International Relations. A self-Study Guide to Theory

causal and constitutive” (Wendt 1998: 104, my italics).
BOTH are necessary elements of a complete explanation of social action 
(Wendt 1987: 362). This epistemological position will be elaborated in the 
next section in more detail. 
4.2. Causal and constitutive explanation 
In order to understand the question-oriented approach to explanation, it is 
important to keep in mind that scientific realism emphasizes “ontology before 
epistemology”. In other words, as opposed to positivism, it is ontology and 
not epistemology that legitimates scientific practice (Wendt 1999: 91, see al-
so Step 1). For the purposes of explanation, this reversal implies that the form 
of scientific explanation depends on the nature and causal properties of enti-
ties. In short, the type and form of explanation depends on ontology. Expla-
nation in turn depends on the object of the question, on “what is taken to be 
problematic” (Wendt 1987: 362). 

---

215 
The position of an ontological and conceptual interdependence of agents 
and structures as “mutually constitutive” (structuration theory) thus has im-
plications for explanation. These implications emerge because, ontologically, 
both (agency and structure) are involved in the production of the social world 
(mutual constitution, co-determination). 
Wendt makes a distinction between two types of questions:
1) “Why (did X happen rather than Y)?”
2) “How (is action X possible) and what?” (Wendt 1987: 362) 
The former is a causal question to be answered by causal theory, the latter a 
constitutive question to be answered by constitutive theory. 
Causal theory: historical explanation 
Causal theory offers answers to Why-questions and requires a type of expla-
nation in the format “X causes Y”. It is the classical form of explanation that 
rests on independent and dependent variables and establishes a causal relation 
between them. That is, X and Y exist independently of each other, X pre-
cedes Y in time, and without X, Y would not have occurred (Wendt 1998: 
105). This type of causal explanation is an explanation of changes in the state 
of some variables. Factors of a change in the dependent variable exist inde-
pendently and temporally prior to the transition. Causality is a relation of log-
ical necessity to initial conditions and laws: if A, then B. It establishes a 
causal relation of temporally sequenced observed events. Explanation is thus 
a generalization about observable, sequenced behavior. Wendt calls it a “his-
torical explanation” (see also Wendt 1998: 105). 
Constitutive theory: structural explanation 
The type of questions answered by constitutive theories is different; at the core 
are how-possible-questions and what-questions. Remember, underneath is the 
ontological position of a mutual constitution of agency and structure. An an-
swer to a how-possible-question thus is to show how the properties of a social 
system are constituted. Constitutive theory offers knowledge about the condi-
tions of possible natural and social kinds (Wendt 1998: 105). How-possible 
questions explain “how the elements of a social kind are composed and orga-
nized so that it has the properties that it does” (Wendt 1998: 112; also called 
“morphological” explanation, a term borrowed from Haugeland 1978).
In contrast to how-possible-questions, what-questions are requests for 
what it is that “instantiates” a phenomenon (not why). It is also called “ex-
planation by concept”. An example for a what-question is, What kind of po-

---

216 
litical system is the EU? (Wendt 1998: 105; Wendt 1999: 110). In the ab-
sence of the structures, the properties of a phenomenon would not exist. This 
is a “conceptual” necessity, NOT a causal one as described above for the case 
of causal theories and explanation. Wendt presents the example of the Cold 
War: the factors constituting the social kind “Cold War” define what a Cold 
War is (but no causal determination) (Wendt 1998: 106). Here the assump-
tions of causal explanation, i.e. dependent and independent variable plus 
temporal sequence (if A, then B), are not applicable. The factors constituting 
the Cold War do not exist apart from a Cold War nor do they ontologically 
precede the “Cold War” in time. Rather, according to Wendt, “when they 
come into being, a Cold War comes into being with them, by definition and at 
the same time” (Wendt 1998: 106; my italics). When the constituting condi-
tions vary, the effects of constitutive structures vary. However, in this case the 
dependency in this variation is conceptual and not causal (Wendt 1998: 106).
Answers to constitutive questions of the what-type are descriptive but ex-
planatory: They classify observations and “unify” them as parts of a whole 
(Wendt 1998: 110). “Explanations – what explain by subsuming observations 
under a concept – as opposed to a law” (Wendt 1998: 110, emphasis orig.). 
Explanation by concepts is thus about achieving explanatory power by “uni-
fication” (Wendt 1998: 111-112).