PHIL2606: Knowledge, Reason and Action
Week 12: Perception and Knowledge

    Perception as a source of knowledge

  1. We are considering the epistemological question:

    EQ: How do we get knowledge via visual perception?

  2. Our approach is to consider the closely related metaphysical question:

    MQ: What is the nature of visual perception?

  3. A space of answers

  4. So what is the nature of visual perception? What is it for someone, say Rob, to see something, say a bottle?
  5. One very natural answer is this: for Rob to see a bottle is for Rob to stand in a certain relation of direct awareness with the bottle, and also with its various properties (its colour, its shape, etc.). Why natural? Because that's what seems to be going on when we see.

    This view is often called direct realism. Why ‘direct’? Because it claims that for Rob to see a bottle is for him to be directly aware of the bottle, rather than indirectly, in virtue of being aware of something else. Why ‘realism’? Because it allows that the bottle of which Rob is directly aware is a mind-independent object (real) rather than a mind-dependent object (ideal).

  6. Does this help with EQ? Not obviously: Rob might be directly aware of a bottle, and of its whiteness, without getting knowledge that the bottle is white. Direct awareness is not enough – something else is needed to get knowledge via perception that the bottle is white.
  7. Furthermore, suppose that direct awareness were enough to get knowledge. Rob might know via perception that the bottle has been used (he might see that it has been used). Is this because he is directly aware of the bottle's usedness? If not, then how does he get that knowledge via perception?
  8. There are also some famous problems for direct realism, to do with the nature of the visual experiences that accompany visual perception:
  9. We will consider five alternatives to direct realism, that purport to avoid these problems:
  10. Phenomenalism. When Rob sees a bottle, he is directly aware of certain mind-dependent things. But he is also directly aware of the bottle, because these mind-dependent things are the bottle. The bottle is a mind-dependent object.
  11. Indirect realism. When Rob sees a bottle, he is aware of the bottle, and of its properties, but not directly. He is directly aware of certain mind-dependent objects, ‘sense-data’, and thereby indirectly aware of the bottle and its properties (because the sense-data are appropriately connected to the bottle and its properties).
  12. Adverbialism. When Rob sees a bottle, he is involved in a certain kind of event (a seeing event) that is occurring in a certain way. If the bottle looks white to him, then the event is occurring whitely (he sees it whitely); if it looks red to him, then it is occurring redly (he sees it redly). These ‘ways of seeing’ are properties of seeing events.

    Adverbialism might allow that Rob is directly aware of a bottle, but does it allow that he is directly aware of its properties? It seems, rather, that he is directly aware of ways of seeing the bottle (and they are not properties of the bottle).

  13. Disjunctivism. When Rob veridically sees a bottle, he is directly aware of the bottle and its properties. In every other case he is not. This retains some of the flavour of direct realism.
  14. Representationalism. When Rob sees a bottle, his visual experience represents the bottle and its properties (which it may do correctly or incorrectly).
  15. Phenomenalism

  16. We will consider the version of phenomenalism argued for in Berkeley, G. (1713), Three Dialogues concerning Hylas and Philonous, first dialogue.
  17. Berkeley argues that everything is mind-dependent (including things like trees). In outline, his argument goes something like this:
    1. The things that we immediately perceive are mind-dependent.
    2. We have no reason to think that there is anything else, and some reason to think that there is not.
    3. In fact, we cannot even conceive of there being anything that is not in some mind.
    4. Therefore, everything is mind-dependent.
  18. He starts by arguing for the first premise (the things that we immediately perceive are mind-dependent). He argues that all qualities we immediately perceive are mind-dependent. This includes so-called secondary qualities – colours, sounds, tastes, smells, and feels; but also so-called primary qualities – size, shape, solidity, and motion.
  19. He runs two kinds of argument. Here they are in the case of heat:

    He runs these arguments for other secondary qualities: colours, sounds, tastes, smells, feels.

    He runs them for primary qualities as well (at least the second argument): size, shape, solidity, motion. He notes that since size is presupposed by shape, solidity, and motion, if the former is mind-dependent then so are the latter.

  20. He argues against there being a ‘material substratum’ that supports the primary and secondary qualities. To support these qualities, it must have extension; but then there must be a second substratum that supports its extension; and so on, ad infinitum (a vicious regress).
  21. Berkeley then argues for the second premise (we have no reason to think that there is anything other than what we immediately perceive, and some reason to think that there is not).

    He questions: what reason have we to think that there is anything other than what we immediately perceive? It is natural to think that there are such things, that cause us to perceive as we do. And it is possible that there are such things. But is that good enough reason to think that there are such things?

    He gives positive reason to think that there aren’t. We posit their existence, to explain the way we perceive as we do. The qualities we perceive are supposed to resemble that which causes them. But nothing can be like a sensation or idea, but a sensation or idea. And no sensation or idea can exist without the mind. So these supposed causes are also mind-dependent.

  22. Berkeley then turns to the third premise. He argues that we cannot even conceive of something existing without being in some mind. It seems easy to do: I can conceive of a tree existing without being thought of or perceived by any mind. But I thereby conceive the tree, so it is in my mind. Thus, I cannot conceive of a tree that is not in some mind. (Analogy: I cannot see an unseen thing.) If we cannot conceive it, then why think it is possible?
  23. Representative realism

  24. We will consider the version of representative realism presented in Russell, B. (1912), The Problems of Philosophy, chapters 1-4.
  25. Russell argues for four major claims (one per chapter):
  26. When we perceive, we are immediately aware of sense data, not physical objects.
    1. Different people see different colour in a table – there is no one colour that appears to be the colour of the table.
    2. So, the colour that we see is not in the table. Rather, it is something that depends on us (and other factors) – a ‘sense datum’.
    3. The ‘real’ colour is not what we see – it is something inferred from what we see. By ‘the colour of the table’ we mean: ‘the colour it will seem to have to a normal spectator from an ordinary point of view under usual conditions of light’. In itself, the table does not have any one particular colour.
    4. The same goes for smoothness, shape, etc.
    5. Our senses do not give us truth about the table itself, but only about the appearance of the table.
    6. The real table, if there is one, is not what we immediately experience by sight or touch or hearing. It is not immediately known to us at all, but must be an inference from what is known.
    7. If we know anything about the table, it must be by means of sense data. But the table is not the sense data, and they are not properties of the table.
  27. Nevertheless, there are such things as physical objects.
    1. If we are only immediately aware of our sense data, then do we have any reason to think that there are physical objects – things that persist when we stop sensing?
    2. It is possible that there are not. But the most simple hypothesis is that there are. It is the most simple way of explaining why, for example, our current ‘table’ sense data are similar to each other’s, and to the ‘table’ sense data we had last week.
  28. Physical objects correspond to our sense data.
    1. What is the nature of the real table?
    2. It cannot be exactly like our sense data. Its real shape, for example, must be in real space, not anybody’s apparent space. Real space is public, our apparent spaces are private. So the real space, in which it has its real shape, must be different from any private space.
    3. Nevertheless, it might correspond to our sense data. (Like library books correspond to their catalog.)
    4. But all we can know about the table is whatever is required to secure the correspondence – we can know nothing of what the table is like in itself. We can know, for example, that the Earth, Moon and Sun are in a straight line during eclipse, but we cannot know what a straight line is in itself.
    5. It is gratuitous to suppose that physical objects have colours, and therefore there is no justification for making such a supposition. The same goes for other sense data.
  29. Although sense data are mind-dependent, physical objects are not.
    1. Russell argues against Berkeley’s idealism, that whatever exists, or at any rate whatever can be known to exist, must be in some sense mental.
    2. He puts Berkeley’s argument this way: to be known is to be in a mind, and therefore to be mental. Hence nothing can be known but what is in some mind.
    3. Russell claims that this is fallacious – it trades on an act/object ambiguity. When something is known, the act of the knowing is in a mind, but it does not follow that the object of the knowing is in a mind, because the act and the object are distinct.
  30. Adverbialism

  31. Two of the key figures in the debate about adverbialism are Roderick Chisholm and Frank Jackson (see Chisholm, R. M. (1966), ‘The Status of Appearances’, in D. M. Rosenthal (ed.) (1986), The Nature of Mind (New York: Oxford University Press), ch. 40, and Jackson, F. (1976), ‘The existence of Mental Objects’, in Rosenthal, ch. 41).
  32. Chisholm

  33. Chisholm argues in support of adverbialism.
  34. He was not the first to do so. Earlier adverbial theorists include G. F. Stout (1909), ‘Are Presentations Mental or Physical?’, Proceedings of the Aristotelian Society 9; and C. J. Ducasse (1949), Nature, Mind and Death (La Salle, Ill: Open Court Publishing), ch. 13.
  35. A couple of preliminaries:
  36. According to (many) sense datum theorists, for an object to appear white is for it to present an appearance (or sense datum) which is white.
  37. Chisholm argues against this:
  38. Chisholm agrees with earlier adverbialists that we use ‘white’ in ‘Patch appears white’, as an adverb (hence the name ‘adverbialism’).

    We use adverbs to attribute properties, not to objects, but to properties or events/processes/state of affairs. We use ‘slowly’ in ‘John is walking slowly’, for example, to attribute a property to John’s walking (an event).

    In a similar way, we use ‘white’ in ‘Patch appears white’ to attribute a property to Patch’s appearing (an event).

  39. We can paraphrase ‘Patch appears white to John’ as ‘Patch appears whitely to John’, or better, as ‘John is appeared to whitely’ (to remove commitment to there being any Patch).
  40. There is no need to think (as sense-datum theorists do) that there are such things as appearances (or sense-data), and that we use sentences such as ‘Patch appears white to John’ to attribute properties to them. All we need are people, and events of being appeared to.
  41. Jackson

  42. Jackson argues against adverbialism.
  43. He is concerned with the adverbialist claim that by ‘I have a red after-image’ we mean ‘I am appeared to redly’. (Compare: ‘Mary has a nice smile’ and ‘Mary smiles nicely’.)
  44. He raises two objections, which he calls the many property problem, and the complement problem. We will consider just the first.
  45. The objection goes as follows. Jackson asks: what, according to adverbialists, do we mean by ‘I have a red and square after-image’? He recognizes three possibilities, and argues against each:
  46. Disjunctivism

  47. See the section about disjunctivism in the SEP article, ‘The Problem of Perception’.
  48. Disjunctivism was first proposed in Hinton, J. M. (1973), Experiences (Oxford: Clarendon Press), and later developed in Snowdon, P. F. (1979), ‘Perception, Vision and Causation’, Proceedings of the Aristotelian Society; McDowell, J. (1982), ‘Criteria, Defeasibility and Knowledge’, Proceedings of the British Academy; and Martin, M. G. F. (2002), ‘The Transparency of Experience’, Mind and Language 17, 376-425.
  49. Recall two of the arguments that pose a problem for direct realism:
  50. Disjunctivists think that these arguments are invalid. Just because the colour of which Rob is directly aware in the illusory case is not in the bottle, it does not follow that the colour of which he is directly aware in the veridical case is not in the bottle. And just because Macbeth is not directly aware of a dagger in the hallucinatory case, it does not follow that he is not directly aware of a dagger in the veridical case.
  51. What they claim is that different kinds of perception are involved in the veridical, illusory, and hallucinatory cases.

    So, we cannot draw conclusions about the nature of perception in the veridical case, from the nature of perception in the other two cases.

    Just because the colours of which we are directly aware in illusory cases are not in the objects that we perceive, it does not follow that they are not in the objects that we perceive in veridical cases.

    And just because we are not directly aware of objects in hallucinatory cases, it does not follow that we are not directly aware of objects in veridical cases.

  52. This allows disjunctivists to be direct realists about veridical perception, while avoiding the problems for direct realism posed by illusions and hallucinations. They can say that in veridical perception we are directly aware of objects and their properties, even though in cases of illusion and hallucination we are not.

    Veridical perception is essentially a relation between an observer and an object and its properties. Illusory and hallucinatory perception are not. This is possible because they are all different kinds of perception.

  53. Why ‘disjunctivism’? Because we can think of the claim as this: to perceive is to have either a veridical perception, or an illusion, or an hallucination.
  54. The main question to ask of disjunctivism is this: is it plausible to deny that veridical perception, illusion, and hallucination are distinct kinds – to deny that perception is whatever it is that they have in common?
  55. Representationalism

  56. For a good account see Crane, T. (2001), Elements of Mind (Oxford: OUP), ch. 5. Also see the section called ‘The Intentionalist Theory’ in the SEP article, ‘The Problem of Perception’ (also by Crane).
  57. Crane argues for representationalism (he calls it ‘intentionalism’).
  58. According to direct realists, what we are directly aware of when we perceive are mind-independent material objects and their properties.

    We have seen that the phenomena of perspectival variation, illusion, and hallucination pose a problem for direct realism.

  59. According to representative realists, what we are directly aware of when we perceive are mind-dependent sense data.

    Perspectival variation, illusion, and hallucination do not pose the same problem for representative realists, but they are committed to the existence of ‘weird’ sense data.

  60. According to disjunctivists, they are distinct kinds of perception that are involved in veridical perception, illusion, and hallucination – not a single kind.

    They avoid the problems of perspectival variation, illusion, and hallucination by claiming that the arguments that present the supposed problems are invalid. But they are left with the counterintuitive claim that these are distinct kinds of perception.

  61. According to representationalists (intentionalists), what we are directly aware of when we perceive are representational (intentional) contents: when we perceive, our experiences represent the world to be a certain way – that the bottle is white, that there is a dagger over there, and so on.

    According to this view, perceiving is like believing: when we believe, our beliefs represent the world to be a certain way – that the bottle is white, that there is a dagger over there, and so on. Perceiving is not believing, but it is like believing in that our experiences have representational content (as do beliefs).

  62. Representationalists uphold the claim that veridical perception, illusion, and hallucination are all the same kind of perception (unlike disjunctivists). They also (typically) reject the appeal to sense data.
  63. So how do they avoid the problems posed by perspectival variation, illusion, and hallucination?

    This way: just as I can believe that Patch is red without Patch actually being red, so too my visual experience can represent that Patch is red without Patch actually being red (illusion). And just as I can believe that there is a dagger before me without there actually being a dagger before me, so too my visual experience can represent that there is a dagger before me without there actually being a dagger before me. The key fact here is that there is such a thing as misrepresentation.

  64. Crane defends ‘strong intentionalism’, the view that all features of visual experience are aspects of its content. That is: two visual experiences cannot be different unless they differ in content – unless they represent the world to be different ways. (But note: he thinks that visual experiences and, say, tactile experiences are different and yet might have the same content.)
  65. There are some problems for this view (strong intentionalism). We will consider two, and how Crane responds:

Exercise

No exercise this week - time to catch up/polish up.