Introducing Sense Data

Ways of using "sees"
Straight stick in water
	Veridical	Delusive
Object does exist	I see a straight stick. Straight stick (material object or (part 2) sense data) --- I see a bent stick. "Bent stick" ((part 1) sense data)	I see a bent stick. Straight stick (material object) I see a straight stick that looks bent.
Object doesn't exist		I see a bent stick. No object: There is no bent stick. Mirage is (supposedly) a good case. Pink elephants (hallucinations) are better.

Becky asks about cases in which something is there, but we can't (or at least some of us can't—like color blindness) see it. It won't come up much this semester, since we are concentrating on perception-based knowledge of physical objects, but it is a very important and much-discussed issue, discussed under the heading of "theoretical terms." After all, science tells us that there are all sorts of things around us no one sees: electrons, cells, …. Usually that discussion starts by assuming we know about some things, either sense data or physical objects, more directly, and so that discussion logically comes after the course topic.

There is no phenomenon that can be described using sense data that we cannot describe perfectly well using our ordinary vocabulary:

• I see a straight stick that looks (appears to be) bent.
• I see a bent stick.
• I see a bent stick even though there isn't one.

We have no trouble saying what our experience is like in our ordinary vocabulary, without ever mentioning sense data. So, why bother? As the table makes clear, we use the word "sees" in (at least) three different ways, and we rely on surrounding cues to disambiguate which sense is being employed at any given time. That, ordinarily, works perfectly well. But, since we are using the same words in different ways, it can lead to confusion. The argument from illusion is a case of such confusion: We start with, "I see a bent stick," which is ordinarily used in the sense (delusive, no object), and compare it with "I see a straight stick," used in the sense (veridical, object), and conclude that even though there is no bent stick I must be seeing something, and wind up with sense data. There is, on occasion, confusion outside rarified philosophical contexts as well: "I saw a cat on the sofa." "No you didn't—I don't have a cat." "I know. It was probably just a shadow."

To avoid such confusion as led to the argument from illusion, we will not speak in the ordinary way. We agree, for present peculiar purposes, to only use perceptual vocabulary in one way. What way should we select? The obvious way is, first of all, the normal or ordinary way in which we say, for example, "I see a straight stick," since,

• that is the basic use, the others are alterations of it
• we want to account for our knowledge that there is a straight stick, that is, that is the use of primary interest
• it is simplest use

That is, we agree to always stay in the (veridical, object exists) square. In that square, we need sense data for the bent stick, and we don't need anything other than sense data for the straight stick. If we see a material object in the straight stick case, that is what we are trying to account for or explain, and so we shouldn't start with it. Thus, for our purposes, we agree to say that we always and only see sense data.

Kevin asks about a (straight stick) in water that looks bent but feels straight. How do we describe that in the present mode? We see "bent stick" sense data, but we feel "straight stick" sense data. Seeing and feeling are different senses, and therefore have different sense data as their objects. Since our project is to see how and to what extent we are justified in assembling our perceptual evidence into a picture of an external world, we start this way and investigate the methods rules and justifications by which we assemble the sense data into knowledge of the material world. First, we shall need to investigate what sense data are: are they mental or external to us; can two people experience the same sense data, ….

---

There is no empirical evidence for (or against) the existence of sense data

The traditional Cartesian view is that there are sense data, and that it is a substantive discovery made by philosophers that there are such things: physicists discovered electrons; philosophers discovered sense data. What made the discovery one for philosopher instead of physicists? It isn't based on any experiment, on any empirical discovery. The discoverers of electrons performed experiments that are hard to explain without electrons. The "discoverers" of sense data, though they didn't perform any experiments, do rely on some facts. The facts are just so well known that formal experiment isn't necessary. However, unlike the case of electrons, the data are perfectly easy to explain without sense data. The cases are not analogous. The facts that suggest sense data can be explained perfectly well without sense data. To say that there are sense data is not an empirical discovery, but a linguistic convention we have decided to adopt for certain purposes. Therefore, anyone who tries to argue that there are no sense data is the victim of a misunderstanding. One might reasonably argue that adopting sense data language is a bad idea for our purposes, but that is a very different thing.

The history is that until roughly 1920 most disputants took it to be a question of fact whether or not there are sense data. That won't do, and so the whole issue of how we obtain knowledge of the external world needed to be fundamentally recast. That is an important part of Ayer's project. Carnap did much the same thing in a rather different way around the same time.

-- ShaughanLavine - 18 Jan 2007