To support an institutionalist causal segment we need two kinds of evi- dence. First, in a temporally and conceptually proximate sense, we must document how and to what extent actions flow directly from institutional conditions. This task is similar to the steps that support a structural claim. We need to document the pattern of institutional constraints or incen- tives around the time of the action, provide a logic by which position in this pattern would dictate certain actions, show that the pattern of action corresponds to the institutional patterns, and (since this is a rationalist logic) offer at least some evidence that actors exhibited broadly rational decision-making and actually followed the logic we have posited.
The other set of steps to support an institutionalist claim reflects the man-made, particularistic nature of this kind of cause. As the introductory chapters stressed, institutions are consequences of earlier actions. Even once we demonstrate their apparent proximate causal importance in a pattern of action, their effects only become distinctly institutional to the extent that we show that the institutions do not reduce to other condi- tions. We must show that at some point in the past, extra-institutional conditions were insufficient to cause people to create or maintain this institutional pattern of action rather than some range of alternatives. In other words, we must document the contingency which was resolved by the particularistic logic of institutional path dependence. Only given this sense of historical alternatives can we substantiate and specify a distinctly institutionalist causal claim.
Documenting contingency may seem a tall order, but it is no more difficult than documenting causality. Arguing that a certain set of struc- tural conditions left a certain range of options open, for example, is just the mirror image of arguing that structural conditions constrained
or propelled people toward a certain course of action. At first glance it may seem harder to support the negative claim that no causal vectors shut down contingency than to argue positively for one causal vector, but these arguments follow the same process. Arguing positively that some range of variation traces to one cause requires us to show, in principle, that no other cause accounted for some of that variation. In other words, positive causal arguments themselves depend directly on negative claims about the contingency of competing causes. In neither kind of claim can we ever address all potential causes. There is always room to suspect that some cause has escaped our notice. We usually truncate our search for causes by focusing pragmatically on the fairly small set of competing claims that scholars have advanced on similar topics. While one of the concluding points of this book is that we must extend this set further than most scholars do—working harder to debate a range of abstractly plausible alternative claims—we can never come close to chasing down every imaginable hypothesis. This is fine: ultimately we are debating other scholars, not Truth itself. Respectable support for causal claims forces us to carve out a range of causal effect for one cause vis-à-vis other active hypotheses (plus, perhaps, a few arguments that may not be present in a specific debate but which have substantial grounding in broader the- oretical approaches). Respectable support for claims about contingency forces us to mobilize evidence that a similar range of actively competing or broadly legitimate hypotheses fail to explain across a certain range of actions (Mahoney 2000).
Conclusion
This chapter began by noting an implicit consensus on how to define
institution but considerable disagreement on how to link institutions to
action. To allow institutionalist to designate a clear and distinct logic of explanation, I first considered in the abstract what kind of claims would give the most direct and irreducible force to institutions. To separate institutionalist claims from ideational or psychological interpretations, they must employ a logic where rational individuals confront intersubjec- tively present, man-made organizations, rules, or conventions. To separate institutionalist claims from structural causality, these constraints must be unintended legacies of past choices made amid structural ambiguity or unpredictability, not intentional solutions or adaptations to structural conditions.
Next I argued that this kind of institutionalist causal segment does not match cleanly to any of the schools that claim the institutionalist mantle. Sociological institutionalists focus on ideational dynamics in which institutions affect action by shaping interpretations of legitimate or conceivable behavior. The most common logic in rationalist institution- alist work is one of ‘structure-induced equilibrium’, with institutions cast as by-products of structural conditions. Other rationalist institutionalists take institutions seriously as constraints but ignore path dependence, effectively treating inherited institutions as a loose extension of struc- ture. Only to the extent that rationalist institutionalists include unin- tended consequences—which many do to some degree—do they employ distinctly institutionalist logic. Historical institutionalists are the most consistent advocates of institutionalist claims, but often mix them with ideational logic. To realize the promise of this compound logic, historical institutionalists must be more careful about the relative contributions of its components.
The last section noted that institutionalist claims demand a two-step demonstration. First they must link institutional positions to action, using evidence much like that needed to link structural positions to action. Then they must do the reverse for extra-institutional conditions— showing that at some point structures, ideational elements, and psychol- ogy (and perhaps other preexisting institutions) did not link clearly to patterns of action, thereby establishing a range of contingency that insti- tutions later resolved. The product is an argument with a very distinctive feel, in which rational people unintentionally construct their own future. It exhibits the combination of positional thinking with particularity that defines the second box of my master matrix.