4 - Parsimony in psychology – chimpanzee mind-reading

Summary

Are chimpanzees mind-readers? That is, do they form mental representations of the mental states of others? Or are they just behavior-readers, forming mental representations of the behaviors of others?1 I considered these questions in the previous chapter as a problem of phylogenetic inference. Given that the most recent common ancestor of chimpanzees and human beings existed about 6 million years ago, what can you conclude about whether chimpanzees are mind-readers from the fact that human beings are mind-readers? The answer I defended in the last chapter was: only a little. Modest Reichenbachian assumptions show that there is an evidential connection between us and them, but those modest assumptions do not show that the evidential connection is strong. Can psychological experiments on chimpanzees build a stronger case for the hypothesis that chimpanzees are mind-readers or for the hypothesis that they are not? And how, if at all, is Ockham's razor relevant to the interpretation of those experiments? My goal in this chapter is to answer these questions.

For experimental psychologists, the problem of chimpanzee mind-reading is a problem of blackbox inference (Sober 1998). You observe the environments that chimpanzees occupy and the behaviors they produce. Your task is to figure out what the psychological mechanisms are that mediate the connection between stimulus and response. These internal mechanisms are intervening variables. You don't observe internal processes directly; rather, you need to infer what they are like from what you do observe.

Corresponding to the two inference problems, phylogenetic and blackbox, there are two types of parsimony. In assessing hypotheses for their cladistic parsimony, you see which hypothesis requires the smallest number of evolutionary changes in character state to explain the data at hand. If human beings and chimpanzees both produce a given behavior, and human beings do this by mind-reading, then the most parsimonious hypothesis (in the sense of cladistic parsimony) is that chimpanzees do the same. The hypotheses to which cladistic parsimony applies get represented by phylogenetic trees. Blackbox parsimony doesn't address this evolutionary issue; it sets aside the genealogical relationship of human beings and chimpanzees and evaluates hypotheses about chimpanzee minds just on the basis of what chimpanzees do. Blackbox parsimony applies to this inference problem by evaluating hypotheses that are represented by flow charts that connect stimulus conditions to intervening variables, which in turn are connected to behavioral responses.