sociology might pursue, and where this agenda fits with the main concerns of sociology and economics.
In doing so, we should keep in mind that the production and distribution of goods and services is just one institutional complex of activities, and that the arguments appropriate to them should have some generic similarity to arguments we might develop to explain political action, science and knowledge, family and kinship, and other persistent social patterns. Thinking about how the sociology of the economy is similar to and different from that of other institutions helps us see what kinds of arguments will work best.
We may begin by asking what is distinctive about economic sociology as a way to explain the economy. In part this depends on one’s concept of “distinctive”. One way analysis of the economy is different from that of some other institutions is that it is largely dominated by a particular academic discipline, economics, which is focused theoretically on concepts of rational or instrumental action, and where “methodological individualism” roots all explanation in the activity of concrete persons. Though sociology should develop its own agenda and argument, rather than react to neoclassical economic analysis, concepts can be sharpened by clarifying where they stand in relation to those developed by economists. A unified theory should build on what both have accomplished.
I argue that there are two very general ways in which the instrumental-reductionist vision is theoretically incomplete, that suggest what distinctive explanatory improvements economic sociology can offer. The first is that any account of human interaction which limits explanation to individual interests abstracts away from fundamental aspects of relationships which characterize economic as well as any other action. In particular, horizontal relationships may involve trust and cooperation, and vertical relationships power and compliance, well beyond what individuals’ incentives can explain. Trust and power drive a wedge between interests and action. And this happens in part because norms and identities result from and structure interaction in cognitive and emotional ways that escape reduction to self-interest, and indeed are key in actors’ definitions of what their interests are.
The second problem for reductionist accounts is that even though we see some spaces where one may adequately explain outcomes by a purely interest-driven model, there is rarely any simple reduction to individual action that can explain how such spaces evolved as they did, with the constraints and incentives that individuals find themselves acting out. In fact, this is a corollary to a more general argument that action driven by interests as well as that driven by trust or power, occur and have outcomes in ways determined by larger contexts than those in which they are located. I mention the situation of interest-driven behavior first only because it is more typically analyzed as context-free.
This second point does not privilege structure over agency, as individuals who find themselves in situations determined by forces beyond their control, and often far beyond their lifespan, may nevertheless turn these situations to advantage and make a deep imprint on future actions and institutions. For example, though it was quite late in the game before Cosimo de Medici could “suddenly apprehend the political capacity of the social network machine that lay at his fingertips”, which he had done little to create, once he did, he dramatically changed the course of Florentine history for many generations.
In practice, these two problems typically occur in the same cases, and though one should separate them analytically, it is hard and somewhat artificial to do so for any particular instance, and I do not succeed very well at it in what follows. But I will try to focus first on the mixed sources of action within confined social spaces of the sort that White (1992) referred to as “molecules”, before moving in the following section to how such molecules are constituted.
To illustrate the first point — inadequacy of a purely interest-based argument — I begin with information flow. Economic sociology has made major contributions to understanding this flow through social networks in labor markets, and within and between organizations. One way to apply this understanding is to adapt it for instrumental argument about how best to manage one’s networks. Not only economists, but also sociologists have done so. These models follow a rational choice approach to understand information flow through social networks, and make contributions that are valuable but not wholly distinctive from that of economics.
But even for this apparently tractable case, it is difficult to stay within a simple framework of instrumental rationality. My study of job information flow, for example, made clear that it is often profoundly misleading to think of the acquisition of such information as the result of “investment” in contacts. People want sociability and hope to be liked, approved and admired by others. Insincere approval is better than none (as those who encourage sycophants well know), but pales in comparison to approval without ulterior motive. Though some “investors” in social relations may achieve great skill in simulating sincerity, as shown by the success of “confidence rackets”, the desire of recipients for true approval, and the vigilance of most in ferreting out its opposite, sharply bound the role of calculated instrumentality in social life.
So economic sociology can make a first contribution to understanding the economy by calling attention to the mixture of economic and social motives that people pursue while engaged in production, consumption or distribution. But there is more to say here, that involves the contexts of social interaction, and how they arise. People typically pursue multiple purposes simultaneously in intersecting social formations. For example, they go to parties with nothing more in mind than a good time. It seems implausible to consider this economically instrumental behavior, as the component of expected economic gain from loud and intense socializing is small to vanishing, thus unlikely to be anyone’s main reason for attending. And yet, information about jobs does pass among partygoers. The point is that the separate institutions of labor markets and expressive socialization routines intersect in ways that cannot be accounted for by the incentives of individuals. This links to the second problem for instrumental theory that I mentioned in the previous section, on how the contexts of action arise, and I will take up the point in more generality below when talking about the intersection of spheres and institutions.
We may summarize the argument so far by saying that in social interaction, people have mixtures of motives and consequently act in ways difficult to describe in terms of pure self-interest. Sociology has expanded on this point by considering how particular kinds of social relations make behavior diverge from the narrowly instrumental. To cut through a vast theoretical underbrush, I simply distinguish here between horizontal and vertical relations, and their impact on this divergence.
Analysis of horizontal (non-hierarchical) relations leads to discussions of “trust” or “solidarity”– states of relationships or groups that lead to cooperation beyond that to be expected from decision dilemmas such as the “free-rider problem” or the “Prisoners’ Dilemma”. Vertical (hierarchical) ties are defined by a quality of these relations that we refer to as “power”, to be distinguished from solidarity or trust. The behavioral consequences of power are domination and compliance; these are parallel to cooperation, the behavioral consequence of trust or solidarity.
Trust and power open a wedge between behavior and incentives that instrumental theorists try hard to close. Their efforts are strenuous, because problems of trust and cooperation, and of power and compliance, pose difficult challenges for any theory based wholly on rational choice and self-interest. I will challenge attempts to bring these within the orbit of such theory, but these attempts, even if successful, would still leave much of social life and economic action unexplained, since it would neglect how the larger social setting determines the parameters within which self-interest was defined, a matter that I take up later.
Perhaps the most ink has been spilled on problems of trust and the cooperation that flows from it. The issue became especially pertinent to arguments about rational choice when theorists pointed out paradoxes of rationality – interacting individuals, rationally pursuing their own goals, achieved results worse than if each had adopted a suboptimal strategy, as in the famous “prisoners’ dilemmas”. Mancur Olson’s The Logic of Collective Action (1965) applied this argument to political theory by pointing out that cooperation to achieve mutually shared goals would be derailed by genuinely rational actors, since each would try to “free ride”. This chilling discovery parallels that ten years later by Oliver Williamson of the likelihood in market relations of “opportunism” – the alloying of simple self-interest with “guile”. These discoveries ended the long era in instrumental theory dominated by what is called the idea of doux commerce, stemming from the time of Montesquieu, that rational action and exchange transformed people into gentlemen, who automatically followed the rules of the game and were trustworthy despite incentives to the contrary. This oversocialized conception gave way rather suddenly to a neo-Hobbesian conception of market relations as nasty, brutish and short – likely only in fact if people were anonymous atoms in relation to one another, and highly undersocialized
For in practice, decision dilemmas and opportunism are overcome if the participants, to use the usual language of everyday life, “trust” one another. Separated prisoners both deny the crime despite the dominant solution, because each trusts the other to do the same. Trust thus leads to an outcome better for the collectivity. But no rational account explains why prisoners would do such a thing; trust means precisely that each expects the other to act against her own interest, as defined by the payoff matrix.
Most instrumentalist literature on trust consists of elaborate efforts to deny the data of everyday life and rescue “trust” from its usual meaning, by explaining that actors only trust each other when incentives are properly aligned so that the trust is reasonable. It is said that we “trust” companies’ promise to repay a debt if a rating company has assigned a AAA to its debt obligation, or that we can trust the other prisoner if the game is repeated ad infinitum. Companies keep good faith and avoid default because loss of a high rating is expensive, and prisoners deny because in the long run , they would be wiped out by their own malfeasance if they didn’t.
While all these phenomena are undeniable, they are hardly conclusive. People often act with confidence because they expect others’ incentives to point them in the right direction, but the commonsense meaning of “trust” is that we expect good behavior of others in spite of their incentives; and such trust is vital for the conduct of social and economic life. If everyone assumed that others merely “did the right thing” because of incentives, economic life would be poisoned by incessant attempts to conceal the true incentive situation for one’s own advantage. In fact, the well is not invariably poisoned. In most situations economic actors drink freely despite incentives for their counterparts to act worse than they actually do. Thus, one central task of economic sociology is to lay bare the circumstances under which people may safely set aside suspicions that rational action would require them to have. By definition, such a task cannot be conceived or implemented from within a theory of behavior that admits only of rational action.
Though I emphasize the divergence from self-interest resulting from trust in horizontal relations, dislike and corresponding distrust and failure to cooperate are equally important and just the flip side of this argument, even though negative relations are rarely integrated into social theory. These divergent networks resulted from low intermarriage rates among different groups of supporters, and they comment that there is “no particular mystery” about this, since “patrician and new men supporters despised each other. Status-conscious patricians …usually would not dream of sullying their own honor by marrying into new men families”. Yet, it was just this separation that gave the Medicis so much leverage in relation to supporters who could not unite.
The concepts of “trust” or “distrust” refer to horizontal relations in which neither party can dictate to the other what she must do. Much of the discussion of trust and the cooperation that flows from it has a parallel in discussions of power and the compliance that flows from it, where the issues are transposed from horizontal and symmetrical to vertical and asymmetrical relations. At all scales in economic action and institutions, people comply, at times, with what they understand others want them to do. Unlike the language of “trust”, however, we have no clearcut usage to demarcate compliance based on incentives from that rooted in other elements of relationships and institutions.
Yet this issue is commonly understood to be important. In trying to carve out a distinctive niche for the concept of “social exchange”, rules out situations where compliance is not plausibly construed as voluntary – as when a thief offers the choice: “your money or your life”. Max Weber classifies types of power in similar ways. For him the least interesting case, which he discusses only briefly, is power based on a “constellation of interests”, such as a monopoly position in the economy, though it obviously is important in its own right. Correspondingly, he notes almost in passing that to run a civil administration on the basis of incessant coercion is too expensive and unwieldy for any but the most unimaginative to pursue. Instead, most of his analysis of distinct historical formations dissects the different circumstances under which people consider it appropriate to follow instructions given by someone in an authority position over them – the “types of legitimate authority”.
This distinctively sociological argument about compliance and legitimacy, which can be made in industrial organizations as well as states leads us to observe that one reason it is artificial to consider either cooperation or subordination as always reflecting pursuit of self-interest is that in most circumstances, actors have definite conceptions of what action is appropriate, and these shared norms or conventions of action, constructed, learned and absorbed within social groups, are explicitly construed by actors as not being matters of self-interest. This is a core part of the meaning of such norms.
Most sociologists have veered away from theoretical argument based on actors’ shared value commitments because of the excesses of mid-twentieth century sociology. This view, which has been called “oversocialized”, leaped from observing that such commitments were a significant force in social life to the conclusion that all social action flowed from them. The opposite extreme is to imagine that moral sense about the economy is entirely subordinated to and derivative from some teleological quest for efficiency pursued by social systems, so that observed norms, though admitted to be important, can be assumed to have been selected out for their economic efficiency. The time has come to find a balanced account, to acknowledge the importance of such norms and conventions, while fitting them into a broader frame of social theory.
For the economy, a beginning of this more balanced account is offered by historian E.P. Thompson in his landmark 1971 essay on 18th century English crowds, and the meaning of their frequent collective action to protest practices concerning the movement of essential foods such as grains. Thompson’s point was twofold. One was that what appeared to be mass hysteria and highly non-rational crowd action, could upon closer observation be seen as part of an organized and sensible campaign with goals easily understood in instrumental terms. But Thompson did not stop at this, which could be well fit into a rational choice framework; instead he insisted that quite aside from sensible goals, people in these crowds were also heavily animated by outrage at the way economic actors pursued their activity. They had definite beliefs about what were legitimate and non-legitimate actions, based on some sense of what individuals owed to the collectivities in which they were embedded – what he called their sense of “moral economy”.
Note that in this formulation, Thompson, dealing as he was with hungry people, was not likely to fall into the expansionist reductionism of normative hegemony – to claim that norms and values alone motivated his actors. The purely instrumental aspect of food riots is tough to miss, as crowds overturned and looted carts of bread headed for distant markets, and this led him implicitly to a formulation in which norms, identities and instrumental rationality jointly motivate and shape action. To discover what the process is by which these perhaps incommensurable motivations act together is not part of his analysis, but would have to be part of any general explanatory scheme in economic sociology.
