This post started off as a reply to a comment under Robert Kelly’s post on historical institutionalism, but it got so long I thought it deserved its own post.

There is a great deal of ambiguity in how we use the term “decision” in contemporary IR, an ambiguity that also infects the closely related concept of “choice.” Briefly, we use these terms to refer both to objects of explanation and to means of explanation — and the ambiguity of our usage leads the the misleading conclusion that to explain decisions or choices necessarily involves a micro-reductionist account of interests, beliefs, and other “internal” factors. But this does not follow, something that would be made clearer if we were more precise about our use of terms and concepts.

When we use “decision” to indicate an object of explanation, we are in effect equating decision or choice with the more general issue of why we have some particular social arrangement or why we see some particular action or behavior — in each case, why we see what we see and not something else. So what Moravscik calls “the choice for Europe” means, if cashed out this way, “why do we have European integration as opposed to something else, presumably a Europe of completely independent traditionally sovereign states.” The “decision to divide Germany” means that we have a divided Germany after the Second World War, rather than a unified Germany. And so on. The important thing here is that designating these outcomes “decisions” doesn’t tell us much of anything about how we should explain them, except that any such explanation has to account for the fact that one possibility was a utilized while others were not. It is even unclear who or what “decides” or “chooses” in this formulation; what is important is that designating the outcome as a choice or decision highlights the contingency of that outcome, because of the outcome was something we thought inevitable we wouldn’t call it a choice or a decision.

But when we use “decision” or “choice” to indicate a means of explanation, we are tacitly adopting a micro-reductionist explanatory strategy. Here we have both an individualist scientific ontology — individual actors precede the social arrangements and institutions in which they are involved — and the idea that we explain what those individuals do by pointing to internal processes and factors, “decision-making.” Under this description, “the choice for Europe” is equivalent to an account of why particular states decided to combine and coordinate their efforts under the rubric of European integration, and “the decision to divide Germany” is equivalent to a specification of the motives that various involved actors had for dividing Germany rather than keeping the country more or less intact. There is no open-ended lack of clarity about who decides; indeed, making a micro-reductionist account like this work requires us to specify the individuals involved (whether those are states, interest groups, particular human beings, or whatever) as part of our explanatory strategy.

Let me suggest that only the second usage is meaningfully about either decision or choice. The first one might be, and in certain cases — Moravcsik leaps to mind — to say “decision” or “choice” is to deploy both usages simultaneously. The ambiguity of the way that we use these terms obscures the fact that there is no necessary logical or theoretical connection between describing an outcome as historically contingent (usage #1) and explaining that outcome as the result of internal decision-making procedures and factors like interests and beliefs (usage #2). If we just called the first usage “historical contingency” instead of “choice” or “decision,” the ambiguity vanishes, along with the erroneous implication that the only way to account for contingent outcomes is to reduce them to decisions that individuals make.

This relates immediately back to the issue of rationalist vs. historical institutionalism(s), and to the question of whether one needs or can meaningfully have both kinds of argument in a single account. Adopting the micro-reductionist usage of “decision” and “choice” would place us close to the rationalist camp, albeit not exclusively so because a belief-driven model of individual decision (i.e. one that did not rely on a strong ex ante specification of interests) would also fit. (Here is a good place to recall that “logics of appropriateness” and “logics of consequences” are equally micro-reductionist modes of explanation, because the causal action takes places inside of individuals. Ole Jacob Sending nailed this point years ago.) On this account, social outcomes have to be reduced to individual decisions in order to be explained, which is why I initially quipped that “rationalist institutionalism” is a contradiction in terms because step one of a rationalist explanation is to make the institution disappear into a morass of individual decisions…the institution itself has no independent explanatory status, and only derives its efficacy from the ongoing choices that individuals make. Logically, one can’t both have an account that accords institutions no independent causal power, and an account that does, so the micro-reductionist explanation is strictly speaking incompatible with other ways of analyzing social outcomes. The ambiguity of the way we use terms like “choice” and “decision” obscures this incompatibility, because it reduces social outcomes to individual decisions by definitional fiat.

The same applies to the kind of explanation of “decisions” that something like a rhetorical coercion account (or other meso-level relational theories involving positionality, and arguably macro-level structural theories too) sets forth. If we are using the term in the second sense, then there’s a pernicious contradiction at the heart of any explanation of decisions or choices: if we can provide a theory that explains why individual I made choice C, then individual I couldn’t really have made any choice other than C — the theory tells us why I chose C, after all — so it is unclear whether the choice for C was actually a choice at all. But this contradiction only applies if we stick to that second micro-reductionist usage of “decision” or “choice,” and it vanishes if we adopt the first usage, the one that I suggested we just call “historical contingency.” If our object of explanation is the fact that we have something rather than something else, there is no need to get metaphysically tied up in knots about free will or anything else along those lines; a theoretical explanation that tells us why we got (e.g.) European integration rather than something other than European integration has precisely zero implications for what individuals chose. Ditto individual action, the contingency of which can be built into explanations that don’t involve individual choices pretty easily. A rhetorical coercion account does not suggest that anyone chose or decided anything, since its object of explanation is neither a choice nor a decision, but instead answers the question why we have one action rather than another.

For these reasons I remain convinced that the best strategy is to keep these two kinds of accounts separate. A less ambiguous use of a term like “decision” would be a great help in this endeavor.