Friday, November 16, 2007

The Metaphysics of Ghosts

I've been thinking (again) about why Chalmers's dualism about consciousness dissatisfies me. Apropos of this, a metaphysics of ghosts.

First, let's distinguish between experiential and non-experiential ghosts.

Non-experiential ghosts are, lets say (for now), constituted of non-physical stuff, ectoplasm. They engage in stereotyped, repetitive actions (shaking shackles, gliding down hallways) but don't think and have no conscious experiences. The person whose living form they resemble is dead and gone with no personal psychic connection to the ghost.

Experiential ghosts, in contrast, think and have experiences, maybe engaging in more complex behavior. While there's something it's like to be an experiential ghost, there's nothing it's like to be a non-experiential ghost -- just as there's nothing it's like to be a mirror image or a shadow or a footprint. The following discussion is confined to non-experiential ghosts.

How do non-experiential ghosts come into being and how are they perceived? Here's a theory-sketch: A person (a projector) has a psychological trauma that creates a non-physical ectoplasmic entity resembling her. Once created, this ectoplasmic entity exists independently of the projector's experience. Such ghosts are seen and heard not by reflecting or creating photons or producing sonic vibrations in the air. Rather, they work directly on the perceiver's visual and auditory cortex. This direct action on the brain explains why ghosts cannot be photographed or audiotaped. Call this the ectoplasmic theory of ghosts.

Here's a competing theory -- the materialist theory. There is no ectoplasm. Rather, when a certain sort of trauma occurs, it directly affects the brains of ghost perceivers, through "paranormal" but perfectly physical action at a distance, cutting out the ectoplasmic middleman. A traumatic event in Schnerdfoot's brain as he is murdered directly causes visual and auditory cortical activity in the brains of other perceivers who walk by the scene of Schnerdfoot's death years later.

Now suppose, further, that the materialist theory of ghosts turns out to be true. It seems right to say, then, something like this: Ghosts are really nothing but effects on our brains from earlier trauma in other people's brains. There are no immaterial entities or properties (setting aside any qualms about whether consciousness itself might be immaterial).

Question: Could a philosopher (Chalmers's counterpart?) in Materialist Ghost World run the following argument? I know that it's a law of nature that whenever there's a ghost it's produced by trauma of such-and-such a sort and has such-and-such effects on perceivers' brains. Yet I can conceive of those causes and effects without the presence of a ghost. Therefore, "being haunted" is not the same property as "being a place in which past trauma causes certain effects in perceivers' brains". There's a possible world in which those properties come apart. Furthermore, since I know Schnerdfoot's house is haunted, I know that materialism is false: There are non-physical properties instantiated in my world!

Since it seems wrong to say of the Materialist Ghost World that it is a world in which non-physical properties are instantiated, there must be some flaw in the argument.

I initially conceived this post as a challenge to Chalmers, but now that I've arrived at the end, I've come to think it fails as a challenge. Here's why: Our Materialist Ghost World philosopher must, it seems, either conceive of ghosts ectoplasmically or conceive of them functionally (in terms of their causes and effects). If the first, it's false to say that he knows that ghosts exist in his world. If the second, it's false to say that there is a physically identical possible world that doesn't contain ghosts. It's not clear that Chalmers's argument for dualism fails in the same way, since it's not clear that he has to choose between an ectoplasmic and a functional conception of consciousness.

Next week I'll try another crack at Chalmers -- without the ghosts!


Anonymous said...

Great post, Eric.

Re: "it's not clear that he has to choose between an ectoplasmic and a functional conception of consciousness."

What other choices do you think he has?

Eric Schwitzgebel said...

Thanks, Pete!

I think he can say phenomenal concepts are sui generis, produced by direct acquaintance with the phenomena. (I'm not sure, though, that I fully understand Chalmers's theory of self-knowledge!) It seems less plausible that the concept of a ghost is similarly sui generis.

If I were less sympathetic with Chalmers on this point, I could push the ghost analogy harder.

Anonymous said...


Here's a functional concept of x: the x that produces, by direct acquaintance, y. Here's another one functional concept of x: the x producing by direct acquaintance a sui generous y. I'm having a hard time seeing, then, how this direct acquaintance stuff yields a concept that's non-functional.

Add something about the non-functional essence of x, and maybe you're out of the functional frying pan, but insofar as ectoplasm is simply defined negatively in constrast with functionalizable things, then you land in an ectoplasmic fire.

I'm not sure if the above makes any sense, but I take myself to have blabbed about a similar point over here: [link]

Eric Schwitzgebel said...

Thanks, Pete! Well, here Chalmers gets tricky. Acquaintance is not supposed to be a causal notion and therefore not a functional notion. (How exactly that works, I'm not sure.)

And since ectoplasm is a substance, maybe one can dodge it if you insist there's no distinct substance, just distinct properties? I'm not sure though.

Thanks for the link. You do put the issue very nicely there! That's a tight little dilemma. One possibility, though, is that even if knowledge of the external world depends on stimulation of the sensory periphery, so also might phenomenal knowledge require functional causation -- and in both cases there may still be something beyond that a skeptical scenario involves denying.