have a place in our normative world as objects in the mind of God or in nature
from which particular practices are a falling away. This means that the philoso-
pher of law will have to possess intellectual virtues of which political judgment is
one. Thus there will be at the theoretical level something like the perception and
discrimination that we expected from the judge and the statesman. Sheer intel-
ligence and analytical virtuosity will not identify the most adequate philosophy
One of the most central traditional contributions of philosophy to our under-
standing of the social world has come in the reformulations of the notion of truth
implicit in the actual practices of actors in different social spheres. Thus Kantā™s cri-
tique sought to identify the notion of truth that was implicit in scientiļ¬c practices,
as did John Deweyā™s philosophy in quite a different way. Implicit in the preceding is
the imperative for a philosophy of law to identify forms of truth that are adequate
to legal practices. Again, the key task will not be to offer an understanding of
doctrine or rules but of valid decisions in particular cases. In resolving particular
cases, the decision maker will always be asked to dwell in the logical gaps that
always separate circumstantial evidence from fully interpreted narratives (theo-
ries of the case) from legal rules from norms embedded in narrative and drama
from just outcomes. The law in action is made in those gaps. An important task
for an adequate philosophy of law is thus to explain what may and does occur
there. The philosophical task is to show how what occurs there can be practically
intelligent. The realists made important starts in seeking to understand the event
of legal decision as something more than the subsumption of value-free factual
narratives under legal rules. The notion of a āsituation senseā was an attempt to
identify concretely the kind of grasp that a decision maker might have in those
grasps. Their notion that cases were decided in the judgeā™s or juryā™s encounter with
meaningful factual patterns was basically correct. At least the sociological wing of
realism was thrown off track by the scientism of their age in its pursue of a truly
scientiļ¬c account of judicial decision making.
We now have richer philosophies of practical intelligence than the realists knew.
We can illuminate decisions in particular cases by exploring and applying in the
legal context what we can learn from a number of different philosophical traditions.
An adequate philosophy of law will rely on philosophically sophisticated concepts
such as Kantian reļ¬‚ective judgment as explained by Arendt, Aristotelian practi-
cal wisdom interpreted by contemporary Aristotelians such as Martha Nussbaum
(Nussbaum 1986), and interpretive understanding as interpreted by philosophers
in the hermeneutical tradition such as Gadamer and Taylor. Those philosophies
allow us to approach the conditions of the possibility of valid results that I men-
tioned at the beginning as a major task of the philosophy of law at this time. There
are ļ¬ve such conditions that can support a concept of legal validity, the elabora-
tion of which in the legal context should be an important task of a philosophy of
law. The ļ¬rst condition is that the narratives we employ at the trial and appellate
levels really do cut at the joints of the human events they seek to represent. This
The Tasks of a Philosophy of Law 237
question has empirical, interpretive, and normative aspects. The second condition
is that our practical intelligence, as mediated through our linguistic practices, can
converge on the place between the individual features of the particular case and
our inevitably overgeneralized public norms where the practical truth of the situ-
ation lies. This raises an old question, mentioned by Plato in the Phaedrus, where
Socrates ļ¬rst asserts, and then retracts, the notion that all we can know about
human events are the probabilities that they exhibit, and not the concrete truth of
the situation. This is a central issue in the philosophy of legal rhetoric. The third
condition that awaits philosophical explication in the context of our adversarial
legalism is a notion described by Charles Taylor as a central tenet of modernist
thought. This notion is especially important as philosophy shifts its focus from
the nature and status of legal rules to the event of judgment in particular cases.
The idea is that we can achieve an insight (epiphany or illumination) by dwelling
in the tensions that are created āby juxtaposing words.ā āThe epiphany comes
from between in the words or images, as it were, from the force ļ¬eld they set up
between them, and not through a central reference which they describeā (Taylor
1989: 465ā“6). The modernist notion is that we can know objective practical truths
through subjective resonance. The fourth condition for the validity of legal judg-
ments, which philosophers of law will continue to develop in the context of legal
practices, is that fundamentally interpretive methods can converge of the truth of a
human situation. This condition is closely related to the second condition. It asserts
that we have forms of intelligence that are not instrumental, that reveal rather than
control. The ļ¬nal condition is that the forms of legal practice we actually employ
can be shown to achieve the human purposes of a legal order. This will require
the dialectical tacking described above and will try to establish that the legal order
actually does allow us creatively to renew our forms of life by reļ¬‚ecting in our
linguistic practices their actual tensions with one another. Because an adequate
philosophy of law will be both interpretive and critical, it will not solely seek to
reconcile us with our practices. The owl of Minerva need not take ļ¬‚ight quite
at dusk. But before philosophy offers broad proposals for reform, it should have
actually seen and thought what we do with the keenness of vision that an owl is
thought to possess.
Arendt, Hannah. āTruth and Politics.ā In Between Past and Future: Eight Exercises in Political
Thought. New York: Viking Press, 1968: 227ā“64.
Bernstein, Richard. The Restructuring of Social and Political Theory. Philadelphia: Univ. of
Pennsylvania Press, 1976.
Burns, Robert P. A Theory of the Trial. Princeton, NJ: Princeton Univ. Press, 1999.
Dreyfus, Herbert. āHolism and Hermeneutics.ā Review of Metaphysics 34.1 (1980): 3ā“23.
Geertz, Clifford. Available Light: Anthropological Reļ¬‚ections on Philosophical Topics. Prince-
ton, NJ: Princeton Univ. Press, 2000.
. āFrom the Nativeā™s Point of View: On the Nature of Anthropological Understand-
ing.ā In Interpretive Social Science: A Reader. Eds. Paul Rabinow and William M. Sullivan.
Berkeley: Univ. of California Press, 1979.
Hampshire, Stuart. Justice Is Conļ¬‚ict. Princeton, NJ: Princeton Univ. Press, 2000.
238 Robert P. Burns
Kolb, David. The Critique of Pure Modernity: Hegel, Heidegger and After. Chicago: Univ. of
Chicago Press, 1986.
Llewellyn, K. N. (Untitled chapter). In My Philosophy of Law: Credos of Sixteen American
Scholars. Boston: Boston Law Book, 1941, 181ā“97.
Nussbaum, Martha. The Fragility of Goodness: Luck and Ethics in Greek Tragedy and Philos-
ophy. Cambridge: Cambridge Univ. Press, 1986.
Pitkin, Hanna Fenichel. Wittgenstein and Justice. Berkeley: Univ. of California Press, 1972.
Rawls, John. A Theory of Justice. Cambridge, MA: Harvard Univ. Press, 1971.
Taylor, Charles. Hegel. Cambridge: Cambridge Univ. Press, 1975.
. āInterpretation and the Sciences of Man.ā Review of Metaphysics 25.1 (1971): 3ā“51.
. Sources of the Self: The Making of the Modern Identity. Cambridge, MA: Harvard
Univ. Press, 1989.
part six. questioning the relationship between
philosophy and american law
28 Law and Philosophy at Odds
larry alexander and emily sherwin
In addressing the topic of this symposium ā“ philosophy in American law ā“ we give
the term philosophy a meaning different from that given by Llewellyn in his article
āOn Philosophy in American Law.ā Llewellyn described philosophy as something
akin to an ideology or decision-making ethos, which will vary over time according
to the interests of decision makers and the needs of society. In contrast, we have in
mind a discipline that employs reason to bring clarity, depth, and precision to the
process of thinking about a given subject.
Given our understanding of the term philosophy, the role of philosophy in law
may seem obvious. Legal decision making is supposed to be a reasoned enterprise.
Judges, in particular, are supposed to decide cases for reasons and to give reasons
for their decisions. Surely philosophical methods, employed by judges and the
lawyers who argue before them, will improve the quality of the process.
We shall argue that, in fact, legal decision makers may sometimes do better
without the insights of philosophy. Law is a reasoned enterprise, but it is also a
rule-governed enterprise. For reasons we shall develop, thorough and meticulous
reasoning can undermine the beneļ¬ts of legal rules. Philosophy is ļ¬ne for observers
of law but not necessarily good for its participants.
In sections that follow, we shall elaborate this point in the context of two practices
that are integral to law: general compliance with legal rules and judicial decision
making by analogy.
Let us assume that the object of a legal system is to settle controversy over the
practical meaning of generally accepted values (Alexander and Sherwin 2001: 11ā“
15; Eisenberg 1988: 4ā“7; Raz 1994: 187ā“92). The values in question may be moral
values or some combination of moral and nonmoral values, as you like. Law settles
disagreement and uncertainty about what actions the prevailing values require
by translating those values into more determinate rules1 (Alexander and Sherwin
Members of the society in question may sometimes prefer freedom of action or case-by-case decision
making by ofļ¬cial adjudicators to settlement by rules. In other contexts, however, they will give ļ¬rst
priority to settlement of uncertainty and controversy, and accordingly will authorize ofļ¬cials to
make authoritative rules (Hart 1961: 121ā“3).
| 241 |
242 Larry Alexander and Emily Sherwin
2001: 53ā“95; Raz 1979: 16ā“19, 22ā“3, 30ā“3; 1986: 57ā“62; Schauer 1991: 42ā“52,
Because legal rules translate indeterminate values into determinate prescriptions
for classes of situations, they introduce a certain amount of error to decision mak-
ing. Most rules will sometimes dictate results that are wrong, judged by the values
the rules were designed to implement (Schauer 1991: 31ā“34, 48ā“54).2 Accord-
ingly, if actors follow rules without exception, some of their acts will be wrong,
and if judges enforce the rules without exception, some of their decisions will be
Yet rules can also prevent error, in several ways. First, rules facilitate coordination.
One personā™s reasons for action often depend on the actions of others. Determinate,
generally applicable legal rules make it easier for actors to predict what others will do
and so reduce the actorā™s uncertainty about his or her own reasons for action (Hurd
1999: 214ā“21; Postema 1982: 172ā“86; Raz 1986: 49ā“50; Schauer 1991: 163ā“6).
Second, rules prevent faulty reasoning. Individual reasoners deciding what course
of action is best, all things considered, may suffer from bias or lack of information. A
well-designed rule dictates outcomes that are correct in most cases (Schauer 1991:
150ā“2, 158ā“9). In effect, it eliminates reasoning errors by eliminating reasoning.
Thus, the test of a good legal rule is not whether it always dictates correct
results but whether it prevents more decision-making errors than it causes. A
sound legal rule is one that, if followed without exception, will result in a lesser
sum of errors than will result from unconstrained reasoning by individual decision
makers (Alexander and Sherwin 2001: 58ā“9; Schauer 1991: 149ā“55; Goldman 2006:
453ā“56). This is true at the level of private conduct and also at the level of judicial
decision making. If judges as a group will make fewer errors overall by enforcing
a legal rule than by deciding each case as they believe they should (taking into
account the effects of nonenforcement on private conduct), then it is better that
all judges follow the rule all the time than that they decide each case by consulting
the balance of reasons.
To take an example from property law: suppose that a law-making authority seeks
to maximize human welfare, and that welfare maximization is a morally important
objective. The authority concludes that under a given set of geological conditions,
private property rights in water will advance welfare. Accordingly, the authority
enacts a prior appropriation rule for nonnavigable surface water: individuals who
divert water for uses of their choosing have priority over subsequent appropriators
in a quantity of water measured by the initial diversion.3
Assume that overall, the authorityā™s rule is morally sound. Water is scarce and
essential to life and commerce. In the absence of well-deļ¬ned property rights,
potential users of water ā“ even those who accept the principle of welfare max-
imization and do their best to implement it ā“ are likely to err in judging the
beneļ¬ts of different water uses and estimating what other water users are likely
The exception is a pure coordination rule, in which no one has a reason to act one way or the
other apart from the coordination beneļ¬ts of the rule. āDrive on the rightā is a possible example
(Alexander and Sherwin 2001: 56; Goldman 2006: 453).
See Cofļ¬n v. Left Hand Ditch Co., 6 Colo. 443 (1882); Restatement (Second) of Torts, chap. 41, scope
Law and Philosophy at Odds 243
to do.4 At the same time, the prior appropriation rule will cause a loss of welfare
in some situations, as when the ļ¬rst appropriator of water uses it less efļ¬ciently
than subsequent appropriators might have done. The rule is justiļ¬ed if, despite its
imperfections, it will produce better results overall than unregulated case-by-case
The difļ¬culty is that it is not rational for individual decision makers (actors
or judges) to follow a rule when they believe the outcome it calls for is wrong
(Hurd 1999: 62ā“94; Kavka 1983). Sensitivity to the potential beneļ¬ts of rules will
not solve this problem, because even a reasoner who understands the reasons for
rules may miscalculate the effects of disregarding a rule (Alexander and Sherwin
2001: 61ā“8; Schauer 1991: 94ā“100). Sanctions for disobeying rules will not solve
the problem because sanctions are imposed by human decision makers, who may
(rightly or wrongly) conclude that no sanction is deserved.5 There is, in other
words, a persistent gap between the best course of action for all actors (obey the
rule) and the rational course of action for any individual actor (disobey the rule if
disobedience appears to the be right thing to do; Alexander 1991; Alexander and
Sherwin 2001: 53ā“95).
Nor can reason itself ā“ philosophy ā“ solve the problem of rule following. Actors
may understand that they (and society at large) will fare better if all people, includ-
ing themselves, follow justiļ¬ed rules in all cases. Understanding this, however,
does not make it any more rational to follow a rule in a given instance if the actor
believes, all things considered, that the action it requires is wrong. Committing
or consenting in advance to follow the rule in all cases does not make such rule
following rational because commitment or consent to act against the balance of
reasons is not, without more, a reason for so acting (Gauthier 1996; Murphy 1997;
Shapiro 1998). Comparing the beneļ¬ts of universal rule following to the beneļ¬ts
of universal disregard of rules does not make rule following rational because only
a single act of disobedience is immediately at stake (McClennen 2004; McClennen
1997; McMahon 2001: 6ā“30) Nor is the actor rationally obliged to comply in order
to fulļ¬ll his or her fair share of a collective obligation to realize the beneļ¬ts of
the rule (Goldman 2006: 462ā“70). As long as the beneļ¬ts of disregarding the rule
appear to outweigh the rule beneļ¬ts attributable to the actorā™s contribution to
upholding the rule, the rational choice, all things considered, is to disregard the
rule. Thus, none of these proposed solutions can overcome the basic difļ¬culty, that
a rational actor cannot take a statistical view of his or her own decision making in
a particular case.
Experience suggests that, in fact, we do follow rules without reļ¬‚ection. Assuming
this is true, we must follow the rules without the assistance of philosophy, which
tells us that unreļ¬‚ective rule-following is not rational. If we could clear away
Determinate rules are necessary even when rule subjects endorse the same abstract moral principles
and are motivated to act on them. Even under the best conditions of moral harmony, reasoning
errors and lack of coordination will lead to morally incorrect actions (Alexander and Sherwin 2001:
232n4; Kavka 1995).
This may be because the actor was right to disobey the rule, or because the actor believed he or
she should disobey, and hence was not culpable (Alexander and Sherwin 2001: 77ā“86; Hurd 1999:
244 Larry Alexander and Emily Sherwin
the effects of habit, training, and convenience, and subject our behavior to the
discipline of philosophy, we would no longer be able to treat rules as if they were
authoritative in themselves.
REASONING BY ANALOGY
It is widely assumed that common law judges reason by analogy (Brewer 1996;
Burton 1995: 25ā“41; Horty 2004; Lamond 2005; Levi 1948: 1ā“6; Llewellyn 1960a:
77ā“87; 1960b: 66ā“9; Raz 1979: 185ā“7; Sunstein 1996: 62ā“100; Weinreb 2005). Faced
with a new case that does not fall within the terms of existing legal rules, the judge
searches for past cases with similar facts, then reaches a parallel decision in the new
case. Faced with a case that appears to be governed by an existing legal rule, but in
which the rule produces a questionable result, the judge searches the facts of prior
cases that applied the rule. If the facts of those cases are dissimilar to the facts of
the new case, the judge may distinguish the rule and reach a contrary result. In
both cases, judgesā™ reasoning is thought to be constrained by precedent cases, even
when their decisions do not conform to posited precedent rules.
Philosophical analysis suggests that analogical reasoning is a spurious idea
(Alexander 1996: 80ā“6; Alexander and Sherwin 2008: 66ā“88; Posner 1990: 86ā“98;
Schauer 1991: 183ā“87). Judges cannot reason directly from one case to another,
because particular disputes are like and unlike in innumerable ways. To determine
whether a new case is relevantly similar to a prior case, the judge must refer, at least
implicitly, to a general proposition that makes certain features of the cases impor-
tant to their legal outcomes. For example, a nuisance claim against a neighboring
property owner who keeps a bear is not simply like a nuisance claim against an
owner who keeps a snake. Their likeness, if any, depends on a proposition about
the acceptability of placing certain types of risks on other owners in a residen-
tial setting. That proposition, rather than the brute facts of the cases, determines
whether the two claims should be decided alike.6
Intuitive ļ¬ndings of similarity, without the aid of general propositions, may
be psychologically possible (Haidt 2001; Sloman 2002: 380ā“4). They are not,
however, a form of legal reasoning in which precedent cases determine the
outcome of current cases (Alexander and Sherwin 2008: 72ā“73; Haidt 2001: 818).
An intuitive determination of similarity cannot be explained in terms accessible to
others. Moreover, there are no criteria by which it can be judged right or wrong.
If one conclusion about similarity is as good as another, past decisions impose no
A narrower form of analogical decision making might seem more promising as
a method of legal reasoning: if the facts of a new case provide at least as much
support for the outcome of a precedent case as the facts of the precedent case
itself, then, a fortiori, the current judge should reach a parallel outcome in the
new case (Horty 2004; Lamond 2005 also defends what appears to be a form of a
fortiori decision making). A fortiori analogies occur less frequently than intuitive
similarities among cases. Unlike intuitions of similarity, however, they appear to
be reasoned, and the outcomes they yield appear to be constrained.
Lloyd Weinreb (2005: 12ā“13, 44ā“5, 77ā“103, 107ā“16) offers a contrary view.
Law and Philosophy at Odds 245
Yet even this limited form of analogy collapses on close examination. To compare
the relative strength of two cases, the judge must determine what weight particular
facts bear in favor of an outcome. In doing so, the judge must rely either on a
general proposition that assigns both relevance and gravity to the features of each
case or on some principle such as utility that provides a metric for numerical
comparisons. Once again, the general proposition or the principle of utility, not
the facts themselves, determines the likeness of the cases.
Another approach to analogical reasoning is to formulate the best legal principles
capable of explaining the bulk of precedents, then to decide new cases according to
those principles. This method of decision making is notably defended by Ronald
Dworkin (1986: 228ā“32, 240ā“50, 254ā“8; Dworkin 1978: 22ā“31). Legal principles
drawn from precedents become the general propositions that link cases and pick
out relevant similarities and differences among their myriad facts. Decision making
based on legal principles is a form of reasoning. Moreover, it appears to be a form
of reasoning constrained by precedent: because legal principles are shaped by the
facts and outcomes of prior cases, the analogy arises from the precedent cases
The difļ¬culty is that legal principles are too malleable to impose effective con-
straint, for several reasons. First, the content of legal principles is largely within the
control of the current judge. Legal principles are not rules or rationales announced
by precedent courts. Instead, the current judge postulates the most morally attrac-
tive principle consistent with an unspeciļ¬ed proportion of prior outcomes. Both
criteria ā“ moral attractiveness and ļ¬t with prior decisions ā“ are matters of individ-
ual judgment (Dworkin 1986: 254ā“8).
Second, legal principles require evaluative judgment at the point of application.
Legal principles do not dictate results in the manner of determinate legal rules.
Several principles may contend for priority in a given case, in which case the
outcome depends on the weight each principle exerts in the context of the case
(Dworkin 1978: 26ā“7). The task of weighing principles, like the task of construct-
ing principles, falls to the current judge. This creates a further opportunity for
discretionary decision making.
Third, and perhaps most important, Dworkinian principles are morally
unattractive, if not incoherent (Alexander and Sherwin 2008: chap. 4, n53ā“88).
They are not moral principles, the application of which must take into account
reliance on past decisions. Rather, they are the most attractive principles available
that would account for (some percentage of ) past decisions, some of which are
sure to have been morally infelicitous. They are the answer to the question, What
would be the morally correct principles in a world where certain morally incorrect
decisions were actually morally correct? If that question is even coherent, it is
nonetheless unclear why anyone should seek guidance from norms that are neither
determinate, like rules, nor morally optimal.
If the foregoing analysis is correct, analogical methods impose no meaningful
constraint on judicial decision making. Nevertheless, the widespread practice of
searching for analogies in past cases as if they were source of law for current judges
may have signiļ¬cant indirect beneļ¬ts for the functioning of the legal system. In
particular, a perceived obligation to study the facts and outcomes of prior cases
can improve the quality of judicial rule making.
246 Larry Alexander and Emily Sherwin
In a legal system that treats rules posited in judicial opinions as binding on
later judges, judges perform a dual rule, as rule makers and adjudicators. This
is unavoidable, but the vantage point and responsibility that judges assume as
adjudicators of particular disputes may impair their judgment as rule makers
for future cases (Alexander and Sherwin 2001: 132ā“3; Alexander and Sherwin
2008: 109ā“17; Rachlinski 2006; Schauer 2006; Sherwin 1999: 1192; 2000: 315;
2006). When a judge formulates a general rule in the context of a speciļ¬c dispute,
the judgeā™s proximity to the facts of that dispute can obscure the full consequences
of the rule. In the terminology of behavioral psychology, the facts of the live dispute
the judge must resolve are available ā“ they come more readily to mind than more
distant facts or statistical probabilities (Gilovich, Grifļ¬n, and Kahneman 1982;
Plous 1993: 121ā“30; Tversky and Kahneman 1982).7 As a result, the outcome of
the rule in the case at hand may appear more representative than it is of the overall
operation of the rule.8 This in turn may lead the judge to adopt a rule that produces
a satisfying result but is not justiļ¬ed in the long run or to modify a rule that is
justiļ¬ed in the long run but produces an unsatisfying result in the current case
(Schauer 2006: 906ā“11).9
Given the circumstances of judicial rule making, a judicial commitment to
analogical reasoning has the potential to improve the quality of judicial rules.
Close attention to the facts of prior cases will bring to light other applications of
a tentative rule, beyond the case at hand. Concrete examples from past cases may
expose ļ¬‚aws in a momentarily appealing rule or may show that an established rule
that appears to misļ¬re in the present case works well in most of its applications.
These positive effects have nothing to do with analogical decision making, but
they are likely consequences of the practice of seeking analogies in prior cases
(Alexander and Sherwin 2008: 118ā“20; Sherwin 2006: 927ā“9).
Analogical reasoning, therefore, presents another instance in which imperfect
reasoners ā“ in this case, judges ā“ may do better without the discipline of philosophy.
Philosophical rigor calls into question the tradition of analogical reasoning in law;
therefore, judges thinking philosophically may no longer feel obliged to consult
the facts and outcomes of prior cases. In this way, philosophy may undermine a set
of habits that is beneļ¬cial to the development of common law, even if analogical
reasoning itself is not.
We have suggested that philosophical analysis casts doubt on pervasive and possibly
salutary features of legal practice. Does this suggestion imply that philosophers
This is also true in the context of legislation (Kuran and Sunstein 1999; Sunstein 2003).
Affect bias can have a similar effect (Rachlinski 2006: 941ā“2; Slovic et al. 2002). Other heuristics
that may affect rule making in the context of adjudication include anchoring, contrast effects, and
attribution errors (Plous 1993: 38ā“41, 180ā“2; Rachlinski 2006: 938ā“40; Ross 1977: 174; Schauer
2006: 896ā“98; Tversky and Kahneman 1974: 1128ā“30).
We do not mean to suggest that legislative rule making is superior to judicial rule making; legislatures
have their problems as well (Rachlinski 2006: 951; Schauer 2006: 912ā“13). Noncognitive problems
also affect legislation (Farber and Frickey 1987; āSymposiumā 1988).
Law and Philosophy at Odds 247