Thursday, November 01, 2007

Virtual models.

This is a statement from Richard Carrier's critique of me:

Cognitive science has established that the brain is a computer that constructs and runs virtual models. All conscious states of mind consist of or connect with one or more virtual models. The relation these virtual models have to the world is that of corresponding or not corresponding to actual systems in the world.

Let's set aside the question of whether the brain has really been shown to be a computer. I take it that this is what a materialist position on intentionality is going to have to look like. But now I have to ask one simple question. In virtue of what is a something in the brain a model of something else, either concrete, or, to make things even harder, abstract. One way in which something can be a model of something else is it is resembles it. Then perhaps a mind is needed to recognize the resemblance, but perhaps one could argue that the modeling relationship exists even with no one there to recognize the modeling. (If a tree models in the forest...) But the trouble is that there is nothing in the squishy grey matter of my brain that resembles, say, a tree in the forest. If I see a green pine tree, where is the green thing in my brain the represents the green pine tree?

Now it is true that some things represent others without physical resemblance. So the word "red" is not red, but it represents red nonetheless. But that relationship seems relative to a mind who uses the term "red" for red things. But this relatioship does not seem to require an independently existing mind.

So, in virtue of what is a virtual model a model.

The link is to Darek Barefoot's excellent reply to Richard Carrier.



At 11/01/2007 11:15:00 PM , Blogger Blue Devil Knight said...

I wish I had Carrier's confidence.

I don't know of any naturalist who holds a resemblance theory. Typically the representation is cashed out in terms of informational or causal interactions (with all sorts of caveats to ensure we don't just represent our eardrums). Like in the honeybee the neural state that represents where the nectar is likely doesn't resemble the location of the nectar at all!

Churchland has a map-based theory of representation, but it quite explicitly not a resemblance based theory. He trashes them, and gives his alternative neuronal map model, in his new book Neurophilosophy at work. If memory serves, this is something like what Carrier has in mind.

I don't know why Carrier calls them virtual models. The models are real, not virtual. Perhaps by 'virtual model' me means that it isn't bound by what is happening in the world, that as in virtual reality we can consider counterfactual worlds or something. Or he is thinking of virtual machines a la Slomaan. It isn't worth the time to figure out what he meant, is my hunch.

At 11/05/2007 03:17:00 PM , Blogger Victor Reppert said...

It seems to me that I am still left with the question "in virtue of what is a model a model?"

Let's think for a minute about maps. Now a map seems to represent the objects mapped in virtue of a structural similarity--thus there seems to have to be a similarity between the structure of the streets of San Diego and the map of San Diego in order for the correspondence to work.

At 11/05/2007 03:38:00 PM , Blogger Blue Devil Knight said...

Victor: you can cash the referential aspect out in terms of information. The neurons that represent two points close in space don't have to be close together as long as they carry information about those points in the world. Information is not sufficient for maps, but for the theorists it is necessary. (You also have to use them to steer behavior at the very least).

Churchland's view is more nuanced, and I recommend his treatment in the book. He distances himself from informational views, for bad reasons I think, and I believe his view is ultimately informational.


Post a Comment

Subscribe to Post Comments [Atom]

<< Home