Anecdote. Austin once gave a lecture in which he attempted to explain why, in many languages, a double negation is an affirmation, but that there is no language in which a double affirmation is a negation. Sidney Morgenbesser piped up, in a sarcastic tone from the back of the room, "yeah, yeah," and then walked out.

Ayer, in section 10, Sentences, Propositions, and Facts, denies that there are meanings. He doesn’t deny that words have meanings, rather he denies that there is a single kind of thing that is the meaning of a word.

“I shall not cast about for further answers to the question, What do symbols mean? For the view that I am taking is that the reason why this problem appears to defy solution is that there is really no problem to solve. … I cannot deal &helllip; with the general case, for the simple reason that there is no general usage to explain. There is no one thing that all symbols mean.” pp. 97–98

The argument is of the same type that Austin is using against Ayer: There is no one thing that we perceive. In many cases, the question, What is being perceived has an answer, but there is, according to Austin, no general usage to explain.

I think this is worth pointing out because, first of all, it may seem that what Austin is doing is radically different from and incomparable to what Ayer is doing, and second, that impression is reinforced by the story that Ayer is an Analytic Philosopher and Austin is an Ordinary Language Philosopher. There are differences in temperament and method, but they are not so extreme as they have often been made to appear. I suspect that most of the more contemporary philosophy you have read in various classes is more in the analytic mode, and so it is important to recognize that, while there is nothing wrong with such philosophy (It’s what I do too.), one must be very careful that the distinctions being made are appropriate distinctions and, especially, not to fall into the trap of false dichotomies, that is into thinking that everything that is not like the main case or cases you have in mind is as a result all similar. That is rarely the case, but very easy to unconsciously assume.

“Ordinary language philosophy” is really best defined as doing philosophy as Austin did it, but it is more usually characterized as the philosophical attempt to discover how words are ordinarily used, instead of imposing technical uses on them.

There Are Sense Data

Last class, we discussed "material objects" the foil of sense data. Now, we turn to what Austin calls the argument from illusion, which Ayer calls the first part of the argument from illusion, which is intended to show that there are sense data. The second part, which Austin just takes to be a second argument is showing that all we see is sense data. We discuss that next class.

The main topic of Chapter 3 of Sense and Sensibilia is how the words "illusion" and "delusion" get abused in the argument from illusion. The abuse is by no means original with Ayer, but he provides a good example. We are told, in the course of the presentation of the argument from illusion that when we are subject to a perceptual illusion our perception is “delusive.” That is odd in many ways—delusive is a strange word, and if you’re willing to live with it, why not say, more plausibly, that our perception is “illusive.” It seems like a possible extension of the way we ordinarily coin words to go from, "that is a perceptual illusion" to, "that perception is illusive," but "delusive is a leap." Why does that matter?

We are familiar with visual illusions, and maybe auditory illusions, but it is not even clear what an olfactory illusion would be. Serena suggests that smelling perfume that smells like oranges and thinking you were smelling oranges might be a case. I agree that that is probably the most plausible kind of case (it’s better than any I came up with), but it still isn’t clearly an illusion—one has a mistaken belief about the source of the smell, but not there is no perceptual mistake about how it smells. Compare the Mueller-Lyer illusion:

More optical illusions.

Here, the lengths, which are the kind of thing we see, look different, although one can discover that they are the same. There is nothing like that in the smell case. Well, here’s an even better example: Becky gives the example of someone who is losing their sense of smell and gets smells wrong. Maybe that’s a case of illusion, I’m inclined to think it probably is, but the point remains that in the case of vision we have a well-established usage and standard examples, in hearing, less so, and in the case of smell it isn’t quite obvious what to say. (It is also not clear what would count as a tactile illusion. Ayer gives that example that a coin feels larger on your tongue than on your palm.) My lingering doubts about Becky’s example have to do, not with the example, but with what an illusion is—we don’t have a shared definition that readily applies to such cases. Illusions are most characteristically set up by magicians and perceptual psychologists, not something that occurs naturally.

Ayer takes lots of ordinary types of experience to involve illusion when they do not—coins rarely if ever look elliptical—and so he makes a certain kind of phenomenon, illusion, look much more important than it is, and then he makes matters worse by calling the illusions delusions. In the case of an illusion you are actually seeing something, though you have something wrong, but in the case of delusion, you aren’t seeing anything at all, at least in the standard cases of delusion ("delusions of grandeur," "paranoid delusions"). Ayer therefore slips in, without argument or discussion, the idea that we frequently see things that aren’t there. That is necessary to make the argument from illusion look plausible, but it is just a rhetorical trick. The argument is not at all convincing when looked at carefully.

There is an additional trick, which is moving from the rather long list of cases Ayer uses to introduce the idea that we are frequently subject to illusions to a few cases that are particularly good for his purposes.

The straight stick that, in water, looks bent is one of the examples Ayer treats in detail and comes back to through the whole book. As Austin says, Ayer "makes the assumptions" that the stick does not really change its shape when placed in water and that it is not both crooked and straight. That is bizarre. Those aren't "assumptions," those are the plain, well-known, uncontested facts. He takes it to follow that one of the ways the stick looks is "delusive." Thus, at least one of the times, the appearance is "delusive." The suggestion, made without argument, that there is a delusion, that we are seeing something that isn't there, softens us up for the claim that what we see is sense data. Not only is there no delusion involved in seeing a straight stick in water, there isn't even an illusion: you can see the water. Even little kids don't get fooled, because the water ripples, and the part under the water is obviously distorted.

Once the con has worked on us, and we are willing to accept that all we see is sense data, Ayer analyzes different uses of perceptual verbs. (We sometimes use "to see" in such a way that it must be veridical, —.) That brings us to the next Chapter.

Looking, seeing, appearing, seeming, like

Ayer, p. 58: The general rule … is that the propositions we ordinarily express by saying that a person A is perceiving a material thing M, which appears to him to have quality x, may be expressed in sense datum terminology by saying that A is seeing a sense datum s, which really has the quality x, and which belongs to M.

That translation method is supposed by Ayer to tell us everything it is necessary to know about what sense data are. It isn’t too hard to see that there are problems:

translates, according to Ayer, into

That translation is not at all what Ayer wants. He needs, at least, to exclude being a material object from the list of qualities he allows. It isn’t too hard to come up with other examples: located about a foot in front of my eyes.

Aside: Husserl and the phenomenologists do take such things to be qualities of our perceptions.

So the “definition” of sense data in fact relies on some unstated understanding of what "appears" means and what sorts of qualities a sense datum may have. The definition at first looks pretty clear because of the introduction of sense data via the argument from illusion. Austin’s point that the argument plays a real role in Ayer’s philosophy begins to become clear.

I’ve just attacked the idea that we know what an appropriate ("perceptual") quality is. Austin suggests a problem of that sort with his examples: Is chicness a quality? But he doesn’t tackle that issue head on, probably because quality is a philosophical term, and Ayer is entitled to torture it into any meaning he likes. But, “appears to him to have” is a term of ordinary language, and Ayer seems to rely on our ordinary understanding of it, which he takes to be something like this:

Austin attacks the idea that that provides a clear criterion for what sense data are like by pointing out that the way something is, the way it really is, the way it looks, what it seems to be, what it looks like, and how it appears are 6 different things, not two. Therefore Ayer’s “characterization” of sense data is, from Austin’s perspective, hopeless.

Austin employs an important rhetorical strategy: since he never explicitly formulates his conclusions, they are virtually impossible to attack. Many philosophers have become very influential that way. Wittgenstein is perhaps the most important modern example in Anglo-American philosophy. The tactic isn’t always cheating: sometimes a well-targeted example really is more valuable than a precisely formulated conclusion.

You can look guilty without seeming to be guilty. You can say, for example, "he doesn’t seem to be guilty, but its likely that he’ll be convicted because he looked guilty during his testimony." “Seemed” has to do with evidence or background expectations in a way that “looks” does not. Appearances can be deceiving (indeed that is the main motivation for the argument from illusion), but can looks be deceiving? It appears to be elliptical though it is round. It looks elliptical though it is round. The first suggests that you might have thought that it actually was elliptical, the second does not. Indeed, you might explain to someone who thought it was elliptical that though it looks elliptical it is round. The way something appears might or might not be how it is. The way something looks is distinct from how it is. "He appears honest" suggests wondering whether or not he is honest. "She looks honest" suggests that she is not.

Here is clear case that shows that the words are used differently, though it doesn’t seem to me to be helpful in figuring out how:

"She looks good" is a comment on her appearance. "She appears good" is not.

Why do such questions of ordinary usage matter for philosophical purposes? It pretty clearly matters for purposes of composition, but so what? If we aren’t clear on what appearances are, we aren’t clear on what sense data are being presented to us, and so the idea that descriptions in sense datum language are incorrigible becomes much less plausible. It isn’t immediate what the sense data are in a given perceptual situation.

Example: Looking at a duck-rabbit, look first at the duck, then at the rabbit, then at the lines and dots. Duck Rabbit Are these perceptions of similar sense data or not? Well, they are perceptions of the same thing. They seem different. Do they look different or appear different? I don’t know. If I could describe to you what I saw looking at a duck-rabbit, would you be able to tell on the basis of the description alone whether I took myself to be seeing a picture of a duck, a rabbit, or mere lines and dots? I don't think that question has an answer, and, similarly, I don't think the parallel question about a complete description of my sense data while looking at a duck-rabbit has an answer either. In that case, knowing more about what count as perceptual qualities might help, but any answer to that question will get Ayer into trouble. (Thompson Clarke has tried to confront a problem of that sort in a serious way, and he ends up discussing, not sense data, but ostensible objects.)

The conclusion is that the plausibility of the idea that we can take ourselves to be perceiving one kind of appearance is a philosophical trick conjured by, on the one hand, contrasting what we appear to see with one kind of thing that we really see—material objects—and on the other hand, by acting like what things look like, how they seem to be, and how they appear are all one and the same thing.

But isn't there some kind of problem about how we come to know that there are things (of various kinds) on the basis of our senses?

If what you interested in is the cognitive psychological or neurological problem of what we (or at least our brains) do with the data from the retina, that is relatively clear. However, the idea that there is some preconceptual kind of thing that we see, that we do still see from which we make up the world is just wrong.

The question of finding out how we get the theory of material objects (including brains and retinas) in the first place from something prior to the way in which we see the world (as made up of lots of different kinds of things) is just a mistake.

The idea that there is such a question is gripping, but it is historically based on the theory of retinas, optic nerves, and the like: The first person to pose the problem was Descartes, and what he was doing was trying to integrate a new post-Aristotelian science and philosophy of the world.

How things look has a lot to do with what we think they are. It isn’t even clear that this works for firings of the nerves in the retina, or even, though this is comparatively good, the light impinging on the retina. Now try to tell me what sense data are? Even the sense data in a single artificially simplified experience? Ayer talks about pen-like sense data as if they are what lead one to the perception of a pen, but the same configuration of stuff looks different if you don’t think it is a pen.

-- ShaughanLavine - 4 April 2003 - 7 April 2003 - 29 Mar 2007