Explanation in Cognitive Science

The practitioners of the various cognitive sciences generally construe themselves as engaged in explaining the behavior of human agents. This raises the question of what sort of explanation suffices to explain behavior.

The D-N account, however, fares poorly in characterizing the explan­ations biologists and cognitive scientists offer. The central problem with applying the D-N account to research in cognitive science is the paucity of acknowledged laws within the fields of cognitive science. Perhaps, though, there are laws without their being referred to as such. Indeed, as Cummins (2000) has maintained, psychologists often speak of effects where other scientists might refer to laws. Thus, one finds references to the spacing effect (Ebbinghaus 1885), the serial position effect (Glanzer and Cunitz 1966), and the Garcia effect (Garcia, McGowan, Ervin, and Koelling 1968), where each of these provides a generalization about what happens under specified conditions.

In biology, there is a similar paucity of acknowledged specifically bio­logical laws - a textbook or research report might refer to laws (or equa­tions) from physics and chemistry but not ones specific to biology. A few philosophers have recently followed the lead of biologists themselves, who commonly appeal to mechanisms as providing explanations. These philo­sophers have attempted to explicate the nature of mechanistic explanation and how it figures in biology. Although they vary in the vocabulary used to characterize mechanisms, the basic idea is that a mechanism consists of component parts which perform different operations and that these parts are so organized and the operations orchestrated that the whole mechanism, in the appropriate context, realizes the phenomenon of interest (Bechtel and Richardson 1993, Machamer, Darden, and Craver 2000, Bechtel 2006, Darden 2006, Craver 2007). Thus, to explain how a cell makes protein, one identifies the various components of the cell that are involved (DNA, mRNA, RNA polymerase, ribosomes, etc.), the operations each of them performs (e.g., RNA polymerase creates an mRNA strand from a DNA template), specifies the organization of the parts, and shows how the various operations are orchestrated to produce a protein.

Appeals to mechanisms to provide explanations are equally ubiquitous in the cognitive sciences, and philosophers have begun to analyze the mechanistic models offered in research on vision and memory (Bechtel 2008) and emotion (Thagard 2006). Memory researchers, for example, have both differentiated memory operations and developed accounts of how they are related.

3.