What Do Philosophers Think? — Part II

Philosophers tend to develop coherent views of their discipline.

As I wrote in the first part of this essay, philosophy is often accused of never making progress, always asking the same questions in circles. And a popular view is that when philosophers disagree on something, there will be n+1 opinions being put forth about whatever subject matter, where n is the number of philosophers opining about it. My answer to the first charge is coming out next year, in the form of a book on progress in philosophy to be published by the University of Chicago Press. We have been exploring the answer to the second one by way of a close look at a survey of professional philosophers’ opinions conducted by David Bourget and David Chalmers [1].

Last time we have seen some of the basic findings, such as the level of support for competing metaphysical accounts, or ethical frameworks. But also worth noting are what Bourget and Chalmers call “specialization correlations.” Here are some highlights that picked my interest:

  • Philosophy of religion (a somewhat embattled subfield) is more likely to include people who accept theism and who are libertarian (i.e., reject determinism) in matters of free will. The same people are also (slightly) less likely to embrace physicalism in philosophy of mind, or to accept naturalism as a metaphilosophy. None of this is at all surprising, of course.

  • Indeed, most of the strongest correlations between philosophical views and subfields are due to philosophers of religion, with a few others attributable to philosophers of science (who tend to be empiricist rather than rationalists) and scholars interested in ancient philosophy (who tend to adopt virtue ethics rather than deontology or utilitarianism).

Even more fascinating are the pairwise correlations between philosophical views, which means that philosophers tend to develop fairly internally coherent positions across fields. For instance:

  • If one thinks that the distinction between analytic and synthetic truths is solid, then one also tends to accept the idea of a priori knowledge — naturally enough.

  • If a philosopher is a moral realist, she is also likely to be an objectivist about aesthetic value. Interestingly, moral realists also tend to be realists in philosophy of science, and Platonists about abstract objects. It is perfectly sensible to reject moral realism in meta-ethics (44% of philosophers do), but — if one is a moral realist — then one probably should also consistently embrace realism in other areas of philosophy as well, which is exactly what happens according to the data.

  • If one thinks that Star Trek’s Kirk survives teleportation (rather than being killed and replaced by a copy), one also — coherently, again — often adopts a psychological view of personal identity.

As one would find in the natural sciences, there are interesting differences on a given question in the opinions of philosophers who do vs those who do not specialize in the subfield that usually deals with that question. As a scientist, I can certainly have opinions about evolution, climate change and quantum mechanics, but only the first one will be truly informed, since I’m an evolutionary biologist, not an atmospheric or fundamental physicist. So too in philosophy. For instance:

  • Many more philosophers of science adopt a Humean view of natural laws when compared to average philosophers from other disciplines.

  • More metaphysicians are Platonists, though that particular differential is not very high (15%).

  • More epistemologists accept a correspondence theory of truth (again, however, the differential is not high: 12%).

Bourget and Chalmers even explored the relationship between one’s identification with a major philosophical figure and one’s views about certain topics. The results are, again, consistent and perhaps not too surprising, demonstrating that philosophy is not the Wild West of intellectual inquiries:

  • If a philosopher admires Quine, he is less likely to accept the analytic-synthetic distinction (and more likely to reject the possibility of a priori knowledge).

  • Someone who finds kinship with Aristotle is also probably a virtue ethicist.

  • In political philosophy, if John Rawls is your guy, you are more likely to be a communitarian.

  • And it really ought not to be surprising at all that philosophers who like Plato are, well, Platonists about abstract objects.

As I wrote in part I, looking at this data and asking “yes, yes, but which one is the true view of things?” is missing the point entirely. Philosophical accounts, or frameworks (I don’t like the word “theory,” which is more appropriate for what scientists do) aren’t true or false. Rather, they are coherent (or not), useful (or not), illuminating (or not).

Perhaps the most interesting and nuanced approach that Bourget and Chalmers take toward their data unfolds when they move from uni- and bi-variate to multi-variate statistics, in this case factor and principal components analyses. This allows them to examine the many-to-many relationships among variables in their data set.

The first principal component they identify, i.e., the one that explains most of the variance in the sample, they label “Anti-naturalism,” as it groups a number of responses that coherently fall under that position: libertarianism concerning free will, non-physicalism about the mind, theism, non-naturalism as a metaphilosophy, metaphysical possibility of p-zombies, and so-called “further fact” view of personal identity. If one were to plot individual responses along this dimension (which Bourget and Chalmers don’t do, unfortunately), one would see anti-naturalist philosophers clustering at the positive end of it, and naturalist philosophers clustering at the negative end (I would be there, in case you were wondering). It would be interesting to see the actual scatter of data points, to get a better sense of the variation in the sample.

The second-ranked principal component (i.e., the one that explains the second highest amount of variation in the sample) is labeled “Objectivism / Platonism” by the authors, and features positive loadings (i.e., multivariate correlations) of cognitivism in moral judgment, realism in meta-ethics, objectivism about aesthetic value, and of course Platonism about abstract objects.

The third component (i.e., third largest amount of variation explained) is about Rationalism, with positive loadings for the possibility of a priori knowledge, the analytic-synthetic distinction, and rationalism about knowledge.

Two more interesting components (ranked fourth and fifth respectively) concern “Anti-realism” (epistemic conception of truth, anti-realism about scientific theories, idealism or skepticism about the external world, Humean conception of laws of nature, and a Fregean take on proper names) and “Externalism” (externalism about mental content, epistemic justification, and moral motivation, as well as disjunctivism concerning perceptual experience). Finally, we get two additional components that summarize a scatter of other positions.

The overall picture that emerges for the study, then, is very much that of a conceptual landscape with a number of alternative peaks (aporetic clusters, as Nicholas Rescher called them [2]), which are internally coherent and well refined by centuries of philosophical inquiry. I suspect that historically many more “peaks” have been explored and eventually discarded, and that the height of the current peaks (as reflected by the consensus gathered within the relevant epistemic communities) is itself heterogeneous and dynamic, with some becoming more prominent in the landscape and others on their way to secondary status or destined to disappear altogether.

One thing is clear from all this: it is empirically false to say that if there are n philosophers in a room there will be n+1 opinions about a given philosophical issue.


[1] What Do Philosophers Believe?, by D. Bourget and D.J. Chalmers, Philosophical Studies 3:1-36, 2013.

[2] Philosophical disagreements: an essay toward orientational pluralism in metaphilosophy, by N. Rescher, Review of Metaphysics 32:217-251, 1978.