World Politics

http://journals.cambridge.org/WPO

Additional services for World Politics:

Email alerts: Click here

Subscriptions: Click here
Commercial reprints: Click here
Terms of use : Click here

The Constructive Turn in International

Relations Theory

Jeffrey T. Checkel

World Politics / Volume 50 / Issue 02 / January 1998, pp 324 - 348

DOI: 10.1017/S0043887100008133, Published online: 13 June 2011

Link to this article: http://journals.cambridge.org/

abstract_S0043887100008133

How to cite this article:

Jeffrey T. Checkel (1998). The Constructive Turn in International Relations
Theory. World Politics, 50, pp 324-348 doi:10.1017/S0043887100008133

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/WPO, IP address: 132.174.254.155 on 05 Apr 2015

 Review Article

THE CONSTRUCTIVE TURN IN

INTERNATIONAL
RELATIONS THEORY
By JEFFREY T. CHECKEL*

Martha Finnemore. National Interests in International Society. Ithaca, N.Y.:

Cornell University Press, 1996,149 pages.
Peter Katzenstein, ed. The Culture ofNational Security: Norms and Identity in
World Politics. New York: Columbia University Press, 1996, 562 pages.
Audie Klotz. Norms in International Relations: The Struggle against Apartheid.
Ithaca, N.Y.: Cornell University Press, 1995,183 pages.

F OR the past decade a central locus of contention within interna-

tional relations has been the neorealist-neoliberal debate. This ex-
change has been fruitful and cumulative, allowing proponents of the
two research programs to sharpen arguments while simultaneously
shedding light on key issues of world politics, for example, the condi-
tions under which relative or absolute gains-seeking behavior occurs.1
By and large, the constructivists under review would concur with
such a characterization. Their critique of neorealists and neoliberals
concerns not what these scholars do and say but what they ignore: the
content and sources of state interests and the social fabric of world pol-
itics. Reaching back to earlier theoretical traditions (the English school,

* Earlier versions of this essay were presented at the 1996 annual convention of the American Po-
litical Science Association, and at the workship on "Structural Change in International Politics," spon-
sored by the German Political Science Association, February 1997. For comments, I thank Andrew
Cortell, Aaron Hoffman, Jeff Legro, Thomas Risse, and Alex Wendt. The financial support of the
Alexander von Humboldt-Stiftung and German Marshall Fund is gratefully acknowledged.
1
See Robert Keohane, After Hegemony: Cooperation and Discord in the World Political Economy
(Princeton: Princeton University Press, 1984); David Baldwin, ed., Neorea/ism and Neoliberalism: The
Contemporary Debate (New York: Columbia University Press, 1993); Robert Powell, "Anarchy in In-
ternational Relations Theory: The Neorealist-Neoliberal Debate (Review Article)," International Or-
ganization 48 (Spring 1994); and "Promises, Promises: Can Institutions Deliver?" International Security
20 (Summer 1995).

World Politics 50 (January 1998), 324-48

 CONSTRUCTIVIST TURN IN IR THEORY 325
some versions of liberalism) and reaching out to new disciplinary foun-
dations (sociology), constructivists seek to expand theoretical discourse.
Regarding both the books under review and constructivism more
generally, this essay advances three claims. First, I argue that construc-
tivism has succeeded in broadening the theoretical contours of IR. By
exploring issues of identity and interest bracketed by neoliberalism and
neorealism, constructivists have demonstrated that their sociological
approach leads to new and meaningful interpretations of international
politics. Moreover, constructivists have rescued the exploration of iden-
tity from postmodernists. By arguing for its importance using methods
accepted by the majority of scholars, they have been able to challenge
mainstream analysts on their own ground. Second and more critically, I
show that constructivism lacks a theory of agency. As a result, it over-
emphasizes the role of social structures and norms at the expense of the
agents who help create and change them in the first place.
Third, I argue that constructivism remains a method more than any-
thing else. The central challenge for these scholars is theory develop-
ment. Having demonstrated that social construction matters, they must
now address when, how, and why it occurs, clearly specifying the actors
and mechanisms bringing about change, the scope conditions under
which they operate, and how they vary across countries. To accomplish
this task, constructivists must integrate their insights and assumptions
with middle-range theory. Otherwise, the empirical ad hocism that
plagues their current work will remain.
The essay is organized as follows. It begins by defining construc-
tivism and its approach to the study of global politics. Next, it consid-
ers the empirical contribution of constructivists, focusing on the three
books under review. Finally, the review explores several issues construc-
tivists must address if they are to mount a sustained challenge to their
competitors in contemporary IR.

T H E SOCIAL CONSTRUCTION OF INTERNATIONAL POLITICS

The constructivist critique of neorealism and neoliberalism reaches well

beyond the level-of-analysis argument of either Image I (individual) or
Image II (domestic politics) theorists. Constructivism is concerned not
with levels per se but with underlying conceptions of how the social
and political world works. It is not a theory but an approach to social
inquiry based on two assumptions: (1) the environment in which
agents/states take action is social as well as material; and (2) this setting
can provide agents/states with understandings of their interests (it can
326 WORLD POLITICS
"constitute" them). Put differently, these scholars question the materi-
alism and methodological individualism upon which much contempo-
rary IR scholarship has been built.
The first assumption reflects a view that material structures, beyond
certain biological necessities, are given meaning only by the social con-
text through which they are interpreted. Consider nuclear weapons—
the ultimate material capability. Constructivists argue that it is not such
weapons themselves that matter. After all, the United States worries
very little about the large quantity of nuclear weapons held by the
British; however, the possibility that North Korea might come into
possession of even one or two generates tremendous concern.2
The second assumption addresses the basic nature of human agents
and states, in particular, their relation to broader structural environ-
ments. Constructivists emphasize a process of interaction between
agents and structures; the ontology is one of mutual constitution, where
neither unit of analysis—agents or structures—is reduced to the other
and made "ontologically primitive." This opens up what for most theo-
rists is the black box of interest and identity formation; state interests
emerge from and are endogenous to interaction with structures.3
Constructivists thus question the methodological individualism that
underpins both neoliberalism and neorealism. This agent-centered
view asserts that all social phenomena are explicable in ways that in-
volve only individual agents and their goals and actions; the starting
point of the analysis is actors (states) with given properties. Ontologi-
cally, the result is to reduce one unit of analysis—structures—to the
other—agents.4
Also implicit in many constructivist accounts is a model of human
and state behavior where rule-governed action and logics of appropri-
ateness prevail. Such logics involve reasoning by analogy and metaphor
and are not about ends and means. Under them, agents ask "What kind
of situation is this?" and "What should I do now?"—with norms help-
ing to supply the answers. Norms therefore constitute states/agents,
providing them with understandings of their interests.5
2
Alexander Wendt, "Constructing International Politics," International Security 20 (Summer 1995), 73.
3
For an excellent discussion of this black box for neoliberals and neorealists written by a theorist
sympathetic to their enterprise, see Powell (fn. 1), 317-24.
4
On neoliberalism's methodological individualism, see Vblker Rittberger, Andreas Hasendever, and
Peter Mayer, "Interests, Power, Knowledge: The Study of International Regimes," Menbon Interna-
tional Studies Review 40 (October 1996), 183-87. For that of neorealism, see Alexander Wendt, "The
Agent-Structure Problem in International Relations Theory," International Organization 41 (Summer
1987), 340-^4.
s
On logics of appropriateness, see James March and Johan Olsen, Rediscovering Institutions: Tbt
OrganizationalBasis ofPolitics (New York: Free Press, 1989).
 CONSTRUCTIVIST TURN IN IR THEORY 327
Scholars of rational choice, by contrast, use a behavioral model based
on utility maximization: when confronted with various options, an
agent picks the one that best serves its objectives and interests. Much
rational choice research ("thick" rationalism) also makes assumptions
about the content of these interests, typically that they are material
goods such as power or wealth. State (agent) interests are given a priori
and exogenously. Norms and social structures at most constrain the
choices and behavior of self-interested states, which operate according
to a logic of consequences (means-ends calculations).6
It is important to note that constructivists do not reject science or
causal explanation; their quarrel with mainstream theories is ontologi-
cal, not epistemological. The last point is key, for it suggests that con-
structivism has the potential to bridge the still vast divide separating
the majority of IR theorists from postmodernists. With the latter, con-
structivists share many substantive concerns (role of identity and dis-
course, say) and a similar ontological stance; with the former, they share
a largely common epistemology. Constructivists thus occupy a middle
ground between rational choice theorists and postmodern scholars.7
To illuminate these differences between constructivists and other
schools, it is helpful to explore their understanding of central terms.
Consider "norms," a concept that has gained much currency in IR schol-
arship over the past decade. While realists see norms as lacking causal
force, neoliberal regime theory argues that they play an influential rule
in certain issue-areas. However, even for neoliberals, norms are still a
superstructure built on a material base: they serve a regulative function,
helping actors with given interests maximize utility. Agents (states) cre-
ate structures (norms and institutions).8 For constructivists, by contrast,
norms are collective understandings that make behavioral claims on

6
On the last point, see Barry Weingast, "A Rational Choice Perspective on the Role of Ideas:
Shared Belief Systems and State Sovereignty in International Cooperation," Politics and Society 23
(December 1995); and Dennis Chong, "Rational Choice Theory's Mysterious Rivals," in Jeffrey Fried-
man, ed., The Rational Choice Controversy: Economic Models of Politics Reconsidered (New Haven: Yale
University Press, 1996). Useful introductions to rational choice are Jon Elster, "The Market and the
Forum," in Elster, ed., Foundations of Social Choice Theory (Cambridge: Cambridge University Press,
1986); James Morrow, Game Theory for Political Scientists (Princeton: Princeton University Press, 1994),
chap. 2; and Donald Green and Ian Shapiro, Pathologies ofRational Choice Theory: A Critique of Appli-
cations in Political Science (New Haven: Yale University Press, 1994), chap. 2.
7
See, among others, Alexander Wendt, Social Theory of International Politics (New York: Cambridge
University Press, forthcoming), chaps. 1-2. There is a good bit of confusion regarding these central
tenets of constructivism; see, for example, John Mearsheimer, "The False Promise of International In-
stitutions," International Security 19 (Winter 1994-95), 37-47.
8
For example, Stephen Krasner, ed., International Regimes (Ithaca, N.Y.: Cornell University Press,
1983). My comparisons here are limited to mainstream IR, since it has been vastly more influential than
postmodern work in shaping the field.
328 WORLD POLITICS
actors. Their effects reach deeper: they constitute actor identities and
interests and do not simply regulate behavior. As explanatory variables,
their status moves from intervening to independent (Finnemore, chaps.
3,4; Klotz, chap. 6, for example). Norms are no longer a superstructure
on a material base; rather, they help to create and define that base. For
constructivists, agents (states) and structures (global norms) are inter-
acting; they are mutually constituted.9
Taken together, these moves by constructivists—their questioning of
methodological individualism and materialism, along with a continu-
ing commitment to the scientific enterprise—have brought a breath of
fresh air to thinking about world politics, in ways accessible to nearly
all scholars. A key issue, however, is whether such new perspectives
allow these researchers to explain important international puzzles
and phenomena and thereby demonstrate the empirical value of their
approach.

PUZZLES AND ANOMALIES IN WORLD POLITICS:

T H E CONSTRUCTIVIST CONTRIBUTION

The books under review seek to make empirical contributions in three

areas: the role of international institutions and organizations (Finne-
more); international security (Katzenstein volume); and the effects of
international norms (Klotz). To evaluate their success, it is necessary to
establish a baseline for comparison.
On international institutions, the dominant school for well over a
decade has been neoliberal institutionalism. Since the publication of
Keohants After Hegemony, these scholars have shown increasing sophis-
tication in exploring the conditions under which institutions are created
in the first place and the various roles they play in world politics.10
Partly out of a concern for theoretical parsimony, neoliberal institu-
tionalists have purposely bracketed several issues, including the sources
of state interests, which are given by assumption. These scholars also
grant only a limited role to institutions, considering them to be the cre-
ation of self-interested states that at most constrain choices and strate-
' Strictly speaking, my discussion of norms as intervening or independent variables is not correct, as
constitutive effects (A enables or makes possible B) are not captured by standard causal terminology (A
causes B). See Wendt (fn. 2), 72. In practice, however, empirical constructivists use the terms inter-
changeably, see, for example, Mlada Bukovansky, "American Identity and Neutral Rights: From Inde-
pendence to the War of 1812," International Organization 51 (Spring 1997).
10
See, among others, Keohane (fn. 1); Lisa Martin, Coercive Cooperation: Explaining Multilateral
Economic Sanctions (Princeton: Princeton University Press, 1992); and Beth A. Simmons, "Why Inno-
vate? Founding the Bank for International Settlements," World Politics 45 (April 1993).
 CONSTRUCTIVIST TURN IN IR THEORY 329
gies. Virtually ignored is the possibility that the effects of institutions
reach deeper, to the level of interests and identity.
The baseline for the second issue-area—international security—is
difficult to establish with precision, given the turbulence stirred up
within this subfield by the end of the cold war. Certainly realism and
rationalism have been and remain dominant here, but scholars have re-
fined their analyses by paying more attention to domestic politics.
Important studies have enriched our understanding of security by
exploring the role of ideology and threat perception, coalition politics,
cognitive variables, and perceptions. While some accuse these scholars
of smuggling into their analyses sociological and cultural variables em-
phasized by constructivists, they are nonetheless still united in a com-
mon commitment to rationalism and materialism. On the former, key
actors (elite decision makers or groups within the state) make cost/ben-
efit calculations and choose strategies designed to maximize certain in-
terests; on the latter, perceptual, ideational, and cultural factors are
ultimately parasitic on a material base.11
Research on international norms, the third area addressed by the
books under review, has been heavily influenced by regime analysis.
These scholars have typically demonstrated that regime norms con-
strain the behavior of states; they are an explanatory variable that inter-
venes between underlying power distributions and outcomes.12
Work on epistemic communities and, more recently, on transna-
tional policy networks has brought research on international regimes
closer to the insights offered by constructivists. It does so by suggesting
that regime norms have deeper cognitive effects. Studies of this sort are
arguably still a minor current within regime theory; they are also beset
by a number of problems. Moreover, these scholars, especially those
working on epistemic communities, embrace a largely agent-centered
view, where state decision makers calculate and reason in response to a
changing material environment.13
11
See, among others, Stephen Walt, The Origins of Alliances (Ithaca, N.Y.: Cornell University Press,
1987); Jack Snyder, Myths ofEmpire: Domestic Politics and InternationalAmbition (Ithaca, N.Y.: Cornell
University Press, 1991); and William Wohlforth, "Realism and the End of the Cold War," Interna-
tional Security 19 (Winter 1994-95).
u
Mark Zacher, Governing Global Networks: International Regimesfor Transportation and Communi-
cations (New Yorlc Cambridge University Press, 1996), for example. An excellent, synthetic review of
the regime literature is Rittberger, Hasenclever, and Mayer (fh. 4).
13
See Peter Haas, Saving the Mediterranean: The Politics ofInternational Environmental Cooperation
(New York: Columbia University Press, 1990); and Kathryn Sikkink, "Human Rights, Principled
Issue-Networks and Sovereignty in Latin America," International Organization 47 (Summer 1993).
For critiques, see Checkel, Ideas and International Political Change: Soviet/Russian Behavior and the End
of the Cold War (New Haven: Yale University Press, 1997), chaps. 1, 7; and Helen Milner, "Interna-
tional Theories of Cooperation: Strengths and Weaknesses," World Politics 44 (April 1992).
330 WORLD POLITICS

CONSTRUCTING NATIONAL INTERESTS

With this background, the task is to assess the contribution of the con-
structivists, beginning with the book by Finnemore. She questions two
assumptions upon which most work on international institutions and
IR more generally rests: the definition of state interests and rational
means-ends calculations as the dominant mode of human interaction
(p. x). In ontological terms, she seeks to move scholarship away from
agent-oriented approaches (neoliberalism, for example) by paying more
attention to the structure side of the agent-structure debate (p. 7).
In an excellent opening chapter, Finnemore argues that a construc-
tivist logic of appropriateness is just as plausible a predictor of human
and state behavior as the rationalists' logic of consequences. When one
makes actor and state interests the dependent variable, as she does, such
logics of appropriateness can be key in determining their content. From
where do such logics come? Systemic norms propagated by interna-
tional organizations are one possible answer; they provide states with
direction and goals for action.
The core of the book is three case studies of how international insti-
tutions (and, in one case, an international nongovernmental organiza-
tion) were able to reconstitute state interests. These not only make for
fascinating reading, but they also offer fresh insights into how institu-
tions matter in world politics. They are also carefully argued, typically
using two streams of evidence: (1) correlations between the emergence
of new systemic norms and changes in state interests and practice; and
(2) analysis of discourse to see if actions are justified in ways consistent
with the values and rules embedded in the norms. These data, along
with attention to alternative explanations, allow Finnemore to build a
plausible case.
Her study of the United Nations Educational, Scientific and Cul-
tural Organization (UNESCO) is representative. Finnemore's puzzle is to
explain why so many countries—developing and developed—created
similarly stuctured science policy bureaucracies in a relatively brief pe-
riod. She begins with a rigorous consideration of alternative explana-
tions for their creation, for example, that they were established at the
behest of powerful domestic constituencies. After testing these quanti-
tatively and finding them lacking, Finnemore advances her own norms-
based argument.
She starts at the international level, documenting how a norm pre-
scribing the creation of national science units initially took hold at UN-
ESCO and was later consolidated there. O n the latter, part of the
evidence is a careful study of the changing discourse within UNESCO
 CONSTRUCTIVIST TURN IN IR THEORY 331
and among its member states. In particular, she notes how, over time,
the notion that such units were needed took on a prescriptive status and
came to be taken for granted.14
Finnemore then turns to the state level, establishing correlations be-
tween the norms promoted by UNESCO and the creation of science bu-
reaucracies by a number of states. To move beyond correlations,
however, she considers several cases (Lebanon, East Africa) in more
detail, analyzing the personal and organizational pathways through
which the UNESCO norms diffused to these states. While the evidence
here is a bit weaker (Finnemore conducted nofieldworkin the respec-
tive countries), it is nonetheless sufficient to allow her to make a plau-
sible case that the norms were causally important for the change in
science policy. Put differently, norms embodying certain logics of ap-
propriateness had provided states with a new understanding of their in-
terests (chap. 2).
Analysis of this sort moves one beyond the understanding of insti-
tutions provided by neoliberal institutionalists in at least two ways.
First, by endogenizing interest formation, Finnemore sheds much-
needed light on a crucial issue ignored by neoliberals: how states come
to define their interests in certain ways. International organizations can
teach states to value certain goals: national science bureaucracies in the
case of UNESCO and poverty alleviation as a policy objective in the case
of the World Bank. Finnemore carefully argues that these new interests
arose in the absence of domestic constituencies or powerful countries
favoring them. Instead, they were diffused to states by systemic norms,
from the outside, as it were. Materialist and rationalist explanations
cannot account for such value and behavioral change.
Second, the book demonstrates that international organizations are
not empty vessels that simply reduce transaction costs, as portrayed by
neoliberals. They are purposive entities that are able, in some cases, to
trump states and their power. Indeed, Finnemore's rich source material
at the international level gives her cases a sense of dynamism and his-
tory in the making that is typically absent from neoliberal accounts of
institutions. She has thus provided a theoretically informed and empir-
ically substantiated argument for how institutions not only constrain
but also constitute states and their interests, solving what is a puzzle for
other theorists.15

" T h e documentation and data come chiefly from archives at UNESCO's Paris headquarters.
15
For a similar argument, see David Strang and Patricia Mei Yin Chang, "The International Labor
Organization and the Welfare State: Institutional Effects on National Welfare Spending, 1960-80,"
International Organization 47 (Spring 1993).
332 WORLD POLITICS
The book also fills a gap in constructivism: failure to tell us why cer-
tain norms arise at particular times. Finnemore provides an answer by
exploring the role of moral entrepreneurs: committed individuals who
happen to be in the right place at the right time to instill their beliefs in
larger global social structures (pp. 24-28, chap. 4, pp. 137-39). 16
Finnemore's account is not without weaknesses, however. Most im-
portant, it is not clear what one does with her argument, with so much
resting on contingencies and idiosyncratic variables. While Finnemore
has demonstrated that social construction is causally important, she has
failed to specify systematically when, how, and why this occurs. To be
fair, one book cannot do everything. All the same, the critical next step
should be the development of a specifically constructivist theory of in-
ternational institutions, one that would elaborate such scope conditions.
A second weakness is the degree to which Finnemore's analysis is con-
sistent with constructivism's mutual constitution of agents and structures.
Now, exactly how one operationalizes mutual constitution is a dilemma
for all empirical constructivists. Finnemore's solution is a bracketing
strategy, where she first brackets agency and then, structures; her case
studies are broadly faithful to this approach (pp. 24-25, chaps. 2-4).
The problem is the wrong choice of agents: the entrepreneurs who
are responsible for the creation of norms in the first place. To analyze
the process of mutual constitution that led to a change of national in-
terests within particular states (her dependent variable), the agents she
should be exploring, especially given her emphasis on global norms as
the structures, are groups and individuals in those same states. If
Finnemore had focused on these agents, it would have led her to ex-
plore several important issues, for example, the feedback effects of state
(agent) behavior on the norms themselves.
A final difficulty is unavoidable given Finnemore's emphasis on sys-
temic social structures: the neglect of domestic politics. A question that
immediately comes to mind when reading her analysis is why norms
diffuse differentially, that is, why they have so much greater impact in
some countries than in others. Through what mechanisms do global
norms work their effects domestically? Finnemore alludes to these is-
sues at several points but provides no clear answers (pp. 125,137). This
is odd, since it is the constructivists, with their attention to practice and
interaction, who should be keying upon process and mechanisms.17
16
On moral entrepreneurs and the development of norms, see also Ann Florini, "The Evolution of
International Norms," International Studies Quarterly 40 (September 1996).
17
Indeed, Wendt himself stresses the importance of mechanisms and process in causal construc-
tivist theorizing, Wendt (fn. 7), chap. 2,91-96.
 C O N S T R U C T I V I S T TURN IN IR T H E O R Y 333

CULTURE AND SECURITY

In a curious way, the Katzenstein volume is both very ambitious and

very cautious. The former is seen in its willingness to question, from a
sociological perspective, the very microeconomic disciplinary founda-
tions of IR, and to do so on empirical issues that realists will recognize
as their own. At the same time, Katzenstein and his contributors do not
advance an alternative theory of national security; in contrast to many
of the better edited volumes, this one does not even provide a common
theoretical framework used by all contributors.18
It does offer extraordinarily fresh thinking about security, however,
along with richly detailed case studies. Among the familiar security
questions explored in a new way are the proliferation of conventional
weaponry, the role of deterrence in the nonuse of nuclear and chemical
weapons, the sources of military doctrine, the Soviet cold war endgame,
and alliance dynamics in both the North Atlantic and the Middle East.
Chapter 1 (Katzenstein) and especially chapter 2 (Jepperson,
Wendt, and Katzenstein) should be required reading in any graduate
seminar on security or IR more generally. This is not because Katzen-
stein et al. have decisively trumped mainstream theorists or because
they have everything right. Rather, the essays are extremely helpful in
explaining how the theoretical schools (neorealism, neoliberalism, con-
structivism) differ and why it matters (chap. 1) and for making sense of
a sociological approach to national security (chap. 2). Moreover, these
scholars are interested in dialogue; the goal is not to demonize existing
approaches but to note their limitations.19
The volume's sociological approach to national security involves re-
laxing the two core assumptions of neorealism and neoliberalism,
which are (1) that the environment of states can be conceived solely in
terms of physical capabilities and (2) that institutions and structures
only constrain the behavior of states with fixed interests. Relaxing the
first assumption opens the possibility of social structures being causally
important in world politics, while relaxing the second suggests that the
effects of these structures may reach beyond behavioral constraint to
identity and interest formation. In other words, just like Finnemore,
this is a challenge to the materialism and methodological individualism
that dominate the discourse in mainstream IR (Katzenstein, 16-17).

18
In addition to the edited volume, Katzenstein has published a monograph that makes many sim-
ilar sociological claims. See Peter Katzenstein, Cultural Norms and National Security: Police and Military
in PostwarJapan (Ithaca, N.Y.: Cornell University Press, 1996).
" See also ibid., chap. 2.
334 WORLD POLITICS
Given this stance, the volume needs to address two key questions:
(1) the content and properties of the social structures having such pro-
found effects on agents; and (2) the causal mechanisms through which
these structures have their affects. The social structures doing the ex-
planatory work are norms and, to a lesser extent, culture. The former
are defined as collective expectations about proper behavior for a given
identity (Jepperson, Wendt, and Katzenstein, 54). That this is the same
definition as used by Finnemore and Klotz is one indicator, among oth-
ers, that a constructivist research program is beginning to consolidate
itself in IR.20
The presence of these normative structures is established through a
variety of well-established and standard methodological techniques, for
example, interview data, qualitative content analysis of primary sources,
statistical studies. The research strategy is broadly similar to Finne-
more's: document the presence of the social structures; note a correla-
tion between these and new state interests; examine changing discourse
as further evidence of these normative effects; and,finally,strengthen
the case by considering alternative explanations, usually drawn from
neorealist and neoliberal theories.
Risse-Kappen's chapter on the North Atlantic Treaty Organization
is a good example of the general approach. His puzzle is to explain
NATO's initial formation and endurance—events that are anomalous, he
argues, from the standpoint of both traditional and more sophisticated
realist theories of alliances. In the first part of his essay Risse-Kappen
discusses these likely alternative explanations and carefully documents
their shortcomings.
Next, he develops his own liberal constructivist approach, where the
norms that govern the domestic decision-making process within liberal
systems come to regulate the interactions among democracies in inter-
national institutions such as NATO. Democracies, Risse-Kappen argues,
"externalize their internal norms when cooperating with each other.
Power asymmetries will be mediated by norms of democratic decision-
making among equals emphasizing persuasion, compromise and the
non-use of force or coercive power." He then deduces four different
ways such norms will influence the interaction process among demo-
cratic allies (pp. 268-71).21

20
Definitional congruence in key concepts of a research program is often seen as a sign of its grow-
ing maturity. See Milner (fh. 13).
21
Students of the democratic peace literature will recognize this as a constructivist extension of their
domestic norms argument. See Zeev Maoz and Bruce Russett, "Normative and Structural Causes of
the Democratic Peace," American Political Science Review 87 (September 1993).
 CONSTRUCTIVIST TURN IN IR THEORY 335
Risse-Kappen illustrates the argument by showing how the interests
in play, both in the formation of NATO and during several key crises
(Suez 1956, Cuba 1962), were shaped by the democratic normative
context in which they evolved. In other words, the interests of states
and alliance decision makers (the agents) were being constituted by
these democratic norms (the structures). His evidence is carefully culled
from secondary and, especially, primary sources, for example, the U.S.
government's Foreign Relations of the United States series and materials
in the National Security Archive. This allows him to dissect the deci-
sion-making process, snowing how norms affected the preferences and
interests of various alliance partners.22
The essay by Risse-Kappen is not at all atypical for the Katzenstein
volume, which contains a number of carefully argued studies docu-
menting the impact of norms. Unfortunately, the volume is much
weaker at theorizing the causal mechanisms that give these social struc-
tures such powerful constitutive effects. This is a fair criticism to make,
as the authors clearly commit themselves to a largely causal epistemol-
ogy (Katzenstein, 4-5, 7; Jepperson, Wendt, and Katzenstein, 52-53,
65-68). However, as Katzenstein himself admits in the book's conclud-
ing essay, structural theories such as sociological institutionalism, which
is accorded a central role in the volume, neglect important processes
that translate structural effects (pp. 512—13).23
One result is that the role of agency, while highlighted empirically in
many of the chapters, is neglected theoretically. The volume short-cir-
cuits one loop in the constructivist method: the causal arrows flow pri-
marily from structures to agents. Mutual constitution, however, implies
they also flow from agents to structures. Some constructivists might
object that such sequential (structures to agents then agents to struc-
tures) causal language misconstrues the essence of their ontology: the
simultaneous, mutual constitution of agents and structures. However,
the empirical application of mutual constitution by these scholars fol-
lows precisely the sequential logic outlined here.24
Despite such shortcomings, this is a very important volume. Its com-
mitment to causal analysis and standard methodologies contributes to a
productive dialogue with neorealists and neoliberals; for the most part,
these scholars are all talking the same language. In addition, the case
22
Risse-Kappen has elaborated these arguments in a separate monograph; Risse-Kappen, Coopera-
tion among Democracies: The European Influence on U.S. Foreign Policy (Princeton: Princeton University
Press, 1995).
23
In his own book, Katzenstein pays much greater attention to mapping such processes, although a
lack of explicit theorizing about them is still evident. Katzenstein (fh. 18), chaps. 3-6.
24
See Finnemore's bracketing strategy (p. 25); Wendt (fh. 4), 364-65; and fh. 9 above.
336 WORLD POLITICS
studies (chaps. 3-11) offer new and meaningful insights. For some au-
thors, this means demonstrating that particular security outcomes can
be explained only when realist analyses are supplemented with con-
structivist approaches (Hermans chapter on Soviet foreign policy under
Gorbachev; Risse-Kappen's essay on NATO).
Other contributors, however, go a step further and argue that their
constructivist approach supplants rationalist and materialist accounts.
For example, in a superbly argued essay, Alastair Johnston shows that
the persistence of China's realpolitik over several centuries can be un-
derstood only in terms of a constructivist explanation that subsumes
structural realism (Katzenstein, chap. 7).
Finally, in an innovation rare in any edited volume, Katzenstein has
included an essay (chap. 12 by Kowert and Legro) that reflects critically
on the book as a whole. This excellent chapter provides the sense of
cumulation and summary that is missing when one reads across the
various contributions. It achieves this not by championing the con-
structivist cause but by critically evaluating the volume's shortcomings.
For developing a more coherent constructivist research program, this is
precisely what is needed. Katzenstein et al. are to be applauded for in-
cluding such a chapter.

GLOBAL NORMS AND THE DEMISE OF APARTHEID

The puzzle Audie Klotz seeks to explain is why a large number of in-
ternational organizations and states adopted sanctions against the
Apartheid regime in South Africa despite strategic and economic in-
terests that had fostered strong ties with it in the past. Klotz argues that
the emergence of a global norm of racial equality is at the heart of the
explanation: it led states to redefine interests even though they had ma-
terial incentives not to do so. This demonstrates a constitutive role for
norms, she argues, where they affect state identity and do not simply
regulate behavior (chaps. 1-2).
The case studies on the United States, Britain, and Zimbabwe
(chaps. 6-8) make for especially fascinating reading. Klotz's extensive
empirical research and attention to domestic politics allow her to ex-
plore how this global norm first reached the national level and the ef-
fects it then had on the interests of various groups and individuals. In
contrast to Finnemore and many contributors to the Katzenstein vol-
ume, Klotz offers much more process-level evidence on how norms ac-
tually reconstituted state interests.
The book thus fills in important gaps in both regime theory and
constructivism. Concerning the former, Klotz demonstrates in a nicely
 CONSTRUCTIVIST TURN IN IR THEORY 337
argued section that neoliberal regime analysis shortchanges the role
norms play in international politics. This is not to argue that the neo-
liberals have it all wrong (Klotz does not say this); rather, their view of
norms as constraints on states, as opposed to constituting them, is only
half the story. Empirically, she shows how this theoretical move can ac-
tually be carried out (pp. 13—33). In an important sense, Klotz is em-
piricizing the abstract critiques of regime theory advanced by Friedrich
Kratochwil and John Ruggie over the past decade.25
The author is also to be commended for flagging an issue that con-
structivist research must address. As Klotz puts it: "The crucial ques-
tion is then how a contested norm . . . becomes institutionalized, both
globally and domestically" (pp. 24-25). Indeed, after reading enough of
this work, one senses that there are all too many norms floating around
"out there" that somehow insinuate themselves "in here," that is, in the
domestic arena. (While Finnemore furthers our understanding of how
norm institutionalization works in international institutions, she ne-
glects the question of domestic diffusion mechanisms and processes.)
Klotz addresses this issue by elaborating three transmission mecha-
nisms that link norms and policy choice: community and identity; rep-
utation and communication; and discourse and institutions. While
these are ultimately underspecified (one would want to know what
mechanisms under what conditions are likely to be at work in a partic-
ular national setting), this is nonetheless a foundation upon which
other scholars should build. By elaborating causal mechanisms that
specify diffusion pathways, constructivists will move away from the cor-
relational analyses too often evident in their work; process tracing of
this sort is a method whose time has come for constructivism.
Three weaknesses limit the impact of Klotz's book, however. First,
the ontology is not one of mutual constitution, not even in its bracket-
ing form—comments to the contrary notwithstanding (Klotz, 168-69,
172). Instead, like both Finnemore and many of the Katzenstein case
studies, this is a study of how social structures, a global norm of racial
equality, reconstituted agents.
Second, the analysis is too often correlational (pp. 158-61, for exam-
ple). In part, this results from a failure to specify more systematically
the causal mechanisms operating at the domestic level (Klotz, 24-33).
However, it is also an artifact of the source material, which is primarily
secondary. Given the arguments that Klotz wishes to make about the
25
Friedrich Kratochwil and John Ruggie, "International Organization: A State of the Art on an Art
of the State," International Organization 40 (Autumn 1986); and Friedrich Kratochwil, Rules, Norms
ami Decisions: On the Conditions of Practical and Legal Reasoning in International Relations and Domes-
tic Affairs (Cambridge: Cambridge University Press, 1989).
338 WORLD POLITICS
effects of global norms on various groups within states, archival, mem-
oir, or interview data would seem essential.
For the congressional representatives in her U.S. study, to take one
case, it matters tremendously for the argument whether their views, in
the presence of the global norm, were changing because they feared ad-
verse electoral consequences (the rationalists' means-ends calculations)
or because they learned new values and beliefs (the constructivists' logic
of appropriateness) (Klotz, chap. 6). Klotz's correlations tell us that the
views were changing but not why this occurred; the necessary process
tracing is never fully carried out in the substantive chapters.
Third, the theory-building potential inherent in the book's ambi-
tious cross-national design goes unrealized. Klotz presents no theory
that might predict her results or explain similar dynamics in other
countries, if one wished to extend the study. This is unfortunate: in the
end, one is still left wondering why regimes and norms have such pow-
erful constitutive effects in some states but not in others.26
SUMMARY
Two conclusions follow from the above. Most important, construc-
tivists have convincingly shown the empirical value of their approach,
providing new and meaningful interpretations on a range of issues of
central concern to students of world politics. At the same time, con-
structivist theorizing is in a state of disarray. These researchers, much
like the rational choice scholars they criticize, have made too rapid a
leap from ontology and methods to empirics, to the neglect of theory
development. This matters tremendously. As a central architect of con-
structivism has recently put it: "If parsimony is over-rated as a theoreti-
cal virtue... cumulation is surely underappreciated." And cumulation, it
might be added, if it is to be efficient and productive, requires theory.27

AGENCY, THEORY BUILDING, AND THE

CONSTRUCTIVIST ENTERPRISE
My purpose in this last section is twofold. I begin by highlighting three
issues that should be easy for constructivists to fix. Two, more difficult
26
By "theory" I mean middle-range theory and its development, which should be the goal of prob-
lem-driven empirical research. See, for example, Thomas Risse-Kappen, ed., Bringing Transnational
Relations Back In: Non-State Actors, Domestic Structures and International Institutions (Cambridge: Cam-
bridge University Press, 1995).
27
For the quote, see Wendt (fn. 7), chap. 1,15. A central message of one recent and influential cri-
tique of rational choice is precisely its neglect of theory development, particularly of the middle-range
sort. See Green and Shapiro (fn. 6), 188; and idem, "Pathologies Revisited; Reflections on Our Crit-
ics," in Friedman (fn. 6).
 CONSTRUCTIVIST TURN IN IR THEORY 339
questions are then explored: the role of agency and the need for theory.
Without more sustained attention to agency, these scholars will find
themselves unable to explain where their powerful social structures
(norms) come from in the first place and, equally important, why and
how they change over time. Without theory, especially at the domestic
level, constructivists will not be able to explain in a systematic way how
social construction actually occurs or why it varies cross nationally.
THE THREE EASY FIXES
Constructivists need, first, to pay greater attention to research design.
As noted, much of the empirical work examines single countries or is-
sues. Cross-national or longitudinal designs would help reduce the
problem of overdetermination that is evident in many constructivist
analyses, where social structures, usually norms, are invoked as one of
several causal variables with little or no insight given on how much of
the outcome they explain (Katzenstein, chaps. 4, 8,10; and Klotz, 114,
162, passim). It would also be useful to consider cases when the "dog
doesn't bark," that is, where state identity/interests, in the presence of a
norm, do not change.28
Second, these scholars should give equal attention to the bad things
in world politics that are socially constructed. There is a tendency in the
recent work to consider only ethically good norms, such as those im-
posing a stigma on the use of nuclear or chemical weapons, those that
helped bring the cold war to an end, or the global norms that facilitated
the demise of Apartheid. Some constructivists are aware of this prob-
lem (Finnemore, 6,31—32; Kowert and Legro, in Katzenstein 485—86),
but future work must address it. It will not only protect these scholars
from getting caricatured as peaceniks by theoretical opponents, but it
will also direct their attention to important unexplored issues such as
the role of social construction in ethnic conflict and war.29
28
On the last point, Klotz's cross-national focus is an important step in this direction. For addi-
tional constructivist research utilizing single-country/issue designs, see Ray Koslowski and Friedrich
Kratochwil, "Understanding Change in International Politics: The Soviet Empire's Demise and the
International System," International Organization 48 (Spring 1994); Richard Price, "A Genealogy of
the Chemical Weapons Taboo," International Organization 49 (Winter 1995); Thomas Biersteker and
Cynthia Weber, eds., State Sovereignty as Social Construct (New York: Cambridge University Press,
1996); Bukovansky (fn. 9); Nina Tannenwald, "The Nuclear Taboo: The Normative Basis of Deter-
rence" (Manuscript, Watson Institute for International Studies, Brown University, April 1996); and
Jutta Weldes, "Constructing National Interests," European Journal ofInternational Relations 2 (Sep-
tember 1996).
w
On the last point, see Lars-Erik Cederman, "From Primordialism to Constructivism: The Quest
for Flexible Models of Ethnic Conflict" (Paper presented at the annual meeting of the American Po-
litical Science Association, San Francisco, September 1996). A particularly egregious example of the
caricaturing is Mearsheimer (fh. 7).
340 WORLD POLITICS
Third, constructivists must take greater care in defining key terms,
for example, institutionalization. This word is invoked in nearly every
analysis of norms (Finnemore, 126; Katzenstein, 56, 96-97,129,143,
161,472,484; Klotz, 24-26), but the reader is given no explanation of
what the process entails. In what institutions—or individuals—do
norms reside? Must norms be internalized first by individuals through
a socialization and learning process? If so, constructivists should pay
greater attention to developing the often implicit cognitive models in
their analyses. Or, does institutionalization occur at a higher level of ag-
gregation, through bureaucratic and legal processes that affect society
as a whole. If this is the level under examination, constructivists could
benefit from the insights of historical institutionalists and of those in
the ideas literature who have studied such dynamics.30
T H E CHALLENGES AHEAD
Ontology and theory building are the central challenges for construc-
tivists.
BRINGING AGENCY BACK IN
This move is necessary if mutual constitution is to be taken seriously as
a way of thinking about the social world. I appreciate the reasoning of
some that a neglect of agency is legitimate, at present, as a corrective to
the extreme agent orientation of most mainstream IR (Finnemore,
chap. 1). Moreover, it has proved very difficult to apply mutual consti-
tution in empirical research.
All the same, constructivists should want to avoid the charge that
they are reducing one unit of analysis—agents (states, decision mak-
ers)—to the other—structures (norms). One result of this reduction is
a failure to explore how norms arise in the first place (and the role of
agency and power in this process), and how, through interactions with
particular agents, norms change over time.31
An example clarifies the importance of the last point. Post-cold war
Europe has witnessed the emergence of norms advancing more inclu-
sive conceptions of national membership (citizenship laws, rights of na-
30
See F r a n k L o n g s t r e t h e t al., Structuring Politics: Historical Institutionalism in Comparative Analy-
sis ( N e w Y o r k C a m b r i d g e University Press, 1992); and J u d i t h Goldstein, Ideas, Interests and American
Trade Policy (Ithaca, N.Y.: Cornell University Press, 1993). N o t surprisingly, it is Katzenstein, the c o m -
parativist, who has offered the most careful constructivist account of domestic norm institutionaliza-
tion. See Katzenstein (fn. 18), chaps. 1-3,5, 7.
31
Dessler's transformative model of international structure should be especially relevant to con-
structivists as they rethink the role of agency in their analyses. See David Dessler, "What's at Stake in
the Agent-Structure Debate?" International Organization 43 (Summer 1989).
 CONSTRUCTIVIST TURN IN IR THEORY 341
tional minorities). Promoted initially by nongovernmental actors and
more recently by the Strasbourg-based Council of Europe, the content
of these norms has now been modified significantly as a result of Rus-
sia's instrumental exploitation of them in a bid to reassert its dominance
among the former Soviet states. The constructivists' normative struc-
tures are themselves being reshaped by the activities of purposeful
agents.32
Three reasons explain why agency has fallen through the ontological
cracks for constructivists. First, many constructivists rely upon the in-
sights of sociological institutionalism for their thinking about the so-
cial world. Those insights, however, are based upon a particular branch
of organization theory that systematically excludes questions of agency,
interest, and power.33
Second, because of their focus on collectively held, intersubjective
understandings (norms), most constructivists, not surprisingly, are less
interested in questions of individual agency. Yet the evolutionary devel-
opment of norms suggests that, at some early point in their life histo-
ries, they may not be collective in any meaningful sense; particular
individuals (Finnemore's moral entrepreneurs, for example) may play
key roles at early stages. Thus, social construction at the level of agents
is—or rather, should be—a relevant concern for these scholars.34
Finally, Wendt, who has been so influential in developing construc-
tivism, has explicitly bracketed individual agency as a factor to be ex-
plained by mutual constitution. For Wendt, a key distinction is between
the corporate and social identity of states, with the former deempha-
sized because "its roots [are] in domestic politics." Since he assumes a
unitary state, corporate identity includes and subsumes that of the in-
dividual. The result is that social construction at the level of individual
agents or, more generally, at any domestic level is neglected. While sev-
eral theorists have criticized Wendt for this stance, no clear under-
standing of how to rectify it has emerged.35
32
Checkel, "Norms, Institutions and National Identity in Contemporary Europe" (Manuscript, Oc-
tober 1997).
33
See Paul DiMaggio, "Interest and Agency in Institutional Theory," in Lynne Zucker, ed., Institu-
tional Patterns and Organizations: Culture and Environment (Cambridge: Ballinger Publishing, 1988);
Paul DiMaggio and Walter Powell, eds., Tie New Institutionalism in OrganizationalAnalysis (Chicago:
University of Chicago Press, 1991), chaps. 1,4; Frank Dobbin, "Cultural Models of Organization: The
Social Construction of Rational Organizing Principles," in Diana Crane, ed., The Sociology of Culture:
Emerging Theoretical Perspectives (Cambridge: Blackwell Publishers, 1994).
34
See also the discussion of norm reproduction in Florini (fn. 16), 374-75,377-80.
35
Alexander Wendt, "Collective Identity Formation and the International State," American Political
Science Review 88 (June 1994); and, for the quote, idem, "Identity and Structural Change in Interna-
tional Politics," in Friedrich Kratochwil and Yosef Lapid, eds., The Return of Culture and Identity in
342 WORLD POLITICS
It is ironic that constructivists therefore find themselves in a predica-
ment all too familiar to rational choice scholars: their ontology has led
them to neglect key issues. The agent-centered approach of rational
choice provides a clear perspective on the microfoundations of human
behavior, but much less clarity on how this connects with the broader
institutional and social context. The dilemma then is how to get from
microfoundations to outcomes.36
Constructivists, despite their arguments about mutually constituting
agents and structures, have advanced a structure-centered approach in
their empirical work. Moreover, Wendt's theoretical stance has led to a
neglect of domestic agency. The result is that constructivism, while
good at the macrofoundations of behavior and identity (norms, social
context), is very weak on the microlevel. It fails to explore systemati-
cally how norms connect with agents.37
SOCIAL CONSTRUCTION AND THEORY BUILDING
To explore such connections, constructivists will need to engage in the-
ory development. At present, constructivism is, like rational choice,
nothing more than a method. It leads one to ask certain questions and
make certain assumptions. However, constructivists should surely want
more. In fact, in the volumes under review, there is a persistent call pre-
cisely for greater specification of constructivism (Finnemore, 130-31;
Kowert and Legro, in Katzenstein, 469-83; Klotz, 26-33).38
The missing element is substantive, middle-range theory, which
would provide constructivists with a set (or better, competing sets) of
research questions and hypotheses that could be tested in various cross-
national and longitudinal studies. The need for theory is especially ev-
ident at the domestic level, where the constructivist "norm" is empirical
ad hocism with all sorts of implicit models of domestic politics and key
actors being invoked.39
IR Theory (London: Lynne Rienner, 1996), 50-51. For critiques, see Sujuta Pasic, "Culturing Interna-
tional Relations Theory: A Call for Extension," in Kratochwil and Lapid, 87-90; and Cederman (fh.
29), 13-19.
36
Rational choice instirutionalism represents an effort to address this dilemma. See Norman Scho-
field, "Rational Choice and Political Economy," in Friedman (fn. 6), 192-93,207-8; and Peter Hall
and Rosemary Taylor, "Political Science and the Three New Institutionalisms," Political Studies AA
(December 1996), 958-62.
37
On the micro versus the macrofoundations of behavior and identity and the tensions between the two,
see "Symposium: The Role ofTheory in Comparative Politics," World Politics 48 (October 1995), 13—15.
38
After earlier confusion, Wendt also now argues that constructivism is not a theory. Wendt (fh. 7),
chap. 1.
39
All the books reviewed are strongest, theoretically, at the systems level, in large part because they
draw upon an already well developed sociological literature that is systemic in orientation. See Martha
Finnemore, "Norms, Culture and World Politics: Insights from Sociology's Institutionalism," Interna-
tional Organization 50 (Spring 1996).
 CONSTRUCTIVIST TURN IN IR THEORY 343
If constructivists are to theorize at the domestic level, they will need
to address three issues. How deep within a policy does one need to go
with a constructivist analysis? How is such an analysis actually carried
out? Under what conditions is a constructivist approach, as opposed to a
rationalist one, even necessary to explain the effects of social structures?
Progress on the first issue requires specification of political actors,
that is, some model of domestic politics within the state. There are all
sorts of domestic frameworks available (pluralist, institutional, and so
on), but these are well known and need not be discussed. Rather, I wish
to suggest that constructivists have already uncovered abundant evi-
dence that the state-level penetration of international social structures
varies cross nationally. The how deep question clearly matters.40
A few examples will clarify the point. Although Finnemore is not
explicit on this score, one can infer from her empirical chapters that
normative effects are limited to state bureaucrats (Finnemore, chaps. 2,
4). In the Katzenstein volume, some authors find norms held broadly
within a polity (Berger on postwar Germany and Japan), while others
see their effects confined to political and academic elites (Herman on
the USSR) or to state decision makers (Risse-Kappen on NATO;
Katzenstein, chaps. 9, 8,10). Klotz's cross-national design uncovers ev-
idence of normative effects at the level of political elites in one instance
(the U.S.); in her British case, however, such influences are partly
blocked by deeper, historically constructed national discourses (Klotz,
chaps. 6, 7).
To make sense of and explain such diversity, constructivists will need
to theorize the varying processes through which social construction oc-
curs. The insights gained from Klotz s partial move in this direction in-
dicate its importance. Here, constructivism would benefit greatly from
utilizing methods developed by IR scholars seeking to place greater em-
phasis on process.41
Having established that social construction occurs at various levels
within the state, the second question can be addressed: how does one
conduct such an analysis? For present purposes, assume three domestic
levels: society, state institutions, and individual decision makers. Fur-
thermore, due to space limitations, consider only the individual level.
What does it mean to explore the social construction of individual de-
cision makers? Theoretically, it is to explore how social structures inter-
* Milner's (fh. 13) advice to mainstream IR theorists on how to conceptualize domestic politics is
relevant here as well.
11
Peter Evans, ed., Double-Edged Diplomacy: International Bargaining and Domestic Politics (Berke-
ley: University of California Press, 1993); Haas (fh. 13); Sikkink (fn. 13); and Risse-Kappen (fh. 26),
among others.
344 WORLD POLITICS
act with and fundamentally affect the identities of these agents, how
certain logics of appropriateness come to govern their behavior.
For constructivists, this means being able to explain how the inter-
ests and identities of particular agents, in the presence of norms,
change—or, equally important, do not change over time. Despite its
centrality, this issue, which directly addresses the cognitive microfoun-
dations of constructivism, has not received the attention it should, es-
pecially in the empirical literature. However, a review of this work
suggests three possibilities.
One is a learning argument drawn from cognitive psychology. Just
such a dynamic is implicit in Finnemore's book, where agents (state
elites), through exposure to norms, are taught new identities and inter-
ests. Because interests are changing, one can infer that this is a con-
structivist claim about complex, rather than simple, learning. (In the
latter, new information allows actors to pursue given interests more ef-
fectively; it can be accommodated within a rationalist framework.)42
The problem for such arguments is that when one introduces the re-
ality and friction of domestic politics, complex learning typically breaks
down. Absent such processes, one is back in the rationalists' world of
simple learning. This politics-learning tension is well established both
theoretically and empirically, with the basic insight being that learning
becomes less likely as the circle of actors grows.43
Symbolic interactionist theory in sociology provides a second possi-
ble way to probe these constructivist microfoundations. Here, individ-
ual identities and interests are formed through a process of interaction,
with two mechanisms being key: imitation and social learning. Since
imitation does not involve interaction (and, thus, mutual constitution), it
is the social learning dynamic that plays a more central role in the con-
structivist accounts. Social learning, much like the cognitive/individual
sort just discussed, can be simple or complex, but given the constructivist
emphasis on identity change, the focus is again on the latter. Specifically,
complex social learning occurs when identities and interests are learned
in response to how actors are treated by significant others.44
42
Personal communication, Martha Finnemore, September 1996. See also Thomas Risse-Kappen,
"Democratic Peace—Warlike Democracies? A Social Constructivist Interpretation of the Liberal Ar-
gument," European Journal ofInternational Relations 1 (December 1995). On the learning literature
more generally, see Jack Levy, "Learning and Foreign Policy: Sweeping a Conceptual Minefield (Re-
view Article)," International Organization 48 (Spring 1994).
Ai
On the learning theory—politics connection, see Richard Anderson, "Why Competitive Politics
Inhibits Learning in Soviet Foreign Policy," in George Breslauer and Philip Tedock, eds., Learning in
U.S. and Soviet Foreign Policy (Boulder, Colo.: Westview Press, 1991).
44
See Wendt (fn. 7), chap. 7; and idem, "Anarchy Is What States Make of It: The Social Construc-
tion of Power Politics," International Organization 46 (Spring 1992). Symbolic interactionist theory
 CONSTRUCTIVIST TURN IN IR THEORY 345
While intriguing, this line of reasoning has not yet been integrated
with empirical research. When and if this occurs, the same problem as dis-
cussed above will confront constructivists: how to maintain complex learn-
ing in settings where the static created by domestic politics hinders it.
Social psychology provides a final possible tool for exploring social
construction at the individual level. Here, the theoretical foundations
are provided by Turner's self-categorization theory, where the focus is
on individual-group interactions. For constructivists, the key process in
Turner's work is depersonalization, for this is how individual identities
and interests change through interaction with a larger social group.45
Unfortunately, this process is so context dependent and unclear (does
depersonalization occur through social learning? through coercion?), it
is not at all certain how constructivists might integrate its insights into
their work. Nor surprisingly, when these scholars have used variants of
self-categorization theory, it has led to unresolved theoretical disputes,
as well as to sloppy empirical work.46
The criticisms and questions raised above should not be viewed as dis-
missive. In addressing an issue of central importance—how to connect
social structures to agents—these scholars are building much-needed
bridges to other literatures. In fairness to constructivists, scholars of ra-
tional choice have been harshly criticized in similar ways for their at-
tempts at the reverse process: connecting their sparse microfoundations
to broader social and normative structures.47
These last comments lead directly to the third question construc-
tivism needs to address more systematically at the domestic level: when
is such an approach, as opposed to a rationalist one, even necessary to
explain the effects of social structures? Because most of the construc-
tivist work to date has been method driven, these scholars have failed to
appreciate that the domestic effects of norms are at times best captured
and explained by rational choice.48
Klotz's U.S. study, for example, suggests that global norms were not
so much transforming the identities of congressional representatives as

has been developed primarily at the individual level, which is why I discuss it here. Wendt, uncon-
vincingly in my view, argues that it can be applied at the level of (unitary) states as well.
<s
See John Turner, Rediscovering the Social Group: A Self-Categorization Theory (Oxford: Basil Black-
well, 1987), chap. 3; and Penelope Oakes et al., eds., Stereotyping and Social Reality (Oxford: Blackwell,
1994), chaps. 1,4.
* On the former, compare Jonathan Mercer, "Anarchy and Identity," International Organization 49
(Spring 1995); and Wendt (fn. 7), chap 7. For the sloppy empirical work, see Glenn Chafetz, "The Po-
litical Psychology of the Nuclear Nonproliferation Regime," Journal of Politics 57 (August 1995).
47
For example, Robert Lane, "What Rational Choice Explains," in Friedman (fn. 6).
48
For details, see Checkel, "International Norms and Domestic Politics: Bridging the Rationalist-
Constructivist Divide," European Journal of International Relations 3 (December 1997).
346 WORLD POLITICS
creating constraints on their behavior (Klotz, chap. 6). In other words, one
is back in the rationalist s world of means-ends calculations (in this in-
stance, a political survival calculus of how best to secure reelection). Now,
Klotz, as well as many contributors to the Katzenstein volume, does rec-
ognize that norms can have instrumental effects such as these. Nonethe-
less, one would want clear indicators of when one dynamic or the other is
likely to prevail. The challenge, then, is to develop scope conditions.49
One is temporal. This is the division-of-labor argument briefly men-
tioned in the Katzenstein volume (Jepperson, Wendt, and Katzenstein,
70; Kowert and Legro, 490-91). Constructivism might be best at ex-
plaining identity and interest formation, but as some later time, when
interests were stable, rationalism might be the right method. Such a so-
lution would have the benefit of making everyone happy: there would
be a legitimate place and time for all approaches. However, the devil
is in the details. Empirically, how does one know a priori when a state
is likely to be in a period of identity formation, where constructivism is
appropriate, as opposed to a time when identities and interests are al-
ready fixed?
A second scope condition is a density-of-interactions argument,
which has been applied primarily to international bargaining. At some
stage in this process, actors may switch from the rationalists' conse-
quential, means-ends logic to a situation in which their preferences are
in genuine flux and open to change through persuasion and communi-
cation. However, the key question is how one predicts such a switch.
What needs to happen and when? Cognitive uncertainty by individual
negotiators? The establishment, through communication and speech,
of some level of collective trust among them? Lacking this specifica-
tion, the same problems arise as with the division-of-labor argument.50
A final scope condition explores the role of domestic institutions.
"Institution," in this case, refers to the bureaucracies, organizations, and
groups that channel and define policy-making within states. In the
three books under review, one sees two very different normative effects
at the domestic level. In some instances, decision makers and elites are
49
For other constructivist accounts portraying similar rationalist logics, see Price and Tannenwald,
in Katzenstein, 138,148-50; and Bukovansky (fn. 9), 21-51. Very similar questions of scope and do-
main are now being asked by several rational choice analysts. See the discussion of "segmented univer-
salism," in Green and Shapiro (fn. 6), 192—93, 204; Michael Taylor, "When Rationality Fails," in
Friedman (fn. 6), 230-33; and Powell (fn. 1), 324.
50
Thomas Risse, "The Cold War's Endgame and German Unification" (A Review Essay), Interna-
tional Security 21 (Spring 1997). This constructivist conception of communication thus extends well
beyond the rationalists' "cheap talk." For an excellent discussion, see James Johnson, "Is Talk Really
Cheap: Prompting Conversation between Critical Theory and Rational Choice," American Political
Science Review 87 (March 1993).
 CONSTRUCTIVIST TURN IN IR THEORY 347
essentially taught (Finnemore) or learn (Herman, in Katzenstein, chap.
8) new beliefs and values in the absence of any obvious domestic pres-
sures; that is, new (constructivist) logics of appropriateness come to
govern their behavior. At other times, norms do not have individual ef-
fects; instead, they mobilize domestic groups that pressure elites to
change policy in ways consistent with the norms (for example, Klotz,
chap. 6). That is, normative effects are operating through (rationalist)
means-ends calculations.
Perhaps this variation is explained and predicted by differences in
political institutions across states. In liberal polities such as the U.S.,
where decision makers have little autonomy from societal groups, the
rationalists' instrumental logic more often captures the domestic effect
of systemic social structures. In states with greater autonomy and insu-
lation from society (say, the former USSR), constructivist logics may
more often capture the unit-level affects of norms.51

CONCLUSIONS

An IBM ad in a recent issue of the Economist shows a well-heeled exec-

utive holding his head and shaking it in despair: "Oh no, another para-
digm shift," he laments. The good news for IR theorists is they face no
such threat from the constructivists reviewed in this essay. However,
this attests not to their failures but to the nature of their goals: dialogue,
a widening of disciplinary foundations, and a commitment to causal
analysis. These scholars are out not to colonize and deconstruct IR but
to revitalize and expand its conceptual lenses.
That one can make so many critical observations about this work
suggests, paradoxically, its achievements. The publication of the books
discussed here, along with the work of scholars such as Wendt, Ruggie,
and Kratochwil, has for the first time given constructivism a critical
mass of research that is both theoretical and empirical. This allows a re-
viewer to probe for lacunae and tensions, as well as synergies in it.
At this point, instead of summarizing, a broader issue needs to be
raised: what kind of constructivism do we want? Some constructivists
might feel this review "mainstreams" them too much. The criticisms on
research design, better specification of key terms, developing middle-

51
For a full theoretical elaboration, see Checkel, "Between Norms and Power: Identity Politics in
the New Europe" (Book manuscript in progress), chap. 2. Recent work on the role of international
norms in U.S. policy-making is consistent with the argument made here. See Andrew Cortell and
James Davis, "How Do International Institutions Matter? The Domestic Impact of International
Rules and Norms," International Studies Quarterly 40 (December 1996).
348 WORLD POLITICS
range theory, taking domestic politics and agency seriously, after all,
sound like a primer for building a more coherent research program.
There are two reasons for constructivists to move in this direction.
First, judging by many comments to this effect, it is the direction in
which they wish to move. Their emphasis on dialogue and causal analy-
sis suggests a fairly standard concern with building a rigorous and co-
herent body of research that speaks to and plays off other literatures
within IR.
Second, in its present form, it is not clear what one does with con-
structivism. How could Finnemore's insights be applied to other inter-
national institutions—NATO, for example? Why do the transnational
norms, which figure so prominently in Klotz's study, have seemingly no
impact in contemporary China? Answers to such puzzles will come only
when constructivists specify more clearly the actors—structures and
agents—and causal mechanisms bringing about change, the scope condi-
tions under which they operate and how they vary cross nationally. Ab-
sent this theorizing, the "what do we do with it" question will remain.