Constitution, Choice and Change: Problems with the `Logic of Appropriateness' and its Use in Constructivist Theory

Abstract

The debate between a moderate version of constructivist theory and rationalist theory centres primarily on the rationality of individual action. The article consists of an in-depth analysis of the `logic of appropriateness' (LoA) invoked in constructivist theory. The analysis reveals that the LoA is a structural explanation and understanding of individual action. As such, it is untenable as a theory of individual action. The implications of this structural bias are discussed in relation to three core claims of constructivist theory. Moderate constructivist theory claims, first, that norms are constitutive for actors' identities. Second, it claims that agents and structures are mutually constitutive. Third, it claims that changes in ideational structures do occur and lead to changes in political practice. I conclude that the LoA is able to account for the first of these claims, but that by virtue of being able to account for this claim, it is, at the level of a theory of individual action, inconsistent with the second, and unable to effectively account for the third.

Full text

Constitution, Choice and Change:
Problems with the ‘Logic of
Appropriateness’ and its Use in
Constructivist Theory
OLE JACOB SENDING
Norwegian Institute of International Affairs
The debate between a moderate version of constructivist theory and
rationalist theory centres primarily on the rationality of individual
action. The article consists of an in-depth analysis of the ‘logic of
appropriateness’ (LoA) invoked in constructivist theory. The analysis
reveals that the LoA is a structural explanation and understanding of
individual action. As such, it is untenable as a theory of individual
action. The implications of this structural bias are discussed in relation
to three core claims of constructivist theory. Moderate constructivist
theory claims, first, that norms are constitutive for actors’ identities.
Second, it claims that agents and structures are mutually constitutive.
Third, it claims that changes in ideational structures do occur and lead
to changes in political practice. I conclude that the LoA is able to
account for the first of these claims, but that by virtue of being able to
account for this claim, it is, at the level of a theory of individual action,
inconsistent with the second, and unable to effectively account for the
third.
K
EY
W
ORDS
♦choice ♦constructivism ♦identity ♦‘logic of
appropriateness’ ♦‘logic of consequences’ ♦norms ♦rationalism ♦
theory of individual action
Introduction
There is currently a significant debate in International Relations theory
between a specific moderate version of constructivist theory and rationalist
theory that centres on the issue of what logic of action best accounts for
European Journal of International Relations Copyright © 2002
SAGE Publications and ECPR, Vol. 8(4): 443–470
(1354–0661 [200212] 8:4; 443–470; 029173)

international political processes (see Checkel, 1999, 2001; Finnemore and
Sikkink, 1998; Kahler, 1998; Schimmelfennig, 2000, 2001).1Indeed, in a
recent article Risse (2000) suggests that the central dividing line between
rationalist and constructivist or sociological theories is exactly that they
bring to bear different conceptions of the rationality of action (p. 3).
In this debate about the explanatory power and empirical scope of
different logics of action, constructivist theorists typically invoke the ‘logic
of appropriateness’ (LoA) developed by March and Olsen (1989, 1995)
(see, for example, Checkel, 1998; Finnemore, 1996; Finnemore and
Sikkink, 1998; Katzenstein, 1996).2Rationalist theory, by contrast, is based
on rational choice, or what March and Olsen call a ‘logic of consequences’
(see Cortell and Davis, 1996; Keohane, 1988; Moravcsik, 1998, 1999;
Schimmelfennig, 2001). It is indicative of the prominence of the LoA in
constructivist theory that when Risse (2000) seeks to show that another
logic of action is also consistent with constructivism, he does so by asserting
that his Habermas-based ‘logic of arguing’ operates within the framework of
the ‘logic of appropriateness’ (LoA). Noting that the LoA is central to
constructivist theory since it accounts for the normative rationality of action
(2000: 4–5), Risse asserts that within the LoA, actors must determine what
kinds of rules and norms must be followed to act appropriately. He then
asks: ‘But how do actors adjudicate which norm applies? They argue. I
suggest, therefore, that social constructivism encompasses not only the logic
of appropriateness but also what we could call a “logic of truth seeking or
arguing” . . .’ (2000: 6; emphasis added).
This debate between rationalist and constructivist theory about the
rationality of individual action rests on certain epistemological and onto-
logical assumptions that are subject to criticism from other versions of
constructivist theory.3These are arguably more genuinely constructivist in
orientation as they address epistemological issues that are not in focus in this
debate about different logics of action. Nevertheless, the debate about the
rationality of individual action represents an important and constructive
dialogue between two different theoretical perspectives that seem to
constitute a central vehicle by which new theoretical and empirical insights
about international politics can be produced. In this article, I shall accept the
terms of this debate for the purpose of analysing in more detail the
theoretical constructs and assumptions that are contained in the LoA.
Rational choice theory has been subject to much discussion, both in IR
theory and in the social sciences in general (see Elster, 1989; Hollis, 1987;
Hollis and Smith, 1990). This has not been done in the case of the LoA.
This seems important, as most constructivist theorists in this debate have
accepted the validity of the LoA as their action-theoretical foundation
without an in-depth exploration of its core theoretical building blocks.
European Journal of International Relations 8(4)
444

My purpose is twofold. First, I discuss whether the LoA is tenable as a
theory of individual action. After a somewhat extended argument, I show
that — contrary to its claims — the LoA has a structural bias both regarding
the understanding and the explanation of individual action. As such, it is
untenable as a theory of individual action. Second, I discuss whether the
LoA is consistent with moderate constructivist theory. I identify three
defining claims of moderate constructivist theory. Moderate constructivist
theory claims, first, that norms are constitutive for actors’ identities. Second,
it claims that agents and structure are mutually constitutive. Third, it claims
that changes in ideational or normative structures do take place and lead to
changes in political practice. Based on an analysis of these claims, I hold that
the LoA is able to account for the first claim, but that by virtue of this, it is,
at the level of a theory of individual action, inconsistent with the second, and
unable to effectively account for the third.
The Theoretical Context of the Logic of Appropriateness:
Communitarianism
As a first step, we have to grasp the theoretical context in which the LoA is
formulated. It is based on studies of modern, formal organizations. A central
insight from these studies is that individual behaviour within organizations is
structured according to the rules, routines and standard operating proce-
dures that define the objectives and functions of these organizations. March
and Olsen apply this insight to political life in general by noting that ‘It can
be extended to the institutions of politics. Much of the behaviour we
observe in political institutions reflects the routine way in which people do
what they are supposed to do’ (1989: 21).
When these administrative-organizational insights are transported to
political life, March and Olsen locate the source of what defines the standard
of appropriateness in something other than functionally ordered and task-
specific rules, as is the case in modern, formal organizations. As they do so,
they invoke a particular version of a communitarian perspective. Rather than
conceptualizing the political community as based on the convergence of
inherently individual interests, March and Olsen invoke the ideal of ‘civic
identity’:
The civic identity ideal presumes, on the other hand, that action is rule-based,
that it involves matching the obligations of an identity to a situation. Pursuit
of the common good is not so much a personal value as a constitutive part of
democratic political identities and the construction of a meaningful person. The
community is created by its rules, not by its intentions. (March and Olsen,
1995: 38; emphasis added)4
Sending: Constitution, Choice and Change
445

This paragraph contains two central arguments, both of which reflect a
communitarian reading of political life. One is that the pursuit of the
common good is constitutive for a democratic political identity and for the
construction of a meaningful person. The other is that the political
community is created by its rules and not by its intentions. Addressing each
in turn, we will see how this particular idea of the political community is
central to the LoA.
The common good in question must, if its pursuit is to be constitutive for
a democratic political identity, be of a kind that enters into the definition of
what a democracy is seen to be. This type of common good is what Charles
Taylor (1995) refers to as ‘irreducibly social goods’. Such goods are goods
by virtue of being common. Hence, Taylor notes about these goods that ‘it
is essential to its being a good that its goodness be the object of common
understanding’ (1995: 139). This (irreducibly) common good is for March
and Olsen the belonging to, and the participation in, a political community.
This is seen as a necessary requirement for a well-functioning democracy. It
is within this line of thought that March and Olsen note that an individual
who is unable to achieve what they call ‘constitutive attachments’ must be
characterized as a ‘nonperson’, or a person without ‘moral depth’ (March
and Olsen, 1995: 37). The common good, then, is to be part of or belong
to a democratic polity, and the pursuit of that common good is constitutive
for our identity as democratic citizens and for the democratic polity.
The second issue concerns what kinds of rules March and Olsen refer to
when they say that the political community is ‘created by its rules, not by its
intentions’. This seems to be based on Searle’s (1995) distinction between
regulative and constitutive rules. Regulative rules are rules that regulate
‘antecedently existing activities’, whereas constitutive rules ‘create the very
possibility of certain activities’ (1995: 27). An example of the former is the
rule of driving on the right-hand side of the road (driving exists independ-
ently of this rule), whereas the latter is found in the case of the rules of chess
(chess is created, constituted, by these rules) (see also Ruggie, 1998: 24).
Constitutive rules, then, define what a social practice is.5It seems to be this
line of thought that lurks behind March and Olsen’s assertion that:
The core notion is that life is organized by sets of shared meanings and
practices that come to be taken as given. Political actors act and organize
themselves in accordance with rules and practices that are socially constructed,
publicly known, anticipated and accepted. Actions of individuals and collectiv-
ities occur within these shared meanings and practices, which can be called
identities and institutions. . . . (March and Olsen, 1995: 30)
Rules here take on an unmistakable hermeneutical dimension: any social
practice is constituted by its rules because others’ actions would be
unintelligible without them being in accordance with public and inter-
European Journal of International Relations 8(4)
446

subjectively shared rules. This hermeneutical aspect of rules can be clearly
identified in post-Wittgensteinian philosophy, as found in Peter Winch’s The
Idea of Social Science (1963), where rules designate ‘forms of life’. As we will
see later, the implications following from this line of thought are directly
connected to the idea that actions are directed towards being appropriate,
since standards of appropriateness cannot be established apart from the
institutional context in which these actions take place, the meaning of which
is defined by certain constitutive rules.6
It is against the background of this communitarian reading of the political
community, expressed through the irreducibly common feature of social
goods, and the constitutive aspect of institutional rules, that institutions
emerge as central in defining the cognitive and normative outlook of a
political community in which actors operate. It is thus through inquiring
into how institutions constitute actors’ identities and specify rules and norms
for appropriate actions, so the arguments goes, that we will grasp the
essential features of political processes. And it is in this setting that the LoA
makes sense as it identifies a contextual or normative rationality of political
action.
The Logic of Appropriateness
For March and Olsen, the LoA is meant to capture something fundamental
about public and civic action, i.e. actions within institutions that have
bearing upon the organization and direction of a polity (March and Olsen,
1995: 251–2). The LoA portrays political action as ‘obligatory action’ and
as being rule- and identity-based. According to March and Olsen, it is
equally individualistic in nature as the logic of consequences (LoC) inherent
in rational choice theory (March and Olsen, 1998: 951). As expressed in a
practical syllogism, the LoA takes the following form (March and Olsen,
1989: 23):
‘Obligatory Action’
1. What kind of situation is this?
2. Who am I?
3. How appropriate are different actions for me in this situation?
4. Do what is most appropriate.
The LoA thus comprises three key elements: (1) situation, (2) identity or
role, and (3) rules. A decision to act in a certain way follows from, is
explained by, 1, 2 and 3. I address each element in turn.
Sending: Constitution, Choice and Change
447

Situation
The relevance of situations in the LoA is related to its function as a bridge
between the moment of action and its institutional setting. Since institutions
are independent variables in a new-institutional perspective, institutions
define the setting or context of action. That is, institutions create order
through organizing principles and define certain ways of doing things
through rules (March and Olsen, 1989: 24). But within institutions, a whole
range of rules and norms exists, and, consequently, a reference to institutions
as the setting for action becomes too vague and unspecified. It is here that
situations come in.
The interpretation of situations allows for individual differences in action
as situations can be interpreted differently. March and Olsen portray the
process of interpretation as one that entails complex reasoning, uses of
knowledge and experience (1989: 30–1). This is an important point as it
directs attention to the ways in which individual actors with similar identities
or roles may act differently by applying different rules, because they interpret
and understand the situation differently (1989: 24). If one actor defines a
situation as a crisis, he or she will apply and follow other rules when acting
than an actor who defines the same situation as uncritical and normal.
March and Olsen hold that individual actors receive the primary tools
through which situations and rules are interpreted from the institutions in
which they are operating. A central objective of the new institutional
framework, and for constructivist theory in IR, is to make endogenous the
formation and changes of the interests that actors pursue (1989: 40–1). A
key argument for the centrality of institutions in political life is thus that they
organize and structure meaning and cognition for individual actors, thus
producing similar behaviour from dissimilar actors due to the central role
of institutions in shaping their identities and the tools by which they inter-
pret situations. Invoking an explicit communitarian conception, they note
that:
The political community is based on a shared history, a valued way of life, a
shared definition of the common good, and a shared interpretation and
common understanding embodied in rules for appropriate behavior. The rules
provide criteria for what is worth striving for, and for what is accounted as
good reasons for action. (March and Olsen, 1989: 161)
In commenting upon the centrality of institutions in politics, March and
Olsen consequently note that ‘Rules and understandings frame thought,
shape behavior, and constrain interpretation’ (1995: 31). We shall return to
this issue later in discussing the understanding dimension of the LoA.
European Journal of International Relations 8(4)
448

Identity
An identity denotes something more fundamental than a role. March and
Olsen use the terms interchangeably, but emphasize mostly the concept of
identity.7We saw earlier that an actor asks ‘Who am I?’ This is not a question
concerning one’s private reflections on the self. It is rather a question
concerned with the official and professional identity of an actor. The identity
that March and Olsen talk about is institutionally defined (democratic
citizen, civil servant, prime minister, priest, etc.). The communitarian
understanding of the political community is here transported directly to the
LoA via the conception that institutionally defined identities account for the
motivation for action specified through the concepts of duties and obliga-
tions (March and Olsen, 1989: 23, 59).
By determining an identity, the actor slips into a framework in which the
duties and obligations of that identity provide the motivations for appro-
priate action, the performance of which comes through the application of a
rule that specifies what is appropriate in the particular situation. This is
captured in the following statement:
In a logic of appropriateness . . . behaviors (beliefs as well as actions) are
intentional but not willful. They involve fulfilling the obligations of a role in a
situation, and so of trying to determine the imperatives of holding a position.
Actions stem from a conception of necessity, rather than preference. Within a
logic of appropriateness, a sane person is one who is ‘in touch with identity’ in
the sense of maintaining consistency between behavior and a conception of self
in a social role. (March and Olsen, 1989: 160–1; emphasis added)
The underlying idea here is that the self becomes social through acquiring
and fulfilling institutional identities. An institutionally defined identity thus
demands certain actions in order for that identity to be fulfilled and
maintained. When actors act according to appropriate rules, they do not do
so because there are external sanctions that compel them to do so, nor is it
some perception of self-interest. Rather, it is because the individual actor has
internalized the duties and obligations that define an institutional identity
that the actor acts as he or she does. Indeed, failure to meet the duties and
obligations is here understood as a non-option:
Anyone incapable of achieving an identity based on constitutive attachments
— if such a person could be imagined — should be described not as a free and
rational agent but as being without character and moral depth, a nonperson.
(March and Olsen, 1995: 37)
Rules
Rules perform the function of storing information about institutional
practices, routines and norms that tell the actor what to do in a specific
Sending: Constitution, Choice and Change
449

situation in order to behave as his or her institutional identity demands. An
action involves ‘evoking an identity or role and matching the obligations of
that identity or role to a specific situation’ (March and Olsen, 1998: 951). In
‘matching the obligations’ of an identity to a specific situation, rules specify
that match by encoding institutionally defined appropriate actions.
Rules contain a repertoire of appropriate actions fitted to a situation by an
actor. In this sense, each actor decides to pursue a certain line of action by
interpreting a situation, determining one’s identity and then searching for a
rule that is subsequently followed. Within this logic of action, actors act not
with reference to self-interests as defined by expected consequences of
action, as in a logic of consequences. Rather, actors act in accordance with
institutional rules that define what is appropriate in different situations,
given one’s identity. A rule thus specifies what action to perform in order for
it to conform to institutionally defined norms of appropriateness.
The LoA explains political action, then, by using a situation to account for
the specificity of the institutional context of action, by placing the motivation
to act in the demands of identity defined in terms of duties and obligations,
and by placing the performance of that act in the following of rules that
specify what is appropriate.
The Logic of Appropriateness: Holism or Individualism?
The LoA contrasts with the individual utility-maximizing picture of action
laid down in rational choice, what March and Olsen label a ‘logic of
consequences’ (LoC). The LoC seems overly individualistic in its omission
of the social and institutional elements that are relevant in explaining and
understanding action, and the LoA captures these dimensions by emphasiz-
ing the contextual-institutional or normative rationality of political action.
Nevertheless, March and Olsen understand the LoA as an individualistic
explanation of political action:
Like the logic of consequences, the logic of appropriateness is explicitly a logic
of individual action. It is specified as a mode of action or justification for an
individual actor. Thus it is as individualistic in structure as is the logic of
consequences. (March and Olsen, 1998: 952)
I will show that this characterization is misplaced, and that because of a
rather strong structural bias, the LoA becomes problematic as an action-
theoretical foundation for constructivist theory. To structure my analysis of
where to locate the LoA, I make use of Martin Hollis’s matrix of four
approaches to analysing social action (Hollis, 1996: 359).8
Although I am here primarily interested in the explanation dimension of
action as it relates most directly to how the LoA is used in constructivist
theory and in the debate with rationalist theory, I will also indicate how the
European Journal of International Relations 8(4)
450

LoA captures the understanding dimension of action as it has bearing upon
how individual action is conceptualized.
Understanding Action: Meaning, Interpretation and Rules
If the LoA is to be individualistic in structure, the individual actor must be
left with a reasonable degree of choice (or agency) concerning how to act in
different situations not specified by the institutionally defined identity, or the
institutionally defined (and defining) rules for appropriate behaviour. One
such ‘escape route’ for the individual actor in the LoA is the process of
interpreting rules and situations — ‘The elements of openness in inter-
pretation of rules mean that while institutions structure politics, they
ordinarily do not determine political behaviour precisely’ (March and Olsen,
1995: 33).
We saw above, however, that a core feature of the new institutional
framework is that institutions provide the very tool-kit through which
individuals interpret rules and situations, and moreover that this is produced
by a homogeneous political community, characterized by a set of shared
interpretations and conceptions of the common good (March and Olsen,
1989: 161). This substantially reduces the degree to which the process of
interpretation can enable individuals to interpret things differently. This is
readily expressed in how March and Olsen conceptualize individual actors in
relation to identities and institutions — ‘Institutions and identities con-
stitute and legitimise political actors and provide them with consistent
behavioural rules, conceptions of reality, standards of assessments, affective ties,
and endowments’ (March and Olsen, 1995: 30; emphasis added. See also
March and Olsen, 1989: 41).
Institutions are in this sense the prime supplier of and partly constitutive
for the very frames of understanding through which individuals come to
interpret and understand the world. This is clearly expressed in how
Figure 1
Sending: Constitution, Choice and Change
451

identities, as a central source for the generation of meaning and orientation
in the world, are here institutionally defined and acquired within institu-
tions. It thus seems that the indeterminacy of institutions on individual
interpretations of rules and situations is not sufficient to make the individual
actor capable of reflecting upon and interpreting beyond or somewhat
independently of institutional schemes. This would be less problematic if
rules specifying appropriateness were understood as some sort of external
constraint in which individual actors could act with reference to some set of
individual self-interests, typical of rationalist theory (see Schimmelfennig,
2000, 2001). But this is not so. In fact, rules here take on a hermeneutical
aspect, which renders them relatively immune from diverging inter-
pretations.
The rules defining appropriateness, we have seen, are constitutive for the
political community in which action takes place. The regulative dimension of
rules, then, is parasitic upon and cannot be understood in the absence of
their constitutive dimension that defines the very meaning of the social
practice within which action is performed. As noted by Giddens, the focus
on this constitutive and fundamental meaning-generating function of rules
implies that, ‘as expressed in forms of life, institutions are analysed only in so
far as they form a consensual backdrop against which action is negotiated
and its meaning formed’ (1979: 50). The central idea in this line of
theorizing is, according to White (1988), one that follows from Peter Winch
(1963) in which ‘the meaning and rationality of an action are derived from
understanding its role in relation to the prevailing norms and beliefs of the
form of life of which it is a part’ (p. 18).
The upshot of this argument is directly related to what kind of rationality
is implied in the LoA. It is one that adequately captures the institutional-
contextual or normative rationality of action in relation to practices within a
political community in the sense that the appropriateness of an action cannot
be established apart from or prior to the constitutive rules that define the
normative space (understood as what is appropriate) of the particular political
community. Hence, both the regulative (as it constrains interpretation) and
the constitutive (as it defines practices and, by extension, the political
community) aspect of rules both in their own way paint a picture of actors
as being more or less hermeneutically ‘programmed’ by the institutions in
which they are located. In this sense, March and Olsen appear to fail in
trying to establish a firm footing from which individual actors can interpret
rules and situations differently, a key requirement for a theory of individual
action. It thus seems reasonable to conclude that concerning the under-
standing dimension of action, the LoA bears a closer affinity to a holist
orientation than an individual one. We therefore locate the LoA in the upper
left quadrant of the matrix.
European Journal of International Relations 8(4)
452

Explaining Action: Why Follow Rules of Appropriateness?
‘If rules are to have any explanatory importance,’ Bohman (1991) notes, ‘it
must be in some larger context of a theory that explains a whole set of
complex interrelationships’ (p. 65). Further, a theory of individual action
seems to require an explanation of how and why actors may refrain from
following or violate certain rules. This motivates an analysis of the
theoretical constructs that make the LoA explain action by reference to
rules. We have seen that duties and obligations define or constitute
institutional identities, and that it is the assumed demand to fulfil these
duties and obligations that provides the motivation for actors to follow rules
specifying appropriateness. The critical question then becomes — What is it
about these duties and obligations that makes actors follow or comply with
them, thus explaining why they follow rules of appropriateness even in the
absence of external sanctions or a conception of interests defined in terms of
outcomes of actions? What, in other words, is the source of the ‘conception
of necessity’ that underwrites and defines the LoA as ‘Obligatory action’
(March and Olsen, 1989: 23)?
In relying upon the duties and obligations that define an identity to
explain why actors follow rules specifying what is appropriate, the LoA
appears to invoke a higher norm that says that these duties and obligations
shall be met. With basis in the writings of Jean Hampton (1998), this norm
must, if it is to explain why actors follow rules of appropriateness, be
motivationally efficacious.9This is so because March and Olsen rely
exclusively on these duties and obligations to do the job of explaining why
rules of appropriateness are followed: no reasons are provided in the LoA for
why actors follow rules of appropriateness beyond the claim that the duties
and obligations defining an identity demand it. That being the case, the LoA
must be said to ultimately explain action as rule-following by relying on a
norm that must be regarded as having, still with Hampton, objective
authority over the actors, thus relying on a motivational internalist position
to explain action (Hampton, 1998: 133–6).
If a norm generates motivationally efficacious reasons for action, it implies
that ‘to know that x ought to be done is to have a motive to do x’
(Hampton, 1998: 136). The very recognition of the norm is here sufficient,
or motivationally efficacious, to explain action because this norm has
objective authority: the norm constitutes a fundamental requirement for the
actor and is beyond the reach of criticism, reflection and potential violation.
Hampton notes that a norm with objective authority implies that it is
‘motivationally efficacious by virtue of its authority’, and further that the
actor ‘must follow this directive or else be condemned as irrational’
(Hampton, 1998: 133–4; emphasis in original). That it is this position that
Sending: Constitution, Choice and Change
453

underwrites the LoA is brought out by March and Olsen’s conception of
duties and obligations being constitutive for an actor’s identity, as actions are
explained by reference to a ‘conception of necessity’ (1989: 161). Moreover,
failure to fulfil the identity-defining duties and obligations is equated to
approach the insane: ‘Within a logic of appropriateness, a sane person is one
who is “in touch with identity” ’ (March and Olsen, 1989).
The implications of this motivational internalist position are far-reaching,
as an actor is never less than fully motivated to act in terms specified by the
rules defining what is appropriate. If the LoA were to allow that an actor
reflected upon, evaluated and possibly violated the norm that says that
institutionally defined duties and obligations shall be met, the motivation for
rule-following would have to be explained in terms of a motivational
externalist position.10 This implies that the norm that says that duties and
obligations shall be met has only normative authority over actors, which
does not generate motivationally efficacious reasons for action. In such a
motivational externalist position, an explanation for why actors follow rules
of appropriateness would take the following form: to know that x ought to
be done does not give sufficient motive to do x but only indirectly, and
coupled with other reasons, motivates x (see Hampton, 1998: 136).
In such a motivational externalist position, the actor is no longer internally
related to the norm since it only has normative, as opposed to objective,
authority. It is specifically the fact that this externalist position allows for an
external, and thus reflexive and evaluative, relationship between the actor
and the norm that it is no longer motivationally efficacious in explaining
action. Crucially, the norm is no longer constitutive for the actor’s identity, as
in the internalist position, and the normative authority of the norm thus
only takes us half-way in explaining why actors follow rules specifying
appropriateness. In a motivational externalist position, then, an action, or
rule-following in general, would be explained by reference to a combination
of the norm that says that duties and obligations shall be met, and a
reasoning process entailing a weighing of institutional demands, perception
of self, objectives defined as outcomes, and perceived legitimacy of the
relevant rules and norms.
Granted the motivational internalist position of the LoA, we have a theory
of action that substantially reduces the ability of the individual actor to act in
non-conforming ways, to act based on a conception of interests, or to reflect
upon and have a reasonably reflexive distance towards institutional rules and
identities so as to strike a balance between the ‘dictates of identity’ on the
one hand, and other wants or interests, on the other. To the degree that
the LoA can account for agency or action in terms of choice, it concerns how
the ‘dictates of identity’ (duties and obligations) shall be met by interpreting
and deciding what is most appropriate in a given situation. As we saw in the
European Journal of International Relations 8(4)
454

previous section, however, the choice involved in deciding how to act
appropriately is very much structured by the way in which institutions define
a shared and consensual frame of reference and interpretation within which
individual actors interpret different rules and situations.11 Indeed, the LoA
conceives of political action as obligatory action and defines it in terms of ‘an
affirmation of belief and an assertion of virtue’ (March and Olsen, 1989:
89).
From this, it seems a reasonable conclusion that concerning the explana-
tion of action, the LoA is closer to a holist than an individualist position:
actions are explained structurally by reference to a norm with objective
authority over the actor that generates motivationally efficacious reasons
why the duties and obligations that define an identity should be met and
which explains why rules of appropriateness are followed. As a theory of
individual action, the LoA paints a picture of actors as being institutionally
programmed to act in appropriate ways, as expressed through their
identities. Concerning both explanation and understanding, then, the LoA
gives ‘analytical priority’ to the holist or structural level.12
The Tension between Constitution, Choice and Change
We have seen that on its own terms, the LoA seems untenable as a theory of
individual action. The LoA is rather, and contrary to its authors’ claim,
structural in orientation. In the remaining part of this article, I shall analyse
the implications of this conclusion for the internal theoretical consistency of
the moderate constructivist research agenda. This means specifying what,
exactly, is problematic with the LoA for constructivist theory. I shall focus on
how the LoA relates to three core claims of constructivist theory: the first
concerns the claim that norms have constitutive effects on actors’ identities.
The second concerns the claim that structure and agency are mutually
constitutive. The third concerns the claim that changes in ideational or
Figure 2
Sending: Constitution, Choice and Change
455

normative structures do occur and can account for changes in political
practice.
The Logic of Appropriateness and the Constitutive Effect of Norms
A central point of contention between rationalism and constructivism
concerns whether norms constrain behaviour (rationalism) or whether they
are constitutive for actors’ identities (constructivism) (Checkel, 1997: 347;
Hurd, 1999). The argument that norms have a constitutive effect on actors’
identities is a crucial one in constructivist theory as it connects the
ontological claim of the centrality of ideational and normative structures
over material ones directly to individual action — norms are constitutive for
actors’ identities, and identities are in turn seen as defining the action-
orientation of actors as it provides actors with an understanding of their
interests (see Checkel, 1997, 1998; Risse, 2000; Price and Reus-Smit,
1998). This theoretical construction enables constructivist theory to launch
a double criticism of rationalist theory: it identifies a normative rationality of
action that breaks with the strategic and instrumental rationality in rational
choice (see Risse, 2000: 4–5), and it seeks to account for what is exogenous
in rationalist theory, namely the production and changes in the interests or
action-orientation of actors (see Adler, 1997; Checkel, 1998; Finnemore,
1996; Klotz, 1995).
In the rationalist conception, actors comply with rules and norms if they
perceive the costs (material and non-material) of non-compliance as higher
than those of compliance (Elster, 1989; Hurd, 1999). Absent the com-
pliance produced by self-interests, there needs to be some kind of external
sanctions. The underlying logic of action is rational choice, or what March
and Olsen call a ‘logic of consequences’. In the constructivist conception, by
contrast, the term ‘comply with’ is misleading as norms have, through
processes of either learning, persuasion or socialization, become intern-
alized. As internalized and constitutive for an actor’s identity, these norms
attain a motivationally efficacious power in explaining action in the sense
that there is what Hurd (1999) refers to as an ‘internal sense of moral
obligation’ that explains action (p. 387).
This theoretical conception is specified and accounted for by the LoA and
forms, I believe, the central reason for its adoption within constructivist
theory. The central issue concerns, first, how norms are related to identities,
and, second, how a particular identity can explain individual action. The
LoA specifies the micro-level process of both. The LoA accounts, first, for
how a norm can be constitutive for an actor’s identity by way of specifying
that a norm is internalized by an actor in the form of becoming linked to the
duties and obligations that define the actor’s identity. Action as appropriate
European Journal of International Relations 8(4)
456

rule-following is then explained by reference to the objective authority of
the higher norm that says that these duties and obligations must always be
met, which effectively motivates actors to act appropriately as specified by
certain rules.
Were the LoA to rely on a motivational externalist position, where the
norm saying that the duties and obligations of an identity should be met is
not motivationally efficacious, then it would not explain action by reference
to the constitutive effect of norms on actors’ identities. Rather, the
explanation of action in a motivational externalist position constitutes a
combination of the causal motivation generated by the normative authority
of this norm and other, non-internalized factors subject to an actor’s
reasoning, judgement and choice. As such, the LoA seems to provide, by
virtue of its motivational internalist position, exactly that micro-level
account of this key claim of constructivist theory: actors comply with norms
because these have been internalized and form part of the actor’s identity.
Individual action is thus explained by reference to how actors respond to the
‘dictates of their identities’ (March and Olsen, 1995: 36). Typically, the
action-orientation generated by the identity-defining duties and obligation is
normative (as defined by standards of appropriateness of the political
community) and not instrumental or strategic.
The Logic of Appropriateness and the Idea of ‘Mutual Constitution’
In seeking to establish constructivism as a middle ground between rationalist
and ‘interpretative’ (postmodern and poststructural) approaches in Inter-
national Relations, Adler (1997) asks the critical question ‘Do we explain
human action on the basis of individual motivation and the causal
interaction of intentional agents, or do we explain individual cognition and
action as a function of social forces or social structures?’ (1997: 324). He
responds by locating constructivism in the middle between Elster’s meth-
odological individualism and Durkheim’s methodological holism. In so
doing, he draws directly on Giddens’s structuration theory (1979, 1984):
Structuration theory, however, argues that the ‘properties of agents and of
structures are both relevant to explanations of social behavior’.... It explains
social institutions and social change as the result of the ‘duality of structure’.
(Adler, 1997: 325)
In fact, structuration theory is typically invoked as the ontological basis of
constructivism (see, for example, Wendt, 1987, 1999). As Checkel notes —
‘Constructivists emphasize a process of interaction between agents and
structures: the ontology is one of mutual constitution, where neither unit of
analysis — agents or structure — is reduced to the other and made
“ontologically primitive” ’ (Checkel, 1998: 326).
Sending: Constitution, Choice and Change
457

How, then, does the structural bias we identified earlier in the LoA fit in
with how agency is defined in structuration theory from which this idea of
‘mutual constitution’ is drawn? A first approximation to answering this
question is to be found in how Giddens defines agency, namely as
‘transformative capacity’:
. . . it is a necessary feature of action that, at any point in time, the agent ‘could
have acted otherwise’: either positively in terms of attempted intervention in
the process of ‘events in the world’, or negatively in terms of forbearance.
(Giddens, 1979: 56)
Furthermore, Giddens consistently uses the term ‘draw upon’ and ‘use’
when he speaks of how agents relate to institutional and structural features
such as rules (1979: 71). Indeed, when Giddens develops his structuration
theory, he explicitly rejects the idea that we can explain or understand action
by referring to how norms have been internalized by actors and so provide
all the motivation necessary to explain action. He notes:
This does not mean that reasons can be linked as directly with norms or
conventions as some philosophers have claimed or implied. Reasons do not
just include the citing of or the appeal to norms: to suppose that such is the
case actually draws the philosophy of action back towards the Parsonian action
frame of reference, since the conduct then becomes driven by ‘internalised’
normative imperatives. (1979: 57; emphasis added)
It is thus a central feature of structuration theory, which is a key building
block of constructivist theory, that the actor is always in a position to
evaluate, reflect upon and choose regarding what rules to follow and how to
act. This is so because the ‘could have acted otherwise’ condition implies
that agency involves choice. This seems to imply a motivational externalist
position in the sense that the actor is never fully motivated to follow a
certain rule by any constitutive and internalized norm.13 It follows from the
idea of agency as ‘transformative capacity’ that there is an external relation
between the actor and the set of rules that defines the structure in which he
or she is implicated since this is a necessary condition for choice. Within this
conception, a norm that motivates actors to follow rules of appropriateness
can only attain a causal motivational force as in a motivational externalist
position, and must be coupled with other reasons — these being the subject
of reflection and choice — to explain action.
In the LoA, by contrast, the actor is motivated to follow rules that specify
appropriate actions only and exclusively by a norm with objective authority
over the agent. Hence, the actor in the LoA does not have, as we have seen
above, a capacity to reflect upon, evaluate and possibly challenge the
structurally or institutionally defined norms explaining why actors act
appropriately.14 Thus, we may say that the LoA, given its motivational
European Journal of International Relations 8(4)
458

internalist position, is inconsistent with the understanding of the actor that
underwrites and makes possible the idea that structure and agency are
mutually constitutive. This is, of course, not to suggest that we should judge
the fruitfulness of the LoA by the standards of structuration theory.15 The
point is rather to identify a problematic inconsistency involved in using both
these theoretical constructs, and that the frequent argument made in
constructivist theory against other approaches concerning mutual constitu-
tion does not sit well with the inclusion of the LoA as the central theory of
action.
The Logic of Appropriateness and the Explanation of Change
A central claim in constructivist theory is that it is able to account for change
in international politics that extends beyond the change in behaviour that
rationalism can account for (Finnemore, 1996; Ruggie, 1998). A key
argument is that because inter-subjectively shared ideational factors such as
norms are constitutive for actors’ identities, changes in such ideational
factors lead to changes in the very identities and interests of the relevant
actors (Adler, 1997; Kratochwil and Ruggie, 1986; Wendt, 1999). Wendt’s
classic ‘Anarchy is What States Make of It’ (1992) makes the case that to the
degree that the international realm is anarchic, it is so only by virtue of its
continual production and reproduction in practice (as consistent with
structuration theory). Hence, the defining structures of the international
realm may change during processes of interaction in practice. During these
processes of interaction, actors’ identities and interests may change (pp.
394–5). This breaks with the position of rationalist theory in which it is only
behaviour that may change, and in which the identities and interests of
actors are exogenously given and stable (p. 392).
We noted above that the LoA accounts for one central part of this process,
namely the action mechanisms implied in the proposition that norms have a
constitutive effect on actors’ identities in terms of being internalized and
followed because of the ‘dictates of identity’. What the LoA cannot account
for, however, is the process by which the changes in ideational structures get
off the ground and are advocated. The LoA can account for how and why
new rules and norms are being followed once internalized, but not for the
process by which certain actors advocate, disseminate and in some way get
others to accept and internalize new norms.
Let me give an example: Finnemore and Sikkink (1998)16 have developed
a model for the study of the relation between ‘International Norms
Dynamics and Political Change’. They formulate a ‘Norm Life Cycle’ model
divided into three stages — ‘Norm Emergence’, ‘Norm Cascade’ and
‘Norms Internalization’. The starting point of analysis is one state of affairs
Sending: Constitution, Choice and Change
459

in which one LoA (LoA0) applies. The end point of the analysis is a state of
affairs in which another standard of appropriateness (LoA1) applies. The
action-theoretical account of what occurs in between these two states of
affairs (LoA0and LoA1) is explained by reference to a logic of action that is
not captured by the LoA. Finnemore and Sikkink note that in order for new
norms to emerge and become accepted by others, norm entrepreneurs or
advocates have to act explicitly inappropriately:
To challenge existing logics of appropriateness, activists may need to be
explicitly ‘inappropriate’. . . . Deliberately inappropriate acts . . . especially
those entailing social ostracism, or legal punishment, can be powerful tools for
norm entrepreneurs seeking to send a message and frame an issue. Thus, at this
emergent stage of a norm’s life cycle, invoking a logic of appropriateness to
explain behavior is complicated by the fact that standards of appropriateness
are exactly what is being contested. (1998: 897–8)17
The critical question then becomes: granted that it is held that the LoA
best describes the action-orientation of different actors before such processes
of normative change start, how does such a process of normative change
ever get off the ground? The LoA can hardly account for such changes — it
cannot convincingly account for the process by which changes in norms
come about, as it portrays actors as being internally related to prevailing
norms as they constitute their very identity. Invoking the ‘dictates of
identity’ to explain action, and excluding the use of judgement and
reflection upon the adequacy of the relevant norms (as in the case of a
motivational externalist position), the LoA cannot account for the agency
implied in changing those norms that it treats as constitutive for actors’
identities. Other action-theoretical accounts thus appear needed to explain
change, as brought out, for example, by Finnemore and Sikkink’s reliance
upon a ‘logic of consequences’ to explain the agency implied in the change
from LoA0to LoA1(1998: 895–9).
In more specific terms, the action-theoretical model that can explain
the change from LoA0to LoA1is one that breaks with the LoA. That is: the
LoA is unable to provide the action-theoretical mechanism implied in
the process by which inter-subjectively shared and constitutive ideational
factors (such as norms) can be challenged and changed over time. What the
LoA can do is to account for how these inter-subjectively shared ideational
factors, once changed, become constitutive for actors’ identities and
explanatory for appropriate action. Precisely because norms are made
constitutive for actors’ identities and are followed because of a ‘conception
of necessity’, however, the LoA cannot account for the action-mechanism
implied in the change from one normative context to another. This is
problematic since the LoA is thus unable to provide a full action-theoretical
account of a core substantive argument about international politics in
European Journal of International Relations 8(4)
460

constructivist theory (see Wendt, 1992, 1999; Kratochwil and Ruggie,
1986).
Conclusion
Our analysis shows, first, that the LoA is untenable as a theory of individual
action. Second, it shows that the LoA is inconsistent with constructivist
theory, as expressed through three of its core claims. The LoA accounts for
and specifies the action-mechanism implied in the key constructivist claim
that norms are constitutive for actors’ identities. The motivational internalist
position of the LoA makes it possible to account for the action-mechanism
implied in this claim. However, precisely because of this motivational
internalist position, the LoA is inconsistent with an equally central claim in
constructivist theory namely that agents and structures are mutually
constitutive, and it is unable to effectively account for the action-mechanism
through which changes in norms occur, which is a central substantive claim
in constructivist theory about international politics.
In light of this conclusion, it is necessary to ask whether there are other
theories of action that seem consistent with this branch of constructivist
theory. We saw at the outset that Risse’s (2000) ‘logic of arguing’ was
subsumed within the normative rationality of action accounted for by the
LoA. Risse tells us why when he notes:
Normative rationality implies constitutive effects of social norms and institu-
tions, since these rules not only regulate behaviour, that is, they have causal
effects, but also define social identities (‘good people do X’). This is where the
‘value added’ of constructivism comes in. (2000: 4–5)
A discussion of whether the ‘logic of arguing’ is tenable as a theory of
individual action and consistent with constructivist theory would require a
separate analysis. However, a brief look indicates that it is consistent with
two of the core constructivist claims discussed above. The central action-
theoretical mechanism, argumentation and deliberation geared towards
reaching a mutual understanding, must be grounded upon a motivational
externalist position since processes of argumentation would break down if
actors were not able to reflect upon and apply the faculty of reason to assess
which norms are, and are not, appropriate and morally justifiable. Hence,
these norms cannot have objective authority over actors. As such, the ‘logic
of arguing’ appears to be consistent with the claim of the mutual
constitution of agency and structure, and able to provide an action-
theoretical account for the explanation of change in constructivist theory
since both these claims require for their action-theoretical specification an
element of reflection, choice and agency not explained by institutional or
structural features.18
Sending: Constitution, Choice and Change
461

A ‘logic of arguing’ accounts, just as processes of socialization and
learning, for the process through which new norms are internalized: An actor
may change his or her identity and interests if he or she is persuaded during
a process of argumentation that previously held norms are not morally
justifiable and cannot be sustained in light of the better argument provided
by others. Exactly because reason and reflection are central to argumentative
processes, however, a ‘logic of arguing’ appears unable to account for the
norm-conform behaviour that is said to result from such a process of
internalization. Rather, since norms are always subject to critique and
challenges in the ‘logic of arguing’, it seems to be the combined causal power
of norms with normative authority and reasoning processes involving
argumentation and reflection that explain action. Action is explained by
reference to the necessary (but not sufficient) structural features embedded
in a ‘common lifeworld’ that produces a common frame of reference, and
the argumentative processes that take place within the horizon of this
lifeworld (Risse, 2000: 10).
A central question that emerges from this brief discussion is how
theoretically robust the claim of the constitutive effect of norms on actors’
identities is in explaining action.19 To my mind, constructivist theory
currently over-emphasizes the relation that evidently exists between norms
and identities when the former are made constitutive for the latter and
connected directly to the explanation of action.20 In this sense, the LoA is
quite similar to the much-criticized norm-sociology of Parsons in which
internalized norms explain action-orientations (see Giddens, 1979).21 Early
constructivist works in IR theory used Searle’s (1979, 1995) central
distinction between regulative and constitutive rules in identifying the
constructed, and thus changeable, character of the rules of the game of
international politics (see Kratochwil, 1989: 21–8. See also Ruggie, 1998:
13; Wendt, 1992). However, Searle’s conception of constitutive rules and
norms does not relate to the constitution of identities. It relates to the
constitution of practices.
During the latter half of the 1990s, a call for a theory of individual action
within constructivist theory seemed to emerge in order to make constructiv-
ism into a full-fledged and moderate research programme (see Finnemore,
1996: 5–30; Ruggie, 1998: 13–16). It is in this context, I believe, that the
central idea of constitutive rules and norms became attached directly to
identities rather than practices as it enabled the identification and specifica-
tion, at the level of a theory of action, of a normative rationality in the
international realm, and an account of interest-formation as expressed by the
tight connection between norms, identity and action in the LoA. This also
led to less emphasis on the central epistemological dimension of constructivist
thought, and less interest in developing a theory of action linked to a theory
European Journal of International Relations 8(4)
462

of knowledge (see Guzzini, 2000). To ensure theoretical progress and
consistency, more analysis seems to be required of how the constitution of
practices relates to the constitution of identities, and how actors reflect
upon, deliberate on, and are structured in their action-orientations by
different and competing norms and other ideational factors.
Notes
Earlier versions of this paper were presented at a PhD course in ‘International
Organizations and Institutions’ at the University of Oslo, June 2000, and at the
Theory Seminar at the Norwegian Institute of International Affairs, August 2001. I
thank Jeffrey Checkel, Stein S. Eriksen, Tor Halvorsen, Thomas Hornburg, Anders
Molander, Iver Neumann, Anne Julie Semb, Eilert Struksnes, the editors and the
anonymous reviewers of the European Journal of International Relations for valuable
comments on earlier drafts. The financial support of the Norwegian Research
Council and the Norwegian Institute of International Affairs is gratefully acknowl-
edged.
1. In the early phase of the debate between proponents of rationalism and
constructivism in International Relations (IR) theory, epistemological and
ontological issues were at the centre of the debate (see Onuf, 1989; Kratochwil,
1989).
2. To be sure, other action-theoretical accounts, such as socialization and learning,
are used in constructivist theory. The LoA subsumes these accounts in the sense
that it consists of a comprehensive and consistent formulation of the logic of
action that is said to result from these processes in the sense that actors
internalize, and are thus motivated to act as defined by, the constitutive norms
and rules of appropriateness embedded in the political community (see
Finnemore, 1996: 28–31, Finnemore and Sikkink, 1998: 913; Schimmelfennig,
2001: 58).
3. For an overview of different versions of constructivism, see Adler (1997) and
Ruggie (1998: 35–6). For more radical or critical constructivist approaches, see
Der Derian (1995) and Campbell (1992). Recently, Guzzini (2000) has
suggested a reconstruction of constructivist theory with basis in Pierre Bour-
dieu’s writings. For Guzzini, a central element of a constructivist stance is that a
theory of action must be coupled with a theory of knowledge (2000: 162).
4. The context of March and Olsen’s identification of the ideal of ‘civic identity’
should be properly understood: they discuss how institutional theories of politics
have adopted communitarian notions of democratic politics by way of
giving importance to the idea of the ‘community’. They criticize the idea that a
political community is based on shared preferences since that presumes that
‘individual action is based on individual values and preferences’ (1995: 38). By
invoking the ideal of ‘civic identity’, March and Olsen rely on a different
conception of the political community that holds that interests are genuinely
common, not individual, in their origin and character and are constituted by
inherently and irreducibly common objectives. This is explored below via Charles
Sending: Constitution, Choice and Change
463

Taylor’s conception of ‘irreducibly social goods’. Granted that the LoA is
grounded on this communitarian reading of the political community, it can be
argued that it does not sit well with the alleged anarchical feature and the
relatively weak normative integration that defines the international realm.
Developed for studying democratic politics at the level of an already normatively
integrated nation-state, its application to explain international politics may
appear problematical. For example, March and Olsen’s assertion that the
political community is ‘based on a shared history, a valued way of life, a shared
definition of the common good and a shared interpretation and a common
understanding embodied in rules for appropriate behavior’ (1989: 161) seems,
on most accounts, not to conform to the international realm. I shall hold this
issue in suspense in this article as it does not relate directly to the internal
theoretical consistency of the LoA and how it is used in moderate constructivist
theory.
5. I will not here consider the broader theoretical issue of the concept of rules and
its use in explaining action. See Giddens (1976) and Bohman (1991) for a good
discussion of this issue.
6. The emphasis on the constitutive and not only regulative dimension of rules is
central to IR constructivist theory, and I shall return to it later. See, for example,
Kratochwil (1989) and Ruggie (1998).
7. This conflates the standard usage of the terms in which roles define functions
whereas identities define meaning. The primary reason why these concepts are
conflated in the LoA and why identity arguably assumes a more important role
is probably related to what I noted above — the LoA is formulated through
studies of formal organizations in which roles (functions) take precedence. When
this conception is applied to political life in general, the concept of identity
(meaning) takes precedence as it specifies, at the level of the individual actor, the
normative and cognitive outlook of the political community.
8. Hollis’s matrix identifies what different approaches give ‘analytical priority’ to in
analysing social action. Structural explanations, or what is labelled systems
explanations in the matrix above, acknowledge that there are individual actors
that act and that they may pursue individual interests and a have a degree of
choice concerning their actions (see Haga, 1992: 117–19). The central point of
disagreement concerns whether the structural level of analysis has a more
significant and deeper causal force in explaining action than the individual level
which gives analytical priority to individual properties and interactions (see
Wendt, 1999: 12). Similarly, holistically oriented hermeneutical analyses do not
dismiss the subjective meanings behind action. Rather, the point is that priority
is assigned to, and the analysis ultimately centres on, the broader collective and
inter-subjective context in which actions take place and attain their meaning.
9. March and Olsen refer at some point to the trust that underwrites the rules
specifying appropriate behaviour: ‘we see the network of rules as sustained by
trust, a confidence that appropriate behavior can be expected most of the time.
Trust, like the rules it supports, is based on a conception of appropriateness
more than calculation of reciprocity’ (March and Olsen, 1989: 38). This
European Journal of International Relations 8(4)
464

argument could support a reading of the motivation for acting appropriately as
residing in the existence of external sanctions. Two aspects seem to render this
argument less valid, however. First, March and Olsen introduce appropriateness
as that which underlies the idea of trust. Hence, this trust refers to a standard of
behaviour and therefore a norm. Moreover, March and Olsen repeatedly stress
the demands of identity on action: ‘Citizens and officeholders are presumed to
act according to norms associated with their roles rather than in pursuit of
personal advantage and interests. They are presumed to respond to the dictates
of their identities’ (March and Olsen, 1995: 37; emphasis added). Second, the
trust that underwrites appropriate behaviour needs for its stabilization a norm.
That is, a norm can be seen as a standard or an expectation of behaviour. But
why do we trust that others will act appropriately? It must be because we have
trust or faith or belief that others are motivated to act by the same norm. Hence,
what motivates action is not external sanctions, but the belief or faith that all
actors have respect for and are committed to the same norm.
10. Hampton discusses the terms ‘motivational internalist’ and ‘motivational
externalist’ in the context of the assumed authority implied in hypothetical, not
categorial, imperatives. Nevertheless, the identification of an important differ-
ence in a motivational internalist and a motivational externalist position holds
true also for grasping the explanatory mechanism implicated in the LoA (see
Hampton, 1998: Ch. 4).
11. It can be argued against this that there is frequently a debate over which norms
apply in different contexts, and that, consequently, the interpretative constraints
imposed by the internalization of institutionally defined rules and norms are not
determining individual choice of what is appropriate. However, the claim that
there is often a debate over which norms apply in different contexts is empirical,
not theoretical. My critique against the LoA is immanent in that it focuses on its
internal theoretical structure. It is on this basis that it can be said that the LoA
is not well suited to capture the individual agency involved in deliberating and
debating over which normative frame applies. I thank James W. Davis for
helping me clarify this point.
12. Hollis and Smith (1990), using this matrix and referring to the relation between
explaining and understanding, hold that ‘there are always two stories to tell
(explaining and understanding) and that combinations do not solve the
problem’ (p. 7). My identification of the LoA as covering both the under-
standing and the explaining quadrant of the matrix is not to suggest that the
LoA succeeds in overcoming the tension between explaining and understanding.
Rather, it is to indicate that the LoA in fact addresses both the understanding
and the explaining dimension of action. I am here primarily concerned with the
explanation aspect of the LoA, and analyse the understanding dimension because
it shows how the LoA assigns to institutions and their constitutive rules and
norms a central role in accounting for how individuals form identities and
interpret situations and rules, thus giving analytical priority to the holist level. It
is once again important to stress that Hollis’s categorization is intended to bring
out what different approaches give ‘analytical priority’ to in understanding and
Sending: Constitution, Choice and Change
465

explaining action. The categorization of the LoA as structural or systemic in
character concerns the explanation of how and why action can be explained as
appropriate rule-following. As noted, reflection and choice here concern how to
act appropriately, which is in turn structured by the institutionally defined
scheme of understanding and interpreting the world.
13. Hampton refers to the motivational internalist position as Kantian in the sense
that the insight, the recognition, that something ought to be done itself
motivates action. Durkheim sociologizes this insight in terms of locating the
source of the objective authority of certain moral norms in social institutions.
Durkheim’s sociology, we know, is structural in orientation, and built on a
methodologically holist position in which structures take causal priority over
individual agents. The LoA operates in a Durkheimian tradition as it adopts a
motivational internalist position and makes it institutionally specific in terms of
giving analytical priority to institutions in socio-political life. It is thus not
surprising that structuration theory is inconsistent with the LoA as the former is
explicitly an attempt to move beyond the structuralism in Durkheim’s sociology.
Giddens’s earlier book on these matters bears the title New Rules of Sociological
Method (1976), an explicit reference to, and critique of, the ‘rules’ contained in
Durkheim’s The Rules of Sociological Method (1964).
14. As we saw in the discussion of the understanding dimension of the LoA, the
room available for differences in interpretation of rules and situations is
substantially reduced since actors receive the tools for such interpretation from
the institutions in which their identities are defined.
15. Guzzini (2000: 162) makes the case that a constructivist theory of action must
be coupled with a theory of knowledge, and must, therefore, be inter-subjective
in character rather than individualist. This argument suggests that there is a
problem with my critique of the LoA based on its failure to account for choice,
reasoning and individual differences. The identification of the LoA as an
untenable theory of individual action is not, however, based on the view that a
methodological individualist and rational choice-like theory of action is the
standard to which constructivist theory should look for a theory of action.
Rather, as brought out by the discussion of structuration theory, a theory of
action must be sensitive to the structural or inter-subjective context of action.
However, the LoA draws the explanation and understanding of action too far in
the direction of a Parsons-like norm-sociology that is problematical also for a
theory of action that would be inter-subjectively grounded.
16. See Risse et al. (1999) for a similar approach.
17. Finnemore and Sikkink do acknowledge a structural bias in March and Olsen’s
‘logic of appropriateness’. They note that ‘In this understanding, norm
conformance driven by a logic of appropriateness starts to look deterministic’
(1998: 913). They continue, however, by asserting ‘Yet, as we survey the norms
research that emphasizes appropriateness logic in IR, very little of it looks
deterministic’ (1998: 914). They cite two reasons for this. One is that IR
constructivists have ‘never been imperialistic in their claims’ and are happy to
acknowledge other logics of action (1998: 914). When judged in relation to our
European Journal of International Relations 8(4)
466

discussion above, there is more to the story: the LoA appears to be needed to
account for the micro-level account of why norms can be said to constitute
actors with their identities and explain why they act according to rules specifying
appropriateness. The other reason they cite, and which is held to be more
important, is that the LoA allows for ‘substantial room for agent choice’ (1998:
914). We have seen above that this assertion is problematic in that choice is here
primarily a matter of choosing how to act appropriately, and this choice is
moreover heavily structured by the institutional scheme that defines the
framework and tools by which situations and rules are interpreted.
18. According to Giddens’s theory of structuration, ‘structure is both medium and
outcome of the reproduction of practices. Structures enters simultaneously into
the constitution of the agent and social practices, and “exists” in the generating
moments of this constitution’ (Giddens, 1979: 5). Commenting upon the
centrality of a common lifeworld, Habermas employs a notion similar to that of
Giddens’s ‘duality of structure’ when he notes that ‘The lifeword forms both the
horizon for speech situations and the source of interpretations, while it in turn
reproduces itself only through ongoing communicative actions’ (Habermas, 1998:
22; emphasis added). The ‘logic of arguing’ thus appears consistent with this
claim (see also Risse, 2000: 10). The ‘logic of arguing’ is also able to account for
change in that it brings out how, through processes of argumentation and
deliberation, actors may change or reconstitute their interests and identities
(Risse, 2000). The ‘logic of arguing’ presupposes that all actors have access to
the discourse, and that they treat each other as equals. Further, relations of
power and coercion are assumed absent (2000: 11). Risse discusses in more
detail the validity of these assumptions in the international realm (2000:
14–23).
19. This issue has not been subject to discussion above, as the terms of this
constructivist position have been accepted in order to focus on its action-
theoretical foundation. However, our conclusion suggests that it is precisely this
claim that appears to require the LoA and its motivational internalist position,
which in turn creates an inconsistency with the two other core claims. I therefore
suggest that this should be a central concern for future research.
20. See Brubaker and Cooper (2000) for a critical discussion of the analytical power
of the concept of identity.
21. March and Olsen do not make a distinction between traditional and modern
society. The LoA seems more adequate as an account of action-orientations in
small, pre-modern societies characterized by a homogeneous culture and a
shared value-structure as the action-orientation of actors would probably be
structured to a large extent by a ‘conception of necessity’, generated by the
constitutive effect of certain norms on actors’ identities and upheld by traditions
and non-secular moral authorities. The matters are different in modern society,
at least in terms of a conception of politics, as there is likely to be a
heterogeneous, plural and contradictory mix of norms and values that are
relevant for determining identities and structuring action-orientations (see
Habermas, 1998: 25). It is significant in this respect that it is central to
Sending: Constitution, Choice and Change
467

Habermas’s general theoretical approach that modern society is, by virtue of the
‘disenchantment of the world’ of which Weber spoke, characterized by a
plurality of lifeworlds and overlapping and partly competing identities and
interests on the part of various groups. This forms the basis for moving beyond
the communitarian elements that he adopts from Durkheim and Parsons on the
forces of social integration (1998: 26).
References
Adler, Emanuel (1997) ‘Seizing the Middle Ground: Constructivism in World
Politics’, European Journal of International Relations 3(3): 319–63.
Bohman, James (1991) New Philosophy of Social Science. Cambridge: Polity Press.
Brubaker, Rogers and Frederick Cooper (2000) ‘Beyond “Identity” ’, Theory and
Society 29: 1–47.
Campbell, David (1992) Writing Security: United States Foreign Policy and the
Politics of Identity. Minneapolis: University of Minnesota Press.
Checkel, Jeffrey T. (1997) ‘International Norms and Domestic Politics: Bridging the
Rationalist–Constructivist Divide’, European Journal of International Relations
3(4): 473–95.
Checkel, Jeffrey T. (1998) ‘The Constructivist Turn in International Relations
Theory’, World Politics 50: 324–48.
Checkel, Jeffrey T. (1999) ‘Norms, Institutions, and National Identity in Contem-
porary Europe’, International Studies Quarterly 43: 83–114.
Checkel, Jeffrey T. (2001) ‘Why Comply? Social Learning and European Identity
Change’, International Organization 55(3): 553–88.
Cortell, Andrew P. and James W. Davis, Jr. (1996) ‘How Do International
Institutions Matter? The Domestic Impact of International Rules and Norms’,
International Studies Quarterly 40: 451–78.
Der Derian, James (ed.) (1995) International Theory: Critical Investigations. New
York: New York University Press.
Durkheim, Emile (1964) The Rules of Sociological Method. New York: Free Press.
Elster, Jon (1989) Nuts and Bolts for the Social Sciences. Cambridge: Cambridge
University Press.
Finnemore, Martha (1996) National Interests in International Society. Ithaca:
Cornell University Press.
Finnemore, Martha and Kathryn Sikkink (1998) ‘International Norms Dynamics and
Political Change’, International Organization 52(4): 887–917.
Giddens, Anthony (1976) New Rules of Sociological Method — A Positive Critique of
Interpretative Sociology. London: Hutchinson.
Giddens, Anthony (1979) Central Problems in Social Theory. Los Angeles: University
of California Press.
Giddens, Anthony (1984) The Constitution of Society: Outline of the Theory of
Structuration. Cambridge: Polity Press.
Guzzini, Stefano (2000) ‘A Reconstruction of Constructivism in International
Relations’, European Journal of International Relations 6(2): 147–82.
European Journal of International Relations 8(4)
468

Habermas, J¨urgen (1998) Between Facts and Norms: Contributions to a Discourse
Theory of Law and Democracy. Cambridge, MA: MIT Press.
Haga, Ånund (1992) ‘Samfunnsvitenskap og praktisk filosofi’, in M. Bertilsson and
A. Molander (eds) Handling, norm och rationalitet. Bergen: Ariadne Forlag.
Hampton, Jean (1998) The Authority of Reason. Cambridge: Cambridge University
Press.
Hollis, Martin (1987) The Cunning of Reason. Cambridge: Cambridge University
Press.
Hollis, Martin (1996) ‘The Philosophy of Social Science — An Introduction’, in N.
Bunnin and E.P. Tsui-James (eds) The Blackwell Companion to Philosophy.
Cambridge: Blackwell.
Hollis, Martin and Steve Smith (1990) Explaining and Understanding in Inter-
national Relations. Oxford: Clarendon Press/Oxford University Press.
Hurd, Ian (1999) ‘Legitimacy and Authority in International Politics’, International
Organization 53(2): 379–408.
Kahler, Miles (1998) ‘Rationality in International Relations’, International Organi-
zation 52(4): 919–41.
Katzenstein, Peter J. (ed.) (1996) The Culture of National Security: Norms and
Identity in World Politics. New York: Columbia University Press.
Keohane, Robert (1988) ‘International Institutions: Two Approaches’, Inter-
national Studies Quarterly 32(4): 379–96.
Klotz, Audie (1995) Norms in International Relations. Ithaca: Cornell University
Press.
Kratochwil, Friedrich (1989) Rules, Norms, and Decisions: On the Conditions of
Practical and Legal Reasoning in International Relations and Domestic Affairs.
Cambridge: Cambridge University Press.
Kratochwil, Friedrich and John G. Ruggie (1986) ‘International Organization: A
State of the Art on an Art of the State’, International Organization 40(4):
753–75.
March, James G. and Johan P. Olsen (1989) Rediscovering Institutions. New York:
Free Press.
March, James G. and Johan P. Olsen (1995) Democratic Governance. New York:
Free Press.
March, James G. and Johan P. Olsen (1998) ‘The Institutional Dynamics of
International Political Orders’, International Organization 52(4): 943–69.
Moravcsik, Andrew (1998) The Choice of Europe: Social Purpose and State Power from
Messina to Maastricht. Ithaca: Cornell University Press.
Moravcsik, Andrew (1999) ‘A New Statecraft? Supranational Entrepreneurs and
International Cooperation’, International Organization 53(2): 267–306.
Onuf, Nicholas (1989) World of our Making: Rules and Rule in Social Theory and
International Relations. Columbia, SC: University of South Carolina Press.
Price, Richard and Christian Reus-Smit (1998) ‘Dangerous Liaisons? Critical
International Theory and Constructivism’, European Journal of International
Relations 4(3): 259–94.
Sending: Constitution, Choice and Change
469

Risse, Thomas (2000) ‘ “Let’s Argue!”: Communicative Action in World Politics’,
International Organization 54(1): 1–39.
Risse, Thomas, Stephen C. Ropp and Kathryn Sikkink (eds) (1999) The Power of
Human Rights: International Norms and Domestic Change. Cambridge: Cam-
bridge University Press.
Ruggie, John G. (1998) Constructing the World Polity: Essays on International
Institutionalization. London: Routledge.
Schimmelfennig, Frank (2000) ‘International Socialization in the New Europe:
Rational Action in an Institutional Environment’, European Journal of Inter-
national Relations 6(1): 109–39.
Schimmelfennig, Frank (2001) ‘The Community Trap: Liberal Norms, Rhetorical
Action, and the Eastern Enlargement of the European Union’, International
Organization 55(1): 47–80.
Searle, John R. (1979) Expression and Meaning: Studies in the Theory of Speech Acts.
Cambridge: Cambridge University Press.
Searle, John R. (1995) The Construction of Social Reality. New York: Free Press.
Taylor, Charles (1995) Philosophical Arguments. Cambridge, MA: Harvard Uni-
versity Press.
Wendt, Alexander (1987) ‘The Agent–Structure Problem in International Relations’,
International Organization 41(3): 335–70.
Wendt, Alexander (1992) ‘Anarchy is What States Make of It: The Social
Construction of Power Politics’, International Organization 46(2): 391–425.
Wendt, Alexander (1999) Social Theory of International Politics. Cambridge:
Cambridge University Press.
White, Stephen K. (1988) The Recent Works of J¨urgen Habermas: Reason, Justice and
Modernity. Cambridge: Cambridge University Press.
Winch, Peter (1963) The Idea of Social Science. London: Routledge.
European Journal of International Relations 8(4)
470