INTERNATIONAL LAW IN BLACK AND WHITE

Daniel Bodansky*

I. INTRODUCTION

Is the study of international law an art or a science? Can the role of

international law be explained by general rules, with predictive value? Or does
it require the exercise of judgment, in order to account for the richness and
complexity of international life? Traditionally, international lawyers have
gravitated to the latter view, analyzing issues in an essentially ad hoc and
eclectic manner. In their controversial new book, The Limits of International
Law,1 Jack Goldsmith and Eric Posner argue forcefully for a more scientific
approach, relying on the methodology known as "rational choice theory." The
book makes many specific claims about the limits of international law. But its
ambition to develop an overarching theory of international law, which reduces
the role played by international law to a few simple explanatory models, is
perhaps its most distinctive feature.
In his work on the sociology of law, Max Weber identified three
complementary perspectives on law, which he called the dogmatic, the ethical
and the sociological:2

9 The dogmatic (or as we would now say, "doctrinal")

perspective takes law as a given-a dogma, as Weber put
it-and seeks to ascertain what it says. It focuses on doctrinal
questions: Did the invasion of Iraq violate the U.N. Charter?
Does the treatment by the United States of detainees at
Guantanamo violate customary and treaty obligations
prohibiting torture? Does international law prohibit
significant transboundary pollution, or require states to take
precautionary actions against potentially irreversible threats
to the environment?

* Robert and Emily Woodruff Chair of International Law, University of Georgia School of
Law. This Essay was presented at a symposium on The Limits of InternationalLaw, University
of Georgia Law School, October 28-29, 2005.
' JACK L. GOLDsMrrH & ERic A. POSNER, Ti-E Lamrrs OF INTERNATIONAL LAW (2005).
2 ANTHONY T. KRoNMAN, MAX WEBER 7-14 (1983).
286 GA. J. INT'L & COMP. L. [Vol. 34:285

" The ethicalperspective, in contrast, steps outside the law and

asks, what ought the law to be? Weber's characterization of
this as the "ethical" perspective suggests that the normative
standpoint from which we evaluate law is that of morality.
But, viewed more broadly, this perspective could encompass
other normative standards such as efficiency or order.
" Finally, the sociological perspectives interrogates law not
from a normative but from an explanatory standpoint. It asks
causal questions about why legal rules emerge and what
effects, if any, they have on behavior.

Although international lawyers often speculate or make implicit

assumptions about the causes and effects of international law, and although
they frequently advance normative arguments for particular legal rules, I think
it is fair to say that, traditionally, international law scholarship has been largely
doctrinal-or as Weber would say, dogmatic-in its orientation. Most of the
leading treatises on international law, if they address issues of explanation at
all, relegate them to an introductory chapter about whether international law
is really "law." 3 And the explanations typically given for how international
law arises and influences state behavior have been unsystematic and
conjectural. Writers posit a host of causal factors, with little, if any, attempt
to investigate the actual effects of international law or its relative importance
in influencing behavior as compared with other factors such as power or self-
interest.4 Most international lawyers instead focus on the content of
international law. Indeed, even when they have a normative agenda, they often
couch their prescriptive arguments about what the law should be as doctrinal
claims about the law actually is.5
In recent years, international law has begun to awaken from its dogmatic
slumber.6 The principal catalyst has been the introduction of social scientific
approaches that focus on issues of explanation. The Limits of International

3 See, e.g., IAN BROWNLIE, PRINCIPLES OF PUBLIC INTERNATIONAL LAW (6th ed. 2003);
MARK W. JANIS, AN INTRODUCTION TO INTERNATIONAL LAW (4th ed. 2003); MALCOLM N.
SHAW, INTERNATIONAL LAW (5th ed. 2003).
4 See, e.g., LOUIS HENKIN, HOW NATIONS BEHAVE (2d ed. 1979); Oscar Schachter, Towards
a Theory of InternationalObligations,8 VA. J. INT'L L. 300 (1968).
' For a notable exception, see TERRY NARDIN, LAW, MORAlrY AND THE RELATIONS OF
STATES (1983).
6 Cf IMMANUEL KANT, PROLEGOMENA TO ANY FUTURE METAPHYSICS (1950) (describing
how Hume awakened Kant from his dogmatic slumber).
2006] INTERNATIONAL LAW IN BLACK AND WHITE

Law is the latest contribution to this now burgeoning literature. In general, it

has little interest in doctrinal questions per se. And, except in the last two
chapters, its arguments are not ostensibly normative (although they have an
important normative implication, since if international law has little potential
to constrain state behavior, then we have little reason to try to develop it).
Instead, Limits focuses on explanatory issues: Why do rules of behavior for
states emerge (or not emerge)? Is international law merely epiphenomenal,
reflecting behavioral regularities that have other causes? Or does it sometimes
play an independent causal role, and if so, when and why?
The turn to explanatory issues opens up the possibility of a more scientific
approach to the study of international law. At least among Anglo-American
lawyers, few since the days of Langdell believe that doctrinal questions can be
answered scientifically. But explanatory issues are amenable to scientific
study. Hypotheses can be developed and then tested against the empirical
evidence. This is the methodology that Goldsmith and Posner propound in The
Limits of InternationalLaw.
The development and testing of explanatory hypotheses about the role of
international law is not, in itself, unique to Goldsmith and Posner's study.
Others working at the intersection of international relations and international
law use the same methodology.7 Nor is the causal model that they employ
original. Indeed, for those focusing on explanatory issues in international law,
rational choice theory-and, in particular game theory-has become the
paradigm de jour. Instead, what makes Limits unusual is its ambition to
develop a general theory of international law, covering the entire range of
sources and subjects-custom as well as treaties, human rights as well as trade
law. It puts forward rational choice theory not merely as one among several
explanatory models, but rather as the exclusive model, which can account
generally for both the development and influence of international law.
Whether one likes Goldsmith and Posner's approach or not, it has two
notable virtues: it poses important questions and it offers provocative answers.
Provocative ideas are often wrong, but they have the benefit of stimulating
debate, shaking people out of their dogmatic assumptions. And this has been
exactly the effect of Goldsmith and Posner's work. Already, their skeptical
claims about the role of customary international law have stimulated a range

7 See, e.g., Symposium on Legalization and World Politics, 54 INT'L ORG. 385 (2000);
COMMITMENT AND COMpLtANCE: THE ROLE OF NON-BINDING NORMS IN THE INTERNATIONAL
LEGAL SYSTEM (Dinah Shelton ed., 2000).
 GA. J. INT'L & COMP. L. [Vol. 34:285

of responses, some from a traditional international law perspective;' others

using the same rational choice methodology that they employ. 9 The result has
been to reinvigorate the study of custom.
This brief Essay will begin by discussing what, I think, is uncontroversial
in Limits, then try to isolate its distinctive and controversial claims, and end
with a brief appraisal of its approach on both substantive and methodological
grounds.

I. LIMITS AND CONVENTIONAL WISDOM

Limits seeks to establish its revisionist credentials by sharply contrasting

its approach with that of "mainstream" or "traditional" international lawyers,
who are portrayed as uncritical believers in the normative force of international
law. Yet it rarely identifies any actual international lawyer who holds the
views that it describes as "mainstream." In fact, I think many international
lawyers-while rejecting the book's more extreme claims-would not
fundamentally disagree with either its rationalist methodology or its skeptical
attitude.
At least in part, Limits reaches different answers from mainstream
scholarship because it asks different questions. In discussing custom, for
example, Goldsmith and Posner criticize traditional analyses for not being able
to explain why custom arises and changes, and why states do or do not
comply.' But most accounts of custom do not focus on explanation; instead,
they seek to determine which norms meet the tests of valid custom. At first
glance, this may not be apparent, since discussions of custom are usually
framed in terms of the "sources" of international law. But when international
lawyers speak of "sources," they usually refer not to the causes of a customary
rule, but rather to the conditions of its legal validity. The difference between
the traditional approach and that of Goldsmith and Posner reflects the
difference between the doctrinal and explanatory perspectives that I noted
earlier. International lawyers typically operate as actors within the realm of
international law, engaging in exegesis of legal texts; they view international
law from the inside, as participants in the legal process. International relations

' Detlev F. Vagts, International Relations Looks at Customary International Law: A

Traditionalist'sDefense, 15 EUR. J. INT'L L. 1031 (2004).
9 George Norman & Joel P. Trachtman, The Customary InternationalLaw Game, 99 AM.
J.INT'LL. 541 (2005); Andrew T. Guzman, A Compliance-BasedTheory ofInternationalLaw,
90 CAL L. REv. 1823 (2002); Edward T. Swaine, Rational Custom, 52 DUKE L.J. 599 (2002).
10GOLDSMrTH & POSNER, supra note 1, at 25.
20061 INTERNATIONAL LAW IN BLACK AND WHITE

theorists, in contrast, look at international law from the outside, as a

phenomenon to be explained. The two perspectives, at least in principle, are
complementary rather than conflicting.
To the extent that international lawyers do consider explanatory questions,
most start from many of the same premises as Goldsmith and Posner:

* first, that, broadly speaking, states can be said to have

interests, which they rationally pursue;
* second, that the rules of international law to a significant
degree reflect the interests of states;"
* third, that state interests play an important role not only in the
creation of international
2
law, but also in determining whether
states comply;'
* fourth, that states often assert "changing and inconsistent
readings of . . . international law, consistent with their
interests," 3
" finally, that power also plays a significant role both in the
development and enforcement of international law.'

Limits often suggests that international lawyers focus exclusively on

noninstrumental accounts of state behavior-that international lawyers are as
5
one-dimensionally normative as Limits is one-dimensionally instrumental.

" See, e.g., ANTONIO CASSESE, INTERNATIONAL LAW 6 (2002) ("[W]hen making law,
settling disputes, or enforcing the law, States do not act in the interest and on behalf of the
international community; they do not fulfil an obligation, but primarily pursue their own
interests.").
12 HENKIN, supra note 4, at 50 (analyzing compliance with international law in terms of the
costs and benefits of observance versus violation); OSCAR SCHACHTER, INTERNATIONALLAW IN
THEORY AND PRACTICE 5-9 (1995).
13 GOLDSMrrH & POSNER, supra note 1, at 63.
14 See, e.g., CHARLES DE VISSCHER, THEORY AND REALrrY IN PUBLIC INTERNATIONAL LAW

149 (1957) (comparing development of custom to gradual formation of a road across vacant
land, in which some users mark the soil more deeply with their footprints, "either because of
their weight, which is to say their power in this world, or because their interests bring them more
frequently this way"). Indeed, international law has often been described as the handmaiden of
power. DAVID HUNTER, JAMES SALZMAN & DURWOOD ZAELKE, INTERNATIONAL
ENVIRONMENTAL LAW AND PoLICY 273 (2d ed. 2001) (describing international law as "the
handmaiden of power, following rather than leading, facilitating rather than constraining").
"5For example, they assume that the old GATT dispute settlement system must have been
a "puzzle for traditional international lawyers' thinking," since it cannot be explained solely in
terms of states' preference for complying with international law. GOLDSMITH & POSNER, supra
 GA. J. INT'L & COMP. L. [Vol. 34:285

In reality, most international lawyers see the world in a multi-dimensional way.

They recognize the role not just of norms, but of interests and power in
shaping state behavior. Inadvertently, Limits itself recognizes this, by citing
traditional lawyers in support of its instrumental explanations. For example,
in discussing the putative norm protecting coastal fishing vessels from seizure
during time of war, Goldsmith and Posner quote extensively from the 1924
British treatise on international law by William Hall, which they say, "gets the
logic of the fishing vessels exemption rule exactly right" 6 -although they add,
somewhat gratuitously, that Hall did so "perhaps inadvertently," presumably
in order to square his apparent acuity with their portrayal of international
lawyers as narrowly normative.
International lawyers also display a much wider variety of views about the
effectiveness of international law in constraining state behavior than
Goldsmith and Posner's account suggests. Some are true believers, to be sure;
but many are skeptics, who, I think, would find much in Limits
unexceptionable. Posner and Goldsmith, for example, argue that compliance
with international law does not demonstrate that international law constrains
state behavior, since compliance may simply reflect a coincidence of interests.
In other words, it may simply reflect what states would have done otherwise.
This is, of course, true, but hardly novel. Indeed, in international
environmental law, it has become almost commonplace to distinguish between
the concepts of "compliance" and "effectiveness."
A skeptical attitude is, in my view, particularly appropriate with respect to
claims about customary international law. 7 As Goldsmith and Posner show,
even the venerable case of PaqueteHabana-oftenseen as an exemplar of the
customary law methodology-rests on limited evidence, much of it
inconsistent or of dubious value. 8 For me, the difficulties of ascertaining
custom first became apparent when I was a junior attorney at the Department
of State, and was asked to investigate whether customary international law
prohibits thejuvenile death penalty. In an effort to ascertain state practice, we
sent a cable to all U.S. embassies worldwide, asking them about local law and
practice on the death penalty. What is more, we enlisted the assistance of the
comparative law division of the Library of Congress for a comprehensive

note 1, at 152. But, of course, no international lawyers in his or her right mind thinks that states
are concerned solely about legal compliance.
16 GoLDsMITH & POSNER, supra note 1, at 76.
" See Daniel Bodansky, Customary (and Not So Customary) InternationalEnvironmental
Law, 3 IND. J. GLOBAL LEGAL STUD. 105 (1995).
18 175 U.S. 677 (1900).
2006] INTERNATIONAL LAW IN BLACK AND WHITE

survey of national legislation. But, as we, perhaps, should have expected,

much of what we got back was of little use-isolated snippets on national
legislation, with little information about actual practice; statements that no
juveniles had been executed in recent years, with no indication as to whether
there had been any cases in which the juvenile execution issue had been raised;
decisions not to impose the death penalty on juveniles, with no evidence of
opinio juris. In the end, even after an extraordinary effort to ascertain state
practice, my conclusion was that little could be said about whether there was
consistent and uniform state practice on the juvenile death penalty, backed up
by a sense of legal obligation. Similarly skeptical conclusions about custom
are not uncommon among international lawyers. More than twenty years ago,
no less an authority than Judge Robert Jennings, President of the International
Court of Justice, wrote, "most of what we perversely persist in calling
customary international law is not only not customary law; it does not even
faintly resemble customary law."' 9

III. APPRAISING THE DISTINCTIVE CLAIMS OF LIMITS

What makes Limits controversial, then, is neither its instrumentalist

approach nor its skeptical attitude per se, but rather the extremes to which it
pushes both of these features. According to Goldsmith and Posner, state
interests are not merely one among several explanatory factors; they are the
"sole determinants" of state behavior.20 And, as a result, international law is
not merely limited in its effectiveness; it has no "exogenous effect on state
behavior."'"
Interestingly, in making these arguments, Goldsmith and Posner come to
exactly the opposite conclusion from neoconservative critics of international
law such as Robert Bork and Charles Krauthammer, who contend that
international law is often contrary to U.S. self-interest and worry that U.S.
policymakers will give undue credence to it.22 In doing so, Bork and

,9 Robert Y. Jennings, The Identification of InternationalLaw, in INTERNATIONAL LAW:

TEACHING AND PRACTICE 3, 5 (Bin Cheng ed., 1982) (emphasis in original). For similar
sentiments by two leading scholars on international human rights law (one of whom now sits as
a judge on the International Court of Justice), see Bruno Simma & Philip Alston, The Sources
of Human Rights Law: Custom, Jus Cogens and GeneralPrinciples, 12 AUSTRL. Y.B. INT'LL.
82 (1992).
20 GOLDSMITH & POSNER, supra note 1, at 39.
21 Id. at 43.
22 Robert Bork, The Limits of InternationalLaw, NAT'L INT., Winter 1989/1990, at 1-10;
 GA. J. INT'L & COMP. L. [Vol. 34:285

Krauthammer buy into the "usual view"-which Goldsmith and Posner

reject-that "international law is a check on state interests, causing a state to
behave in a way contrary to its interests. 23 In Goldsmith and Posner's view,
this gets things backwards. "The causal relationship between international law
and state interests runs in the opposite direction," they argue. "International
law emerges from states' pursuit of self-interested policies on the international
stage. International law is, in this sense, endogenous to state interests. It is not
a check on state self-interest; it is a product of state self-interest., 24 As a
result, there would seem to be no danger that policymakers will adopt what
George Kennan once disparagingly described as the "legalistic-moralistic"
approach to international relations.25
The relentlessly instrumentalist approach of Limits raises two fundamental
questions, which I will consider in turn. First, is an exclusively instrumentalist
account of state behavior compatible with international law qua law? Second,
do Goldsmith and Posner present a convincing case for pursuing their approach
to the exclusion of others? In particular, does their theory have
methodological advantages over noninstrumental approaches?

A. Self-Interest and Legal Obligation

Although Goldsmith and Posner acknowledge that international law is a

"real phenomenon," they draw a sharp contrast between acting out of self-
interest and acting out of a sense of legal obligation. This raises the question:
Does their rational choice theory of international law leave any room for
international law to operate as law?
In considering the relationship of self-interest to legal obligation, it is
important to distinguish between two very different issues: first, why states
develop international rules, and second, why they comply with (or violate)
those rules? Goldsmith and Posner do not clearly separate these questions
because they think the answer to both is the same: Calculations of rational self-
interest determine decisions by states about compliance as well as legal
development. But there is no reason, a priori, to suppose that the explanation
of lawmaking and compliance is the same. For example, states might negotiate

Charles Krauthammer, The Curse of Legalism: InternationalLaw? It's Purely Advisory, 201
NEW REPUBLIC 44 (Nov. 6, 1989).
23 GOLDSMrrH & POSNER, supra note 1, at 13.
24 Id.
2 GEORGE F. KENNAN, AMERICAN DIPLOMACY 95 (pbk ed. 1970).
2006] INTERNATIONAL LAW IN BLACK AND WHITE

a treaty based on their perceived self-interest, as Goldsmith and Posner argue,

but then continue to comply even when their self-interest changes because they
feel this is the "right" thing to do, or because of domestic political or legal
factors.
From the standpoint of a traditional international lawyer, much more is at
stake with respect to the issue of compliance than lawmaking. In general,
international law does not concern itself with the question of why states
negotiate treaties or engage in behavior that helps create customary norms.
Mainstream international law is thus compatible with virtually any explanatory
approach-purely instrumental accounts along the lines that Goldsmith and
Posner elaborate; realist approaches that emphasize the role of power;2 6 and
constructivist accounts that focus on the role of ideas, values and learning. But
for international law to matter, for it to be a reality rather than merely
epiphenomenal, it must exert some independent influence on state behavior;
it must have some effect on compliance.
At times, Goldsmith and Posner seem to suggest that international law, as
such, exerts no independent influence. "A state's compliance.., has nothing
to do with acting from a sense of legal obligation," they argue; it is determined
solely by rational calculations of self-interest.2 "States do not act in
accordance with a rule that they feel obliged to follow; they act because it is
in their interest to do so. The rule does not cause the states' behavior; it
reflects their behavior."28 The assumption here seems to be that self-interest
and a sense of legal obligation offer two mutually exclusive accounts of state
behavior.
But this reflects an unduly cramped view of what it means to act out of a
sense of legal obligation. Acting out of a sense of legal obligation requires
merely that a rule is entrenched-that states take the rule as a reason for action
independent from the reasons that led them to adopt the rule in the first place.29
The idea was nicely expressed in a poster I once saw showing Isaac Newton
sitting beneath the apple tree, with an apple just beginning to fall. The poster
proclaimed at the bottom, "Gravity: It's not just a good idea. It's the law!" Of
course, with gravity, its status as "law" adds nothing to its force-that is the

26 Although if a treaty results from fraud, corruption or coercion, then this deprives the
theory of any legal effect. Vienna Convention on the Law of Treaties arts. 49-52, May 22, 1969,
1155 U.N.T.S. 331.
27 GOLDSMITH & POSNER, supra note 1, at 39.
28 id.
29 See FREDERICK SCHAUER, PLAYING BY THE RULES: A PHILOSOPHICAL EXAMINATION OF
RULE-BASED DECISION-MAKING IN LAW AND IN LIFE (1993).
294 GA. J. INT'L & COMP. L. [Vol. 34:285

joke. But the joke depends on our understanding the term, "law," not merely
in terms of physical regularities, but in a second legal sense, as providing a
reason for action in and of itself, because of its status as "law." Like the law
of gravity, an international law may be adopted by states because they consider
it a good idea. But once adopted, its legal force does not depend on states
continuing to accept the rule as a good idea; its status as law constitutes an
independent reason for action.
Just as mainstream international law is neutral as to why states develop
international rules, it is also neutral as to why states might accept international
rules as reasons for action. Goldsmith and Posner seem to assume that the only
possible basis of legal obligation is morality. The idea that states obey
international law out of a sense of legal obligation, they argue, "reduce[s] to
the idea that a state is drawn toward compliance with international law because
compliance is the morally right or legitimate thing to do."3 ° But a sense of
legal obligation can have other bases than morality. 3 It does not depend on
states having pure transcendental wills, obeying international law simply
because that is the right thing to do, without regard to consequences. Instead,
actors might accept rules as reasons for action based on prudential grounds.
Hobbes, for example, suggests that individuals have a long-term self-interest
in the maintenance of order, which serves as a basis for legal compliance.
Similarly, international lawyers have attempted to explain compliance with
international legal obligations instrumentally-in terms of reputational factors,
for example, or states' long-term self-interest in a functioning legal system.
Do Goldsmith and Posner accept the possibility of an instrumental account
of legal compliance? Do they agree that states may accept international rules
as reasons for action based on self-interest? Their argument is unclear on this

30GoLDsMrrH & POSNER, supra note 1, at 15. Goldsmith and Posner advance both
descriptive and normative arguments against this theory oflegal compliance. Descriptively, they
deny that states, even when acting consistently with international law, are influenced by moral
considerations as opposed to self-interest. Normatively, they argue that states have no moral
duty to comply with international law. Although Goldsmith and Posner evidently believe their
descriptive and normative arguments to be mutually reinforcing, the normative argument tends
to undercut the credibility of the descriptive claim. For, if their descriptive claim is true-that
states never comply with international law for moral reasons-then their normative argument is
beside the point. That they nevertheless advance it, in essence as an argument in the alternative
against the moral force of international law, suggests both a lack confidence in their descriptive
claim and a polemical agenda that may cloud the objectivity of their descriptive analysis.
31 SeegenerallyA. JOHNStMMONS, MORALPRNCIPLES ANDPOLITICALOBLIGATIONS (1979);
Leslie Green, Law and Obligations,in OXFORD HANDBOOKOFJURISPRUDENCE AND PHILOSOPHY
OF LAW (Jules Coleman & Scott Shapiro eds., 2002).
2006] INTERNATIONAL LAW IN BLACK AND WHITE 295

point. They say that "states comply with [international law] when it is in their
rational self-interest to do so, and not otherwise. 32 But this statement contains
a fundamental ambiguity about whether the calculation of self-interest remains
the same before and after the development of a legal rule. To the extent that
the existence of a legal rule changes a state's calculation of self-interest, by
providing it with an additional interest to do what the rule directs (or to
retaliate against others for failing to do so), then the rule has an effect and is
not merely epiphenomenal.
At times, Goldsmith and Posner seem to deny this possibility, for example,
when claiming that customary international law is not an "exogenous influence
on state behavior."3 3 But this statement can be read in a more limited way, as
merely a corollary of their earlier argument that international law is a product
of-and hence endogenous to-state interests.3 4 On this reading, international
law cannot be an exogenous influence on state behavior for the simple reason
that it has already been made endogenous.
Given Goldsmith and Posner's reluctance to face this issue squarely, any
conclusions regarding their views must remain tentative. But I think the book
as a whole leaves room for international law to play a distinctively legal role.
Despite its claim to exclude from its analysis "a preference for complying with
international law,"35 Limits does not appear to reject the possibility that a legal
rule may change a state's calculation of self-interest. It downplays reputational
factors, for example, but does not deny their existence.36 Yes, Limits is
skeptical about the role of international law-as its title suggests, very strongly
so. But it acknowledges that international law plays a causal role by
communicating information about the strength of a state's commitment, what
counts as cooperation, and so forth. As the authors recognize, in some subject
areas, such as diplomatic immunity law, compliance with customary rules is
very high, even when violations of those rules would be in a state's short-term
self-interest.37 If this reading is correct, the issue is not whether, on a purely
instrumental account, international law can be consequential, but rather, how
much difference it makes and in what contexts.

32 GOLDSMIH & POSNER, supra note 1, at 100.

33 Id. at 43.
34 Id. at 13.
" Id. at 9.
36 Id. at 102 ("Both [retaliation and reputation stories] are consistent with rational choice

premises.").
37 Id. at 55.
 GA. J. INT'L & COMP. L. [Vol. 34:285

B. A Scientific Methodology of InternationalLaw?

In his classic nineteenth century fantasy, Flatland,Edwin Abbot portrays

a world of only two dimensions, whose inhabitants are unable to comprehend
the arrival of a stranger from the three-dimensional world of spaceland. 38
Reading Limits often brought to mind Flatland,for Limits gives a similarly
flattened account of international law, reduced to the dimensions of self-
interest and, to a subsidiary extent, power. Or, to change the metaphor, Limits
depicts a world of black and white, with few shades of gray, much less of
color. In the very simple model of international law it elaborates, unitary
states rationally pursue their self-interest. Gone from this account are the
innumerable features that give richness and texture to international
affairs--domestic politics, nongovernmental groups, intergovernmental
organizations, leaders, ideas and ideology. Gone too are the normative
accounts of state behavior dear to international lawyers, which focus on
noninstrumental factors such as morality and legitimacy.
Although Limits has been criticized for the incompleteness of its initial
assumptions-for instance, that states are not the only important actors, that
they are not unitary actors, and that they do not always behave rationally-the
simplified model of state behavior depicted in Limits is not, in itself,
problematic. A hallmark of the scientific method is abstraction from reality.
Few scientific models fully take account of the complexities of the real
world-that is why so much effort is necessary in the laboratory to create ideal
conditions in which to test a theory. If Galileo had indeed dropped a cannon
ball and a much lighter musket ball from the Leaning Tower of Pisa, the
musket ball would have dropped slightly more slowly due to air resistance.
The abstraction of Goldsmith and Posner's theory does not distinguish it from,
say, Newton's law of gravity, which does not take into account friction.
In science, the rationale for simplification is methodological. In order to
develop testable hypotheses, we need to isolate different causal factors.
Simplification requires abstracting from reality, but has a significant payoff,
namely, scientific rigor. So long as we distinguish between simplifying
assumptions and reality-so long as we do not confuse flatland with
spaceland-then there is no problem.
In Limits, however, it is not always clear that Goldsmith and Posner do
maintain this distinction-and that is what, I think, undermines their account

38 EDwIN ABBOTr, FLATLAND: A ROMANCE OF MANY DIMENSIONS (Little Brown 1939)

(1890).
2006] INTERNATIONAL LAW IN BLACK AND WHITE

of international law. True, they describe their theory modestly at times, as a

"pragmatic tool" to "organize our ideas and intuitions and to clarify
assumptions."3 9 As they acknowledge: "No theory predicts all phenomena
with perfect accuracy. And we do not deny that states act irrationally, because
their leaders make mistakes, because of institutional failures, and so forth."'
But, despite these qualifications, the book has more hegemonic aims. It seeks
not simply to provide an instrumental account of international law, but to
exclude noninstrumental approaches as methodologically deficient.41 And it
attempts not merely to explain particular features of international law, but to
develop a comprehensive theory that can account for the entire range of
42
international legal phenomena.
In relentlessly focusing on a single explanatory factor, Goldsmith and
Posner remind one more of international relations theorists than of
international lawyers. International relations scholars tend to subscribe to a
particular theory; they are realists, or institutionalists, or constructivists, but
seldom all three. International lawyers, by contrast, typically take a more
eclectic approach, drawing on a variety of methodologies and theories as the
occasion suits. Power, interests, ideology, domestic politics may all figure into
their accounts of the international legal process. I first encountered the
difference between the two mindsets many years ago, when participating in a
program populated by international relations scholars. I was perplexed, at the
time, by the imperative the political scientists felt to adopt a single theory, and
asked why they could not simply draw the best from each, since each appeared
to contain an element of the truth. The response was always the same: an
eclectic approach would be unrigorous and unscientific, a view that Goldsmith
and Posner evidently share.4 3
The claim to scientific rigor deserves careful consideration. How should
we appraise it? Are there good methodological reasons to exclude normative
factors from our explanatory toolbox, as Goldsmith and Posner argue? What
are the putative benefits of Goldsmith and Posner's methodology and what are
the costs?
Three virtues often associated with the scientific method are falsifiability,
predictive value, and empiricism. Let us consider each of these in turn.

" GoLDsMITH & POSNER, supra note 1, at 7.

4 Id.
41 Id. at 10 (arguing that a preference for complying with international law should be
excluded for methodological reasons).
42 Id. at 17 ("Ours is a comprehensive analysis of international law.").
41 Id. at 9-10.
 GA. J. INT'L & COMP. L. [Vol. 34:285

Goldsmith and Posner themselves propose a test of falsifiability for their

theory." But their theory has sufficient play in the joints that falsification is
no easy matter. Apparently contrary evidence can always be explained away.
The European Convention on Human Rights, for example, seems highly
effective, contrary to Goldsmith and Posner' s claim that "modem human rights
treaties have had no significant impact on human rights protection."45 But they
dismiss this as a sui generis result, resulting from regional integration rather
than international law.4 6 Or they note that the "only rigorous empirical" study
of GATT dispute resolution comes to the conclusion that states "care about
complying with GATT rulings," but then proceed to find various faults with
the study.47 For critics, the difficulty of falsifying rational choice theory
represents a serious indictment. But modem philosophy of science teaches that
scientific theories, in general, cannot be falsified in any straightforward way.
At the extreme, contrary evidence can always be dismissed as "experimental
error."4 In the web of our beliefs, we practice what the philosopher, Willard
van Orman Quine, has called a "principle of minimum mutilation., 49 When
evidence appears to falsify a theory, we first try to make adjustments at the
margins, rather than change our core beliefs. Ptolemy added epicycle upon
epicycle to explain the planetary motions, in order to preserve his core belief
that the planetary bodies revolved in circular orbits, just as Goldsmith and
Posner propose successive refinements to preserve their core belief in rational
choice. Theories are not generally falsified by disproving particular
propositions; they are displaced as a whole when a better theory comes along.5"
Perhaps a better measure of scientific theories than falsifiability is
predictive value. Generally, this is the payoff for abstraction in science: It
yields interesting predictions. It tells us things we would not otherwise have
expected: Maxwell's theory of electromagnetism predicts the existence of
electromagnetic waves; Newton's theory of gravity predicts that feathers and

4 Id. at 10.
45 Id. at 121.
46 Id. at 126.
47 Id. at 156-57.
" For example, in his oil-drop experiment to determine the charge of the electron, for which
he won a Nobel Prize, Robert Millikan apparently discarded experimental results that did not fit
his theory, on the grounds that they must have involved instrumental errors. See GERALD
HOLTON, THE SCIENTIFIC IMAGINATION 25 (1978).
49 Willard van Orman Quine, Two Dogmas of Belief, in FROM A LOGICAL POINT OF VIEW
(1953); see also WILLARD VAN ORMAN QUINE, THE WEB OF BELIEF (2d ed. 1978).
'oTHOMAS KUHN, THE STRUCTURE OFScIErNTmc REVOLUTIONS (3d ed. 1996) (discussing
the role of anomalies in ultimately producing a "paradigm shift").
20061 INTERNATIONAL LAW IN BLACK AND WHITE

cannon balls will fall at the same rate in a vacuum, regardless of their mass;
Adam Smith's theory of the invisible hand predicts that the competitive,
egoistic behavior of individuals can produce collectively desirable results,
through the operation of the market; and trade theory predicts that lowering
tariff barriers will generally make a state better off, regardless of what other
countries do. Common sense would not lead us to expect any of these things;
they are all non-intuitive or even counterintuitive. In my view, one of the
disappointments of Limits is that it yields relatively few such novel or
unexpected predictions. In the field of diplomatic immunities, for example, it
predicts that rogue states are more likely to violate rules than civilized states,5 1
and that states are more likely to violate international law when the stakes are
high.52 Similarly, with respect to the law of war, the theory predicts that states
will make an exemption to the general rule prohibiting seizure of coastal
fishing vessels, if the fishing vessel serves a military purpose.5 3 But far from
being unexpected, these predictions are exactly what one would have
anticipated, even if one had never heard of rational choice theory.
A final virtue often associated with science is empiricism. Although Limits
does not yield a rich trove of predictions, does it at least have significant
empirical support? The answer is uncertain. For despite occasional claims by
Goldsmith and Posner that Limits is empirical in nature, its empiricism is quite
thin. In general, its approach might more accurately be characterized as
conceptual. The structure of its arguments tends to take the following form:
suppose that one state has preference x and another state preference y, then this
is what we would expect to happen. But whether actual states in the real world
actually have these preferences, and whether, when they do, the results are
what the book predicts are questions not systematically addressed.
Of course, Goldsmith and Posner do claim to test their conceptual analysis
on the basis of a number of case studies. But even accepting Goldsmith and
Posner's versions of these case studies, 54 they are essentially anecdotal in
nature. And even when Goldsmith and Posner discuss actual rather than purely
hypothetical states, their identification of state interests remains largely

51 GOLDSMrrH & POSNER, supra note 1, at 57.

52 Id. at 58.
51 Id. at 75-76 (in commenting on the exemption allowing seizure of a fishing vessel when
it serves a military purpose, Goldsmith and Posner proclaim in triumph, this is "just the sort of
exemption our theory would predict").
5 See David M. Golove, Leaving Customary InternationalLaw Where It Is: Goldsmith and
Posner'sThe Limits of International Law, 34 GA. J. INT'L & COMP. L. 333 (2006) (challenging
the historical account provided in Limits of the "free ships, free goods" rule).
300 GA. J. INT'L & COMP. L. [Vol. 34:285

conjectural. For example, in discussing British and French practice in the

nineteenth century, they say that Britain and France "might" both have had an
interest in seizing each other's fishing vessels, but that "perhaps" they would
have been better off if they had saved their money for other purposes.5 5 These
are not empirical statements; they are conjectures. If one did a concordance
of the verbs used in Limits, I suspect that "might" would appear near the top
of the list.
In describing the methodology of Limits as conceptual rather than
empirical, I do not mean to disparage the book or suggest it is without value.
Despite its shortcomings, it succeeds admirably in its goal of providing "a
simple but plausible descriptive account" 6 of various international regimes.
Although the configurations of state interests and power that it
elaborates-coincidence of interests, coordination, cooperation and
coercion-are familiar to international relations theorists,57 Goldsmith and
Posner deserve credit for articulating them in a clear and coherent manner,
thereby making them accessible to a wide audience of international lawyers.
Whether one agrees with the particular way in which they apply game theory
to analyze particular issues, their discussion of customary international law
amply demonstrates the value of rational choice analysis. Not only does it help
illuminate particular issues, it offers a progressive research agenda, one of the
hallmarks of the scientific method. 8
But while I have little doubt that rational choice theory represents, to a
significant degree, the future of international legal scholarship, it does not tell
the whole story. It helps us understand certain features of international law,
but not others-it comes at a cost, which we might broadly call "reality."
Consider, for example, the failure by the U.S. to ratify the U.N. Convention on
the Law of the Sea (UNCLOS). UNCLOS is currently supported by the Bush
Administration, a large majority of the Senate, the Defense Department, and
most segments of industry. It was reported on favorably by the Senate Foreign
Relations Committee. But, as of December 2005, the Senate had not voted on

55 GOLDSMITH & POSNER, supra note 1, at 76.

56 Id. at 10; see also id. at 13.
57 See, e.g., ARTHUR A. STEIN, WHY STATES COOPERATE: CIRCUMSTANCE AND CHOICE IN
INTERNATIONAL RELATIONS (1993); SCOTT BARRETT, ENVIRONMENT AND STATECRAFT: THE
STRATEGY OF ENVIRONMENTAL TREATY-MAKING (2003).
58 See, e.g., lmre Lakatos, Falsification and the Methodology of Scientific Research
Programmes, in CRITICISM AND THE GROWTH OF SCIENTIFIC KNOWLEDGE 91 (Irte Lakatos &
Alan Musgrave eds., 1970); LARRY LAUDAN, PROGRESS AND ITS PROBLEMS: TOWARDS A
THEORY OF SCIENTIFIC GROWTH (1977).
2006] INTERNATIONAL LAW IN BLACK AND WHITE

advice and consent to ratification due to a "hold" put on the treaty by a handful
of senators. Here, it makes little sense to think of the United States as a
unitary actor, rationally pursuing its self-interest. Instead, the failure to ratify
UNCLOS is the result of domestic politics, which could easily change-for
example, if elections resulted in new Senate leadership less deferential to the
senators who are blocking the treaty. To understand these twists and turns, we
need the methodology of history, not political science.
Consider similarly the growing role of both international and domestic
tribunals.59 State interests might be able to provide an explanation for why
states establish international tribunals in the first place and comply with their
decisions, or why they give domestic courts authority to apply international
law. But it is difficult to explain the actual practice of international tribunals
such as the WTO Appellate Body, or the British House of Lords in the
Pinochet case, 60 in purely instrumental terms. The most parsimonious way to
account for the behavior of most if not all tribunals is in terms, not of state
interests, but of legal rules and legal reasoning.
Where Goldsmith and Posner get into trouble is in mistaking their
simplifying assumptions for reality. This happens at both the micro and macro
levels. Sometimes, in a move reminiscent of President Reagan's occasional
confusion of movies and reality, Goldsmith and Posner start by expounding a
hypothetical, then later treat the hypothetical as a statement about the real
world. For example, they posit a CEO who follows contract law out of self-
interest rather than a preference to comply with the law. Then, on the basis of
this hypothetical, they draw an empirical conclusion, claiming that "this is
surely the case for international law as well."'" Similarly, after speculating
about the possible views of different U.S. agencies on compliance (using the
verbs "may" or "might" seven times), they then draw the factual conclusion
that bureaucracies in reality "have competing preferences" and that, "when
bureaucracies differ on compliance issues, the compliance view does not
always prevail. 62 In doing so, they transmute speculations about what "may"
be the case into declarative statements about what "is" the case; they confuse
plausible conjectures with empirical truths.

" Laurence R. Helfer& Anne-Marie Slaughter, Toward a Theory ofEffective Supranational

Adjudication, 107 YALE L.J. 273 (1997); Laurence R. Heifer & Anne-Marie Slaughter, Why
States Create InternationalTribunals:A Response to ProfessorsPosner and Yoo, 93 CAL. L.
REv. 899 (2005).
60 See PHILIPPE SANDS, LAwLEss WORLD (2005).
61 GOLDSMITH & POSNER, supra note 1, at 105.
62 Id. at 106.
 GA. J. INT'L & COMP. L. [Vol. 34:285

The same problem occurs at the macro level, in their rejection of

noninstrumental accounts of international law. The fact that we can plausibly
understand some international rules in terms of unitary states acting rationally
to further their interests does not mean that states are the only relevant actors
or that interests are the only relevant motivation. Goldsmith and Posner admit
that whether states care about compliance with international law is an
empirical question, which they "do not purport to resolve."63 Instead, they
criticize reliance on noninstrumental factors on methodological grounds,
claiming that "the assumption of a tendency towards compliance [with
international law] has little if any explanatory value." 64 But this seems plainly
false. I promise my daughter that I will take her to the zoo on a particular
day-a day that unbeknownst to me is Super Bowl Sunday. I believe that I
should keep my promises, if possible. I take her to the zoo, even though this
means foregoing a Super Bowl party to which I have been invited. In what
sense does my normative belief in the importance of keeping promises have
"little if any explanatory value"? To be sure, it does not explain why I
sometimes break my promises. Nor does it exclude the possibility that
instrumental factors may also play a role-for example, my desire to "look
good" to my daughter and others. So it does not offer a full account of my
behavior with respect to promises. But this is not the same as saying that it
lacks explanatory value. Goldsmith and Posner's argument assumes that
normative factors are like intelligent design, which is inferred from the
complexity of life and hence cannot explain that complexity in a noncircular
way. 65 But to the extent that we have empirical evidence that, on actual
international law issues, important state actors are influenced by a logic of
appropriateness as well as a logic of consequences, and that these actors
influence the state's decisions, there is no methodological reason to exclude
this noninstrumental account.
How important a factor in explaining state behavior is the preference for
compliance with international law? In the best selling book Freakonomics,the
authors note that people comply with the moral norm against stealing 87% of
the time, even in the absence of any enforcement.' They attribute this to the
moral sentiments, which Adam Smith wrote about more than 200 years ago.

63 Id. at 10.
'4 Id. at 15.
63 For that matter, rational choice theory has the same problem: to the extent it infers
preferences from behavior, then using those preferences to explain behavior is circular.
66 STEVEN D. LEVITT & STEPHEN J. DUBNER, FREAKONOMICS: A ROGUE ECONOMIST
EXPLORES THE HIDDEN SIDE OF EVERYTHING 50-51 (2005).
2006] INTERNATIONAL LAW IN BLACK AND WHITE

But if people indeed have moral sentiments, as I think our experience of

everyday life confirms, then why should these stop at the water's edge? Why
should they play no role in decisions by governmental actors about
international affairs? In recent years, significant efforts have been made to
address the various factors that contribute to compliance with international
law, although assessing their relative importance has proved exceedingly
difficult.6 7 My own limited experience is that, in international environmental
regimes, U.S. officials take compliance quite seriously, and are unwilling to
join a treaty unless they know exactly how the United States will be able to
fulfil its obligations." But, of course, in other areas, for example, payment of
United Nations dues, U.S. compliance with international law has been poor.
The institutions with the most obvious preference to comply with
international law are those responsible for dispute settlement: courts, arbitral
tribunals, WTO panels, and so forth. Although judicial decision making
involves many factors, not all of which are necessarily legal in nature (a
judge's politics, for example, or country of nationality), I do not think it is
possible to understand judicial behavior absent an assumption that judges are
trying to apply the law. So, as international tribunals proliferate and become
more influential, the preference for compliance with international law will gain
in importance.
The reductionist quality of Goldsmith and Posner's account makes it
particularly inappropriate for the practitioner of international law, who, unlike
the theorist, operates in the real world, and therefore needs to try to account for
the full range of causal factors that influence the behavior of states and other
international actors. The theory is of relatively little value, pragmatically, in
helping us think through actual issues.
Consider, for example, the problem of global warming. Climate change has
been subjected to considerable analysis in the manner that Limits suggests, as
a multi-party prisoner's dilemma. This is fine insofar as it goes. But
Goldsmith and Posner's approach stops just at the point where things get
interesting. Although in the long run, climate change may be a prisoner's
dilemma game, in the short term it is difficult to understand the issue in these
terms. On the one hand, some actors are pushing ahead with action to combat
climate change, such as the European Union, the state of California, a large

67 Shelton, supranote 7.
For example, the President has been unwilling to ratify the Basel Convention on
Transboundary Movements of Hazardous Wastes, despite Senate consent, because Congress has
been unable to enact implementing legislation.
 GA. J. INT'L & COMP. L. [Vol. 34:285

number of cities, and businesses such as British Petroleum and General

Electric, even though their actions are not being reciprocated by others, and
even though they fail to constitute a "minimum viable coalition," in the
parlance of game theory. On the other hand, the United States under President
Bush opposes such action not just by itself, but by others as well. It is not
trying to free ride, as game theory would predict; rather, it is trying to stop the
bus altogether.
To understand all of this, we need to explore an issue that Limits leaves to
one side, as exogenous to its theory, namely how preferences form and change.
And to do this, we need to bring in a variety of causal factors that are outside
Goldsmith and Posner's model-leadership, domestic politics, and even the
role of values. For if one asks officials in northeastern states of the United
States why they are spearheading a regional initiative to cap CO 2 emissions,
the answer they most often give is, because it is the right thing to do.
The simple model set forth in Limits does no better in helping us assess the
potential international responses to climate change. In what forum should
negotiations proceed? What types of commitments should be used? How can
we restructure incentives through institutional design in order to promote
compliance? The simple game theoretic models that Goldsmith and Posner
elaborate do not even begin to address these more complex issues of
institutional design.69
In the end, Limits makes a convincing case that rational choice theory can
help us better understand the development and effectiveness of international
law. But it provides no compelling reason why noninstrumental factors might
not also play a role. It presents a flattened picture of the world, drained of
texture and nuance and color. It illustrates that to understand international law
we need not only science, but also art.

69 For a more sophisticated application of game theory to international environmental issues,

see BARRETT, supra note 57.