Case Studies and Theory Development in the Social Sciences - Alexander L. George [77]
The best examples of theories about social causal mechanisms at the frontiers of the observable world are theories on cognitive mechanisms based on experiments on individual decision-making. These include prospect theory, schema theory, and other cognitive theories, as well as rational choice theory. More generally, theories about social mechanisms can be classified as positing agent-to-agent mechanisms (such as theories of persuasive communication, emulation, strategic interaction, collective action, or principle-agent relations) or structure-to-agent theories (evolutionary selection, socialization, and so on). Structural theories must ultimately work through or be consistent with the actions of individuals, but they can be modeled and tested at the macrolevel.284 Often, as we note in Chapter 11, it is useful to develop models that incorporate both agent-centered and structure-centered mechanisms, so that theories can address how certain kinds of agents (personality types, for example) operate in certain kinds of social structures.
Causal Mechanisms, Contexts, and Complexity
Our definition of causal mechanisms states that these mechanisms operate only under certain conditions and that their effects depend on interactions with the other mechanisms that make up these contexts. In other words, a causal mechanism may be necessary, but not sufficient, in an explanation. 285
In this regard, causal mechanisms are consonant with what Paul Humphreys has termed his “aleatory theory” of explanation. In this view, effects are brought about by bundles or configurations of mechanisms, some of which contribute to the effect and some of which may operate to counteract the effect or reduce its magnitude. Aleatory explanations take the form of “Y occurred because of A, despite B,” where A is a set of contributing causes and B is a potentially empty set of counteracting causes. (The set A cannot be empty or we would not have an explanation for the occurrence of Y.) Salmon gives an example, modified from Humphreys, in which a car went off a road at a curve because of excessive speed and the presence of sand on the road and despite clear visibility and an alert driver. He notes that the addition of another mechanism or contextual factor can change a contributing cause to a counteracting one, or vice versa: sand decreases traction on a dry road, but increases traction when there is ice on a road.286 Here again, typological theorizing allows for this kind of interaction, as it can incorporate causal mechanisms that offset one another in some contexts and complement one another in others.
Similarly, Jon Elster discusses a number of psychological theories which posit mechanisms that are in tension with one another, such as the “sour grapes syndrome” in which one’s desires are adjusted in accordance with the means of achieving them, and “the opposite mechanism,” when one wants what one cannot have, precisely because one cannot have it.287 Elster recognizes the challenge presented by such contradictory mechanisms and suggests the need to for identify the different conditions under which each applies: “Moving from a plurality of mechanisms to a unified theory would mean that we should be able to identify in advance the conditions in which one or the other mechanism would be triggered… . My own view is that the social sciences are currently unable to identify such conditions and are likely to remain so forever.”288
This statement underscores that many scholars equate the context-dependence of causal mechanisms with complexity in social relations. Indeed, Elster’s pessimism about the ability to identify the conditions under which mechanisms are triggered is similar to his skepticism on the usefulness of general theories in the social sciences. He argues that “the aim of such theories