly
less fragilethancascadesin marketsor societygenerally.
Changesin motivation or incentives- in therewardstructure ofindi-
vidualaction- can also dispelcascadesor preventtheirarisingin thefirst
place. Whereeach personis attempting, not to make the bestpossible
individualguess,but to maketheguessthatwillprovemostusefulto the
group,each guesserwillignorethe earlierguessesand recordher un-
aided viewof theproblem.51Whenthishappens,each guesserdiscloses
moreprivateinformation to be aggregatedintothegroupjudgment,and
the qualityof the groupjudgmentimproves.The same effectcan arise
notonlyfromincentives, butalso fromcognitive quirks.Irrationallyover-
confidentindividualscontributea greatdeal to the group;preciselybe-
cause theyare overconfident, theytendto contribute privateinformation
thatcan help to blockor shattercascades.52
The significance ofall thisis to reveala tensionat thefoundations of
the Burkeanview- whatI willcall the Burkean paradox. The paradoxis
thatif manyparticipants in the line of precedentor traditionfollowed
theprecedentor tradition(ratherthanexercisingtheirindependentrea-
son) because doingso wasa wayto conserveon decisionmaking costsor
improve their information, then the informational value of the line of
precedent or traditionis lower to that extent; there are fewer indepen-
dentmindscontributing to thecollectivewisdom.The line ofprecedent
or tradition itselfbecomesa kindof informational cascade,whichlowers
the totalinformational value of the whole. The stingin the problemis
49. Bikhchandani et al., Behaviorof Others,supranote 44, at 157-58.

50. Daughety& Reinganum,supranote 47, at 159.
51. Cass R. Sunstein,WhySocietiesNeed Dissent66-70 (2003).
52. Cf.AntonioE. Bernardo8cIvo Welch,On the Evolutionof Overconfidence and
Entrepreneurs,10 J. Econ. & Mgmt.Strategy 301, 326-27 (2001) (showingconditions
underwhichirrational overconfidence ofentrepreneursin notfollowing actionsofothers
improvesqualityof aggregatedinformation).
that a strategythat is individuallyrational forjudges at any given time-

followingprecedent- is harmfulto all iffollowedby all because it drains
precedent or traditionof any epistemicvalue.53 Burkeanismmighteven
be analogous to a kind of informationalpollution, givingrise to severe
problems of collective action: It is individuallyrational but collectively
harmfulfor individual decisionmakersto relyon precedent or tradition,
ratherthan their unaided reason.
Obviously much depends on the specifics of the cascade at issue.
Some cascades can cause stampedes in the rightdirection,not the wrong
direction (according to whateverexogenous definitionof rightanswers
we have assumed). And cascades are fragile in many ways,tending to
shatteras circumstanceschange. But to the extent thatprecedent or tra-
dition reston cascades, ratherthan the independent and unaided contri-
butions of manyminds, Burkean praise forprecedent or traditionis self-
defeating. The best contributionsto the streamof precedent or tradition
are those in which individualjudges, or small groups ofjudges, do exer-
cise theirunaided reason. Those who relyon precedent or traditionon
the ground that it is the "bank and capital of ages" make withdrawals
from the common pool of information,for their private benefit; those
who exercise their unaided reason contributeto the common pool, for
the good of all futurejudges.
5. Randomor Correlated Biases.54- Another strandof Bentham's cri-
tique of the common law centered on the professional self-dealingof
"Judgeand Company."55 The judges and lawyerswho administeredthe
common law of England, on thisview,constitutedan elite guild who ar-
ranged the legal systemin waysthatbenefitedthe legal profession,to the
harm of the polityat large. Bentham interpretedthisclaim bydiagnosing
Judge and Company as promotersof their"sinisterinterests]" (a motiva-
53. In principle,judges mightpreventthe Burkeanparadoxfromarisingbyissuing

opinionsthatrecordtheirindependent judgmentswhilebasingtheiractualdecisionson
precedent.In thismodel,latercourtswillhaveboththeprecedentbased decision,which
is of lowor nil informationalvalue,and theearliercourt'sindependent judgment,which
has positivevalue. Althoughthismaysometimeshappen,I believeit is rare,forbothan
economicreasonand a psychological one. The economicreasonis thatjudges decide on
thebasisof precedentin partto conserveon thecostsof decisionmaking forthemselves.
To incurthecostsofforming an independentjudgmentthat,byhypothesis, willnotaffect
thecurrentdecisionis to providea benefitonlyto futurejudges and is thusa contribution
to an informational publicgood; manyjudges willfreeride,and theinformational public
good willbe underproduced.The psychological reasonis thatit is hard to keep one's
precedent-independent judgmentfrombeing infectedby whatmanyjudges and other
people in the past have said. Althoughthe claim thatthe viewsof people in the past
constituteour presentidentityis mysterious, the claim thatthose viewsstructureour
presentthinking is not.
54. Partsof thissubsectiondraw upon a similardiscussionin AdrianVermeule,
ShouldWe Have LayJustices?, 59 Stan.L. Rev.1569,1589-90 (2007).
55. See JeremyBentham,Introductory Viewof the Rationaleof Evidence,in 6 The
WorksofJeremyBentham22-24 (JohnBowringed., London, Simkin,Marshall& Co.
1843).
tional distortion)or as in the grip of "interest-begotten prejudice" (a cog-

nitivedistortionarising froma motivationaldistortion).56
However,we mayalso interpretBentham's core concern in less sinis-
ter terms,and in a way that fitscomfortablywithinthe frameworkof the
JuryTheorem. In this interpretation,the decisions made byJudge and
Company are less accurate than might otherwise be the case because
judges sufferfrom correlated biases. We must distinguishtwo issues, com-
petence and the directionof error. Competence is the likelihood thatan
individual or an average group member will err. But if group members
err,it is best if theirerrorsare uncorrelated- if theyare as likelyto err in
one direction as another. Where error is randomlydistributed,biased
guesses in either direction wash out, on average, and true guesses
prevail.57
Random distribution of bias is a major force behind the Jury
Theorem. Where errorsare correlated- where group members are sub-
ject to a common and systematicbias thatcauses them to err in only one
direction- group performance can fall dramatically. The chance of a
majorityselectingthe rightanswervaries inverselywiththe correlationof
bias across the group.58 This implies that under plausible conditions, a
group of lower average competence, but with uncorrelated biases, will
outperforma group with higher average competence but highlycorre-
lated biases. Counterintuitively, the effectof correlatedbiases is so strong
that addingworse-than-random guessersto a groupcan improve overallperform-
ance if theirbiases are negatively correlatedwiththoseof highlycompetent ex-
perts.59 Here too, of course, everythingdepends upon the precise num-
bers, but the crucial point is that u[t]he' uninformed voters drive the
average correlationdown, thusmore than compensatingfortheirrelative
ignorance."60
From a Benthamiteperspective,the problem is thatJudge and Com-
pany are likelyto have highlycorrelated biases, arisingboth froma com-
mon socioeconomic background and from their common professional
training. On one view,"[m]ost lawyers,whatevertheirbackground,have
56. Bentham,Handbook,supranote 30, at 34-36.

57. There is a similarity ofjudges
here to Cardozo's idea that"[t]he eccentricities
balance one another"so that "out of the attritionof diverseminds there is beaten
somethingwhichhas a[n] . . . averagevalue greaterthan its componentelements."
BenjaminN. Cardozo,The Natureof theJudicialProcess177 (1921).
58. See Krishna K. Ladha, The CondorcetJuryTheorem, Free Speech, and
CorrelatedVotes,36 Am.J. Pol. Sci. 617, 625-30 (1992).
59. See id. at 628-29.
60. Id. at 629. In Ladha's example,a groupof threedecisionmakers each competent
at the .8 level cannotbe certainof gettingthe rightanswerwhateverthe votingrule.
However,if the group adds two uninformed,worse-than-random guessers,whose
competenceis a mere.3, thenthegroupis certainto choose therightanswerso long as
theguessesof theuninformed are negatively
correlatedwiththoseof theinformed.See
id.
a narrow,professionally inflectedperspectiveon governance."61Even

morebroadly,thevaluesoflawyers as lawyers oftenhavethe"smellofthe
lamp about them."62 As compared either to ordinary people or to other
professionals, are
lawyers generally more introverted, rationalas op-
more
to
posed emotional, more judgmental, more competitive, aggressive,and
materialistic;63they also have distinctivepolicy biases.64
Suppose thatlawyers do have correlatedbiases. Suppose,forexam-
ple, thatlawyers - and hencejudges- are likelyto systematically erron
the high side when theyattemptto estimatethe costsand benefitsof
additionallegal processin governmentdecisionmaking, and thuswill
tendto requiregovernment to affordextraincrements of processwhose
social costsare greaterthan theirsocial benefits.65As comparedto a
groupwhoseerrorsare randomlydistributed, the guessesof a groupof
past lawyers- thejudges decidingthe cases now relied upon as prece-
dent- are less likelyto trackrightanswersabout the value of process.
This reducesthe informational value of the precedentsto future judges
who are trying to drawaid frommanyjudicial mindsin the past.
The keypointis not thatjudges are likelyto get thingswrong;it is
thatwhen theydo get thingswrong,theyare likelyto errin systematic
ratherthanrandomways.The correlationof biasesacrossthe decision-
makinggrouptradesoffagainstthecompetenceofthegroup'smembers.
Even ifjudges are elite experts,of veryhighaveragecompetence,their
likemindedness reducesgroupperformance overall,narrowing the mar-
of
gin superiority between the performance singlejudge groupof
of a or
judges exercising theirunaided reason today,and a judge or groupof
on
judges drawing precedent. This is an updatingof Bentham'scritique
of legal and judicial professionalism: Ratherthanpointto motivational
distortions on the partof lawyersand judges, one mayalso claim that
commonprofessional trainingreducesthe informational qualityofjudi-
cial decisionmaking, even ifalljudges and lawyers are sincere.
61. RichardA. Posner,Law,Pragmatism, and Democracy128 (2003).

62. JohnHartEly,Democracyand Distrust:A TheoryofJudicialReview59 (1980).
63. See Susan SwaimDaicoff,Lawyer,KnowThyself40-42 (2004).
64. Three relevanthypotheses, in roughlydescendingorderof rigor,are that: (1)
legal traininggiveslawyersa strongstatusquo orientationand a bias to conventional
morality, educatedadults,see LawrenceJ. Landwehr,Lawyers
as comparedto similarly as
Social Progressivesor Reactionaries:The Law and Order CognitiveOrientationof
Lawyers, 7 Law & Psychol.Rev.39, 44-48 (1982); (2) legal trainingreduceslawstudents'
generalconcernforsocialjustice,see Susan Ann Kay,Socializingthe FutureElite: The
Nonimpactof a Law School,59 Soc. Sci. Q. 347 (1978) (notingthishypothesis, but then
conductingstudythatsuggeststhatlawschooldoes notaffectstudents'outlookon social
in publicserviceor publicinterest
justice),and reducestheirinterest lawyering,see Robert
Granfield & ThomasKoenig,LearningCollectiveEminence:HarvardLawSchooland the
SocialProductionofEliteLawyers, 33 Soc. Q. 503,517-18 (1992); (3) legaltrainingcauses
lawyersto favorcumbersomeand complicatedprocessesforgeneratingpolicy,see, e.g.,
leroldS. Auerbach,A Plagueof Lawyers, Harper's,Oct. 1976,at 37, 40-42.
65. Cf. Auerbach,supra note 64, at 41-42 (suggestingthatlegalizationof society
on individualproceduralrightsto detriment
places priority of substantive
norms).
C. Burke,Condorcet,
and Constitutional
Law: Comparative
Institutional
Problems
So farwe haveexaminedsomeproblemsthatarisewithinthelogicof
theCondorcetianinterpretation ofBurke,bothas appliedto socialtradi-
tionsin generaland as appliedtojudicial precedentin particular.Here I
willturnmorespecifically to problemswithconstitutional precedentand
tradition.In both cases, comparativeinstitutional problemsbecome
critical.
A judge or groupofjudges decidingconstitutional cases at a given
timemayturnto the information embodied in precedentto improve
theirdecisionmaking.The Condorcetianinterpretation of Burke de-
scribesthe conditionsunderwhichusingprecedentin thiswaywillim-
proveupon the unaided reason of currentjudges. Yet in the constitu-
tional setting,there are always alternativesources of information
available:thejudgmentsofpastconstitutional framers, ofcurrentlegisla-
tures,and of the currentexecutive. Evenif thecollectivewisdomofmany
past judges is superior to that of one or a few presentjudges, on
Condorcetiangrounds,itdoes notat all followthatthecollectivewisdom
of manypastjudges is superiorto thatof otherinstitutions.
The same basic analysisapplies to traditions, albeitwithmodifica-
tions. Social traditionsstandon somewhatbetterepistemicgroundthan
butwheresuchtraditions
judicial traditions;
strictly are refracted through
judicial decisions,theirCondorcetiancredentialscan be no betterthan
thatofthejudges themselves, whomustidentify and applythetraditions.
On the otherside of the equation,legislatorstoo can drawupon social
traditions,and will have Condorcetianadvantagesin doing so. Ulti-
the
mately, same comparative institutionalproblemsafflictcommonlaw
constitutionalism whetherthe focusis on precedentor tradition.
1. PastJudgesand Framers. - In manycases,the principaltargetof
commonlaw constitutionalists is originalism,whichis the viewthatthe
Constitution shouldbe interpreted accordingto itspublicmeaningat the
timeof enactment.66 As againsttheoriginalist view,commonlawconsti-
tutionalistsoffera visionofcasesunfolding overtimeand embodying the
aggregatedwisdomof manyminds. Commonlaw constitutionalists see
thetextas a starting pointfora processofincremental traditionalism,in
Burkeanfashion. In this opposition,originalismsuffersfroma dead
hand problem - whyshould the viewsof past framerscontrolthe liv-
-
ing? while common law constitutionalismis a form of living
constitutionalism.
But whichapproachis betteron informational grounds?Here we
mustdistinguish originalismas a first-best
view from originalism as a sec-
ond-orderdecisionmaking strategy. On the former version,originalists
66. I take this"publicmeaning"originalismto have become the standardversion,

supersedingthe older "framers'intentions"versionof originalism.The differences
betweenthemmakeno difference to theanalysishere.
claim thatthe original understandingjust is whatthe Constitutionmeans,

perhaps what any writtenconstitutionnecessarilymeans. To engage in
interpretationis, necessarily,to ask about what the framersor ratifiersof
a document understood it to mean.67 On the latterversion,originalism
is a strategyof second-order rationality, just like common law constitu-
tionalism. "Given a sufficiently good constitutional text,originalistsmain-
tain that better resultswill be reached overall if governmentofficials -
includingjudges - must stickto the originalmeaning rather than empow-
ering them to trump that meaning with one that they prefer."68Here
too, differentoriginalistsmighthave differentaccounts of whatfirst-order
value- welfare,justice, and so on- is served by this strategy;originalism
would representa lower-levelconvergingagreement across these camps
on a method of constitutionalinterpretation.
The relevantpoint, however,is thatoriginalistssuggestthatthe high-
est source of latentwisdom is the textand original understanding.69For
second-orderoriginalists - those who thinkthatoriginalismis a good sec-
ond-orderstrategy for constitutional adjudication, ratherthan an account
of the essence of interpretation - a keyclaim is that textand original un-
derstanding are informationallysuperior to other sources of law that
judges mightdraw upon in constitutionalcases. Below, I will offersome
reason to doubt the truthof this claim, but it is theoreticallycrucial. It
emphasizes that common law constitutionalistsmust consider the infor-
mational credentials of all relevant sources of law before declaring in
favorof a regime of strongprecedent.
We must be clear about what comparison is being made, exactly.
Sometimes,perhaps often,original textsand understandingsare ambigu-
ous or general; in such cases common law constitutionalists invoke prece-
dent to supplement the original understanding,rendering it sufficiently
specificto determine the outcome of litigatedcases. When thisis so, the
informationalcontent of originalismand the informationalcontent of
common law constitutionalismare additive; text,historyand precedent
are supplements,not competitors.
Still,this is not alwaysso. The testingcase for both originalismand
common law constitutionalismoccurs when (let us stipulate) there is a
clear original understanding,a clear line of precedent,and the twoare in
conflict. Should the views of framersand ratifiersor the views of later
67. See GaryLawson,On ReadingRecipes. . . and Constitutions,85 Geo. L J. 1823,

1833-34 (1997); SaikrishnaPrakash,Radicalsin TweedJackets:WhyExtremeLeft-Wing
Law Professors Are WrongforAmerica,106 Colum. L. Rev. 2207, 2224 (2006) (book
.
review)
68. Cass Sunstein& RandyBarnett,Constitutionin Exile?,LegalAffairs
DebateClub,
(on filewith
May3, 2005,at http://legalaffairs.org/webexclusive/debateclub_cie0505.msp
the ColumbiaLaw Review).
Some Originalist
69. StevenG. Calabresi,Text,Precedent,and theConstitution: and
Pennsylvaniav.
Normative Arguments for Overruling Planned Parenthoodof Southeastern
Casey,22 Const.Comment.311, 345-46 (2005) ("[T]he SupremeCourtoughtto follow
textoverprecedent").
theconstitutional
judges prevail? Let us suppose that, despite the internal problems sur-
veyed above, a streamof precedents does embody aggregated wisdom or
informationthat is superior to that of a singlejudge or small group of
judges today. Still,it is a verydifferentquestion whetherthat stream of
precedent is informationally superior to the viewsof framersand ratifiers
in the past.
We will assume that the necessary conditions for a Condorcetian
analysis are satisfied;in other words, that all members of the relevant
groups (framers,ratifiers,and laterjudges) were addressing a common
question, voted sincerely,and so on. What is the relativeaverage compe-
tence of the decisionmakinggroups being compared? Which group will
be more likelyto answerthe relevantquestions correctly?From an infor-
mational perspective,there are many difficultissues to be untangled, be-
cause the tradeoffsare numerous and cross-cutting.
On Benthamitegrounds,but contraryto Bentham's own prejudices,
pastjudges have an informationaladvantage over framersand ratifiersof
the even more remote past, all else equal. This is a centralpoint of com-
mon law constitutionalism:Post-framing judges have had the opportu-
nity to learn from experience70 or to acquire new informationabout the
costs and benefitsof framing-erachoices. In Condorcetian terms,more
informationyields higher average competence.
However, other considerationsare relevantas well. On some views,
the Framerswere superior statesmen,possessed of extremelyhigh com-
petence that mightcompensate or more than compensate for theirrela-
tive lack of information. One might explain this higher wisdom of the
Framers (if it exists) either by a selection mechanism or by referenceto
the circumstancesof the founding era. As for selection, one mighthold
that the Framers were the elite of the nation, selected fromamong the
notables of their states, as opposed to laterjudges who often took the
bench through cronyism or party service. As for circumstances,one
might believe that emergencies and national crises produce sharpened
cognition in lawmakers,because the costs of making mistakesare higher,
and that crises produce emotional states- of solidaritywith contempo-
rarycitizensand withfuturegenerations- thatinduce more effortforthe
public good.
Even if neither selection nor unusual emotional states sufficeto in-
duce extraordinarycompetence in constitutionalframers,the high costs
of informationin foundingeras can help along other Condorcetian mar-
gins. Behind the "veil of uncertainty,"lack of informationmakes it
equally easy to err in one direction or another.71 Sincere framersat-
70. Cf. Strauss,Common Law Genius,supra note 27, at 8 (arguingthat under

commonlawapproachone can breakwithpastifone can "showthatthe . . . old regime
did notworkverywell").
71. John C. Harsanyi,Moralityand the Theory of Rational Behaviour, in
Utilitarianismand Beyond 39, 44-46 (AmartyaSen & BernardWilliamseds., 1982)
modelof moraljudgments).
(explainingequiprobability
temptingto guess the welfare-maximizing constitutionalrule will know

less than later generations,but will also be less biased, preciselybecause
theyare less likelyto know the rule's distributiveconsequences for them-
selves, their friends,family,profession,or social class. There is thus a
tradeoffbetween the framers'relativelack of informationabout the con-
sequences of the rules theyadopt, on the one hand, and their relative
impartiality, on the other.72 In the JuryTheorem's terms,lack of infor-
mation reduces competence, but pervasiveuncertaintyreduces the corre-
lation of biases and pushes the whole group towardrandomlydistributed
error,which is desirable.
Furthermore,there is a crucial problem of numbers,rarelydiscussed
by common law constitutionalists.Burke did not clearlydistinguishtradi-
tion fromjudicial precedent, but we have seen that a major distinction
between the two is that the numbers involved in the latter are much
smaller, especially at the Supreme Court level. The problem bites hard
when a Condorcetian interpretationof precedent is compared to a
Condorcetian interpretationof originalism. The Framers and ratifiers,
taken as a group, were a large body,even in the small population of eight-
eenth centuryAmerica. To be sure, theywere also an internallyhetero-
geneous group, and it seems plausible that the average competence of
the Framerswas higher than the average competence of the ratifiers.But
so long as the latterwere more likelyto get the answerrightthan wrong,
on average, theirnumbersvirtuallyguarantee that theywould have done
so. Witha votingpopulation of,say,one thousand framersand ratifiers, a
very low average competence (so long as it is greater than .5) suffices
to
all but guarantee a correctanswer.
No such guarantee is possible for the relative handful of Justices,
slightlymore than one hundred, who have ever sat on the Supreme
Court; and of course forany givenprecedent, the number ofJusticeswho
have ever voted in the relevantsequence of cases is much smaller than
that. This comparison looks betterif the lens is expanded beyond prece-
dent, properlyso-called, to include lower courtjudges, academics provid-
ing background commentary,and other participantsin the ongoing legal
communitythat surrounds the Court. That expansion, however,is itself
problematic on several grounds. The firstis that it strainsthe require-
ment of a common question. Lower courts, for example, often do not
ask the same question as the Supreme Court, preciselybecause theyare
lower courts and thus have a hierarchical obligation to respect the
Court's precedent, an obligation with no analogues on the Court itself.
The second is that behind the Framers and ratifiersthere also stand
broader intellectualcommunities;whyshould not their"votes"be aggre-
gated as well?
DesignWritSmall
ofDemocracy:Institutional
72. See AdrianVermeule,Mechanisms
Ch. 27-54 (2007).
In the hardest case, where subsequent precedent contradicts text

and original understanding,it is not obvious thatframersand ratifiersas
a decisionmaking group are inferior,along dimensions of information
and effectiveaggregation,to a series of laterjudges. Everythingdepends
on particulars,and few generalizations are possible. This means that if
the main targetof common law constitutionalismis originalism,and the
main weapon for strikingthe targetis the Condorcetian theoryof infor-
mation aggregation,then the strikelargelymisses its mark.
None of this is intended to defend originalism,which is presump-
tivelyobjectionable on Benthamitegrounds: All else equal, changing cir-
cumstancesdemand that decisions be made not by the framersof centu-
ries ago, but by decisionmakers with better current information,
especiallyas the founding era and the last major episodes of formalcon-
stitutionalrevision (Reconstructionand the Progressiveera) recede into
the distantpast. The only point here is that common law constitutional-
ism has yet to come to grips with the problem that an informationalde-
fense of precedent must consider, and compete with,an informational
defense of text and original understanding.
2. PastJudgesand Current and Thayer.-
Legislatures:Burke,Condorcet,
A differentcomparison is between (1) a line of judicial precedent that
announces some rule and (2) a statutethat contradictsthat rule. In or-
der to bias the inquiryin favorof common law constitutionalism,I will
assume thatwhen the statuteis enacted there is already a clear and con-
sistentline of Supreme Court precedent to the contrary,and that the
underlyingconstitutionalprovisions are ambiguous, vague or general.
This must be the best case for common law constitutionalism,insofaras
overridingthe view of currentlegislaturesis concerned.
However, in cases of this type, the same Condorcetian arguments
that common law constitutionalistsdeploy against originalismalso sup-
port judicial deference to the views of current legislatures. Precedent
mightbeat original text,but be beaten in turnby currentstatutes. If that
is so, then the logical consequence of the informationalinterpretationof
Burke is not robustjudicial review,in the style of the Warren Court;
rather it is deference to legislativejudgments. On this account, Burke
leads via Condorcet toJames BradleyThayer,who argued forjudicial def-
erence to current legislaturesin all but the clearest cases of legislative
mistake.73
To make progress on these issues, let us temporarilystipulate to
some crucial assumptions that I shall examine in the subsequent discus-
sion. First,suppose that the legislatorswho enact the statute and the
courts who reviewit both address the common question of its constitu-
73. See JamesBradleyThayer,The Originand Scope of the AmericanDoctrineof

ConstitutionalLaw, 7 Harv. L. Rev. 129, 144 (1893) (arguingthat courts"can only
disregard[an] Actwhenthosewho have the rightto makelawshavenot merelymade a
mistake,but have made a veryclear one,- so clear that it is not open to rational
question").
tionality(whetheror not the legislatorsalso address its meritsas policy).

Second, suppose that the legislatorswho voted for a statuteare on aver-
age more likelyto get that question rightthan wrong,as are thejudges
who reviewit. Third, suppose that both legislatorsand judges sincerely
vote their views of the statute's constitutionality.Obviously all of these
assumptionsoftenfail to hold, but thatis irrelevant.Legislatorsmay not
vote sincerely,butjudges maynot either. Theoristswhojustifyprecedent
on jury-theoreticgrounds assume sincere votingamong judges, so we will
do the same among legislators,in order to be able to compare apples
withapples.
Given these assumptions, several stylized facts about legislatures
threaten to give them an overwhelmingCondorcetian advantage: the
sheer numerosityof theirmembers,theirvotingpractices,and theirpow-
erfultools foracquiring information.I will take up each of these in turn,
beginning with numbers. It is highly likely that legislatureswith hun-
dreds of members proceeding under majorityrule will vote correctly,
even if their average competence is near-random,just because of their
sheer numbers. Witha mere 301 voters- a smallergroup than the House
of Representatives - and an average competence of only .6, a majoritywill
be almost certainto choose correctlybetween twooptions.74 In compara-
tive terms,even if the average competence of Justices is much higher
than that of legislators,it is extremelyunlikelythat the nine Justicessit-
tingat any one time,or even the fewdozens ofJusticeswho participatein
a line of precedent,can do betterthan a highlynumerous legislature,just
because it is hard to improveon a vote thatis almost certainto be correct.
Suppose that the Justicesare much betterlawyersthan, say, the average
senator or even than the members of the judiciary committees. Not
much follows. The distinctively Condorcetian contributionis that num-
bers matter;manynear-incompetentscan do betterthan a small panel of
elite experts.
Quite obviously,the crucial question here is whetherlegislators'av-
erage competence is greater or less than one-half (in the two-option
case). I put aside the argument that legislatorsmight alwaysraise their
competence to .5 by flippinga coin; this overlooks that an incompetent
voter might not realize his own incompetence. The more serious prob-
lem is that, as Condorcet himselfsuggested, average legislativecompe-
tence mightslip below the crucial thresholdof one-halfpreciselybecause
of the large number of members. "A verynumerous assemblycannot be
composed of veryenlightenedmen. . . . [T]he more numerous the assem-
bly,the more it will be exposed to the riskof making false decisions."75
However, we should not overstatethe problem. We need only enough
confidence in the mechanisms of legislativedecisionmaking to believe
74. See List & Goodin,supra note 18, at 287 tbl.l. If averagecompetencein the
groupof 301 fallsto .51, barelybetterthanrandom,the majority's chance of choosing
correctly percent.See id. at 287.
is aboutsixty-four
75. Condorcet,Applicationof Mathematics, supranote 17, at 49.
thatlegislators are on averageonlyslightly betterto gettherelevantques-

tionrightthanwrong.If theundemandingthresholdis met,thennum-
berswilltell.
Legislativeinstitutions help to push individuallegislatorsup to this
threshold.Beginwithvotingpractices.In mostlegislatures on mostis-
sues,votingoccursrelatively simultaneously.76 Benthamobservedthat
the simultaneity of legislativevotingdampened "undue influence,"be-
cause legislators votein ignoranceof howotherlegislators have actually
voted.77Bycontrast, whereall thejudges whoparticipated in a streamof
precedentsovertimeare aggregatedintoa notionalvotinggroup,later
judges knowhow earlierjudges votedand can deferto theirpredeces-
sors- the sortof rationalimitationthatyieldsinformation cascades.
Mostimportant are theinstitutional
determinants oflegislative infor-
mation.Byvirtueoftheextremeinternalspecialization ofthecommittee
system, large legislativestaffs,and the professionalism of the legislative
career (at thefederalleveland in manystates,thoughnot all), modern
legislatures are enginesforgeneratingand processinginformation. The
senioritynorm and the proliferationofsubcommittees imply that a mem-
ber of the modernHouse of Representatives spendsmanyyearsbecom-
ingdeeplyexpertin a narrowsliceofpublicpolicy.78This has twocross-
cuttingeffects.On the one hand,specializationraisesthe averagecom-
petenceofmembers, includingmemberswhohaveinfluenceon constitu-
tionalquestions,such as membersof thejudiciarycommittees.On the
otherhand, floordeferenceto the committeesreducesthe numberof
effectivelyindependentvotes. The upshotis a smallernumberof more
competentvotes,and the resultis unclear. But it is not implausibleto
thinkthataveragecompetencein a modernlegislature,highlyprofes-
sionalizedand specializedbyvirtueofan elaborately reticulated commit-
tee system, is muchhigherthanin legislatures of theeighteenth century.
So too,as comparedto courts,legislators plausiblyhavebetterinfor-
mation thanjudges about the factualcomponentsand causal conse-
quences of theirconstitutional decisions. Here the main tradeoff is be-
tween evenhandednessand information.Judicial procedures are
designedto ensureequalityofinputs,a formofevenhandedness.Thisis
basicallya leveling-down strategyof institutionaldesign: By virtueof
76. When the House leadershipheld open votingforthreehoursto allow timeto

roundup membersto votefora controversial prescription drugbill in 2003, an outcry
resulted.See John K. Iglehart,The New MedicarePrescription-Drug Benefit- A Pure
Power Play,350 New Eng. J. Med. 826, 828-29 (2004) (describinghow "embittered
Democrats"denouncedvoteas "a travesty").
The newDemocraticCongresselectedat the
end of2006limitedthispracticebyprovidingthatvotesmaynotbe heldopen "forthesole
purposeofreversing theoutcome."See House Rule XX, cl. 2(a), availableat http://www.
rules.house.gov/ (on filewiththe Columbia
ruleprec/house_rules.htm Law Review).
11. JeremyBentham,PoliticalTactics106-07 (MichaelJameset al. eds., 1999).
78. See Keith Krehbiel,Informationand LegislativeOrganization142-43 (1991)
(arguingthatdivisionof labor into committeescombinedwithseniority systemfosters
expertise).
prohibitionson ex parte contacts and equal time for briefingand argu-

ment,no partyhas an advantage in presentinginformationto thejudges.
Moreover, the lack of any electoral connection on the part of federal
judges gives them a remoteness fromcurrentpolitics. The price of this
evenhandedness and remoteness,however,is a relativedearth of informa-
tion.79 Legislators hear frommany more interestsand social sectors,in
both formaland informalways. Aggregatingacross some dozens ofjudi-
cial minds ameliorates thisproblem, but does not plausiblyovercome the
massive initial advantage of legislativenumbers and specialization in the
Condorcetian framework.
3. EndogenousInformation and RationalIgnorance.- A furtherset of
considerationsthatbear on the Condorcetian credentialsof currentlegis-
latures involvesincentivesto acquire information. The basic problem is
that the amount of informationeach voter (whetherlegislatoror judge)
decides to acquire is endogenous to the size of the votinggroup. In large
legislatures,each legislatorhas an incentiveto remain rationallyignorant.
The legislatorwill investless in informingherselfabout the questions at
hand if she estimatesthat,because of the legislature's large size, she is
extremelyunlikelyto be the decisive voter in any event. More generally,
informationand deliberation are collective goods in legislatures;each
member may attemptto free ride on the expertise of others.80 As Ben-
tham himselfput it:
The greater the number of voters the less the weight and the
value of each vote, the less its price in the eyes of the voter,and
the less of an incentivehe has in assuringthatit conformsto the
trued end and even in casting it at all.81
All else equal, these considerations tend to undercut the informa-
tional advantagesof legislatures. On thisview,the comparativeadvantage
of courts is theirrelativelysmall size. By raisingeach judge's probability
of being the pivotalvoter,the incentiveto acquire informationis greater,
and it is easier for all members of the group to engage in mutual moni-
toringof freeriders. On a nine member Supreme Court,each Justicehas
a nontrivialchance of being the swingvoter,and colleagues will notice
and disapprove if a Justiceacquires no informationbefore voting.
However,these considerationsare themselvesundercut by other fea-
tures of legislatures. It is not clear, in the abstract,how the endogeneity
79. See Neil Komesar, ImperfectAlternatives:Choosing Institutionsin Law,

Economicsand Public Policy 141 (1994) (noting tradeoffbetweeninformation and
evenhandedness) .
80. Cf. ChristianList & PhilipPettit,An EpistemicFree-Riding Problem?,in Karl
Popper: CriticalAppraisals128, 138-40 (PhilipCatton8cGrahamMacdonaldeds.,2004)
(suggestingthat knowledgeof JuryTheorem among group's voters may increase
temptation to freeride on information betweengroupsize
of others). For the tradeoff
and free ridingon the information of others,see generallyDrora Karotkin& Jacob
Paroush,OptimumCommitteeSize: Quality-Versus-Quantity Dilemma,20 Soc. Choice 8c
Welfare429 (2003).
81. Quoted inJonElster,ExplainingSocial Behavior413 (2007).
of informationnets out against the sheer numerosityof legislators. In

some models of electoral voting,a large number of voterswhose informa-
tion is quite poor due to the rational ignorance problem can nonetheless
be almost certain of getting the correct answer,just because numbers
count.82 The differencebetween a Supreme Court of nine members and
a Congress of more than four hundred is plausiblytoo large- the initial
Condorcetian advantages of the lattergroup are plausiblytoo great- to
be sensitiveto differentialincentivesto acquire information.
Furthermore,legislatorsoften abstain fromvoting,whereasJustices
veryrarelydo so. Rational abstention can allow uninformedlegislators
to, in effect,defer to colleagues who are known to be well-informed, thus
raising the chance that the decisive vote will be cast by a well-informed
voter.83 Of course legislatorswho abstain often do so, in fact,for less
high-mindedreasons, but again we are assuming sincere and public-spir-
ited behavior on all sides in order to focus on the informationalcompari-
son between legislaturesand courts.
Finally,the relativelylarge size of legislaturesmerelyreduces legisla-
tors' incentivesto acquire informationas part of thelegislative process.It
does not remove whateverinformationthey begin with or acquire for
other reasons. Analogously,in electoral models of rational ignorance, it
has been pointed out that "[i]t is reasonable to believe that voters are
involuntarilyexposed to a flowof political informationin the course of
everydayactivities";the consequence is that even under rational igno-
rance, "successfulinformationaggregation is possible because the infor-
mation acquired by each voter goes to zero but it does so slowlyenough
to allow the effectof large numbers to kick in."84 As discussed above, it
seems plausible that legislatorsbegin withmuch more relevantinforma-
tion than do judges. Legislators acquire informationrelevant to judg-
ments of constitutionality - informationabout societal mores and evolv-
ing standards of decency, the consequences of constitutionaldecisions,
and the factson which constitutionaldecisions rest- throughtheirclose
contactswithconstituents,governmentpolicymakers,and the broader so-
ciety. By contrast,informationalimpoverishmentis a principal cost of
thejudges' insulationfromthe hurlyburlyof governmentpolicy. So leg-
islators' relative disincentiveto acquire more informationthrough the
legislativeprocess, ifit exists,maynot matterverymuch in lightof legisla-
82. See Cesar Martinelli,

Would RationalVotersAcquireCostlyInformation?, 129J.
Econ. Theory225, 240-41 (2005).
83. Cf.Timothy J. Feddersen& WolfgangPesendorfer, The SwingVoters'Curse,86
Am.Econ. Rev.408,408-24 (1996) (discussing conditionsunderwhichuninformed voters
willrationallyabstain).
84. Martinelli,supranote82, at 226-27;see also MinoruKitahara& YoheiSekiguchi,
Condorcetjury Theoremor RationalIgnorance?4 (rev.Jan.31, 2006),availableat http://
wakame.econ.hit-u.ac.jp/~riron/Workshop/2006/sekiguchi06.pdf (on file with the
ColumbiaLaw Review)(showingconditionsunder which "[e]ven if the amount of
information acquiredbyeach voteris small,theamountofaggregatedinformation can be
largeenoughto [reach] thecorrectdecision").
tors' initial informationaladvantages,just as a runner who gets a cramp

may stillwin if her lead was large enough to begin with.
4. Legislaturesand Precedent. - So far I have been speaking of the
decisions of "currentlegislatures,"but this understatesthe information
contained in legislation, which also draws on legislativetraditionsand
precedents; there is thus a Condorcetian case for common law legisla-
tures that parallels the case for common law courts. Like courts,legisla-
turesaggregate over many minds, not only byvirtueof legislatures'large
membershipsat any one time,but by using a kind of internalprecedent,
thus including the judgments of past legislators in current
decisionmaking.
It is sometimes suggested that courts use precedent, while legisla-
tures do not. All legislatures,on thisview,act on a blank slate so far as
formallaw is concerned,85while courtsare in some sense legally"bound"
by precedent. Incremental and cumulativedecisionmakingis integralto
the judicial process, but no legislatureis formallybound by past enact-
ments, although past enactments of course set a policy status quo and
thus allocate the burden of inertia in one way or another.
There is a confusionor slippage here between formaland functional
arguments. The Condorcetian argument forjudicial reliance on prece-
dent has nothing to do with the legally binding force of precedent,
whateverthatmeans; it would be exactlythe same in a systemthathad no
such formal norm or rule. The Condorcetian argument is just that
judges will do better,in theircurrentdecisions, by drawingon the latent
informationembodied in precedent. A system of judicial precedent
economizes on the costs of decisionmaking- both process costs and er-
ror costs.
There is a parallel argument for legislatures. Even though legisla-
turesare never formallybound by earlier enactments,and can alwaysre-
peal internal rules, rational legislatorswill use past statutesand internal
legislativeinterpretationsof statutesand legislativerules as informational
aids, economizing on process costs and reducing errors. This is not the
same as the point thatextantlaw creates a defaultstatusquo thatmustbe
overcome to enact new legislation. The informationaluse of legislative
precedents applies even ifthose statutesare no longer legallyin effect.A
lapsed statuteor rule may contain informationrelevantto designingnew
rules or policies in the present.86
85. Undertheconventional maynotentrenchstatutesbymakingit

view,legislatures
harderfora futurelegislature to repealthestatutethanitwasfortheoriginallegislature to
enact it. For a critiqueof thisview,see Eric A. Posner& AdrianVermeule,Legislative
Entrenchment: A Reappraisal,111Yale LJ. 1665 (2002). Forcritiquesofthecritique,see
JohnC. Roberts& ErwinChemerinsky, Entrenchment ofOrdinary Legislation:A Replyto
ProfessorsPosnerand Vermeule,91 Cal. L. Rev.1773 (2003); see alsoJohnO. McGinnis8c
MichaelB. Rappaport, ofLegislative
The Constitutionality Supermajority Requirements: A
Defense,105Yale LJ. 483, 503 (1995).
86. Symmetrically, thisimpliesthaton theCondorcetianinterpretation ofprecedent,
judges should count overruledcases as information, despite their null legal status.
It is thustheroutinestuff businessforlegislators
oflegislative to rely
on past statutes,rules,and internalprecedentsin this informational
sense. Draftersof new legislationoftenuse old legislationas a starting
point,modifying it to fitnew problemsand circumstances.Legislators
willpointto predecessorstatutesto showthatthebodyhas alreadybeen
regulatingin a givenfield,or thatpastpolicyexperiments haveworked
welland shouldbe expanded,or thatpastpolicieshavenotworkedwell,
or thatpolicyin a givenfieldis "settled."Andprecedentsdominateinter-
nal legislativeinterpretation,by house parliamentarians and legislators,
of internalrules.
These pointsapplyin the constitutional arena as well. Whetheror
notone thinksitusefulto attachthelabel of"super-statutes"87 to statutes
thathave somehowacquired a vaguely"constitutional" nimbus,legisla-
turesclearlydo enactstatutes witha viewto liquidatingor construing an
ambiguousconstitutional text,and thesestatutesoftenbecome de facto
entrenchedovertime. The mechanisms of thisentrenchment are poorly
understood,88 but one such mechanismis plausiblyCondorcetian.Pre-
sentlegislators mayrelyupon whatpastlegislators did, not because they
are in anysenseboundbythosedecisions,butbecausetheywishto incor-
porateand use the information latentin thosedecisions. Considerthe
decision,by a Republican-controlled Congress,to renew the Voting
RightsAct89in 2006 withoutsubstantial opposition,and withmanylegis-
latorsfromboth partiesarguingthatthe Act is now part of the legal
fabric.90That theActwasdue to sunset,and actuallyhad to be affirma-
tivelyrenewed,emphasizesthatlegislators'use of past statutesas prece-
dentdoes notdependon whetherthepaststatuteis stillthestatusquo, or
ratherhas lapsed.
We maydescribethislawmakingstrategy as Burkean,in the sense
thatcurrentlegislators do not relysolelyupon theirprivatestockof rea-
son. Even ifjudicial precedent,possiblydrawingupon broadersocietal
Supposethatat Time 1, nineJustices votedunanimously forRule R, butthenat Time 2, a

differentnineJusticesvoted 5-4 to overrulethe Time 1 precedentin favorof rule S.
Puttingaside theproblemthattheJustices at Times1 and 2 werenotreallyaddressingthe
samequestions(becauseonlytheformer wasa case offirst a Justicesitting
impression), at
Time 3 shouldaggregateoverthewholegroupofeighteenJustices, countingthe"vote"as
13-5in favorof R.
87. WilliamN. Eskridge, Jr.& John Ferejohn,Super-Statutes: The New American
Constitutionalism,in The Least Examined Branch 320 (RichardW. Bauman & Tsvi
Kahana eds.,2006).
88. For some speculations, see id.
89. Pub. L. No. 89-110,79 Stat.437 (codifiedas amendedat 42 U.S.C. 1971,1973
to 1973bb-l(2000)).
90. The Act was renewedwithsupportfromthe RepublicanHouse leadership.It
passed the House by a vote of 390-33and the Senate by a vote of 98-0. See Raymond
Hernandez,AfterChallenges,House ApprovesRenewalofVotingAct,N.Y.Times,July14,
2006,at A13 (describingHouse voteand initialoppositiontoAct);CarlHuse,Bya Voteof
98-0,SenateApproves25-YearExtensionofVotingRightsAct,N.Y.Times,July21, 2006,at
Al (describingSenatevote).
traditions,aggregates latent informationrelevant to constitutionallaw,

the episode of the Voting RightAct suggeststhatlegislaturestoo are com-
mon law constitutionalists.Praise forthe latentwisdom contained in pre-
cedent says nothing at all about the crucial comparativequestion, which
is whether common law constitutionalismis best carried out in legisla-
tures or in courts,where the outcomes of these two processes conflict.
The upshot is thatifjudges draw on the contributionsof manyminds
aggregated in past judicial decisions, legislatorsdraw on the contribu-
tions of many minds aggregated in past legislativedecisions. Recall the
counterintuitiveargument that the Warren Court was actually a "com-
mon law court" in the sense that its major decisions were traceable to
background lines of judicial precedent or broader political traditions.91
By the same token,one mightmake out a case thatthe Congress of some
innovativeera was actuallya common law Congress. An example is the
Endangered Species Act of 1973,92sometimesdescribed as a statutethat
made a radical break with the past, but which actually built upon the
Endangered Species Conservation Act of 196993 and the Endangered
Species PreservationAct of 1966,94as well as even earlier legislationand
treatiesapplicable to migratorybirds and federal lands.95 Of course this
sortof case mightbe more or less convincingor strained,but the same is
true of the argumentthatmost of the major Warren Court decisions had
roots in precedent. What the example shows is that any categorical dis-
tinctionbetween precedent-usingcourtsand memorylesslegislatures,for-
ever acting on a blank slate, is a cartoon.
5. Past Judgesand the CurrentExecutive:Rooseveltand Condorcet.-
Some of the same considerations hold, with appropriate modifications,
for executive branch decisionmaking. It is not clear, on Condorcetian
grounds, that the aggregated informationembodied in a line of prece-
dent is superior to the informationcontained in executive branch deci-
sionmaking. In a sense, the executivebranch contains manymore minds,
at any given time, even than Congress, whose staffof twenty-oddthou-
sand employees is dwarfedby the hundreds of thousands of (non-mili-
tary) personnel employed by the executive. This is misleading because
the internal organization of the executive branch is more hierarchical
than thatof American legislatures,which displayrelativelyweak partydis-
cipline. In Condorcetian terms,this hierarchyincreases intragroupdef-
erence and reduces the effectiveindependence of the aggregated"votes."
Nonetheless, the executive branch is a they,not an it; the "unitaryexecu-
91. Strauss,CommonLaw Genius,supranote 27.

92. Pub. L. No. 93-205,87 Stat.884 (codifiedin scatteredsectionsof 16 U.S.C.).
93. Pub. L. No. 91-135,83 Stat.275 (repealed 1973).
94. Pub. L. No. 89-669,80 Stat.926 (repealed 1973).
95. See BradfordC. Mank,Can CongressRegulateIntrastateEndangeredSpecies
Under the CommerceClause?,69 Brook.L. Rev.923, 933-41 (2004). The Endangered
SpeciesActdid breakwiththepastin important ways,especiallybyapplyingto nonfederal
lands,but itwas not createdout of thinair.
tive"is a legal claim about formallines of authorityand removalthatdoes

not describe executive decisionmaking and has no direct relevance to a
Condorcetian analysis.
Because of the internal diversityof the executive branch and the
multiplicityof its officials,it is an open question whether an executive
decision aggregatesmore or less informationthan the decisions ofjudges
in a regime of precedent. Executive branch policyanalystsare, plausibly,
highlycompetent in the sense that theyare experts in particularfields.
The harder question is how well executive expertise is actually used by
decisionmakers higher in the chain; even if higher decisionmakers are
sincere, and thus impose no political distortionson the analysts'conclu-
sions, theymustin manycases select between or among competinganaly-
ses, and maydo so well or poorly. Here too, however,examining only the
current executive understates the case; executive offices routinelyuse
precedents,in the sense of prior decisions thatare informationally useful
whether or not legally binding. There is no basis for a systematicpre-
sumption thatjudicial precedent is informationally superior to executive
decisions where these conflict.
6. Implicationsand Conclusions.- Overall, it is hard to generalize
about the relative informationalproperties of common law precedents
and traditions,the original decisions of framersand ratifiers,the deci-
sions of currentlegislatures,of the currentexecutive, and so forth. All
one can offerare some comparativestaticsand institutionalcautions. As
to competence, the lower the average competence of pastjudges, and the
higher the average competence of past framersor ratifiers,the less likely
it is thatcurrentjudges should overrideclear textbased on contrarysub-
sequent precedent. However,where circumstanceschange over time,the
informationaladvantage of laterjudges will become pronounced. Much
depends on the rate of change in the legal and political environment,a
point I underscore in Part II below.
If the legal and political environmentis changing rapidly,however,
legislation is plausibly the best response of all. Here too some general
comparative staticsare possible. The lower the average competence of
past judges, and the higher the average competence of current legisla-
tors,the less willingcurrentjudges should be to overridecurrentstatutes
on the basis of precedent. On numbers,even verycompetentjudges to-
day, possessed of precedents issued byverycompetentjudges in the past,
will rarelydo well to overrule the decisions of current legislatures,just
because of theirsheer relativenumerosity,which counts for a great deal
in Condorcetian termsifaverage competence is greaterthan one-half(in
the case of two choices) .
These points are not useless, but they are not sharp tools either.
However,the veryfactthatit is hard to make furtherprogressalong these
lines, in the abstract,itselfsuggests some useful generalizations. First,
common law constitutionalists must beware of a fallacious move froman
explicitand valid96comparison - thatthe aggregatedwisdomof many

or in a
judges participants tradition, overtime,is superiorto thatof one
or a fewjudges exercisingtheirunaided reasontoday - to an invalidim-
plicitcomparison - that the aggregated wisdom embodied in precedentis
superior to thatembodied in the constitutionaltextand originalunder-
in
standing, statutes, or in the decisions of executiveofficials.The point
is not to assertthatthe constitutional commonlaw is or is not,in fact,
superiorto thealternative sources;itis thatno amountofcomparingone
judge to manyjudges can supplythe answerto the otherinstitutional
comparisons.
Second,theveryweaponsthatcommonlawconstitutionalists deploy
againstoriginalism maybe used againstthembyoriginalists, and can also
be used againstthemwitheven greatereffectby Thayerians,who can
plausiblyclaimthatan informational analysisplaysdirectlyto thecompar-
ative advantage of legislatures.By invokingCondorcet against the
originalists,thecommonlawconstitutionalists tendto exposetheirflank.
The firstvirtueof anytheoryof constitutional adjudicationis to explain
whycourtsshouldhavethepowerto overridelegislative action. Original-
ism,forall itsgrievoustheoreticalproblemson othermargins,offersa
straightforward accountof thatpower: Courtsmaydo so whenand be-
cause thereis a validlyenactedhighersourceoflawauthorizing or requir-
ing themto do so. Bycontrastthe Condorcetianinterpretation of com-
mon law constitutionalism, and of Burke,leavesit not at all obviouswhy
judges shouldrelyon precedentor even traditionto trumptheviewsof
currentlegislatures.The veryplausibleclaim thatmanyjudicial minds
are superiorto one or a fewdoes not giveanyleveragein makingthat
additional,and moredifficult, claim.
To maketheimplications ofthesepointsmoreconcrete,someobser-
vationsabout different legal and institutionalcontextsare useful. Con-
siderthe following:
as a shield.>ig7An implica-
as a sward"versus"Burkeanism
"Burkeanism
tionoftheforegoing is thatthereis no soundbasisforusingBurkeanism
-
as a sword meaningthatthereis no warrantforoverriding thejudg-
mentsof currentlegislatures or the executive
by referenceto precedent
or judicial traditions. The most plausible consequence of the
Condorcetiananalysisis Thayerism.In thisregime,legislaturesmight
themselves be precedentialists usingthecommonlaw
and traditionalists,
of past legislation,and broadersocial traditions,to informnew legisla-
tion. The overallregimewouldbe one of Thayerism common
forcourts, law
constitutionalism
forlegislatures.
betweenprecedent(ju-
Here we mustbe carefulwiththedistinction
morebroadly.We haveseen that,at the
dicialtraditions)and traditions
SupremeCourtlevel especially,precedentsmaynot embodythejudg-
96. Assumingawaythe internalproblemsdiscussedin I.B.

97. Sunstein,BurkeanMinimalism,supranote 5, at 356.
meritsofverymanyminds,at leastcomparedto legislatures.Social tradi-

tions,on the otherhand,maystandon a betterfooting,iftheyincorpo-
ratethejudgmentsofa largenumberofpeople. However,thereare two
problemswithany attemptto distinguishprecedentfromtraditionas
guides to judicial decisionmaking.One quite fundamentalproblem,
identified byBentham,is thattraditions are notself-definingor self-apply-
ing; if traditionsare refracted through then
judicial decisions, theyhave
no betterepistemiccredentialsthando thejudges themselves, who may
(inadvertently) distortor misconceive them in the processof identifica-
tionand application.98The second problem,relevantespeciallyin con-
stitutional cases,is thatlegislatorstoo can relyon broadersocial tradi-
tions, just on legislative
not orjudicial precedent.Ifjudges can discern
and applysocial traditions thathavestrongCondorcetiancredentials, so
can legislators; and to theextentthereare rightanswersaboutwhattradi-
tions permitor require,highlynumerouslegislatureshave a built-in
Condorcetianadvantageovercourtsin discerningthoseanswers.
If Burkeanismas a swordis infirm, thereis stilla possiblerole for
Burkeanargumentsthatsupportlegislativeand executivejudgments -
Burkeanismas a shield. But thisis not a veryimpressive role. Where
constitutional textsare ambiguousand precedentpointsin the same di-
rectionas enactedlegislationor executiverules,judges willupholdwhat
thelegislature or executivehas done. Here thereis no obviousproblem,
and no realneed to defendcommonlawconstitutionalism as a distinctive
approach to constitutional adjudication.Burkeanismas a swordis the
theoretically crucialcase forcommonlawconstitutionalism, butit is also
theverycase whereCondorcetianreasoningmoststrongly suggeststhat
judges- eithercurrentjudges or a notionaljudiciaryaggregatedover
time- shoulddeferto otherinstitutions.
TheFrankfurter canon.JusticeFrankfurter suggestedthatthe deci-
sions of past legislaturesand presidentsabout how to allocate powers
amongthemselves placesa sortof commonlawglosson ambiguouscon-
stitutional texts,especiallyin areas of separationof powersand foreign
affairs, where the constitutional textureis especiallyopen." On the
Condorcetianview,Frankfurter's pointis eitherimportant or banal,de-
pendingupon whetherit is assumedthatthereis or is not a line ofjudi-
cial precedentthatconstruestheconstitutional ambiguityin a waydiffer-
ent thanthe otherbranches.
Ifthereis no suchlineofprecedent,thenon Condorcetianpremises
it is banal thatcurrent judges shouldrelyon theimplicitcollective judg-
mentof legislatorsand presidentsover time. Those judgmentswillbe
made bya statistical groupfarmorenumerous,and at leastas expert,as
98. Jeremy Waldron,CustomRedeemedbyStatute,51 CurrentLegal Probs.93, 100

(1998) ("Benthamarguedthatonce theycome intothehandsofthejudges,customstend
to be ill-usedand subjectedto arbitrary
and unpredictable
modifications.").
99. YoungstownSheet 8c Tube Co. v. Sawyer,343 U.S. 579, 610-11 (1952)
(Frankfurter,J.,concurring).
the smallgroupof currentjudges. If thereis a line of contrary prece-

dent,however,Frankfurter's idea is more cutting.It suggeststhatthe
aggregated wisdomofmanyjudges,overtime,is stillinferior to theaggre-
gated wisdom of legislatorsand executive over
officials, time. In thislat-
tercase, Frankfurter's viewis a specialcase, in the area of separationof
powers,of theclaimthatBurkeancommonlawconstitutionalism should
-
not be used as a sword at leastnot whenBurkeanismis interpreted in
informational terms.
and re-enactments.
Statelegislatures Here is anotherspecialcase of the
same point. Suppose thatat Time 1, SupremeCourtJusticesissuea pre-
cedentdecisioninvalidating statelaws. At Time 2, manystatesreenact
thoselaws. ShouldtheCourtacquiesceat Time 3, or enforceitscontrary
precedent?100On Condorcetiangrounds,the answeris (almost cer-
tainly)thattheCourtshouldacquiesce. Assume,howeverheroically, that
statelegislatures and the Courtare addressingthe same questionat all
times - whethera givenstate statutory rule complieswiththe federal
Constitution - and thatthereis a rightanswerto thisquestion. The ag-
gregatedcurrent judgmentsofa massivenumberof statelegislatures are
in
superior, terms of the JuryTheorem, to thoseof the groupcomprising
thejudges and Justiceswho sat in the precedentcase(s). It is irrelevant
thatthe legislators' judgmentsare themselves sortedinto groups (state
legislatures)who make separatecollective decisions underinternalma-
jorityrule. The JuryTheoremapplies even to the purelynotionaland
groupcomprisedof all the statelegislatorstakentogether,so
statistical
long as the Theorem'sotherconditionsare met.
In general,Condorcetianreasoningdoes notsupplya convincing in-
terpretation of common law constitutionalism, the
particularly latter'sex-
plicitlyBurkeanclaimthatthe limitsof humanreasonsupportreliance
on constitutional precedentand tradition.The internalproblemis sim-
ply that the JuryTheorem's conditionswill,quiteoften,notapplyto any
realisticdescription oftheprocessofcommonlawconstitutionalism. The
comparative institutionalproblem is that Condorcetian reasoning itself
defangscommonlaw constitutionalism in the mosttheoretically crucial
cases- whereprecedentis used to trumpclear constitutional textand
history, or to trumpthejudgmentsof currentlegislatures and executive
officials.On Condorcetiangrounds,themostplausibleregimewouldbe
one of Thayerian,highlydeferentialcourts,withcommonlaw constitu-
tionalismentrusted to legislatures.The pointthattheviewsofmanypast
judges embody more latentinformation than the viewsof a fewjudges
be
sittingtodaymay correct, but that is not the relevantcomparisonin
the testingcases.
100. Foran exampleof thisissue,compareFurmanv. Georgia,408 U.S. 238,239-40

(1972) (per curiam)(invalidating Georgiaand Texas deathpenaltyschemes),withGregg
of
deathpenaltyafterstatereenactments
v. Georgia,428 U.S. 153, 169 (1976) (validating
newschemes).
II. Burkeas Darwin

I willnowturnfrominformational accountsofcommonlawconstitu-
tionalismto evolutionary ones.101The Condorcetianinterpretation of
Burkeassumesthe existenceof exogenousrightanswers,somehowde-
fined,in constitutionalcases; positsthatjudges are sincerelyattempting
to findthoseanswers;and positsthatjudges are morelikelyto findthose
answersby relyingon commonlaw precedentsthan on their"private
stockofreason." In theevolutionary bycontrast,
interpretation, nojudge
need be seekingto producerightanswers.Different judges have differ-
ent motivations, ideologies, and biases. Nonetheless,the claim runs,
some mechanismcauses the uncoordinatedaction of biasedjudges to
producea bodyof commonlaw thatis economicallyefficient.If such a
mechanismexists,itwouldbe a speciesof invisiblehand explanation,in
whichefficiency would arise fromhuman action but not fromhuman
design.102
Whatis theevolutionary mechanismexactly?A selectionmechanism
requiresvariationin therelevantpopulation(oforganisms or cases); her-
so thatfavoredvariationscan be maintainedin the population
itability,
when theyarise; and differential reproductive advantage(or "fitness"),
meaningthatfavoredvariationsare likelyto replacecompetitors.103 In
evolutionary analysesofthecommonlaw,thefirst requirement is typically
said or assumedto be satisfiedbyvariationin thebiasesor preferences of
judges or litigants,whilethesecondis satisfied bytheforceofprecedent.
However,the thirdrequirement is the mostproblematic;here the anal-
ogyto Darwinianprocessesbecomesparticularly obscure. Whatmecha-
nismcorrespondsto naturalselectionof relatively fitter
organisms?
The mostfamiliarattemptto answerthisquestioninvolvesthe de-
mand-sidemodelsofcommonlawevolutionthathavebeen prominent in
law and economics. Here "demand-side" means thatthe commonlaw's
claimedevolutiontowardefficiency is poweredby the selectionof cases
forlitigation- by litigants'decisionsabout whatcases to appeal or set-
101. For the evolutionary of Burke,see, e.g., Strauss,CommonLaw

interpretation
Interpretation,supranote5, at 879 (arguingfora "formoftraditionalism, characteristic
of
the commonlaw method,[that]calls forrecognizing thevalue of conclusionsthathave
been arrivedat, overtime,byan evolutionary process");Sunstein,BurkeanMinimalism,
supranote5, at 368 ("Burkeansmightstressnotsocialpracticesbut theslowevolutionof
judicial doctrineovertime.... For theseBurkeans,whatis particularly important is the
judiciary'spriorjudgments,whichshouldin turnbe based on a seriesof smallsteps,and
shouldavoidradicaldepartures."); Young,supranote5, at 653-56 (offering evolutionary
interpretationof Burkeand connecting"Burke's theoryof reform"to "thecommon-law
traditionof evolutionarychange").
102. This is a paraphraseof Adam Ferguson'sfamousobservationthat "nations
stumbleupon establishments, whichare indeed the resultof humanaction,but not the
executionofanyhumandesign."AdamFerguson, An Essayon theHistory ofCivilSociety
119 (Fania Oz-Salzberger ed., CambridgeUniv.Press1995) (1767).
103. See RobertN. Brandon,Conceptsand Methodsin Evolutionary Biology47
(1996) (definingnaturalselectionaccordingto thesecriteria).
tie.104I willexamineboth internaland institutional problemswiththe

demand-sidemodels. First,as an internalmatter,these models suffer
froma kindof fragility; theyare not robustagainstslightchangesin the
assumptions.Second, in comparativeinstitutional terms,there is no
straightforward to
way transpose these models to the settingof constitu-
tionalcommonlaw. Even ifconstitutional commonlaw tendsto evolve
towardefficiency, and even ifit is desirablethatconstitutionalcommon
law be efficient,
legislationmay also tend toward efficiency, forthe
and
same reasons. Moreover,in a rapidlychangingenvironment, legislation
mayat anygivenmomentbe moreefficient thanconstitutional common
law, even ifthelattersystematically tendstoward efficiencywhile thefor-
merdoes not. In parallelto theargumentofPartI, theattemptto inter-
pret Burke by way of Darwinleads most naturallyto James Bradley
Thayer'sidea thatcourtsshould systematically deferto legislaturesin
constitutionalmatters.
A. Demand-SideModels: InternalProblems
I willnot attempta comprehensive overviewof thismassivelitera-
ture.105InsteadI willsingleout tworelevantproblemswiththedemand-
side models: the fragilityof theirassumptions, and theirsilenceabout
the rate at whichcommonlaw precedentsconvergeto efficiency. The
formerproblemis betterknown,but the latteris evenmoreserious,be-
cause it suggeststhatin rapidlychangingenvironments the tendencyof
commonlaw to convergeto efficiency - evenassumingit exists - willnot
matterverymuch;in particular, itwillprovideno comparative advantage
overlegislation.
- The basic engineof thedemand-sidemodelsis selec-
1. Fragility.
tionpressurebylitigants:If inefficient precedentsare challengedmore
frequently,while precedents moreoftenleftin place,thelaw
efficient are
willtendto efficiency overtime. The basicaspirationof theliterature is
to proposemechanismsthatwillshowwhy,and underwhatconditions,
litigantswilltend to differentially precedents.In the
litigateinefficient
model thatinspiredmuchof theworkin thisarea, the basic idea is that
inefficientprecedentsinflictdeadweightlosses;the loserswillthushave
an incentiveto paymoreto attacktheprecedentsthantheirbeneficiaries
104. Forsupply-sideanalysisbased on competitionamonginstitutions providinglegal

rules,see generally
ToddJ.Zywicki, The Riseand FallofEfficiencyin theCommonLaw: A
Supply-Side Analysis,97 Nw.U. L. Rev. 1551,1620 (2003) [hereinafter Zywicki,Rise and
Fall] (arguingthatcompetitionamong different courtswithoverlapping jurisdictions
duringformative era ofcommonlaw"led to innovation and incentives to provideefficient
legal rules"). Zywickialso arguesthatan increasein the strength of precedentin the
nineteenthcenturyexacerbatedvariousproblemswiththe commonlaw, such as rent
seeking.See id. at 1565.
105. Fora recentsummary, see Paul H. Rubin,WhyWas theCommonLaw Efficient?
2-9 (EmoryLaw& Econ. ResearchPaperSeries,PaperNo. 04-06,2004),availableat http:/
/papers.ssrn.com/sol3/papers.cfm?abstract_id=498645 (on file with the ColumbiaLaw
Review).
will pay to defend them.106Many second-decimal refinementshave been

proposed, but the underlyingideas are similar,so I willcall thisthe "basic
model" and take it as the subject of discussion.
The basic model has been subjected to severalmajor lines of critique
withina broadlyrational choice framework.107First,the basic model as-
sumes that precedents can only be reaffirmedor overruled. If theycan
also be furtherentrenched, however, then differentiallitigationmight
actuallytightenthe grip of inefficientprecedents on the law.108 Second,
the basic model assumes that all classes of litigantshave, on average,
equal stakes and an equal interestin the futureconsequences of prece-
dent. But if differentclasses of litigantsare differentiallyorganized, per-
haps because they have systematically differentstakes or because some
are repeat playerswhile others are single-shotlitigants,then precedent
can just as well evolve in inefficientdirections.109There maybe eventual
convergence, but to rules favoringorganized groups, not rules favoring
efficiency.
In general, rent-seekinglitigationis as much a problem for models
of common law efficiencyas rent-seekinglobbyingis formodels of legisla-
tiveefficiency,110 a point I willreturnto shortly.According to one histori-
cally-inflectedversion of thishypothesis,common law was efficientin the
eighteenth and nineteenth centuries,but is no longer so, because of the
of
large-scaleorganization rent-seekinglitigationgroups111or (in a sup-
ply-sideversion) because of the decline in competitionby differentcourt
systems.112The basic model can and has been refined to tryto account
106. Paul H. Rubin,Whyis theCommonLawEfficient?, 6J.Legal Stud.51, 53 (1977)

[hereinafter, Rubin,WhyEfficient]; see also GeorgeL. Priest,The CommonLaw Process
and theSelectionof Efficient Rules,6J. Legal Stud.65 (1977).
107. For a critique drawingon behavioral economics, see Adam J. Hirsch,
Evolutionary TheoriesofCommonLawEfficiency: Reasonsfor(Cognitive)Skepticism, 32
Fla. St. U. L. Rev.425 (2005); fora differentcritique,see GillianK. Hadfield,Bias in the
EvolutionofLegalRules,80 Geo. LJ. 583,616 (1992) (arguingthatclaimsaboutefficiency
ofcommonlaware based on "a weakconceptofefficiency"). I willbracketsuchissuesin
thediscussionhere.
108. See WilliamM. Landes8cRichardA. Posner,Adjudication as a PrivateGood,8 J.
Legal Stud.235, 273 (1979).
109. The authorsofthebasicmodelrecognizedthispoint. See Rubin,WhyEfficient,
supranote 106,at 53-57.
110. See GordonTullock,The Case Againstthe CommonLaw 15 (1997) (arguing
thatentirelegal system is influencedbyinterest groups);EinerR. Elhauge,Does Interest
GroupTheoryJustify MoreIntrusive JudicialReview?,101Yale LJ. 31, 34 (1991) (arguing
that"[t]he litigationprocesscannotbe treatedas exogenousto interestgroup theory
because thatprocessis also subjectto formsof interest groupinfluence").
111. See Tullock,supranote 110,at 52; Paul H. Rubin,CommonLaw and Statute
Law,11J. Legal Stud.205, 211-19 (1982) [hereinafter, Rubin,CommonLaw and Statute
Law].
112. See Zywicki, Riseand Fall,supranote 104,at 1620-21.
for these critiques,but onlyby makingmore stringent the conditions

underwhichprecedentconvergesto efficiency.113
Aftersomethirty yearsofdiscussion,thefairest assessment is thatthe
thesisof commonlaw efficiency (at least as drivenby demand-sideef-
fects)shouldbestbe takenas a possibility, no more. Underimaginable
conditionsthe commonlaw willsystematically tend to convergetoward
efficiency. However, slightchanges in the initial assumptionsseem to
producelarge effects, the
driving implications of the modelshitherand
yon and leaving the basic thesis unresolved. The natural responseis to
call forsomefacts;in theabstract, however, itcan at leastbe said thatthe
mechanismsin vieware insufficiently robustto inspireanyconfidence.
2. Environmental -
Change. Suppose the basic demand-sidemecha-
nismis notfragile, butrobust,and ignorecomparative institutional ques-
tions,whichI shall take up shortly.What implicationsdoes the basic
model have? Surprisingly few,for a simplereason thathas not been
muchdiscussedin theliterature.Efficiency is alwaysrelativeto thecon-
straintsestablishedbytherelevantenvironment. In theeconomicsense,
in whichall economicfactorsare put to theirhighest-value use and all
Pareto-improving tradesare consummated, theseconstraints arise from
availableresourcesand technology(parametricconstraints)and from
competition byothereconomicactors(strategic constraints).In thebio-
logicalsense,theoptimaladaptationoforganismsis determinedbynatu-
ral resourcesand competition fromotherorganisms.
The consequence is the following.The claim thatsome popula-
tion- of organismsor cases- evolvestoward efficiency is importantly dif-
ferentthana claimthatthepopulationreaches efficiency. A crucial varia-
ble is the relativerate of change in the population,comparedto the
environment.If the environment changesslowly, relativeto the process
ofadaptation,organisms willbe well-adapted giventheconstraints. Butif
the environment changesquickly, relative to the speed of adaptation,
thenat anygivenmomentmostoftherelevantpopulationmaybe poorly
adapted.114Thus an importantgeneralargumentagainstinstitutional
analoguesof naturalselectionis thatthe rateof changein the political
and social worldis, in general,much fasterthan in the naturalworld.
Againstthebackgroundof a naturalenvironment thatis relatively stable
overperiodsofmillionsofyears,itis plausibleto proceedon theassump-
tionthatmostextantorganisms are well-adapted.In institutional settings
such as the market,however,the higherrate of environmental change
meansthatmanyextantfirmsmaybe poorlyadaptedat anygiventime,
even ifthereis strongpressuretowardefficiency.
An analogousproblemariseshere. Nothingat all in the demand-
side modelsimplies,in and ofitself, anything about (1) therateat which
113. See, e.g.,KeithN. Hylton,Information,

Litigation,and CommonLawEvolution,
8 Am. L. & Econ. Rev.33, 57 (2006).
114. See JonElster,Ulyssesand theSirens6 (rev.ed. 1984).
the common law converges to efficiencyor (2) the rate of change in the
background social, economic, and policy environmentagainstwhich effi-
ciencymustbe judged. If the speed of convergence is slow,relativeto the
rate of change in the background environment,then even a systematic
tendency toward efficiencywill not guarantee that most common law
rules are efficient.Indeed, if the rate of environmentalchange is high,
many or most common law rules mightbe inefficientat any given time,
even if the pressure toward efficiencyis powerful.
I have indicated the conditions under which the common law might
tend toward or reach efficiency;although it is hard to know in the ab-
stractwhetheror not those conditions hold, there are no real grounds for
confidence thattheydo. A centralpoint is thatthe standarddemand-side
models, even if robust,carryno implicationat all that most common law
rules will be efficientat any given time. One must also compare the rate
at which common law tends to efficiencywiththe rate of environmental
change.
The implication is that the thesis of common law efficiencyis most
plausible in and forperiods in which the rate of economic, technological
and political change is relativelyslow, but is less plausible where change
proceeds relativelyrapidly. Above, I mentioned the thesis that the com-
mon law was efficientin the nineteenth centurybut has since ceased to
be so, because of the rise of organized litigationgroups or the decline of
supply-sidecompetition among legal systems.The currentconjecture is
in the same family,although the mechanism involvesthe relativerates of
change in the common law and the background environment. Perhaps
Burke's dictum,interpretedin evolutionaryterms,offereda plausible ac-
count of common law efficiencyin England in the seventeenthand eight-
eenth centuries, because the common law had time to evolve close to
efficiency in a relativelystable environment.Even ifso, thereis no reason
to thinkthat the common law is efficientin other timesand places, such
as today.
B. Demand-SideModels: Constitutional
Implications?
In thissection,I will examine the question whetherthe demand-side
models of common law efficiency, even ifinternallyrobust,have any pay-
offforconstitutionallaw. In general, theydo not. First,it is familiarthat
efficiencyis a dubious normativegoal forconstitutionallaw. Second, the
basic demand-side model uses essentiallythe same mechanism- the will-
ingness of regulated parties to bid more to avoid deadweightlosses- that
powers the standardmodel of legislativeefficiency.The basic model thus
offersno convincingreason to assume thatjudicial precedents will con-
verge to efficiencywhile legislation does not. Third, even if legislation
has no systematictendency to converge to efficiency,legislation can re-
spond more quickly to a changing environment. Perhaps the common
law is alwaystendingtowardefficiency, yetit maynonetheless at any given
momentbe farther awayfromtheoptimumthanis thebodyoflegislation

at the same moment.
1. ShouldConstitutional Law BeEfficient? - Efficiency is a deeplycon-
troversialgeneralgoal forconstitutional law. I willonlytouchupon the
relevantissueshere,bothbecause thepointis familiar,115 and because it
yieldsonlyan externalcritiqueof thepremisesof theevolutionary inter-
pretationof Burke,ratherthan an internalcritiqueof the interpreta-
tion'slogic.
One well-known set of problemsinvolvesthe moral statusof effi-
ciency. efficiencyequatedwithwealthmaximization,
If is thereis thema-
jor problem thatwealth, in is
itself, not a value.116If efficiency is equated
witheitherParetooptimality or satisfactionoftheKaldor-Hicks criterion,
underwhichwinnersgain more than loserslose, distributive concerns
come to thefore.117These questionsaboutthemoralstatusofefficiency
do not apply,at least in the same way,to efficient commonlaw in the
subconstitutional sense. Wherethesubconstitutional commonlawis effi-
cient,but thereare no constitutional constraints,anydesiredredistribu-
tion can take place throughthe tax and transfer system;doing so will
itselfbe the efficient means of redistribution.118 However,if constitu-
tionalcommonlawis used to blockredistributive transfers thatare ineffi-
cientbut (arguably)requiredbyjustice,thenthedubiousmoralstatusof
efficiency is consequential.
One mightsidestepor dilutetheseproblemsby distinguishing be-
tweenor among different constitutional provisions.Perhapsefficiency
shouldbe the goal of constitutional law under the TakingsClause,but
not under the Due ProcessClauses. Yet nothingin the evolutionary
model of the commonlaw is tied to thesetextualand doctrinaldiffer-
ences; the relevantmechanismsare not sufficiently fine-grained to ac-
commodatethem. Those differences could onlybe takeninto account
underan intentionalmodel,in whichjudges purposefully promoteeffi-
ciency, either because they have an intrinsictastefor it,or because public
pressuresforcethemto do so.119An intentional modelofthissort,how-
ever,wouldno longerbe Burkeanin anyobvioussense;itbelongsto the
realmofdeliberatesocialengineering, byjudges,thatBurkeansderideas
115. The bestdiscussionis Elhauge,supranote 110,at 48-66.

116. See generallyRonaldDworkin,Is Wealtha Value?,9 J. Legal Stud. 191 (1980)
(arguingthatwealthcannotbe understoodas a value).
117. See Shaun HargreavesHeap et al., The Theoryof Choice 241-42 (1992).
118. See Louis Kaplow& StevenM. Shavell,Whythe Legal Systemis Less Efficient
thanthe IncomeTax in Redistributing Income,23 J. Legal Stud.667, 674-76 (1994).
119. See RichardA. Posner,Economic Analysisof Law 99 (1st ed. 1972) ("In
searchingfora reasonably objectiveand impartial standard,as thetraditionsof thebench
require himto do, the
judge can hardly failto considerwhether thelosswas theproductof
wasteful,uneconomicalresourceuse."). For a critiqueof the intentional versionof the
commonlawefficiency thesis,especiallyas appliedto publiclaw,see FrankI. Michelman,
Constitutions,Statutes,and the Theoryof Efficient Adjudication,9 J. Legal Stud. 431
(1980).
"innovation." Nor, of course, would an intentionalmodel be Darwinian

either.
Thayeras Darwin. - Suppose, however,that
2. LegislativeEfficiency?
the goal of constitutionallaw is in general to promote efficientlaw. Sup-
pose also that a line of constitutionalcommon law decisions have estab-
lished a certain rule, and that a course of legislation produces a statute
conflictingwith the constitutionalcommon law rule. It does not follow
that thejudges would best promote efficiencyby invalidatingthe statute
on constitutionalgrounds. The verypremises suggestingthat the com-
mon law tends towardefficiencyalso suggestthatlegislationtends toward
efficiency.
Recall that the mechanism drivingthe basic model of common law
efficiencyis thatinefficientprecedents inflictdeadweightlosses, and that
the partieswho sufferthose losses will have greaterincentivesto attempt
to overturnthem than their beneficiarieswill have to defend them (as-
suming equal stakes). Over time, inefficientprecedents will be chal-
lenged more oftenand more frequentlyoverturned. This model is, how-
ever, essentiallythe same as a standard model suggestingthat legislation
will be efficient,and for the same reasons.120 In this model, "pressure
groups" who sufferdeadweight losses from inefficientlegislation will
spend more to overturnsuch laws than their beneficiarieswill spend to
defend them. Despite the statusquo bias of the legislativeprocess,which
favorsgroups defending inefficientlegislation, the systematictendency
over time will be toward an efficientcorpus of legislation. The parallel
between this model and the basic model of common law efficiencywas
clear to all, rightfromthe beginning.121
Both the basic model of common law efficiencyand the standard
model of legislativeefficiencyassume equal distributionof stakes across
groups. Accordingly,the standard model of legislative efficiencyhas
often been questioned by public choice theoristswho posit differential
organizationbygroups withunequally distributedstakesin legislativeout-
comes; such groups will seek and obtain inefficientlegislation,and those
who sufferthe deadweight losses of such legislationwill be insufficiently
organized to resist.122But what is sauce for the goose is sauce for the
gander. The same groups will also enjoy theirorganizational advantages
120. See GaryS. Becker,A Theoryof CompetitionAmong PressureGroups for

PoliticalInfluence,98 QJ. Econ. 371, 375 (1983).
121. See Rubin,Common Law and StatuteLaw, supra note 111 (controllingfor
change over time,commonlaw is not systematically more efficient than statutelaw);
Rubin, Why Efficient, supra note 106, at 61 (discussingparallels betweenefficient
legislationand efficient
commonlaw).
122. See generallyMancurOlson,The Logicof CollectiveAction:PublicGoods and
theTheoryof Groups141-48 (1965).
in the adjudicativeprocess,123 includingthe process of common law

constitutionalism.124
Commonlawconstitutionalists haveattemptedto drawnonarbitrary
distinctions betweenlegislativeand adjudicativeprocessesin waysthat
favorthe latter.To date none are convincing.Considerthe argument
thatthe minimumbid- the minimuma group mustpay to become a
playerin thelawmaking game- is lowerin thecourtsthanin thelegisla-
tiveprocess,while the marginalbenefitsof incrementalexpenditures
tendto be higherin legislatures thanin courts;bothpointsare intended
to show thatwell-funded groups have a greateradvantagein legisla-
tures.125Among other problemswiththis argument,126 however,the
point about marginal benefits of expenditures undermines the basic
mechanismthatis supposed to drivethe commonlaw to efficiency. If
litigationexpendituresquickly reach a point of zero marginal benefit,
thenit willnot matterthatpartiessuffering deadweightlosseswouldbe
to
willing pay more to challenge inefficient precedents thenthe benefi-
ciarieswouldbe willingto payto defendthem,becausetheextraexpend-
itureswillhave no positiveeffecton the likelihoodof overturning the
inefficientprecedents.
3. Efficiencyand SocialChange.- Puttingaside theforegoingpoints,
let us supposenowthatefficiency is a plausiblegeneralgoal forconstitu-
tionallaw, thatjudicial decisionssystematically tend towardefficiency,
and thatlegislative decisionsdo not. It stilldoes not followthatjudges
willbest promoteefficiency by invalidating statutesthatconflictwitha
contrary line of precedent.
The further problemis thatifthebackgroundenvironment is chang-
ing rapidly,the legislationin forceat anygiventimemaywellbe more
efficientthan contraryprecedent. Althoughthe lattersystematically
tendstowardefficiency at all times,itneverattainsitat anypoint,because
theenvironment createsa movingtarget.Newlegislationhas no dynamic
tendencytowardincreasingefficiency overtime(we are supposing), but
maysimplystartat a higherlevelofefficiency thanthecommonlawrules
in place at a givenmoment,whichweredevelopedin a previousenviron-
123. See Elhauge,supranote 110,at 34; see also FrankB. Cross,The Judiciary and
Public Choice, 50 Hastings LJ. 355, 360-68 (1999); Lewis A. Kornhauser,Legal
FoundationsofEconomicAnalysis ofLaw 15-16 (May31,2006) (unpublishedmanuscript,
on filewiththeColumbia Law Review), availableat http://www.law.nyu.edu/clppt/program
2006/readings/D_N_W_colloquium_excerpt.pdf.
124. See Pritchard& Zywicki, supra note 25, at 494-501 (discussingrentseeking
throughSupreme Court and
litigation decisionmaking in constitutional
cases).
PublicChoice,supranote4, at 221-22 (arguingthatthereis greater
125. See Merrill,
"minimumbid limitation" forlegislaturethanjudiciary);cf. Todd. J. Zywicki,Gordon
Tullock'sCritiqueoftheCommonLaw46 (GeorgeMasonL. & Econ. ResearchPaperNo.
07-13,2007),availableat http://papers.ssrn.com/sol3/papers.cfm?abstract_id=964781
(on
filewiththe Columbia Law Review)(discussingdifferences in "rent-seeking
pressures"on
andjudiciary).
legislature
126. For furtherdiscussion,see generallyCross,supranote 123.
ment and now sufferfromobsolescence. This is a short-runadvantage of

legislationthatdissipatesas the common law evolves towardefficiencyin
the new environment;but if furtherenvironmentalchange ensues, the
common law may never catch up. Where environmentalchange is rapid,
a series of shortruns is all there is. This is an evolutionaryparallel to the
Benthamiteargument,discussed in Part I, that legislaturessystematically
tend to have more currentinformation,and thusbetterinformationover-
all, even ifjudicial decisions more effectively aggregate the information
pastjudges possessed.
The superior efficiencyof legislation in a rapidlychanging environ-
ment is merelya possibility.Everythingdepends on the relativerates of
change in three variables: the common law, the body of legislation,and
the background political and economic environment. But the possibility
blocks any directinferencefromthejoint premises that (1) the common
law tends towardefficiencyand (2) legislationdoes not, to the conclusion
that (3) judges who aim to promote efficiencyshould use constitutional
common law to override statutes.
I conclude that the evolutionaryinterpretationof Burke, like the in-
formationalinterpretation,has fewuseful implicationsfor constitutional
law. If this is so, then two conclusions are possible. The minimum con-
clusion is thatthe recent and sophisticatedargumentsforBurkean consti-
tutionalismbased on the limitsof reason are a dead end. If no other
arguments for Burkean constitutionalism are forthcoming, then a
strongerconclusion follows: Burke's dictumis an insightthatcannot use-
fullybe cashed out in constitutionallaw. From the point thatjudges do
betterto draw upon the "bank and capital of ages" ratherthan their"pri-
vate stock of reason," not much follows.
C. Supply-Side
Models: The CardozoTheorem
To round out the picture, I will examine some recent economic
models thatmighteasilybe adapted to support a version of common law
constitutionalism.These models have recentlybeen developed by econo-
mistsNicola Gennaioli and Andrei Shleifer. The foundationfortheirap-
proach is the "Cardozo Theorem":
The eccentricitiesof judges balance one another. One judge
looks at problems from the point of view of history,another
fromthat of philosophy,another fromthat of social utility,one
is a formalist,another a latitudinarian, one is timorous of
change, another dissatisfiedwiththe present;out of the attrition
of diverse minds there is beaten something which has a con-
stancyand uniformity and average value greaterthan its compo-
nent elements.127
127. Cardozo,supranote57, at 177. The CardozoTheoremis named,and discussed,

in Nicola Gennaioli& AndreiShleifer,The Evolutionof CommonLaw, 115J. Pol. Econ.
43, 44, 60-61 (2007) [hereinafter
Gennaioli& Shleifer,Evolutionof CommonLaw]. A
companionpaperis NicolaGennaioli& AndreiShleifer, Overruling of
and theInstability
As Cardozo's dictum focuses on the behavior ofjudges and not that

of litigants,Gennaioli and Shleiferuse it to develop supply-sidemodels of
legal change, in contrast to the demand-side models discussed previ-
ously.128 Their models combine an informationaggregation approach
withan evolutionaryapproach: Their premise is thatconvergence to effi-
ciency occurs only in the restrictedsense that efficientuse is made of all
possible informationpossessed by the relevantpool ofjudges. In the best
case, the unintended byproductof countervailingjudicial biases is thatall
material distinctionsare built into the body of legal rules, and the law
becomes increasinglyprecise.129
Whateveritsmeritsas a pictureof the evolutionof ordinarycommon
law, the Cardozo Theorem cannot be used to support any robustversion
of the constitutionalcommon law, or so I will argue. To date, models
based on Cardozo's dictum posit an excessivelysharp distinctionbetween
overrulingand efficiency, one thatis an artifactof the modeling strategy.
Moreover, convergence to efficiency(and indeed convergence of any
sort) is least likelyto occur in these models when the polarization ofjudi-
cial viewsis high, perhaps because of the politicallycontroversialcharac-
ter of the subject. And ideological polarization acrossjudges is a chronic
circumstanceof modern constitutionaladjudication, at least at the level
of the Supreme Court.
To be clear, neither economists nor legal scholars have yet drawn
upon this new work to argue for the efficiencyof common law constitu-
tionalismas opposed to the subconstitutionalcommon law. What follows
is thereforein the nature of a preemptivediscussion,justifiedby the close
connections between the Burkean and Cardozean approaches. It is inevi-
table thatsooner or later- probablysooner- someone will attemptto ar-
bitrage these models into law generallyand constitutionallaw in particu-
lar. It is sensible to begin to consider theirimplications.
As before, I both raise internal questions about these models and
examine theirimplicationsfor constitutionallaw. Internally,the models
currentlypresuppose a sharp distinctionbetween overrulingand distin-
guishing, but that distinction is untenable. As for implications, the
chronic circumstancesof constitutionallaw- especially its politicallypo-
-
larizing character make it unlikelythat convergence to efficiencywill
occur in constitutionalsettings,according to the models' internallogic.
Law,35J.Comp.Econ. 309 (2007) [hereinafter Gennaioli& Shleifer,Overrulingand the

of Law].
Instability
128. AlthoughGennaioliand Shleiferdo considerhowtheirmodelofoverruling can
be extendedto incorporatedemand-sideeffects, theydo not do so fortheirmodel of
distinguishing.See Gennaioli& Shliefer,Overruling ofLaw,supranote
and theInstability
127, at 320-22. In anyevent,the distinctive and novelfeatureof theirapproachis the
focuson the supplyside- on efficiency as the unintendedbyproduct of uncoordinated
decisionmaking bya seriesof biasedjudges overtime.
129. See Gennaioli& Shleifer, Evolutionof CommonLaw,supranote 127,at 62.
1. Overruling, and Constitutional

Distinguishing, Law. - To date, the
Cardozean models presuppose a distinctionbetween two mechanisms of
common law change: the overrulingof cases and the distinguishingof
cases.130 In these models distinguishing,but not overruling,tends toward
efficiency,131 although in both cases a high degree of polarization across
judges will dampen those tendencies. These resultsarise from the ana-
lyticsetup. In the model of distinguishing,all judges are assumed to be
more or less biased in one direction or another- too much liabilityor
too little,relative to optimal liability. When Case 1 is decided in a way
that conflictswith the biases of the judge sittingto decide Case 2, the
Case 2 judge maydistinguishthe precedent,but onlyon relevantor mate-
rial dimensions. If in Case 1 the court held that dog owners are always
liable when strangersare bitten,then in Case 2 the court maydistinguish
the Case 1 precedent by modifyingthe rule to say that liabilitydoes not
attach when the dog was leashed. But the Case 2 court may not say that
liabilitydoes not attach because the bite occurred on a Tuesday.
The resultis thatdistinguishingcases over timeadds materialdimen-
sions to the law, and this is a kind of extra informationthat adds preci-
sion to legal rules; each judge who distinguishesa prior precedent con-
tributesa portion to the law's increasing precision over time.132Under
certain conditions,judges' differingbiases wash out and the benefitsof
increasing precision are all that is left. Most optimistically,the law may
even converge to an efficientregime in which rules are maximallypre-
cise, so thatthe social losses fromfailingto sortbetween materiallyunlike
cases are minimized. However, we will soon see that the conditions
under which thisoccurs are stringentand especiallyunlikelyto obtain for
constitutional(as opposed to ordinary) common law.
The overrulingmodel is different."Since overruling,unlike distin-
guishing,does not bringnew materialdimensions into the law, it leads to
the volatilityof legal rules withouta tendency to improve the law over
time. With overruling,there is no benefitof legal evolution."133Where
polarization among judges is high, the legal rule will never converge to a
stable state (let alone an efficientstable state); ratherit will fluctuatebe-
tween extremes. Convergence to efficiencycan only occur in the special
case in which unbiased judges are more likelyto vote to overrule prece-
130. Distinguishing is addressedin Gennaioli& Shleifer,EvolutionofCommonLaw,

supranote127,at 43. Overruling is addressedin Gennaioli& Shliefer,Overrulingand the
of Law,supranote 127, at 309.
Instability
131. Tends towardefficiency, butis unlikelyto convergeto fullefficiency.
Gennaioli
and Shleiferdo not claimthatdistinguishing producesthe fullefficiency ofjudge-made
law. Rather,theymake the weakerclaim thatlegal evolutionis beneficialon average,
because "the informational benefitof distinguishingimprovesthe qualityof the law."
Gennaioli& Shliefer, Evolutionof CommonLaw,supranote 127,at 47.
132. See id. at 54-56.
133. Id. at 62-63.
2007] THE UMITS OF REASON 1529
dentsthanbiasedjudges and in whichthedegreeof polarizationis only

intermediate134 - an "implausibly stringent" set of conditions.135
The Cardozeanmodelsthusdrawa sharpcontrastbetweena regime
of distinguishing, in whichcommonlaw evolutionis beneficialon aver-
age under a wide range of conditions,and a regimeof overruling, in
whichcommonlawis volatileand convergeson efficiency onlyundervery
unlikelyconditions.However,thiscontrastis an artifact of themodeling
setup. In thesemodels,materialdistinctions count as "distinguishing,"
whereasimmaterialones do not; theyare automatically classifiedas a
formof overruling.But commonlaw theoryrecognizesthatthe sheer
accumulationof material distinctionscan itselfamountto a formof over-
ruling. Cases are sometimeswidelyrecognizedas havingbeen "distin-
guishedto death,"such thatall lawyers realizetheyhavebeen implicitly
overruledor "limitedto theirfacts."Moreover,commonlaw theorists
showthatjudges oftenimplementtheirbiasesprecisely bydistinguishing
precedentsbased on immaterialor irrelevantdimensions - by holding
thatdog ownersare safeifthe biteoccurredon a Tuesday.136
It is truethatfromthestandpointof theindividual judge, usingma-
terialdistinctions will be strictly to
preferable using irrelevant distinc-
tions;theformer, unlikethe latter,bothallowsthejudge to indulgeher
biasesand also suppliesbeneficialinformation. Butthisholdsonly"until
materialdimensionsare exhausted."137 The qualifieris critical.In the
constitutional commonlaw (and perhapsalso in ordinary commonlaw),
the limitsof materiality are quicklyreached.
Considerthebewildering seriesofdistinctions thatproliferate in the
Supreme Court'sjurisprudence of state sovereignimmunity under the
EleventhAmendment,138 whereat presentstatescan be sued in federal
courtforequitablereliefbutnotfordamages,unlessCongresshas clearly
said otherwise,but only under its power to enforcethe Fourteenth
Amendmentratherthan (say) the CommerceClause, or unless the
UnitedStatesratherthan a privatepartyis the plaintiff, and so on.139
The processthatled to thispatchwork of rulesis one ofJusticesadvanc-
ingdistinctions thatwereimmaterial, giventhepriorlaw,and embedding
themin the new legal rules. Remarkably, proponentsof all competing
viewsof theAmendment seem to agreethatthecurrentrulesare wrong,
134. See Gennaioli8cShleifer, Overruling and theInstability

of Law,supranote 127,
at 317-18 (Proposition4).
135. Id. at 323.
136. See KarlN. Llewellyn, The CommonLawTradition287 (1960) ("'[A] distinction
withouta difference' is a stench. . . .").
137. Gennaioli& Shleifer, Evolutionof CommonLaw,supranote 127,at 61 n.8.
138. I use "EleventhAmendment"only as a shorthand,bracketingthe question
whetherthe Courtis here developinga bodyof law rootedin theAmendment's textor
ratherin backgroundstructural principlesof federalism.
139. Foran overview of these"complexand oftencounterintuitive of
interpretations"
theEleventhAmendment, see 1 LaurenceH. Tribe,AmericanConstitutional Law 519-66
(3ded. 2000).
given their own views; all that keeps the currentrules in place, despite
being no one's firstchoice, is that there is no agreement on what set of
rules should replace them. Perhaps a regime of rules thatare everyone's
second choice satisfiesas many people as much as possible, and is thus
efficientin one sense, but the process of adding cynicaldistinctionsdoes
not plausiblyenrich law's informationalcontent.
2. CardozeanConstitutionalism'? - Apart from the internal logic of
the models, their main implication is that constitutionalcommon law
should be both highlyvolatile and farfromefficient.One keyvariable is
the effectofjudicial polarization- the distance between the viewsof dif-
ferentjudges - on the shape and pace of legal change. Cardozo's hope
that"the eccentricitiesofjudges balance one another" and yield a typeof
"constancy" in the law turns out to be optimistic. As Gennaioli and
Shleiferexplain, in a regime of overruling,when polarization is high, the
common law will simplyfluctuateor vacillate fromone polar view to an-
other; not only will the law not converge to efficiency,it will never con-
verge at all.140
Vacillation between polar viewsis a well-knowneffectof overrulingin
the constitutionalcommon law. Consider the sequence from National
League ofCitiesv. Usery,141which announced constitutionalprotectionfor
the "essential functions"of state governments,to Garcia v. San Antonio
Metropolitan which overruled the earlier precedent-
TransitAuthority,142
with the dissentingJusticesissuing a warningthat the wheel would turn
yetagain.143 As it turnedout the dissenters,when theyobtained a major-
ity,did not reinstateNationalLeague of Cities,but theydid press forward
with alternativefederalism-protecting doctrines,such as the law of state
sovereignimmunitydiscussed above, and the principle thatstatescannot
be commandeered into enacting or enforcing a federal regulatory
program.144
In a regime of distinguishingwithoutoverruling,Cardozo's conjec-
turedoes better,because legal change throughdistinctionsadds informa-
tion thatoverrulingdoes not. Even withhigh polarization,distinguishing
alwaysadds precision to the law,and somedegree of polarization is affirm-
ativelybeneficial, because it induces distinctionsin cases where unpo-
larized judges would simplyacquiesce in whateverdecision was firstis-
sued. Although not all extant rules will be efficientat any point, the
average tendencyof common law over time will be toward efficiency.145
140. See Gennaioli& Shleifer, and theInstability

Overruling of Law,supranote 127,
at 324.
141.426U.S.833,850-52(1976).
142.469U.S.528,557 (1985).
143. See 469 U.S. at 580 (Rehnquist,
J.,dissenting);469 U.S. at 589 (O'Connor,J.,
dissenting).
144. Printzv. UnitedStates,521 U.S. 898, 925 (1997) (M[T]heFederalGovernment
may not compel the States to implement,by legislationor executiveaction,federal
regulatoryprograms.").
145. See Gennaioli& Shleifer,Evolutionof CommonLaw,supranote 127,at 60-62.
On theotherhand,highpolarizationdoes reducethebenefits ofthe

distinguishing process, because greater bias causes each judge to use
cruderand less informative distinctions.146 When polarizationpasses a
certainthreshold, thejudge-madelawtendslessstrongly to efficiencyand
may even become less because
efficient, the costs of polarizationoutstrip
thebenefits;in otherwords,thereis "an invertedU-shapedrelationship
betweenpolarizationand the efficiency ofjudge-madelaw."147The up-
shotis that
the evolutionof commonlawwouldproducemostsociallyeffi-
cientresultsin theareas oflawwherethereis roomforchange
and updating,butwherethedisagreement amongjudges is not
extreme.The relatively apoliticalyet stillchangingareas oflaw,
suchas contractand corporatelaw,are thelikelycandidatesfor
relatively efficientoutcomesresultingfromthe decentralized
evolutionofjudge-madelaw. In theextremely politicalareasof
law, by contrast,the likelihoodthat the law gets stuckon a
wrongtrajectory is higher.148
ThisconclusionimpliesthatCardozo'sdictumis leastlikelytojustify
theconstitutional commonlaw- especiallythehighlypolarizingconstitu-
tional common law governingrightsand liberties - whichis precisely
wheremanycommonlaw constitutionalists wantto staketheirclaims.149
Here somequalifications are necessary.A largesetor constitutional con-
troversiesare easy,noncontroversial, or unpolarizing.150 The subsetthat
are litigated(and the sub-subsetlitigatedto the SupremeCourt) are
more likelyto be highlypolarizing,althoughnot uniformly so. Moreo-
ver,structural constitutional lawis often,perhapson average,lesspolariz-
ingthantheconstitutional lawofrightsand liberties.Whenall thequali-
fying is done, however, the comparative pointremains:Taken together
and on average,constitutional cases are more polarizingthanthe ordi-
narycommonlaw. Relativeto ordinary commonlaw,constitutional lawis
morepervasively concernedwithdistribution ratherthancoordinationor
cooperation;the scope forcommoninterest(somehowdefined)in the
rulesgoverning contract, property, and tortis greater, whilethescope for
commoninterest(somehowdefined)in therulesgoverning abortion,or
gaymarriage, or free speech in wartime is less.
AlthoughI believethisclaimwillbe intuitively plausibleto constitu-
tionallawyers, it is hard to finddirectevidenceeitherforor againstit.
However,we mayprovideindirectsupportfortheintuitionbyobserving
thatissuesare oftenconstitutionalized precisely when,and because,they
146. See id.

147. Id.
148. Id. at 61-62.
149. See, e.g., Poe v. Ullmann,367 U.S. 497, 543 (1961) (Harlan,J., dissenting)
guaranteedbyDue ProcessClause of Constitution).
(discussingscope of "liberty"
150. Cf. FrederickSchauer,EasyCases,58 S. Cal. L. Rev.399, 407 (1985) (arguing
that constitutionalprovisionsthat are rarelylitigatedare still importantpart of
constitutionallaw).
are polarizing.Partisanentrenchments, in whichan outgoingcoalition

attempts to constitutionalize
a favored policyto bind thehandsofsucces-
sors,are routine.151 And nonjudicial institutionswillsometimesaffirma-
or
tivelysupport encourage the judges to resolvecontroversialissues
throughconstitutional review,152 preciselybecause thoseissuesare so po-
larizingthatlegislators
on all sideshope to duckresponsibilityor to make
a politicaltargetof thejudges.
Conclusion
Much of thisessayhas been critical,aimingto expressskepticism

about recentargumentsfor common law constitutionalism that draw
upon sophisticated social-scientific
models of precedent. However, I also
hope the critique has constructive byproducts, at both the methodologi-
cal and thesubstantive level. Methodologically,I have offereda positive
interpretation of the projectof commonlaw constitutionalism, as an at-
to
tempt provide social-scientific
underpinnings forBurke 's loose reflec-
tionson thelimitsofhumanreason,whilealso translating Burkeanviews
to the constitutionalsetting. So too I have attemptedto update
Bentham'scritiqueof the constitutional commonlaw,adaptingit to the
distinctivesettingof the constitutionalcommon law. Evenifthecritique
ofcommonlawconstitutionalism offeredherefails,itis affirmatively use-
fulto translatethe commonlaw's historiccompetitorinto the constitu-
tionalsetting.
Substantively,I havesuggestedthattheinformational and evolution-
ary mechanisms canvassed above lead most naturally,not to Burke,butto
JamesBradleyThayer - to extensivejudicial deference to legislativeen-
actments.To theextentthatinformational and evolutionary mechanisms
yieldinsights about the conditions under which courts should trumpleg-
islationon thebasisofconstitutional precedent,thoseconditionsare pre-
dictablyverynarrow.The first virtueofanytheoryofconstitutional adju-
dicationis a theoryof judicial review - of judicial power to override
commands.Buttheverymechanismsthatcommonlawconsti-
legislative
tutionalistsinvokein praiseofconstitutional precedentcastlegislationin
a betterlightstill. If manyjudicial mindsare betterthanone or a few,
manylegislative mindsare plausiblybestof all.
151. See, e.g.,HowardGillman,How PoliticalPartiesCan Use theCourtsto Advance

TheirAgendas:FederalCourtsin theUnitedStates,1875-1891,96 Am.Pol. Sci. Rev.511,
521 (2002) (describingRepublicanParty'ssuccessful effortsto entrenchitspolicyagenda
in courtsduringlate nineteenthcentury).
152. See KeithE. Whittington,"InterposeYourFriendly Hand": PoliticalSupportsfor
theExerciseofJudicialReviewbytheUnitedStatesSupremeCourt,99 Am.Pol. Sci. Rev.
583, 591-93 (2005) ("Whenfacedwithsuch issues,electedofficials mayactively seek to
turnovercontroversial politicalquestionsto the courtsso as to circumvent a paralyzed
legislatureand avoid the politicalfalloutthatwould come withtakingdirectaction
themselves.").

10 2307@40041740 PDF

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

10 2307@40041740 PDF

Hochgeladen von

Copyright:

Verfügbare Formate

Common Law Constitutionalism and the Limits of Reason

Author(s): Adrian Vermeule

You may need to log in to JSTOR to access the linked references.

COMMON LAW CONSTITUTIONALISM AND

A centralclaim of commonlaw constitutionalism has beenthat

* Professor of Law, Harvard Law School. Thanks to Adam Cox,

mon law is not plausiblyseen as a repositoryof latentwisdom. The rele-

1. EdmundBurke,Reflections on the Revolutionin France74 (FrankTurnered.,

The central theme in the followingcritique is this: Argumentsfor

tional law,where litigationis infrequentand stakesare chronicallydistrib-

This Partexplains,and questions,theinformationalversionof com-

4. For example, on the basis of the value of stabilityand protectingsettled

approach to constitutionaladjudication and details a Condorcetian inter-

For other Burkeans, however,traditionhas instrumentalor deriva-

7. See, e.g., Merrill,Bork v. Burke,supra note 5, at 515 (proposingthatchoice

lar cases attempt,on a blank slate, to figureout for themselveswhat is

9. See, e.g.,Strauss,CommonLaw Interpretation,

12. Cf. Cass R. Sunstein,Commentary, IncompletelyTheorizedAgreements,108

3. Precedent,Tradition,and theJuryTheorem. - The CondorcetJury

15. FortheCondorcetian ofBurke,see Sunstein,BurkeanMinimalism,

tainlyexceed the competence even of the group's wisestmembers,if the

20. Sunstein,BurkeanMinimalism, supranote 5, at 371.

mic termsnecessarilyassume the existence of rightanswers,suitablyde-

24. See DavidA. Strauss,Tradition,PrecedentandJusticeScalia, 12 CardozoL. Rev.

vote" standard in reapportionmentcases are justified by pointing to a

29. Id. at 31-32.

knowledge we have gained in the succeeding years. Experience in this

31. This pointdoes not hold forthepollinginterpretation, whichwe haveputaside

38. Trop v. Dulles,356 U.S. 86, 101 (1958).

line of decisions are thought of as deciding, in common, on the same

41. See Estlundet al., supranote 32, at 1326 (Jeremy

peach the Condorcetiancredentialsof otherinstitutions, such as legisla-

49. Bikhchandani et al., Behaviorof Others,supranote 44, at 157-58.

that a strategythat is individuallyrational forjudges at any given time-

53. In principle,judges mightpreventthe Burkeanparadoxfromarisingbyissuing

tional distortion)or as in the grip of "interest-begotten prejudice" (a cog-

56. Bentham,Handbook,supranote 30, at 34-36.

a narrow,professionally inflectedperspectiveon governance."61Even

61. RichardA. Posner,Law,Pragmatism, and Democracy128 (2003).

66. I take this"publicmeaning"originalismto have become the standardversion,

claim thatthe original understandingjust is whatthe Constitutionmeans,

67. See GaryLawson,On ReadingRecipes. . . and Constitutions,85 Geo. L J. 1823,

70. Cf. Strauss,Common Law Genius,supra note 27, at 8 (arguingthat under

temptingto guess the welfare-maximizing constitutionalrule will know

In the hardest case, where subsequent precedent contradicts text

73. See JamesBradleyThayer,The Originand Scope of the AmericanDoctrineof

tionality(whetheror not the legislatorsalso address its meritsas policy).

thatlegislators are on averageonlyslightly betterto gettherelevantques-

76. When the House leadershipheld open votingforthreehoursto allow timeto

prohibitionson ex parte contacts and equal time for briefingand argu-

79. See Neil Komesar, ImperfectAlternatives:Choosing Institutionsin Law,

of informationnets out against the sheer numerosityof legislators. In

82. See Cesar Martinelli,

tors' initial informationaladvantages,just as a runner who gets a cramp

85. Undertheconventional maynotentrenchstatutesbymakingit

Supposethatat Time 1, nineJustices votedunanimously forRule R, butthenat Time 2, a

traditions,aggregates latent informationrelevant to constitutionallaw,

91. Strauss,CommonLaw Genius,supranote 27.

tive"is a legal claim about formallines of authorityand removalthatdoes

explicitand valid96comparison - thatthe aggregatedwisdomof many

96. Assumingawaythe internalproblemsdiscussedin I.B.

meritsofverymanyminds,at leastcomparedto legislatures.Social tradi-

98. Jeremy Waldron,CustomRedeemedbyStatute,51 CurrentLegal Probs.93, 100