Classicists between Wishful Thinking and Neuroscience

At the recent meeting of the Pacific APA, I had an interesting conversation with a prominent philosopher of science who also does some work in cognitive science.  We were talking about whether neuroscience and psychology contrain each other.  I was arguing that they do.  He suggested that perhaps the only constrain is that the algorithms postulated by psychologists must be implemented in the brain, but that is something that we may not need to check, say, for fifty more years.  This is more or less the old Fodorian “autonomy” line, according to which psychologists can do their science without paying attention to neuroscience.

When I objected that you can’t figure out the algorithms without knowing how the brain represents and processes information (because algorithms in the relevant sense are series of operations on representations, so how can you figure out the algorithms unless you know what the representations and operations look like?  But I didn’t get to say that), he interrupted me and said, “I hope it’s not that hard”.

I’m not sure why learning about the brain has to be dismissed as being “that hard”.  I’d say it would make things a bit less hard.  But in any case, his attitude struck me as wishful thinking.

More promising is the merging of classicism (more specifically, Language of Thought theory) and neuroscience proposed by Susan Schneider in her unpublished paper, “The Central System as a Computational System,” which is available on her website and forms a core chapter of her book manuscript, The Language of Thought: New Philosophical Directions.

Schneider argues that you can think of LOT in terms of the “global workspace” proposed by Bernard Baars and other cognitive neuroscientists.  The important point is that Global Workspace Theory (GWT) is supported by neuroscientific evidence, and that Schneider’s version of LOT would find an ally in GWT.  I’d like to see a lot more detail about what the LOT sentences are supposed to look like in terms of neural representations, but at least Schneider’s proposal is a step in the right direction.  If any elements of classicism are going to survive, it’s not going to be by wishing neuroscience to be irrelevant but by integrating classicism with neuroscience.