To: Francis M.
From: Geoffrey Klempner
Subject: The disjunctivist view of perception
Date: 3 November 2005 11:55

Dear Frank,

Thank you for your email of 24 October, with your University of London essay in response to the question, 'Do veridical perception, illusion and hallucination have something in common? Justify your answer.'

I should declare at the start that my own preference would be to defend a simple disjunctivist theory, in Sturgeon's terms. So I am looking for an argument which will persuade me to give up disjunctivism.

Just to keep things simple, I am going to concentrate on veridical perception and hallucinations. I agree with you that 'illusion' is a slippery concept. I can see the round table in the furniture shop as elliptical without being in the slightest doubt that it is round; or, on the basis of the very same experience I can come to believe it is elliptical. (Whether I draw one conclusion or the other depends on the context. E.g. all the other tables in the shop are elliptical, or I mistakenly believe that my vantage point is higher than it in fact is.) In the first case, I not only believe the table is round, I 'see it as' round. Similarly, in the second case, I not only believe that the table is elliptical, I 'see it as' elliptical.

Consider the statement, 'I see a elephant' as an example of veridical perception, and also as an example of a hallucination.

On the sense data theory, as you note, in both cases it is true that 'I see X' for some X. In both cases, I see a elephant-shaped sense datum. But in the first case the sense datum is caused by, and also more or less correctly 'represents' the elephant, while in the second case the sense datum is caused by something entirely different, say, the state of my intoxicated brain.

On the disjunctivist theory, it is simply false that in both cases I see X for some X. If my perception is veridical, then I see an elephant. If I am undergoing a hallucination, on the other hand, then I only seem to see an elephant. End of story.

On the disjunctivist view, 'It seems to A that A sees X' does not contain a component, 'A sees Y, for some Y'. That is how the disjunctivist avoids the argument from illusion.

Sturgeon does not like this 'solution' which he calls 'quietist', because it apparently makes no attempt to explain the nature of our experience when we undergo a hallucination. I seem to see an elephant, when no elephant is there. What is it about my experience which could possibly make this true?

I would certainly be interested in an argument which put pressure on the view that nothing more needs to be said about 'seeming to see', no further explanation needs to be given in terms of the subjective contents of experience. My seeming to see an elephant has a physical cause, to be sure - my inebriation - but why isn't it enough to say that?

I didn't gain from your essay any real understanding of what Sturgeon proposes as a third alternative to the sense datum theory or disjunctivism.

Instead, you offer your own 'take'. You start of by saying, 'Even if the three cases of veridical perception, illusion and hallucination are the same in core nature, I do not think this means that they are all caused by the same thing.' What, then, is the 'core nature' which they share, if not 'experiencing a sense datum'?

Instead of answering that question, however, you go on to explain how it is possible subjectively to tell whether an experience is a hallucination. 'If under some circumstances we can know the difference, there must be a difference to be known.'

My response would be to consider the Matrix scenario. To be sure, there is such a thing as 'waking up'. That would be a way of 'knowing' the difference, but that is not a feature internal to the experience itself, which was what we were looking for. Again, when you have wised up, you can learn to 'spot' whether you are in the Matrix or the real world, but that is no help until you have had the benefit of having woken up.

Once we allow the possibility that hallucinations can be totally convincing, incapable of being distinguished subjectively from veridical perception, then we are thrown back once more on the two alternatives of sense data and disjunctivism. Either there is something subjectively seen (sense data) or we merely 'seem to see'.

You go on to describe a 'framework' where the content of perception is accounted for by the properties of the object and the perceiver. We are not deceived (under illusion) so long as we are aware of the contributions of these two factors in the final result. To take a trivial example, things look rose coloured if you are wearing rose coloured glasses. But if you are aware of this then you can use this knowledge to adjust your perceptual judgements.

So far so good, but then what do we say about hallucinations? According to your 'take' one would say that there will be cases where there is no object at all, just me and my drunken state. In the case where I seem to see an elephant when no elephant is there, it is my own nature that accounts for the experience. But that throws us back on the previous dilemma: either we look for an explanation in terms of content (seeing sense data) or we refuse to give any further elaboration (quietism). Where is the third alternative?

All the best,

Geoffrey