• Choose a category
• All
• Classic-Fiction

By Root 1437 0

must a causal mechanism involve the irreducibly smallest link between one entity and another, and at what point does inquiry into causal mechanisms stop? On this question it is useful to think of the frontiers of research as a potentially movable border between the observable world and the unobservable ontological level where causal mechanisms reside. This is most evident in the physical sciences. At one point in history, there were no microscopes to allow scientists to examine molecules or their observable implications, so it would not have been unreasonable to question the existence of “molecules.” Once new instruments allowed scientists to observe molecules and their implications, it became unreasonable to disbelieve the broad outlines of the molecule model, and debates (and instruments of observation) moved on to examine the nature of atoms and then subatomic particles. Theoretical arguments in physics now focus on whether it is reasonable to believe in the esoteric mathematics of string theory, which posits the existence of additional dimensions for which there are as yet no readily observable implications, or whether inquiry can or must end with the seemingly impenetrable observational barrier that surrounds the odd implications of quantum effects.281

No matter how far down we push the border between the observable and the unobservable, some irreducibly unobservable aspect of causal mechanisms remains. At the frontier of our knowledge at any given time, our theoretical commitment to molecules, atoms, quantum mechanics, or string theory resembles an “as if” assumption about the underlying mechanism at the next level down. In this sense, the causal mechanism view, like the D-N model, ultimately does not offer an explanation of laws themselves at the frontiers of our knowledge. Unlike the D-N model, however, the causal mechanism model, at every point up to the potentially movable border of the unobservable, explains hypotheses or laws with reference to observable implications on underlying processes at a lower level of analysis.

Thus, the commitment to explanation via mechanisms differs from more general “as if” assumptions in that it pushes inquiry to the outer boundaries of what is observable and urges us to expand those boundaries rather than stop with demonstrably false “as if “ assumptions at higher levels of analysis.282 This process is less obvious in the social sci-ences, but even here new instruments of observation and measurement at the macrolevel (public opinion polls, measures of GNP, and so on) and the microlevel (evidence on cognitive processes within the brain) are expanding the boundaries of what we can observe.283

The formulation of hypotheses about a particular causal mechanism, and the decision on whether to model these hypotheses at the micro- or macrolevel, is a theory-building choice of the investigator. At the microlevel, this choice is influenced by the state of knowledge about the causal process that is operating and the limits to observation posed by extant instruments for data collection. The formulation of a given hypothesis is provisional, being reformulated if we acquire additional information about the causal process at a lower microlevel or new tools that provide finer-grained observations. As in the example above of our changing understanding of how smoking causes cancer, “explanations” that are satisfactory at one point will later come to be considered insufficiently precise as new evidence becomes available at lower levels of analysis.

New insights about underlying mechanisms ideally take the form of simple and widely generalizable models, as is sometimes true in the physical sciences. In the social sciences, however, models built on detailed observations often take the form of complex and contingent generalizations (or middle-range theories) that describe a smaller subset of a phenomenon with a higher degree of precision or probability. Rational choice theorists, among others, emphasize the need for microlevel mechanisms at the level of individuals, but often argue that rational