Nature of Sense Data

We Could Be Wrong about Sense Data

I said that Ayer said that Moore (and others) were wrong to take it to be possible that one might be mistaken about ones own sense data for the following tactical reason: If we can be wrong about sense data, sense data are subject to the argument from illusion to exactly the same extent that material objects are. That means Moore, to discuss our knowledge of sense data, would end up needing to introduce something else (hyper sense data?) that we could not be wrong about, which would make his actual use of the term "sense data" pointless.

In fact, that is not what Ayer says. What he says (75) is

He [Moore] is not entitled to assume that because the distinction between veridical and delusive perception applies to material things, it applies also to sense-data. [sic] For this assumption is inconsistent with the way in which he intends that the word "sense-datum" [sic] should be used.

Sense Data and Idealism

Are sense data physical objects? Obviously not. So, what are they? Berkeley (and, it is likely, Moore) answers, mental objects. Why? That is the only plausible alternative. In Berkeley's famous "Esse est percipi" (to be is to be perceived), what is perceived is (though he doesn't use the term) sense data, and, since sense data are not physical, they must be in the mind, from which Idealism follows (that is, that only ideas---things in the mind---are real). Moore agrees that if sense data are in the mind, Idealism would follow, and so, to argue against idealism, he argues that they are external to the mind---objects of mental acts, not parts of mental acts. Moore engages Berkeley on his own terms.

Ayer points out that the whole debate is based on a mistake: In agreeing to take sense data as the basic objects of perception, we have made sense data prior to both physical objects and mental objects (hallucination). We have the problem of determining how we gain knowledge of external physical objects on the basis of sense data, but we equally have the problem of determining how we gain knowledge of mental objects (hallucinations) on the basis of sense data.

Sense data are obviously not physical objects, and they are, for the same reasons and in the same way, not mental objects. We introduced a new vocabulary of sense data for the express purpose of introducing a way of describing what is perceived that is neutral in the wars between the realists and the idealists. We didn't introduce sense data language as the result of an empirical discovery. Then it would make sense to ask how they fit in with other things we learn about on the basis of experience (Are they physical? Do they persist?). We ask such questions about electrons and unconscious drives, but we should not about terminology introduced that is neutral about all such questions. Ayer: a thing is nothing apart from its actual and possible appearances (sense data) (35). That sounded like an idealist claim, but it only can be an idealist claim if it makes sense to argue that sense data are mental.

Knowledge of Sense Data

Like perceptual terms, the word "knowledge" has many different uses. For present purposes, perhaps the three most important are illustrated by

The kind of knowledge of the external world we are trying to account for is knowledge that: I know that I am sitting on an uncomfortable chair. The basis of that knowledge that, we are all agreed, includes in some important way knowledge of sense data. It is always dangerous to use terms in multiple ways, if we don't want to get confused. Ayer therefore says he will only use "know" in the sense of knowledge that, not in the sense of knowledge of.

It immediately follows that whatever it is we do with sense data that leads us to know that various truth bearers are true, it is not knowledge (that is, not knowledge that). Ayer says, not that we know sense data, but that we are aware of sense data.

Right now, I am aware of laptop sense data, as a result of which, I seem to know that there is a laptop. The argument from illusion suggests that how I know that requires some explanation, but the awareness of the sense data does seem to lead to some knowledge of which I am already certain. Perhaps I might say, for example, "I am now aware of laptop sense data." What is in quotes is a truth bearer and Ayer thinks, and I agree, I know that it is true.

It is claimed that such knowledge that certain propositions are true as a result of awareness of sense data are "indubitable" (not subject to doubt) or "incorrigible" (not subject to correction. There is one other kind of knowledge of propositions that has traditionally been taken to be beyond doubt: knowledge of so-called analytic truths. One traditional version is that a truth is analytic if its negation is self-contradictory. Example: I know that "all bachelors are unmarried." The two kinds of cases are quite different: we know that an analytic truth is true without any need for relevant experience, awareness, perception, whatever. How could we possibly be certain of something that we might never have been able to know?

I think the chair I'm sitting on is blue, but I could not only be wrong about that, I could find out that I was wrong about that: If, when I get up, I discover that the back of my shirt is blue, and the chair looks red, I'll conclude that I was wrong, it wasn't blue after all. How could I possibly discover that I was wrong in thinking that it appeared to me that I was sitting on a blue chair? It is possible: I could be wrong about what the word "blue" means. That seems crazy, but what if I had thought I was sitting on a cyan chair?

I can be wrong about supposedly incorrigible reports about sense data, but I can equally be wrong about analytic truths: had I been wrong about what "bachelor" means, I might have, incorrectly, been thinking that "all bachelors are unmarried." The only way to be wrong, in both cases, is, Ayer says, to be wrong about the language, the meanings of the words.

We can be wrong in reports, and even beliefs, about sense data in a variety of ways: we could lie, deceive ourselves, and be wrong about the uses of language. Nonetheless, we say that sense-data reports are incorrigible because there is a certain way in which we cannot be wrong about them (at least, if we use the terms as Ayer suggests): no subsequent experience can lead us to conclude that we were wrong in making a sense-data report (at least in the sense that I mean). If, for example, I see a black tux and, on going into the sun, I find that it is really navy blue (Father of the Bride?), I conclude that I was wrong in thinking (saying, believing, whatever) that that was a black tux. However, I would not have been wrong in thinking, saying, … that I was seeing black tux sense data, or, in more normal language, that it looked like a black tux to me (assuming, of course, that I know the difference in meaning between "black tux" and "navy blue tux). When we make statements based on observation about physical objects, they commit us to the possibility of other observations. Statements about sense data (perhaps, appearances), bring along no such commitments. No subsequent experience could lead us to correct our statements about sense data, and they are, as we therefore say, "incorrigible."

-- ShaughanLavine - 01 Feb 2007 - 06 Feb 2007

http://dingo.sbs.arizona.edu/~slavine