2. Empirical limitations

Boyer and Petersen's selection of cases reveals two distinct sorts of empirical limitations. First, they dramatically overstate the degree of institutional homology or ‘isomorphism’ (DiMaggio and Powell, Reference DiMaggio and Powell1983) in the three domains that they specify. Second, they select three domains where one probably sees more homology than in other domains. The latter is particularly important because it is directly relevant to their suggestion that an evolutionary psychology approach can replace traditional institutional theory.

The fact that the implications of their specific cases are being exaggerated becomes clear once the reader enters into the discussion of their specific contents. The claim that ‘marriage’ is universal ends up meaning that many societies draw a distinction between transient sexual encounters and those that will be regularized and linked to procreation, and that the latter is often ritually enacted. Whether this last specific generalization is true is entirely an empirical question, and the evidence seems to indicate that it is, at the very least, very common. Moreover, Boyer and Petersen's suggestion that this is a function of evolved cognition is intuitively plausible. In certain respects, their argument resembles traditional functionalist theories that attempt to explain the near-universality of this particular institutional feature of society as a response to functional necessities: humans must procreate, procreation requires protection and socialization, and the marriage/family form is well-suited to this. To be clear, I think that Boyer and Petersen's approach is preferable in this instance, since traditional functionalism faces some difficulties when asked to explain why this particular institutional arrangement might be more functional.

I think that marriage is their ‘best case’, in particular because it links so closely to the core mechanism through which evolution operates: the impact of adaptations on reproductive success. In other words, if there is an institutional setting we would think most likely to have partially evolved roots, it would be that institutional setting that regulates sexuality. Just the same, let us be clear that even in their best case Boyer and Petersen overshoot the actual implications in arguing that general institutional theory is not needed given what evolutionary psychology can tell us. First, evolutionary psychology has so far only given us a plausible theory of why ‘marriage’ (defined, as noted above, in very abstract and general terms) might be common. Second, much about the institutional regulation of sexuality and reproduction is not addressed by their claims (indeed, most of such institutional regulation is not addressed). Indeed, there is far more variation in how marriage and families are structured such that one cannot possibly reduce it all to humanity's shared cognitive architecture. Anyone who reviews anthropology's past disciplinary obsession with the mapping of kinship networks (which, despite common features, do vary markedly) will quickly see this point. Again, my claim is not that Boyer and Petersen are wrong in suggesting that evolved cognition plays a role here. Rather, I argue that (a) they plausibly suggest that it plays a role but (b) they overstate that role in suggesting that evolutionary psychology somehow displaces institutional analysis (or ‘general theories of institutions’).

This becomes much clearer in their other cases, which are weaker to begin with. This is not to say that there are no common features of systems of legal norms, but those which they identify are so general that their own treatment of them can hardly constitute ‘institutional analysis’ at all, at least if we want that analysis to be able to detail the concrete content of specific institutions. One thing is quite clear: legal norms are not ‘intuitively obvious and compelling’ (Boyer and Petersen, Reference Boyer and Petersen2012: 1) across distinct cultural contexts.Footnote ² Legal systems vary dramatically in their structure and norms.Footnote ³ To begin with, as the authors acknowledge, we can only speak of true ‘legal systems’ when we have complex, differentiated societies. This means that for most of human history we did not have them. Moreover, even in the narrow time frame in which we have had them – indeed, even within what we could reasonably speak of as specific legal-institutional traditions (e.g., British common law, Islamic Sharia, the Napoleonic Code and the network of similar legal institutional structures to and through which it diffused, and so forth) – we see widespread variation. It is only by reducing these complex institutions to a small series of vague meta-institutional rules or norms that Boyer and Petersen can suggest that our key task in this connection is explaining their shared features. To be clear, I am not denying the possibility that human beings do have common, biologically pre-programmed moral intuitions. I consider this to be one of the more important, open, empirical questions in the contemporary human sciences. The claim that there are such commonalities is plausible and there is some evidence for it. The claim that there is a clear set of such commonalities, though, and the description of the specific content of those common intuitions, is most emphatically not settled science. Thus we are left with the plausible but incomplete suggestion that humans may have evolved some very basic and general moral intuitions or tendencies of moral cognition, which may explain any discernible common institutional features of legal systems. This much is good. However, given the diversity of actual legal institutions, the conclusion that actually follows here is that what evolutionary psychology can explain about legal systems is quite limited (though, I would stress, important, a point to which I return below), even if the strongest versions of the theory that there are robust, biologically universal features of human morality is ultimately established. Institutional variety and contingency cannot be ignored or explained away on the basis of the presence of common moral intuitions.

I will discuss their arguments with respect to ‘commons management systems’ in less detail. Again, though, the story is basically the same. Boyer and Petersen point to some abstract and general common tendencies, and leave all of the actual, empirically discernible, institutional variation out of the discussion. The story of abstract commonality in some institutional domains is indeed an important one, and the suggestion that it is best explained by evolved cognitive and affective traits is, as I have been emphasizing, highly plausible. However, it is not the main story when it comes to institutions and their functioning.

These are the main implications of their ‘best cases’. However, let us note that the broader argument they make is about institutions in general. Here are some institutions for us to think about, just to remind us of what sorts of things we have in mind: the specific rules and practices of baseball (or any other game, rather than general notions of fairness or competition); the organizational practices of non-profit and for-profit firms (including family businesses, publicly traded companies, and many other varieties), public entities, and public–private partnerships; the techniques practised by salespeople such as car dealers, market hagglers, stock traders, door-to-door salespeople, and other merchants meant to pressure, influence and even manipulate consumers; the Roman Catholic mass, an Islamic Khutbah, or any other, quite distinct, liturgical and ritual forms from different religious traditions; the traditions of ‘trainspotters’ and ‘bird watchers’ (and other such institutionalized hobby groups); traffic rules of various kinds; patterns of housing construction (both the geographical distribution of homes and their styles and features); the formal compositional features of distinct traditions of classical and popular music; the argumentative style of analytical philosophy (and other argumentative styles, ranging from the proverbial to casuistry to scientific modes); the ‘normative structure’ of science (Merton, Reference Merton1979) and other forms of knowledge production; the calculus, and other mathematical techniques and modes of analysis; the literary form of the novel, the epic, the poem, the play, and specific literary forms and devices like iambic pentameter or free verse; the system of Arabic numbers (and other such systems); organizational patterns of firefighters and other types of public service of this sort; styles of combat and warfare, including both the martial arts but also the various organizational and strategic repertoires of large armies; forms of dancing such as ballet, tap and hip hop; forms of swimming (such as ‘freestyle’ or ‘backstroke’); distinct culinary traditions and habits; patterns of deference and systems (and subsystems) of stratification, like the feudal society of orders, the caste society, modern class societies, and so forth; types of states; types of specific bureaucracies within states; and so forth. Of course, the pages of this journal could be filled with a list of this sort, reflecting the diversity of social institutions of which our societies are largely composed, and the notion that all of these specific traditions simply reduce to a series of biologically encoded tendencies and intuitions is doubtful. Any such suggestion may be scientistic but, barring its empirical demonstration, is not scientific.

To be fair, this is not Boyer and Petersen's claim. Rather, it is more like an implicit assumption. The variety and complexity of institutional life fade on their description into vague and abstract common features of a shared psychological substratum. However, it need not do so. As I discuss below, in principle there is no reason why we could not integrate a (culturalist and essentially historical) theory of institutions with the sort of insights from evolutionary psychology to which the authors point. Indeed, I take this to be their ultimate aim (the integration of evolutionary psychology and institutional analysis) and I want to contribute to the effort.

3. Theoretical limitations

In my estimation there are two important theoretical limitations to the approach that Boyer and Petersen put forward. The first is that it is unjustifiably reductionistic. The second is that it ignores empirically observable institutional dynamics. Perhaps it would be better to say that these are two aspects of the same basic theoretical limitation.

Boyer and Petersen's approach is clearly reductionistic. Indeed, the ultimate implication that they draw is that there is no need for a ‘theory of institutions’ since (some) institutions seem to exhibit common features across different cultural and historical contexts. The conclusion, of course, would only follow if they could show that the most important features of most institutions actually are universal. Since, however, they only show that some abstract features of some institutions in several domains are common or shared by many cases, the conclusion clearly does not follow. There is simply no way in which one can derive the concrete content of any specific example of the institutions they consider from the psychology of individual actors. Therefore, while evolutionary psychology (not to mention advances in neuroscience) has indeed led us to the conclusion that some Durkheimian assumptions were wrong (Durkheim, or those influenced by him anyway, clearly understated the extent to which biologically determined psychological qualities manifest themselves in social lifeFootnote ⁴), the fundamental Durkheimian point still stands. That is, social facts are still social, and social life is still an emergent reality sui generis (Durkheim, Reference Durkheim, Lukes and Halls1982, Reference Durkheim, Lukes and Halls1995).

Classic and well-established works in organizational analysis and the new institutionalism allow for a ready demonstration. One consistently observed empirical quality of organizational fields (sets of organizations recognized by actors in their midst as of a type) is that they tend towards both isomorphism and structural decoupling (DiMaggio and Powell, Reference DiMaggio and Powell1983; Meyer and Rowan, Reference Meyer, Rowan, Powell and DiMaggio1991). This means that, in any given organizational field (or, if one prefers, within ‘linked ecologies’ as discussed by Abbott, Reference Abbott2005; see also Fourcade and Khurana, Reference Fourcade and Khurana2011), we tend to see (a) increasing organizational-structural similarity and (b) numerous non-functional (but often shared) structural features. Indeed, the emergence and development of such features can be traced over time (Powell and DiMaggio, Reference Powell and DiMaggio1991; Barman, Reference Barman2007). This work is traditionally understood to have critical implications for narrow economic conceptions of institutions (e.g., Williamson, Reference Williamson1985; North, Reference North1990) that treat them as invariably coordinating mechanisms that reduce transaction costs, lower risk, and so forth. Some institutions are decidedly non-functional and actually increase transaction costs, raise risk and uncertainty, and otherwise cause problems to many actors in their midst.Footnote ⁵ More generally, some institutions do not serve the ostensible aims of the organizational and individual actors to which they are linked (Meyer and Rowan, Reference Meyer, Rowan, Powell and DiMaggio1991). Again, this is not new, but a well-established core finding of sociological institutionalism.Footnote ⁶ Its implication for the discussion at hand is that the processes through which such institutions are created are clearly and irreducibly social. The fact that, say, non-profit organizations in Chicago and San Francisco develop different approaches to ‘donor control’ of the uses to which funds may be put (Barman, Reference Barman2007), is a consequence of social processes within those fields, and not some underlying psychological substratum. And of course much institutional analysis focuses precisely on concrete institutions and organizations of this sort. In short, it is not difficult to select examples of institutions and institutional configurations and to show how some of their key features are products of their social–ecological context.

The reductionism implied by Boyer and Petersen's approach is even more clearly visible if we consider what their account suggests about the causes of any given institution or institutional configuration.

In its barest essence, the causal model they put forward is as follows:

While evolutionary context is the ultimate cause, the proximate cause of any given institution is its fit with evolved cognitive and affective tendencies relevant to the particular domain to which it relates. In other words, the account at least implicitly models institutional development and persistence as consequences of fit between institutions and biologically encoded cognitive and affective tendencies. In other words, in Boyer and Petersen's model institutions that closely align with those tendencies are regarded as ‘legitimate’ and those that do not as ‘illegitimate’. The latter, rather than the former, are expected to persist over time.

This model, like their approach in general, has some intuitive plausibility and is certainly worth taking into consideration. In particular, it is plausible that fit between institutional design/structure and our cognitive and emotional architecture does, at the very least, impose limits on what is institutionally possible or at least likely. For example, if we accept that there may be biologically encoded basic human intuitions about fairness (as hard as this may be to believe upon surveying the historical record of human societies), we might expect them to place some pressure on institutions concerning the distribution of goods and services in society. Of course, the myriad inegalitarian forms of stratification systems that have been institutionalized suggest that any such limiting pressure is relatively weak in the face of power.

This raises a point that would undoubtedly be stressed by historical institutionalists (e.g., Mahoney and Thelen, Reference Mahoney and Thelen2009) and representatives of related views (e.g., Galaskiewicz, Reference Galaskiewicz, DiMaggio and Powell1991). The argument made against a ‘general theory of institutions’ suggests that what really matters about institutions can be explained on the basis of shared psychological tendencies. But what about differences in power and the distinct historical–institutional trajectories to which such differences give rise? All sorts of organizational and institutional arrangements are created and persist not because they comport with our shared expectations and intuitions but because powerful actors impose them. Think of the rise of the modern state, with its imposition of taxes, conscription and frequent ‘total wars’ (Tilly, Reference Tilly1992; Ertman, Reference Ertman1997; Herbst, Reference Herbst2000; Centeno, Reference Centeno2002; Gorski, Reference Gorski2003). Even North et al. (Reference North, Wallis and Weingast2009) see ‘open access orders’ as consequences of elite coalitions deciding to expand access to non-elite actors in order to safeguard their own position from predation.Footnote ⁷ Or consider the basic economic and political institutions that diffused by colonialism to much of the world, which many believe continue to generate dramatic variation in economic performance and democratic quality (Acemoglu et al., Reference Acemoglu, Johnson and Robinson2004; Bernard et al., Reference Bernard, Reenock and Nordstrom2004; Robinson and Acemoglu, Reference Robinson and Acemoglu2008; Lange, Reference Lange2009; Acemoglu and Robinson, Reference Acemoglu and Robinson2010; Mahoney, Reference Mahoney2010; Woodberry, Reference Woodberry and McCleary2011, forthcoming). It should be obvious to all concerned that there is no way that shared psychological factors can account for these patterns and developments. Moreover, they appear to be persistent and consequential.

Boyer and Petersen might respond that they are not against institutional analysis as such, but that they are against a ‘general institutional theory’, preferring instead ‘domain-specific accounts of institutional design in different domains of evolved cognition’ (Reference Boyer and Petersen2012: 1). This gets us closer to a potential point of fruitful convergence between evolutionary psychology and existing institutional theory. However, it does not get us all of the way there. There are two issues here.

First, their reductionism is again evident in their preferred approach. That is, they assume that institutions are essentially epiphenomenal to evolved cognitive tendencies. Only if we make this (untenable) assumption could we justify beginning with cognitive domains and moving on to study institutions. This assumption would also cause us to almost immediately run into empirical problems, since outside of a few examples, there is unlikely to be a general correspondence between such domains and sets of institutions. It is one thing to say that the institution of marriage corresponds to a cognitive domain pertaining to sexuality and reproduction. It is another to presume a priori that cognitive domains link to bounded sets of institutions more generally.

The second issue concerns what we mean by a ‘general theory of institutions’.Footnote ⁸ If all that we mean is institutional theory carried out on the assumption that biology and psychology are irrelevant, then we are in agreement that it is unsatisfactory. But ‘general theory of institutions’ would be an odd name for such approaches. Let us briefly consider one actual ‘general theory of institutions’ to see whether Boyer and Petersen's arguments endanger it in any way. Given its importance and clarity, I will focus on a schematic rendering of John Searle's general theory.

As most readers probably recall, Searle (Reference Searle2010) sees human social life as largely ‘institutional’ in character, and indeed he sees institutions as the distinctive mark of human societies. He defines institutions broadly as sets of ‘constitutive rules’ and he argues that they are called into being through ‘declarative speech acts’. For the purposes of the discussion at hand I have stripped down his approach to its essential features, as the point is not to review Searle's theory but to consider whether Boyer and Petersen's approach truly constitutes a problem for general theories of institutions. It clearly does not, at least with respect to Searle's theory. The fact that one or another domain-specific cognitive or affective tendency might influence institutional ‘design’ does not really have an impact on these more general claims at all. Institutions can be thought of as sets of constitutive rules whether or not these rules are thought to link to evolved cognitive tendencies. They can be thought of as consequences of declarative speech acts, even if evolved cognitive tendencies might influence the probability that any such speech act will be accepted and ‘institutionalized’. So the implications of Boyer and Petersen for the most general theories of institutions do not seem to be what they suggest them to be.Footnote ⁹ We have already seen that they do not seem to compete with theories like historical institutionalism that aim to explain the emergence of specific institutional patterns and configurations either. Indeed, they logically cannot compete with such theories, because they have no way to account for institutional variation.

This seems like pretty strong criticism, and it might leave readers wondering what evolutionary psychology has to say about institutions at all. I think that it actually has a good deal to say, and for this reason it is critically important that its proponents precisely specify what evolutionary psychology might contribute to institutionalism. I have already hinted that the key here is that evolutionary psychology can tell us a great deal about the boundary conditions (Greenfeld, Reference Greenfeld2005; see also Greenfeld, Reference Greenfeld2006) within which institutions develop.

The biological and psychological makeup of human individuals is one of the key ‘boundaries’ of institutional development (another is the physical constitution of their environments). Boundaries of this sort matter a great deal, and psychological ‘boundaries’ more than other sorts. This is because institutions act, in part, through the psychology of individuals. If human beings did not have the capacity to remember, for example, institutions could not exist. If they did not have the capacity to produce and transmit symbols (Searle's ‘declarative speech acts’ being a particular type of symbol construction), they likewise could not exist. Such boundary conditions are easy to understand, visible, as it were, from a great distance and requiring little in the way of sophisticated analysis (institutionalism accounts for them simply by acknowledging their existence as facilitating conditions). Boyer and Petersen, though, and the research which they cite, help us to gain a finer-grained perspective on boundary conditions that until recently have largely been hidden. We need to get still clearer on these points.

Let us return to the example of legal frameworks and moral intuitions. We need to know more about the extent to which human beings exhibit common moral intuitions and the nature of these intuitions. Claims about this must be precisely specified and empirically corroborated. If this level of refinement is attained (and my impression is that this is on the horizon), we would expect commonly held moral intuitions to indeed be a sort of boundary condition for the development of legal institutions (among other sets of institutions). Given the wide range of such actual institutions, we should not assume that legal institutions are simply epiphenomena or reflections of those intuitions, however, which they clearly are not. Rather, we would expect moral intuitions to be one of a number of selective pressures in the establishment of institutions, with the usual suspects (formal institutional design, coalitions of actors with different levels of power, varying levels of resources, and cultural patterns) also exerting selective pressure.

What we really need in this connection, if we are to work together, is (a) this common framing of the problem before us and (b) evolutionary psychologists’ efforts to document and catalogue boundary conditions of this sort.