The PhilPapers Fallacy (Part 3 of 9)
Table of contents
Part 2: Relevant expertise matters
Part 4: Selection effects matter
Part 5: How PhilPapers respondents interpreted the survey questions matters
Part 7: Philosophical fashions matter
Part 8: Demographic, social, and cultural forces matter
2.2 Causality matters
Studying philosophy could cause a decrease in belief from an initially high baseline, e.g., it could be that 90% of philosophers start out holding a belief, but studying philosophy causes belief to drop to 60%. Or it could be that 5% of philosophers begin as undergraduates endorsing a particular view, but 45% endorse that view on completion of their PhDs.
Thus, when someone appeals to the fact that most philosophers endorse a particular view, this could be taken, all else being equal, as evidence for that view, but all else is never equal in practice, and considerations about the direction of change induced by studying philosophy must also be factored into our consideration. If, for instance, 100% of nonphilosophers were moral realists, but only 62% of philosophers are realists, it would be a bit strange to point to the fact that it’s still a majority, and declare that this is good evidence of moral realism. On the contrary, it could be good evidence against realism. If almost everyone who didn’t study philosophy endorsed a view, if nearly half of them abandoned that view as a result of studying philosophy, it could be that consideration of the arguments weighs very heavily against realism, but isn’t sufficient to overcome a high baseline that biases people towards realism.
This is important, because many of those who would appeal to the 62% majority in favor of moral realism as evidence of moral realism may also think that most nonphilosophers are moral realists. If that’s the case, then it’s consistent with these two beliefs that studying philosophy tends, on average, to reduce belief in moral realism.
This illustrates a more general concern: the fact that most philosophers endorse a particular view does not entail that they think this way as a result of their training and expertise in philosophy; it’s also possible that many philosophers continue to hold a particular view in spite of studying philosophy. We can’t know what proportion falls into which category merely by looking at the raw proportion of philosophers who endorse a given view.
We could obtain some insights from comparing the results of philosophers to nonphilosophers, but there are significant (in fact, I think, insurmountable) difficulties with doing so. Among other difficulties, it’s hard to know whether, if we attempt to survey nonphilosophers, their responses reflect their genuine stance towards the philosophical issue in question, or are instead the result of an unintended interpretation of the stimuli.
My own research suggests that the latter explanation is so frequent for most if not all paradigms in the psychology of metaethics that we know very little about how nonphilosophers think about moral realism and antirealism.
Another problem with appealing to experimental philosophical data is that we may only care about the intuitions of people who report their intuitions under particular conditions, and the particular conditions that we require of those reflective intuitions may only be obtained by what amounts to engaging in philosophy (see e.g., Kauppinen, 2007, who argues for something like this).