Review Essay

How a Judge Thinks

How Judges Think. By Richard A. Posner, Harvard

University Press, 2008

Reviewed by Michael J. Gerhardt†

The question of how judges think is surprisingly difficult

and contentious. The question requires defining not only the
essential features of judging but also distinguishing them from
the distinctive aspects of what legislators and executive author-
ities do. The question has long vexed many jurists as well as
laypeople, political leaders, and particularly academics, who
have expended enormous time and energy in trying to figure
out precisely how judges make decisions—in the same way as
politicians do or differently?1 Scholars extensively explore, in
other words, whether, or to what extent, judges base their deci-
sions strictly on the law and not on political considerations.

† Samuel Ashe Distinguished Professor in Constitutional Law and Di-

rector of the UNC Center for Law and Government, University of North Caro-
lina at Chapel Hill Law School. B.A. Yale University; M.Sc. London School of
Economics; J.D. University of Chicago. I am grateful for the helpful sugges-
tions of Mitu Gulati and David Klein on an earlier draft. Copyright © 2009 by
Michael J. Gerhardt.
1. For a small sampling of the numerous works by legal academics and
social scientists on how judges and Justices decide cases, see LAWRENCE
BAUM, JUDGES AND THEIR AUDIENCES (2006); RONALD DWORKIN, LAW’S EM-
PIRE (1986); MICHAEL J. GERHARDT, THE POWER OF PRECEDENT (2008); THO-
MAS G. HANSFORD & JAMES F. SPRIGGS II, THE POLITICS OF PRECEDENT IN THE
U.S. SUPREME COURT (2006); JEFFREY A. SEGAL & HAROLD J. SPAETH, THE
SUPREME COURT AND THE ATTITUDINAL MODEL REVISITED (2002); SUPREME
COURT DECISION-MAKING (Cornell W. Clayton & Howard Gillman eds., 1999);
Jack Knight & Lee Epstein, The Norm of Stare Decisis, 40 AM. J. POL. SCI.
1018 (1996); Mark J. Richards & Herbert M. Kritzer, Jurisprudential Regimes
in Supreme Court Decision Making, 96 AM. POL. SCI. REV. 305 (2002); Emer-
son H. Tiller & Frank B. Cross, What Is Legal Doctrine?, 100 NW. U. L. REV.
517 (2006).

2185
2186 MINNESOTA LAW REVIEW [93:2185

The efforts of scholars and others to clarify how judges do

their jobs undoubtedly have had some influence on public un-
derstanding of judging. The recent presidential election was a
dramatic reminder of the efforts of national political leaders to
shape the composition and direction of the federal courts, par-
ticularly the Supreme Court. The election showcased, inter
alia, the major presidential candidates’ efforts to talk about
judging in ways that were primarily designed to fortify their
bases, attract political support, and advance their political
agendas, including their plans for (re-)shaping the federal judi-
ciary.2 For their part, legal scholars have not done as much as
they could to edify the public about what judges do. Indeed,
they have been frequently chastised for their overly abstract
and partisan commentaries on judging.3 It is thus probably no
surprise that law schools have not taught law students as much
they could about how judges actually think. Law schools help to
train law students to read judicial opinions critically and to
construct (and deconstruct) theories of judging, but they rarely
instruct their students on how to write judicial opinions (as ei-
ther trial judges or members of multi-judge tribunals). Legal
academics have failed to develop consensus on what judging en-
tails and the criteria for evaluating judicial performance.
Into this thicket of uncertainty and disagreement comes
the eminent legal scholar and judge, Richard Posner, to clarify
once and for all how judges really think.4 Posner promises to
turn back the curtain to expose what judges do. Posner’s views
on judging should be of interest to anyone who seeks to under-
stand the judicial function. His book provides an excellent
springboard to discuss some of the most important recent ef-
forts to explain and to evaluate judging.
This Essay consists of four parts. In the first Part, I ana-
lyze several of the most prominent models produced by scholars
to explain judging, each of which Posner analyzes in his book.

2. See, e.g., Neil A. Lewis, Stark Contrasts Between McCain and Obama
in Judicial Wars, N.Y. TIMES, May 28, 2008, at A17.
3. Cf., e.g., ROBERT J. SPITZER, SAVING THE CONSTITUTION FROM LAW-
YERS: HOW LEGAL TRAINING AND LAW REVIEWS DISTORT CONSTITUTIONAL
MEANING 33–59 (2008) (describing how student-run law reviews and pressure
on faculty to publish yield dubious scholarship); Harry T. Edwards, The Effects
of Collegiality on Judicial Decision Making, 151 U. PA. L. REV. 1639, 1640–41
(2003) (noting that some academics ignore the importance of collegiality).
4. RICHARD A. POSNER, HOW JUDGES THINK (2008). Although I refer
throughout this Essay to the author as Posner, I do so strictly for the sake of
brevity. I do not mean to convey any disrespect to Judge Posner by this short-
hand.
2009] HOW A JUDGE THINKS 2187

Like Posner, I find each of these models to be somewhat illumi-

nating but on balance to be incomplete and misleading.
In Part II, I examine Posner’s theory of judging (and the
principal models on which it draws), which I ultimately find
unpersuasive, unrealistic, and unsupported. Posner’s account
perhaps works best to describe how Posner himself makes deci-
sions as a judge. While I cannot prove this, Posner’s account
conceivably is his sophisticated effort to defend his conception
of himself as a “good judge” in spite of the political-science lite-
rature (which he largely accepts) demonstrating the wide-
spread impact of judges’ political biases on their decisions. In-
deed, the book restates Posner’s longstanding (mistaken) belief
that other judges are usually nothing more than faint imita-
tions of Posner himself. Moreover, Posner, like many other
theorists, mistakenly presumes that that there is, or should be,
only one correct way for judges to behave. Yet, the Constitution
allows national political leaders to construct a federal judiciary
with diverse approaches to construing the Constitution and
other laws, and thus the Supreme Court and other courts have
been composed of jurists with varying approaches to constitu-
tional interpretation. Since the Constitution vests responsibili-
ty over judicial selection in national political authorities, the
latter serve as a principal check on the composition and direc-
tion of the judiciary and thus help either to keep off the courts,
or to maintain on the courts, particular judicial attitudes and
methodologies.
In Part III, I examine various constraints on judging.
While I agree with Posner that the federal system is not de-
signed to ensure merit appointments (the appointments of the
supposedly “best-qualified” people as judges) and that state
merit-selection systems are undervalued in the academy, I ar-
gue that he has undervalued the significance of judicial
precedent, judges’ concerns about their reputations, and the fil-
ters and incentives within the federal judicial-selection process.
Part IV considers how legal scholars and law schools may
better serve as meaningful checks on judicial performance. I
agree with Posner that some objective measures of judicial me-
rit might be conceivable, but law schools should do more than
what he suggests—offering law students training in social
science methodologies. Law schools should also expand their
resources for training law students to solve problems in all of
the areas in which lawyers function today, including but not
limited to the legislative process, administrative agencies, ex-
2188 MINNESOTA LAW REVIEW [93:2185

ecutive offices, banks, and businesses. Law professors, too,

could modify or rethink their own professional activities. Per-
haps most importantly, they could benefit from some humility.
Law professors usually try—or promise—to shift fundamental
paradigms in legal (and judicial) thinking, but most do not suc-
ceed. I agree with the many other legal scholars who have proc-
laimed that we should speak truth to power, but the truth we
speak to power should not just be a partisan scolding (as it so
often is) but useful and meaningful instruction in the skills and
the discipline that law students—and lawyers—will need in or-
der to be effective problem solvers as judges, advocates, legisla-
tors, or any of the myriad other roles that they are called
upon—or take the initiative to perform—in today’s society.

I. FIVE POPULAR MODELS OF JUDGING

Posner recognizes that, in writing about judging, he is not
writing on a blank slate. He is hardly the first jurist or eminent
scholar to construct a theory of judging. In this Part, I discuss
five models of judging that are particularly popular and in-
fluential within the academy. While Posner does not discuss
these models in great detail, they are useful for developing an
overview of the literature to which Posner’s new book attempts
to make a contribution,
The first prominent model is attitudinalism, which sug-
gests that judges directly vote their policy preferences.5 This
theory is based on extensive empirical analyses, which measure
judicial voting on the basis of whether it affirms a conservative
or liberal outcome or policy. Attitudinalists find, inter alia, that
judges and Justices primarily ground their decisions not on the
law (formally defined as the texts of statutes, the Constitution,
and precedent) but rather on the basis of factors external (or
exogenous) to the law, such as judges’ policy preferences.6
Hence attitudinalists argue, for instance, that if a judge likes
school vouchers then she simply votes to uphold their constitu-
tionality. All the rest is window dressing at best, subterfuge at
worst.
While Posner finds that this model has some merit, he
suggests that it fails to explain the many nonroutine cases in
which legal materials point to relatively clear outcomes and

5. See generally SEGAL & SPAETH, supra note 1 (describing the tenets of
the attitudinal model).
6. Id. at 86–96.
2009] HOW A JUDGE THINKS 2189

thus leave little sway for judges’ policy preferences.7 An addi-

tional problem is that, even in the hard cases of constitutional
law in which no single source points to an obvious answer,
there are many unanimous or nearly unanimous decisions in
which the Justices agree in spite of having arguably different
policy preferences. Moreover, the hard cases of constitutional
law are hard in part because the choices confronting the Justic-
es do not involve easy trade-offs; the cases might involve con-
flicts not between one value a Justice cherishes and another
she does not but rather among several values that the Justice
cherishes.8 Attitudinalism is undercut further by the presump-
tion that the categories of liberal and conservative are fixed
over time (they are not) and that judicial attitudes are imper-
vious to change (they may not be). More importantly, attitudi-
nalists do not rule out legal variables as a basis of, or relevant
to, judicial decision making; instead, the most that they have
shown is the difficulty if not impossibility of ruling out the
claim that political preferences play some role in how judges
think.
Second, Posner addresses the rational choice model,9 which
is very popular in the social sciences. This model suggests that
judges act strategically; they manipulate cases and other legal
materials to maximize their personal policy preferences.10
Posner counters that this model is too simplistic or myopic, and
posits that judges might have preferences other than policy or
political ones, though he fails to demonstrate, at least empiri-
cally, the extent or the nature of these other possible prefe-
rences.11 A more serious problem is that the rational choice
model cannot be disproved. Although common sense suggests
that judges might vote strategically, the model is completely
circular: it presumes that every outcome is strategic and thus
treats every outcome as strategic.12 Even so, this model, like at-
titudunalism, does not rule out the relevance of legal variables;
it suggests, at most, that we cannot yet rule out the possibility

7. See POSNER, supra note 4, at 27–28.

8. Id.
9. See id. at 29–31 (describing the strategic model). For an example of an
application of the rational choice model to the impact of precedent in judicial
decision making, see Knight & Epstein, supra note 1.
10. SEGAL & SPAETH, supra note 1, at 97–109 (discussing the rational
choice model and its relationship to the attitudinal model).
11. See POSNER, supra note 4, at 73–76.
12. Cf. SEGAL & SPAETH, supra note 1, at 97–109.
2190 MINNESOTA LAW REVIEW [93:2185

that strategic considerations are entirely irrelevant in judicial

decision making.
The third model is sociological and posits that panel com-
position influences outcomes.13 In this theory, a panel consist-
ing of judges who do not all belong to the same party is more
likely than a homogenous one to seek compromise among the
disparate views of the judges.14 Posner suggests that one rea-
son for the possible differences between homogenous and di-
verse panels is the general aversion of judges to dissent.15
Another problem is that legal disputes are not perfectly fungi-
ble; the factual differences between cases are quite pertinent to
their disposition. Without the ability to compare how two diffe-
rently constituted panels vote on the very same case, the socio-
logical model lacks a coherent or sensible measure by which to
compare what a homogenous panel does in a particular area of
law to what a differently composed court does in a different
case in the same area of the law.
Organizational theory is a fourth model, which considers
judges to be agents whose function is to serve a principal.16
This theory posits that as agents judges are obliged to be inde-
pendent from partisan influences,17 but this conception of
judges is flawed in at least two respects. First, the theory lacks
criteria for determining the principal whom judges serve. Posn-
er suggests that one possible principal whom judges serve is
the federal government;18 this is, however, problematic since it
would mean that federal judges might have a conflict of inter-
est in every case in which the federal government is a party. A
more plausible principal that federal judges serve is the people
of the United States or the Constitution,19 but organizational
theory does not tell us how to decide which of these is the cor-

13. See, e.g., CASS SUNSTEIN ET AL., ARE JUDGES POLITICAL?: AN EMPIRI-
CAL ANALYSIS OF THE FEDERAL JUDICIARY 17–45 (2006).
14. Id. at 22, 23 fig.2-1.
15. See POSNER, supra note 4, at 32–34, 51 (discussing the various rea-
sons that judges are reluctant to author dissenting opinions).
16. See generally Donald R. Songer et al., The Hierarchy of Justice: Test-
ing a Principal-Agent Model of Supreme Court-Circuit Court Interactions, 38
AM. J. POL. SCI. 673 (1994) (discussing the relationship of inferior courts as
agents of their Supreme Court principal).
17. Cf. POSNER, supra note 4, at 39; Songer et al., supra note 16, at 675
(explaining that a faithful agent-judge would apply the law as interpreted by
the Supreme Court unaffected by other considerations).
18. See POSNER, supra note 4, at 125–26.
19. See id. at 126.
2009] HOW A JUDGE THINKS 2191

rect principal whom judges serve or how judges may properly

serve either or both.
Perhaps the most popular model among legal scholars is
legalism, the classic view that judges and Justices base their
decisions on the law.20 Posner suggests that legalism is over-
rated.21 He suggests that legal materials are rarely so clear
that they provide unobjectionable, easily identifiable solutions
to legal disputes.22 Instead, he suggests, legal materials almost
invariably require judges to exercise discretion to interpret
them and to apply them to particular disputes.23 Since the law
provides no clear mechanisms for constraining this discretion,
it cannot be explained on the basis of legalism. Posner posits
that jurists, particularly those on the Supreme Court, are
moved less by “the law” and more by their own attitudes about
the law.24 Since law “is suffused with ideology,” Posner agrees
with Roscoe Pound that “the law” consists of legal doctrines,
“techniques for deriving and applying doctrines,” and “social
and ethical (in a word, policy) views.”25 The Pound view of judg-
ing seems closely to approximate that of Posner, though, as we
will see, their common view turns out surprisingly to lack em-
pirical verification.

II. POSNER ON JUDGING

There are many accounts of judging by prominent judges,
of which Richard Posner’s is the most recent.26 Posner’s account
is nonetheless significant because of his eminence as a jurist
and scholar and because he has grounded his theory in several
other respectable theories of judging and a distinguished tradi-
tion of judging. Indeed, much of what Posner has tried to do in
his book (and career) is to “modernize” great jurists such as
Benjamin Cardozo and Oliver Wendell Holmes, Jr., to demon-
strate a link (albeit a sophisticated one) between their impres-
sions about judging and his particular understanding and exer-

20. See generally Frederick Schauer, The Limited Domain of the Law, 90
VA. L. REV. 1909 (2004) (analyzing the viability of the view that the law acts
as a “limiting principle”).
21. See POSNER, supra note 4, at 7–9, 41–56.
22. Id. at 47.
23. Id. at 49.
24. Id. at 42–44.
25. Id. at 43 (citing Roscoe Pound, The Theory of Judicial Decision (pt. 3),
36 HARV. L. REV. 940, 945–46 (1923)).
26. E.g., BENJAMIN N. CARDOZO, THE NATURE OF THE JUDICIAL PROCESS
(1921); O.W. HOLMES, JR., THE COMMON LAW (1881).
2192 MINNESOTA LAW REVIEW [93:2185

cise of the judicial function. By relying on other theorists’ mod-

els of judging and linking himself to a distinguished tradition of
judging, Posner attempts to avoid the problem that his account
of judging, like that like of other judges, is self-serving and
nothing more than self-rationalization. In this, however, he on-
ly partially succeeds.
To understand Posner’s theory of judging, it helps to know
the other theories on which it most heavily relies. The first is
psychological, which “focuses on strategies for coping with un-
certainty.”27 Posner believes that uncertainty is widespread in
the law (in spite of what he has said about the extent of non-
routine cases with clear, determinate answers),28 and so legal
disputes allow a wide swath for judges’ personal psychologies to
influence their decision making. Equally important to Posner’s
understanding of judging is the economic theory of judging,29
which is a variation on the rational choice model. It stresses
that judges are, like everyone else, utility maximizers; they de-
rive professional satisfaction in many different ways, including
but not limited to the prestige of their jobs, their power or in-
fluence, and their leisure time.30 Because Posner further be-
lieves that the law eventually runs out—i.e., there is a point at
which legal materials simply do not provide the answers to the
legal questions confronting the courts—it leaves judges with no
alternative but to become pragmatists.31 But, Posner suggests
that once the law runs out, judges and Justices end up deciding
cases on the basis of their assessments of the likely conse-
quences of their decisions.32 As described by Posner, pragmat-
ism, which is the conception of judging he has defended (and
practiced) throughout his career, is forward-looking and “lo-
cal[ized],” meaning that it leads judges to tailor a solution to
the problem in front of them.33 Pragmatism is closely related to
the phenomenological model, which stresses the importance of
both intuitions and predispositions to judicial decision mak-
ing.34 Bayesian decision-making theory emphasizes that “pre-

27. POSNER, supra note 4, at 35.

28. Id. (calling uncertainty “a fundamental characteristic of the U.S. legal
system”).
29. Id. at 35−39.
30. Id. at 35–36.
31. Id. at 231.
32. Id. at 40.
33. Id. at 241.
34. Id. at 40.
2009] HOW A JUDGE THINKS 2193

conceptions play a role in rational thought,”35 while phenome-

nological theory suggests that judging is based on how it must
(or does) feel to judges.36 Posner suggests that many books on
judging written by prominent judges describe their impressions
of what judging entails and that these impressions illuminate
how judges think.
Posner synthesizes these theories into a model that also
takes into account the ramifications of the special institutional
contexts in which different judges operate. In particular, dis-
trict and courts of appeals judges are subject to various internal
constraints (such as dissent aversion37 and legalistic tech-
niques)38 and external constraints (such as appellate review,39
the norm of following the precedents of higher courts,40 and
judicial elections in states and congressional control over fund-
ing and jurisdiction in the federal system).41 But, Supreme
Court Justices are relatively free from all of these constraints
to act upon their personal or policy preferences, and so Posner
suggests that it is this relative freedom from the usual con-
straints on judging that leads Justices to make decisions as
most political scientists find: on the basis of their political bi-
ases more than the law.
In Posner’s view, psychological, phenomenological, econom-
ic, and pragmatic theories have different explanatory force, de-
pending on whether one is trying to explain lower court or Su-
preme Court judging. For instance, he suggests:
[B]ecause an American judge, especially at the appellate level, is an
occasional legislator [who has to fill in the gaps and ambiguities of the
law,] yet with no constituency to answer to, his judging is likely to be
influenced by temperament, emotion, experience, personal back-
ground, and ideology (influenced in turn by temperament and expe-
rience), as well as by an “objective” understanding of what would be
the “best” legislative policy to adopt in order to resolve the issue in
the case.42
Posner further describes the work product of a judge as “re-
flect[ing his or her] preferences,”43 which differ “depending on a

35. Id. at 67.

36. Id. at 40.
37. Id. at 32−34.
38. Id. at 175−202.
39. Id. at 142.
40. Id. at 135.
41. Id. at 156.
42. Id. at 174.
43. See id. at 253 (comparing judges to legislators while simultaneously
distinguishing them).
2194 MINNESOTA LAW REVIEW [93:2185

judge’s background, temperament, training, experience, and

ideology, which shape his preconceptions and thus his response
to arguments and evidence . . . .”44 In addition, Posner suggests
that judges are ultimately “constrained pragmatists” who “ac-
knowledge[] the inevitability that like cases will often be
treated inconsistently not only because different judges weigh
consequences differently . . . but also because, for the same rea-
sons, different judges see different consequences.”45
Posner’s account thus far is, however, unpersuasive for
several reasons. First, it seems to boil down to nothing more
than the proposition that no two judges are alike. This sounds
like common sense, but it is a remarkably anticlimactic insight
given the sophisticated theorizing that led up to it. More impor-
tantly, it eludes verification, for it posits too many variables
that we can never really verify, such as each judge’s tempera-
ment, psychology, political and other preferences, life expe-
rience, and predispositions. It is ultimately a theory that is im-
possible to disprove.
Second, Posner makes many assertions throughout the
book that are in tension with his general description of how
judges think. For instance, he maintains, in spite of the account
set forth above, that courts of appeals judges “are strongly mo-
tivated to adhere to precedent, not only because they want to
encourage adherence to the precedents they create, but also—
and this is more important to most judges—because they want
to limit their workloads.”46 Moreover, Posner has failed to dem-
onstrate the precise circumstances in which—in lower courts—
ideology trumps or is trumped by various other factors, includ-
ing precedent. Indeed, at least one new study suggests, con-
trary to Posner’s suggestion, that ideology is not all that counts
in judging even on the Supreme Court.47 Nor does Posner prove
that judges who do not think of themselves as pragmatists real-
ly are pragmatists. The fact that these judges are trying, but
not succeeding perfectly, to be “legalists” does not transform
them into pragmatists; it makes them imperfect legalists.

44. Id. at 249.

45. Id.
46. Id. at 145.
47. See Paul H. Edelman, et al., Measuring Deviations from Expected Vot-
ing Patterns on Collegial Courts, 5 J. EMPIRICAL LEGAL STUD. 819 (2008) (sug-
gesting and applying a method of measuring how far the votes in a given case
depart from the pattern one would expect to see if Justices divided strictly
along ideological lines and noting nonideological considerations).
2009] HOW A JUDGE THINKS 2195

Posner simply fails to demolish legalism, something he has

tried to do throughout his career. A central feature of legalism
that Judge Posner singles out for criticism is reasoning by
analogy.48 But, the fact that Posner demonstrates some logical
fallacies in reasoning by analogy does not refute its utility or
omnipresence in judicial decision making. The most that Posn-
er has shown is that that he has not been persuaded to employ
reasoning by analogy, but he has not shown that the late
Charles Black was wrong when he asserted more than four
decades ago that “the case [method] remains paramount in im-
portance and in authority. This is the heart-method of Anglo-
American legal reasoning.”49 If the quest is to construct a posi-
tive account of what judges do—a principal objective of Posner’s
book—then reasoning by analogy is what most of them are ac-
tually trying to do most of the time. It seems likely that tempe-
rament and experience, among other factors, might influence
reasoning by analogy, but Posner has not yet proven it.
Third, Posner’s own theorizing undercuts his theory of
judging. Throughout his career, Posner has insisted that the
law is a soft science rather than a hard science, like physics, in
which the practitioners agree on the premises and the stan-
dards by which their work may be judged. The problem, at least
for Posner, is that, in law, judges often operate from different
premises (or preconceptions, as Posner would have it).50 But, if
judges operate from different premises and preconceptions, it is
unclear on what basis they can be fairly compared with each
other. It makes sense to compare their default rules and pre-
conceptions, but when these things are different we are doing
nothing more than comparing apples and oranges.
Consider what we can learn from a more sophisticated
analogy between legislating and judging than the one sug-
gested by Posner. In theory if not fact, legislators can be de-
scribed as one of three different kinds: perfect representatives
(reflections of their constituents), agents (who follow their con-
stituents’ preferences perfectly), or fiduciaries (who try to act in

48. Id. at 180–91.

49. CHARLES L. BLACK, JR., STRUCTURE AND RELATIONSHIP IN CONSTITU-
TIONAL LAW 4 (1969).
50. See, e.g., RICHARD A. POSNER, AN AFFAIR OF STATE: THE INVESTIGA-
TION, IMPEACHMENT, AND TRIAL OF PRESIDENT CLINTON 230–40 (1999) (de-
scribing different approaches of public intellectuals to the Clinton-Lewinsky
investigation and noting that silence “may have been due in part to an inter-
nal war between their professed values . . . and their politics”).
2196 MINNESOTA LAW REVIEW [93:2185

the best interests of their constituents).51 Legislators might ad-

here to one of these models all the time or just some of the time,
and on a given piece of legislative business a legislator might
try to adhere to some combination or all of these models (in at
least some respects). We lack perfect information on which
model or models each legislator adheres to (and when). Indeed,
it makes little sense to compare how one legislator functions
with another without knowing each one’s conception of his job.
The same, as Posner is wont to say, is true for judges.

III. POTENTIAL CONSTRAINTS ON JUDGING

Social scientists and legal scholars devote enormous time
and energy to pondering whether there are any meaningful
constraints on judicial performance.52 Posner’s apparent under-
standing of these constraints is unorthodox. He describes as
“internal” those constraints on judging that naturally originate
in or are organically a part of the minds of judges53 while he
describes every other possible constraint as “external.”54 This
characterization of constraints is confusing if not confused,
since it is the opposite of how social scientists and legal scho-
lars analyze or conceptualize constraints. According to most
people who study judging, internal constraints are those en-
demic to, or part of, the legal process, while external con-
straints are those supposedly imposed from outside the legal
process onto judicial decision making.55 Social scientists and le-
gal scholars debate whether precedent, as something internal
to the law, actually constrains Justices or whether Justices
ground their decisions not on the law but rather factors exter-
nal to the law (such as their own political preferences).

51. See generally WILLIAM N. ESKRIDGE, JR. ET AL., CASES AND MATE-
RIALS ON LEGISLATION 66–76 (4th ed. 2007) (describing theories of the func-
tions performed by legislators).
52. See, e.g., LEE EPSTEIN & JACK KNIGHT, THE CHOICES JUSTICES MAKE
(1998); Lee Epstein & Gary King, The Rules of Inference, 69 U. CHI. L. REV. 1
(2002); Stefanie Lindquist & Frank Cross, Empirically Testing Dworkin’s
Chain Novel Theory: Studying the Path of Precedent, 80 N.Y.U. L. REV. 1156
(2005); sources cited supra note 1.
53. POSNER, supra note 4, at 174 –203.
54. Id. at 125–57.
55. See GERHARDT, supra note 1, at 68 (discussing the dominant social-
science models of precedent and noting that external constraints are “factors
external to the law, such as the justices’ personal or policy preferences” while
internal constraints are factors “internal to the law, including the Constitution
or precedent”).
2009] HOW A JUDGE THINKS 2197

Interestingly, Posner recognizes there are at least two in-

ternal constraints on lower court judging that exert no similar
binding restraint on Supreme Court Justices. The first is
precedent (defined as prior judicial decisions that are arguably
on point), which Posner declares that district and particularly
courts of appeals judges are “strongly motivated” to follow.56
Moreover, he describes appellate judges as generally averse to
dissenting because they prioritize collegiality.57 Posner does
not, however, believe that precedent strongly constrains Su-
preme Court Justices, because they are not accountable to any
higher tribunal; they are the final authority on their own
precedent.58 Precedent, in other words, operates horizontally on
the Supreme Court (as persuasive authority) as opposed to ver-
tically (as an inflexible mandate imposed by a superior court
upon an inferior one). Nor does Posner find Justices to be
averse to dissent or unusually influenced by collegiality con-
cerns; they are, in his judgment, almost entirely free to vote on
the basis of their respective ideologies (theories of the law).
Otherwise, the Justices are relatively immune to public opi-
nion, though political checks, including the continued support
of both Congress and the President, are instrumental to their
stature in the constitutional order.
There are difficulties, however, with Posner’s analyses of
precedent and collegiality as judicial constraints. First, he un-
derstates the importance of precedent as a constraint on the
Supreme Court. Posner mistakes the severely limited extent of
robust path dependency in constitutional adjudication—
conditions under which precedents rigidly mandate particular
outcomes—for its complete absence. My own research indicates
that, in fact, Supreme Court precedent has limited path depen-
dency horizontally, i.e., constitutional decisions rarely robustly
foreclose or compel particular outcomes.59 I have further found
that there is actually substantial stability in constitutional law;
the Court rarely overturns itself and, in any given Term, it is
rare for more than only a few constitutional doctrines to be in
flux.60 Indeed, most constitutional doctrines—and the Court’s

56. POSNER, supra note 4, at 145.

57. Id. at 32–34.
58. See id. at 51–56, 275–77, 345–46 (discussing recent Supreme Court
decisions that have overruled long-standing precedents and concluding that
Supreme Court precedent is not constraining Justices).
59. GERHARDT, supra note 1, at 79–110.
60. See id. at 9–46 (analyzing historical patterns of Supreme Court
precedent and noting the relatively few instances where prior precedent was
2198 MINNESOTA LAW REVIEW [93:2185

attachment to precedent as a preeminent mode of constitution-

al discourse—are enduring.61 Moreover, Justices generally rec-
ognize (and exhibit) the need to reconcile their decisions with
earlier precedents.62 It is a rare (and completely ineffective)
Justice who attempts to build the world anew across the board
in constitutional adjudication.
Second, Posner’s account leaves out several significant
findings made by social scientists (besides those I have already
noted). For instance, he does not discuss the importance of Jus-
tices’ crafting opinions to appeal to different audiences,63 even
though such efforts arguably reveal the possibility that con-
cerns about appealing to certain audiences are a plausible ex-
ternal constraint on at least some judges. Nor does Posner dis-
cuss the significant findings that Justices rarely deviate from
the federal government’s suggested position or the position that
is supported by most of the amicus briefs in a given case.64 If
one were interested in illuminating the internal constraints on
judging, these findings cannot be ignored.
Moreover, Posner barely discusses another prominent
model of judging—the institutionalist model.65 Posner does dis-
cuss the importance of context for understanding judging, but
he tends to discuss each of the features of the institutional con-
text in which judges operate separately rather than as a collec-

overturned); Michael J. Gerhardt, The Irrepressibility of Precedent, 86 N.C. L.

REV. 1279, 1282 (2008).
61. See GERHARDT, supra note 1, at 177–98 (discussing the idea of “super
precedents” and their role in constitutional analysis and discourse).
62. See id.
63. See BAUM, supra note 1, at 43–49 (discussing the possibility that
“judges’ interest in the esteem of their audiences affects their choices as deci-
sion makers”).
64. See, e.g., Andrea McAtee & Kevin T. McGuire, Lawyers, Justices, and
Issue Salience: When and How Do Legal Arguments Affect the U.S. Supreme
Court?, 41 LAW & SOC’Y REV. 259, 268–69 (2007) (noting that research has
demonstrated “over and over again that the U.S. government is far more suc-
cessful than any other party or amicus curiae” and that Justices are “reluctant
to make decisions that will disaffect significant numbers of organized inter-
ests”).
65. See, e.g., Howard Gillman, What’s Law Got to Do With It?: Judicial
Behavioralists Test the “Legal Model” of Judicial Decision Making, 26 LAW &
SOC. INQUIRY 465, 492–94 (2001) (“[S]cholars . . . have found persuasive evi-
dence in support of the view that the institutional setting within which judges
operate shape their behavior . . . .”); Keith E. Whittington, Once More Unto the
Breach: PostBehavioralist Approaches to Judicial Politics, 25 LAW & SOC. IN-
QUIRY 601, 608–32 (2000) (discussing the role of institutionalism in thinking
about the law).
2009] HOW A JUDGE THINKS 2199

tion, aggregation, or system of forces that influence judging.66

Institutionalists posit, inter alia, that the context in which
judges and Justices operate matters: that the thinking and per-
formance of judges and Justices is shaped in part by the com-
plex institutional structures in which they function.67 By sepa-
rately assessing a few of the features of the institutional setting
in which judges operate, Posner fails to provide a comprehen-
sive account of judging. None of the features of this setting op-
erate apart from the others; together, they exert considerable
force on judges and Justices to channel their decision making
along certain lines. They are constrained to function within a
system in which there are strong pressures for them to forge
coalitions to decide cases, to respect the norm of stare decisis,
and to conform to the expectations of judges and the legal
community about how they should behave as judges. Judges
are influenced by what other judges expect from, and of, them.
These expectations reinforce various aspects of judging, such as
temperament, appreciation for collegiality, and the basic task of
judges to decide cases or legal questions on the basis of their
best interpretations of the pertinent legal materials, including
statutes and precedents. The institutional features of judging
might not dictate the answers to particular questions of law,
but these features are likely to shape the means or the process
by which judges decide those questions. Interestingly, Posner
questions many of these well-settled notions and practices, and
thus finds himself weaving a theory of judging that ultimately
suggests more about Posner than it does about how other
judges actually function.
Third, another possible constraint on judging is the judicial
selection process. Posner accurately observes that the federal
system is not designed to ensure the selection of the best-
qualified people to serve as judges; instead, it is designed to
keep off the bench people who are unqualified, dishonest, or
ideologically outside the mainstream of constitutional law.68
The process is designed to weed out nominees who lack the
temperament, the integrity, the skills, or the outlook (and val-

66. See GERHARDT, supra note 1, at 79–82 (providing an overview of insti-

tutionalist research on judging).
67. Id. at 80.
68. See POSNER, supra note 4, at 155–56 (“The pool from which our judges
are chosen is not homogeneous, though neither is it fully representative; it is
limited as a practical matter to upper-echelon lawyers, almost all of whom are
well-socialized, well-behaved, conventionally minded members of the upper
middle class.”).
2200 MINNESOTA LAW REVIEW [93:2185

ues) that we expect federal judges to have and therefore appear

to (or actually do) pose threats to our established notions of
how judges should act or to too much established law. We can
therefore expect the people who become federal judges to come
to the bench with notions or attitudes of judging and the law
that fit within the mainstream of constitutional jurisprudence.
Among the most important of these notions is a healthy respect
for precedent and stability and continuity in the law. But, the
mainstream in constitutional law is not fixed, so that politi-
cians make the critical decisions about which attitudes regard-
ing precedent (or judging more generally) are outside the main-
stream.
There is a lot more to be said about the relevance of federal
judicial selection to judging. While judicial selection oftentimes
seems to place a premium on ideology, it does not preclude ap-
pointments based on merit. It is not, in other words, impossible
to emphasize both ideology and strong professional credentials
in the selection process (evident, for instance, in the recent ap-
pointments of Chief Justice John Roberts and Associate Justice
Samuel Alito, Jr.). Indeed, Lee Epstein and Jeffrey Segal have
found that impressive professional accomplishments streng-
then rather than weaken judicial nominations.69
Nevertheless, it is a mistake to impose upon a system a
purpose that it has not adopted for itself. If a system is not de-
signed to ensure the selection of judges based on merit, it is of
very limited utility to assess that system on how well it selects
judges based on merit. If it does, then we might have to ac-
knowledge that the system designed by the framers for judicial
selection is not necessarily incompatible with merit-based ap-
pointments. But, if, as Posner suggests, the federal system is
not designed to select people based on merit, then it should not
be surprising—and it is not very illuminating—to find that
judges are rarely selected or rejected on the basis of merit. If we
want more merit-based appointments, we need to construct
consensus on judicial merit and then either to create incentives
for political leaders to make judicial appointments on the basis
of merit or to redesign the current system to maximize merit-
based appointments. Until we rethink or redesign the federal
scheme for judicial selection, we should evaluate it on whether
it achieves the objectives of the federal officials whom the Con-
stitution charges with its maintenance.

69. LEE EPSTEIN & JEFFREY A. SEGAL, ADVICE AND CONSENT: THE POLI-
TICS OF JUDICIAL APPOINTMENT 66–70, 102–06 (2005).
2009] HOW A JUDGE THINKS 2201

Posner is, however, respectful of the efforts of nearly twen-

ty states to employ mechanisms for selecting judges on the ba-
sis of merit (a notion Posner curiously avoids defining).70 Un-
like most other academics, he recognizes that while there may
be some politics involved in merit-selection systems, it is a dif-
ferent kind of politics than that which dominates either elec-
tions of judges or the federal judicial selection system. Never-
theless, it would be interesting to know how judges chosen
through merit selection compare with each other (and with
those not chosen through such methods), how lawyers or the
bar rate them, and the extent to which ideology plays a role in
their decision-making.

IV. LEGAL EDUCATION AND SCHOLARSHIP AS

CONSTRAINTS ON JUDGING
A final possible constraint on judging is legal education
and scholarship. Posner recommends in-depth training of law
students in social-science methodologies to better equip them to
assist judges in appreciating the likely benefits and costs of a
given decision.71 If, as Posner suggests, judges are in the busi-
ness of weighing the different possible consequences of particu-
lar decisions, they need to develop more sophisticated tech-
niques for making such calculations.
But Posner’s prescription is problematic for several rea-
sons. First, it derives from his mistaken belief that any devia-
tion from what he regards as the correct methodology or out-
come has to be unprincipled judging—effectively legislating
from the bench. Posner faults law professors for the rigidity
with which they hold their positions, but he is prone to the
same rigidity. Presumably, pragmatists are influenced or take
into account experience, but his own theory of judging has not
been influenced, apparently at all, by his own experiences as a
judge. He says the same things today about judging that he has
been saying for years, even well before he became a judge him-
self. More importantly, Posner seems not to recognize that the
Constitution empowers presidents and senators to decide what
kinds of judges to appoint—and thus what kind of judging is
permissible in our court system. The courts are what national
political leaders have made them. Presidents and senators are
the principal checks on the judiciary, and judges’ varying atti-

70. See POSNER, supra note 4, at 139.

71. See id. at 204 –29.
2202 MINNESOTA LAW REVIEW [93:2185

tudes about the law are a reflection of the choices of the presi-
dents and senators who are responsible for their appointments.
While neither judges nor Justices vote perfectly in accordance
with the preferences of the officials who appointed them, they
are rarely appointed—at least for the past few decades—
without some ideas of their likely judicial philosophies and con-
stitutional commitments. The Constitution does not dictate a
particular interpretive approach, but instead authorizes na-
tional political leaders to make the critical choices of the range
of interpretive approaches that they will allow—or disallow—in
the federal system. Thus, the federal judicial-selection process
may be used not only to fortify certain constitutional attitudes
or commitments within the judiciary but also to filter out the
nominees whose records suggest their constitutional opinions
are outside the mainstream of contemporary constitutional
thought.72
Second, not all judges would agree with Posner that they
are merely pragmatists. The popularity of reasoning by analogy
(or the case method) remains persistent and widespread in ad-
judication, and advancing arguments grounded in the social
sciences, as Posner is wont to do, will not appeal to any judges
except for Posner and the few other jurists who agree with him
on both the utility of social science research in the law and the
absolute futility of reasoning by analogy. Moreover, there are
many different modes of constitutional argumentation, and
Justices may choose which ones they find most compelling or
persuasive. Interestingly, Posner does not measure the kinds of
arguments that are popular on the Court, but in fact the most
popular mode of argumentation on the Court is precedent.73 It
is practically impossible to find an opinion that does not cite or
rely on precedent, whereas other sources, such as original
meaning, are rarely mentioned.74
Third, Posner discounts one of the most important things
law schools could do to train students to better appreciate judg-
ing: actually asking them to draft judicial opinions. Law schools
are rife with classes that call upon students to criticize judges,
but few if any classes are devoted to instructing students in the

72. See MICHAEL J. GERHARDT, THE FEDERAL APPOINTMENTS PROCESS

74 –77, 128–34, 162–79 (rev. ed. 2003) (noting that presidents and senators
tend to base their appointment decisions on philosophical and ideological
bases).
73. Gerhardt, supra note 60, at 1283–84.
74. See id. at 1283–86 (noting that the Roberts Court claimed precedent
as a basis for every one of its constitutional decisions in its first two Terms).
2009] HOW A JUDGE THINKS 2203

art of writing judicial opinions or working together on a collegi-

al court. Granted, such training might be premature, but it
might be a good start for lawyers to begin to understand the
thought processes of judges and the challenges of deciding cas-
es, forging majorities in contentious cases, and crafting judicial
opinions.
Fourth, Posner, like many other academics, ignores lawyer-
ing outside courts. Law schools are long overdue in recognizing
the ramifications of the fact that courts are not the only institu-
tions that routinely confront questions of law. Indeed, teaching
students about how institutions other than courts handle legal
arguments would help them to understand the ways in which
judging differs from legislating or enforcing the law as execu-
tive officials are charged to do. Moreover, a central concern of
the Carnegie Report75 is to urge law schools to do more to train
law students to be effective problem solvers in all of the areas
in which lawyers are called upon to solve problems inside and
outside of courts.76
Fifth, legal scholars could do a lot more to serve as a mea-
ningful check on the judiciary. Posner is probably right that
law professors produce too much scholarship that is either tan-
tamount to a partisan brief or that is designed principally to
get attention in the academy than actually to enrich constitu-
tional argumentation or understanding.77 It is more than ironic
that, even though Posner derides the trends in legal scholar-
ship, those trends are clearly moving in the direction that
Posner urges. Interdisciplinary scholarship is valuable because
it can enrich our understanding of the law’s relationship to oth-
er fields, but its focus is not on elucidating the law for judges or
holding them accountable for their mistakes. It is deliberation
disconnected from much of what judges do, which is to decide
cases on the basis of reasoning by analogy.

75. WILLIAM M. SULLIVAN ET AL., CARNEGIE FOUNDATION FOR THE AD-

VANCEMENT OF TEACHING, EDUCATING LAWYERS: PREPARATION FOR THE PRO-
FESSION OF LAW (2007)
76. Id. at 95–125 (discussing different teaching and learning methods
such as the iterative mode, T-funnel technique, and composition theory in le-
gal writing, all of which could be utilized in law school to develop student
problem solving skills that “legal professionals must master in order to func-
tion well in a variety of legal roles”).
77. See POSNER, supra note 4, at 211–29 (discussing how the legal acade-
my is separating itself from the judiciary and noting how the Court rebuked
the academy for overreaching in its legal analysis in a lawsuit challenging the
Solomon Amendment).
2204 MINNESOTA LAW REVIEW [93:2185

To be sure, some empirical scholarship, as Posner notes,

does attempt to hold judges accountable by rating their perfor-
mances.78 Focusing on citation practices, as Stephen Choi and
Mitu Gulati do, for instance, provides one basis on which to de-
termine what judges think of each other’s work.79 But, citation
practices are not a perfect measure of quality by any means;
they miss many of the things that judges do that fly below the
radar, such as asking questions at oral argument; and they pu-
nish judges who are good, maybe even very good, but are over-
shadowed or obscured by the few judges who are perceived, for
whatever reason, as the very best in particular fields.

CONCLUSION
I have long revered Richard Posner as a scholar and judge.
I am confident that it would be a pleasure—and an intellectual
challenge—to argue a case before him. I am also confident, par-
ticularly after reading his book on judging, of the arguments
that I would make to him, because he has told his readers the
kinds of arguments that would persuade him. The problem is
that these arguments are not likely to have much influence on
most of Posner’s colleagues, and there is the rub: Posner has
told us how he thinks as a judge, but Posner’s matrix of judging
does not provide much help or guidance on how judges other
than Posner think. Posner’s account of judging seems, more
than anything else, to be a defense of his own conception of
judging and an attempt to distinguish him from the Justices
whom political scientists have shown are grounding their deci-
sions primarily in their political biases rather than the law.
But, to know how judges other than Posner actually decide cas-
es, we need to know about them. It is not enough to suggest
they really are something other than what they purport to be,
particularly when so much of what they do and are has not
been fully disclosed. Nothing in Posner’s book negates that the
first and perhaps still best place to look to know how judges
generally think is the books and the opinions that they have
written as judges.

78. See POSNER, supra note 4, at 146–56 (noting, inter alia, the work of
Stephen Choi and Mitu Gulati).
79. See Stephen J. Choi & Mitu Gulati, A Tournament of Judges?, 92 CAL.
L. REV. 299, 305–09 (2004).