Notes from J.L. Austin's Sense and Sensibilia (January 2004): Overview: Austin criticizes sense-data theories of perception, specifically how Ayer's presentation of the argument from illusion need not lead us to the acceptance of sense-data. (Austin also criticizes the arguments of Price and Warnock.) Austin is arguing against the position that: We never directly perceive or sense material objects, but only sense-data (our ideas, impressions, sense-perceptions, etc). Sense-data are supposed to be non-material, not part of the material things being observed, and somehow private. Austin attributes many faults in Ayer's argument to verbal fallacies -- misunderstandings and oversimplifications of a few particular words: 'directly', 'real', 'illusion', 'perception', 'looks', 'appears', 'seems', etc. Austin finds further fault in Ayer's repeated false dichotomies -- carving up things into two categories when Austin believes there are far more than two categories. (In some cases Austin accuses Ayer leaves certain things unclassified, while in other cases Ayer lumps together quite dissimilar things in the same category.) Austin's approach to Ayer seems to be primarily linguistic (dare I say even deconstructionist). Chapter 2: Austin points out a contrast Ayer asks us to make between what the ordinary man and (most) philosophers believe or are prepared to admit: The ordinary man believes he perceives material things. Ayer conveniently gives us the easy examples of the familiar objects which ordinary men believe to perceive (books, pens, flowers), thus leading us to think that the only things perceived are material things. Ayer leaves out many more complicated things we might believe to perceive (voices, shadows, mountains, rainbows, vapors). Even though our senses do sometimes deceive us (metaphorically that is), we think we can generally trust our senses. Important Austin claim: Talk of deception only makes sense against a background of general non-deception. It must be possible to recognize a case of deception by checking the odd case against more normal ones. All cases in which we are deceived are not alike; the many different ways of being deceived (optical illusion, seeing a ghost, the headless woman) should not be lumped together. Some philosophers say that ordinary objects are not directly perceived. But what does directly mean? Austin says 'directly perceived' is understood in contrast to 'indirectly perceived'. For vision 'indirectly perceived' means 'with aided vision' (glasses, telescope, microscope, mirror, etc.) or 'as represented by' (blips on a radar screen representing enemy ships). To say "I see indirectly" is a uselessly imprecise thing to say; there are too many different things it could mean to be meaningful without further explanation. (Note that it's unclear what 'indirectly perceived' means for the other senses.) Chapter 3: What is the motivation of the argument from illusion? It is thought (by some, not Austin) to lead to: (1) In some cases (those of illusion), we perceive sense-data, and (2) In all cases, we perceive sense-data. Sense-data is supposes to be that object of which we are directly aware, which is not a part of any material thing. Ayer's primary examples of illusion are refraction (stick immersed in water looks bent), mirage (thinking one perceives something which does not exist), and reflection (something appearing to be in two places at once, with orientation reversed in one). Ayer says that in all of these instances we do see something, and that the something which we are seeing is not what is actually there. We are seeing illusions. Austin identifies a source of confusion in Ayer's argument: Ayer lumps under illusion (that is, treats the same) instances of both illusion and delusion, which are very different things. Delusion suggests conjuring something up, seeing something which isn't there; illusion suggests seeing something real, but seeing it incorrectly, that is, as other than it is. (Austin also points out that we have names for some other things which neither real nor illusion nor delusion, e.g., dreaming and misreading.) Sense-data are a radical solution to something which may not be a problem. We say the stick "looks bent" but we don't necessarily mean there is in some way a "bent stick" being perceived. A stick can be said to look bent without meaning that it looks like a (non-immersed) bent stick. 'Looks', 'appears', and 'see' mean different things in different contexts, and these shouldn't be confused/conflated. Seeing in a mirror is a fundamentally different kind of seeing than seeing on TV than the more "ordinary" seeing. Chapter 4: Austin criticizes Ayer for his use of imprecise language; in particular, for using the words 'looks', 'appears', and 'seems' interchangeably when they are different. Austin analyzes 'looks', 'appears', and 'seems': 'looks' has to do with having a look; 'appears' is used when considering a certain circumstance; 'seems' implies being based on certain, although not conclusive, evidence. Compare "The hill looks steep", "The hill appears steep (from here)", and "The hill seems steep (to judge from the difficulty the hikers are having)." The construction 'looks like' or sometimes just 'looks' can be used figuratively. Consider "the moon looks no bigger than a sixpence". This does not mean the moon literally looks no bigger than a sixpence; it means that the moon when it is seen at a great distance looks about the size as a sixpence at normal distance looks. When we say "the stick looks bent", we don't mean that it is bent, or even that it is really mistakable for a bent sick. Chapter 5: Austin asks: Even if we assume sense-data in the special case of illusion, must we accept sense-data in the more ordinary illusion-free cases? Austin says no. Some flaws in Ayer's argument: Ayer assumes that all perceptions/experiences can be (meaningfully) categorized as either delusive or veridical, that all delusive perceptions are similar (the same) and all veridical perceptions are similar (the same), and that delusive perceptions are far more common than they actually are. Ayer says there is no intrinsic difference (Price says qualitative difference) in kind between our perception of the veridical and our perception of the delusive. Is this claim true? Again, Austin says no. We may use similar words to describe our perceptions, but the experiences are different (and are often identified with such phrases as "dream-like quality"). Yet another flaw in Ayer's argument stems from him taking our fallibility to make some distinctions to imply that there is no distinction to be made. Again, Austin says this is false. Just because we sometimes don't distinguish between illusion and reality (that is, we can be fooled) doesn't mean that the things we fail to distinguish are in fact indistinguishable (incapable of being distinguished). Consider a profession tea-taster (or wine-taster) who can distinguish tastes which many others cannot. If I'm not a wine connoisseur, my claiming "There can't be any difference between the flavors of two different wines because I regularly fail to distinguish between them" is absurd. Chapter 6: Austin considers the difficulties which Ayer himself identifies with the argument from illusion. In Austin's estimation, the difficulties Ayer addresses are not real difficulties. Austin takes Ayer to believe the false claim that the only alternative to always perceiving sense-data is always perceiving material things. Even if we believe that some perception is not direct, Austin does not think we must accept sense-data. Ayer's Berkelian view (according to Austin) is that there are hard empirical facts about -- and only about -- sense-data. Chapter 7: Austin analyzes the word 'real'. 'Real' is a normal word. Some philosophers seem to think they can assign meanings to words as the like, but many of the words they choose to define (e.g., 'real') already have established meanings which shouldn't be ignored. 'Real' does not correspond to a natural quality. We can't identity something (realness?) which 'real ducks', 'real cream', and 'real progress' have in common. (Contrast this with the ease with which we identity what it is that 'yellow ducks' and 'yellow pigs' have in common.) 'Real' is a complex word. Austin surveys some complicated examples. What is the real color of a deep-water fish which appears multi-colored in its preferred habitat, but grayish white under full light? What is the real taste of saccharine? What is the real color of the sky? Of a chameleon? Of a pointillist painting? Of an after-image? 'Real' is substantive-hungry. It only makes sense to say something is real if the answer to the question "A real what?" is understood. Consider "These are pink". It doesn't matter if we are referring to roses or diamonds. But in "These are real" it does make a difference. (A fake diamond can be a real rock and not a real diamond.) 'Real' is a trouser-word; it is identified by its negation. 'To be a real X' gets it's meaning from what it is to be 'not a real X'. 'Real' is a dimension-word. There are many different kinds of real (and corresponding different kinds of not-real), including: genuine, live, true, authentic, and natural. Consider the differences between "That's not a real diamond; it's a fake", and "That's not her real hair color; it's dyed", "That's not real ammunition; it's dummy ammunition". 'Real' is an adjuster word. It helps us adjust the defining attributes of dynamic categories. (This is a bit hard to understand, but one example is "This cat can talk; of course it's not a real cat, but it's cat-like in just about every way, aside from being able to talk.") Chapter 8: Ayer distinguishes qualitatively delusive and existentially delusive. Qualitatively delusive is (roughly) when one of the qualities of the object is under suspicion; existentially delusive is when the existence of any object is under suspicion. Austin criticizes that this distinction fails to address the many instances in which we think we see something, not where there is nothing at all, but where there is something of another kind than we though there was (e.g., mistaking a decoy duck for a real duck). Ayer distinguishes between sense-data of real qualities and sense-data of illusions based on the reliability (predictive value) of those sense-data to identify reality. Austin says this does not serve to give us a general account of real. (Austin thinks no such account is possible.)