In my last post, I presented some concerns about the institutional/interactional distinction with regards to the subject of justice and sketched a potentially new way to distinguish between the claims labeled (1) and (2) in that post. Some further thought has led me to be less confident that the worries I presented should lead us to question the adequacy of the institutional/interactional distinction. Nevertheless, I still think there’s at least one reason to shift from an institutional/interactional distinction to a structural/interactional distinction. So this post will attempt two things. First, I’ll sketch out a reason to think that institutions and structures aren’t coextensive (by arguing that institutions are instances of structure). Second, I’ll offer a reason for thinking that the important distinction is between structuralism and interactionism rather than institutionalism and interactionism. (Note: in the last post, I suggested that we ought to dump the institutional/interactional distinction wholesale in favour of a structural/non-structural distinction. For now, I’ve changed my mind; I think we should turn in the institutional/interactional distinction for the structural/interactional distinction.)
So, are ‘institution’ and ‘structure’ coextensive? No. Consider a tentative definition of ‘structure’:
Structure: the set of elements that constitute the situation in which agents find themselves, esp. those elements that constitute the way agents are positioned vis-à-vis each other.
To illustrate: In the contract negotiation scenario I presented last post, the contract demands made by the employees or the terms offered by the employer aren’t part of the structure of that situation. The rights each party has and the respective bargaining position of each engendered by those rights are structural features.
Now consider the two broadest definitions of ‘institution’ I could find:
Institution: a persistent and connected set of rules (formal or informal) that prescribe behavioural roles, constrain activity, and shape expectations. (R. O. Keohane, ‘International Institutions: Two Approaches’ (1988), p. 83)
Institution: ‘Organized patterns of socially constructed norms and roles, and socially prescribed behaviors expected of occupants of those roles, which are created and re-created over time.’ (R. E. Goodin, ‘Institutions and Their Design’ (1996), p. 19)
Admittedly, the examples of structural features I’ve given fit under these definitions. That’s fine. I think that institutions are structuring entities; i.e., that institutions play a significant role in constituting the situation of agents vis-à-vis each other. I also think that institutions are structural features; i.e., the presence of an institution mediating the relationship between agents is a constitutive feature of their situation. But I don’t think that institutions are the only structural features; indeed, the absence of any institution could be a feature of some structure. For example, consider a bargaining situation in the state of nature. A state of nature is, by definition, void of institutions, even in the broad sense. But we wouldn’t say that a bargaining situation in the state of nature is unstructured. At the very least, the power (in)equality and the relative bargaining position of the bargaining parties are constitutive features of their situation vis-à-vis each other.
To be clear, what I’m claiming is that institutions are (a) structuring entities and (b) a subset of all structural features. So structuralism and institutionalism don’t have different concerns — the core concern of both is the way in which social relations are structured. But structuralism is more general than institutionalism; the latter concerns itself with the justice of the structure that arises from institutions, whereas the former concerns itself with the justice of structure in the absence of institutions as well.
Given the foregoing, why adopt the structural/interactional distinction instead of the institutional/interactional distinction? Is the distinction between structures and institutions idle? Or can it do some useful work? I think the distinction between structures and institutions is especially useful when thinking about international justice. Typically, when we think of international justice problems, we think of wealth distribution, trade, humanitarian intervention, disease, climate change, etc. For the most part, there is some institutional structure in place to deal with these problems, even if it is only minimal. So it’s natural to think that one ought to be an institutionalist about justice if one is concerned about the way in which international relationships are structured (as, e.g., Pogge is). But international institutions aren’t as robust as domestic ones and we’re living at a time when the global institutional order is still being established. Accordingly, the establishment of international institutions is a significant problem of international justice. And if one is inclined to think that our primary normative concern ought to be the structure of agents’ situation vis-à-vis each other, then, at least with respect to the problem of establishing international institutions, we ought to be concerned about the structure from which institutions emerge. Indeed, it doesn’t make sense to concern ourselves the institutional arrangements from which institutions emerge; at some point, institutions must emerge from a situation where there was none.
There might be some reasons to confine ourselves to a concern about institutions rather than wider structures. In the first place, one of the primary reasons we seem to be concerned about institutions is that they are coercive and that any coercive structure ought to conform to some principles of justice. Coercion implies the intentional use of force or threats to compel certain kinds of behaviour. Thus, it seems wrong to say that preinstitutional structure (or structure in the absence of institutions) is coercive. The structural features are just natural facts: x has more power than y, so it’s a structural feature of x and y’s relationship that x has more bargaining power than y. It’s true that this situation will enable x to coerce y, but it doesn’t seem right to say that x’s capacity to coerce y is itself imposed upon x and y. So if our concern is with coercive structures, we’d be right in confining our interest to institutions. Second, even if we are concerned about the relational structure from which institutions emerge, we might think that there’s little that can be done to change it short of establishing institutions and that we could simply address our structural concerns during the institutional design phase. Thus, our concern would be properly restricted to institutions.
Despite these, I think there are some prima facie reasons for swapping a structural view for an institutional one. First, if one is concerned with making the actual world more just rather than simply theorizing about what constitutes an ideally just state of affairs (i.e., if one is interested in doing so-called non-ideal theory rather than ideal theory), then a concern for establishing just institutions must address the justice of the structure of the bargaining situations from which institutions are emerge. Realistically, the design of institutions is going to reflect the interests of the parties who are able to impose their will during the bargaining process. So non-ideal theorists should concern themselves with thinking about how to bring about more just preinstitutional bargaining structures so that we end up with more just institutions. (There’s lots more that should be said here, but I’m not yet sure what to say and, in any case, it’s peripheral to my central point.) Further, it’s not so unusual to think about the justice of preinstitutional bargaining structures. Consider Rawls’s construction of the original position. The original position is the hypothetical preinstitutional bargaining structure from which principles of just institutional design emerge. Rawls is very careful to situate his hypothetical agents vis-à-vis each other in what he feels is a just way. (For him, this means placing the agents behind a veil of ignorance, thereby eliminating any bargaining advantages that might accrue from morally arbitrary natural and social characteristics of the agents.) Now, unlike Rawls, non-ideal theorists concerned about the situation from which international institutions emerge don’t have the luxury of placing actual agents behind a veil of ignorance as a way to bring about a just preinstitutional bargaining structure. But this doesn’t mean that we can’t be concerned about the justice of preinstitutional structures; it simply means that we are constrained in our theorizing about how to bring about just preinstitutional structures.
One final (though perhaps not very strong) reason for exchanging institutionalism for structuralism is that calling one’s view ‘institutional’ when one’s primary concern is structure (as is ostensibly the case with Rawls and Pogge) is likely to be misleading. Calling the view ‘institutional’ often leads others to believe that the view is concerned exclusively with the justice of particular institutions (e.g., the UN, WTO, World Bank, etc.). While it’s undoubtedly true that a structural view is concerned with the way in which particular institutions structure the situation of agents, particular institutions are only instances of the structuralist’s primary concern. Indeed, it’s not the institutions per se that the ‘institutional’ view is fundamentally concerned with; the ultimate concern is the way in which social relations are structured. Institutions are only instrumentally important in this regard because they are the predominant structuring entity in our world. So, while this is perhaps only a cosmetic reason for abandoning the ‘institutional’ label in favour of the ‘structural’ label, it’s a cosmetic change that could help clarify what precisely is at stake. (Of course, some view called ‘institutionalism’ could serve as an intermediate view between structuralism and interactionism. Ostensibly, it would be the view that is exclusively concerned with institutions because it thinks there is something special about coercive structure. This might even be the right way to read Rawls and Pogge. However, insofar as one’s concern is with structure simpliciter (coercive or not), ‘structuralism’ is a better handle than ‘institutionalism’.)
If all this is adequate, my conclusion is that we ought to scrap the institutional/interactional distinction in favour of a structural distinction, at least in cases when we can’t assume the presence of institutions or where institutions are still in their formative stage, e.g., in the international sphere. (Where robust institutions are present, as in most domestic cases, the structural and institutional views would probably look too similar to distinguish between them.)
Earlier, I said that pt. 3 would offer some reasons for thinking we ought to be structuralists about justice. I guess I’ve gestured in that direction a little already; at least, I’ve offered some reasons for thinking we should be structuralists as opposed to institutionalists. Perhaps pt. 3 will outline some reasons for thinking we ought to be structuralists as opposed to interactionists.