Sense-Data and Mind Dependence

Here is how the Stanford Encyclopedia (SEP) defines sense data theory: Every case in which we see is a case in which we are aware of sense data. Which are defined by the following:

1. we are directly aware of in perception,
2. is dependent on the mind, and
3. has the properties that perceptually appear to us.

We are only aware of non-mind dependent things by being aware of sense data, which are mental things and could not be other than they appear. I think the view characterized above could use some dusting off, some reconstruction.

In this post I propose to begin doing that. I'd like to present what I take the central claim of sense-data theory to be and work toward refining it.

SEP characterizes sense-data theory as a view in the first instance about seeing. But this might be misleading. Sense-data theory is probably best understood as a theory about the metaphysics of experience. As I understand it, the central claim of the sense-data theorist is this: episodes of experience are to be reduced to episodes of seeing. In particular, episodes of experience are to be reduced to episodes of being sensorily aware (in a visual way) of sense data. So, on the sense-data theorist's view, experiences take their structure from episodes of seeing. Note that this inverts the standard view, according to which seeing and experiencing are rather different things and seeings are, as it were, built out of experiences (plus some other stuff).

At any rate, according to sense data theory canonical types of visual experience are identical to canonical types of seeings. The thesis is supposed to have some kind of explanatory punch. Consider a straightforward case. Child: "Why does it seem like this when I see the giraffe, mommy?" Mommy: "Because, dearest, you bear the seeing relation to a particular that possesses such and such features." Child: "Ah yes ...it is all clear now."

Types of Experience

Imagine seeing a giraffe in Tanner Library. In such a case you'd be subject to some kind of visual experience. It is plausible to suppose that it is possible to be subject to the very same experience if you were hallucinating a giraffe, seeing a similar but distinct giraffe, or misperceiving Rob Gressis as a giraffe. That is, it is plausible to suppose that the class of indistinguishable experiences shared by episodes of seeing, hallucinating, and misperceiving witness a canonical type of experience. Suppose that these possible experiences form a canonical type of experience. The sense data theorist will then go on to claim that in each of these cases tokening the experience consists in tokening the same type of seeing.

Types of Seeing

When I see, I see a particular and its features. Suppose that it is essential to the identity of an episode of seeing that it is a seeing of x and x's properties. Every counterpart of a seeing of x must be a seeing of a counterpart of x and that counterpart's features. And so, if canonical types of experience are to be reduced to canonical types of seeings, we get the following view: To have an experience is essentially to bear the seeing relation to a particular and its features.

Now if the hallucination of the giraffe and the seeing of the giraffe, and the seeing of a distinct but similar giraffe token the same canonical type of experience, then that experience cannot consist in a seeing of the giraffe (or not an actual giraffe anyway) and its properties. Sense data are whatever can fill this role. They're particulars that even hallucinators can bear the seeing relation to. Moreover, they're particulars that can possess the features that are revealed in perception.

The Stanford encyclopedia requires that sense data be mental things. This is probably wrong. One is tempted to suppose that sense data are mental in the way that one is tempted to suppose that abstract objects or fictional entities are mental. ...which is to say, at this point in philosophical history, not very tempted.

Some possibilities:

Sense-data are possibilia.

Sense-data are abstract actual particulars.

One might have general scruples about trade in abstracta. I propose to ignore those scruples. In believing, I bear some relation a proposition. In deliberating, I bear some relation to a rule of inference. In being one of five, I bear some relation to a number. If experience turns out to be an awareness relation to some abstracta... great. I don't see any reason right now why awareness relations must take concreta as their relata.

One worry about this view is that it isn't clear to me that abstracta can bear, for example, color properties or location properties.

Another Worry:

I think we could tell a story about how thinkers could get in the business of trade with certain abstracta. Causation, for example, seems to be a relation between states of affairs, or facts, or events. One attraction of causal theories of content is that they provide a way of seeing how certain kinds of abstracta could enter into the best explanation of our mental lives. By being causally related to facts, your thinking gets to be all fact-y.

But if sense-data are not the kind of abstracta that have clear natural analogs -- as facts are natural, but abstract, analogs of propositions -- then we cannot tell the above kind of story about how our visual systems end up being best described in terms of them.