To: Katherine A.
From: Geoffrey Klempner
Subject: 'No-one ever REALLY perceives a tree'
Date: 6 June 2006 11:32

Dear Katherine,

Thank you for your email of 23 May, with your fourth essay for the Possible World Machine, in response to the question, ''It is obvious when you perceive a tree, what your eyes actually register is an upside-down image of a tree on the back of your retina. Therefore, no-one ever REALLY perceives a tree.’ Comment on this argument.'

In your answer to this question, a crucial role is played by the notion of 'belief'. I receive certain sensory impressions, but it is my belief about what a tree looks like - what a tree is - that leads me to interpret this sensory impression as a tree.

We don't need to have been previously acquainted with an object of type F, you argue, provided that we have a mental image of what F's look like. To adapt your example, I would know that I was standing face to face with a stegosaurus if I ever met one.

Mirror images are an interesting case of perception. For example, I think of myself as having a mole on my left cheek whereas in reality the mole is on my right cheek. This impression often goes unchallenged - until, say, we look at a portrait photo of ourselves and wonder what looks 'wrong' about it.

I think that despite this, you can 'perceive' something in a mirror. Perception can be distorted in all sorts of ways and still be perception.

What I am going to do now is describe a thought experiment. I want you to think about what you would say about this.

You have been offered the chance to test drive a giant mechanical 'man', a 'mechanoid', made of steel and plastic and standing fifty feet high. Inside the mechanoid's 'head' is a control tower where you are sitting in front of a steering wheel and a row of screens. There is no window. All your information comes from TV cameras, radar, sonar and other devices.

The environment which the mechanoid is designed for is very hostile, so you need a variety of methods of detecting what is there 'outside'.

Now you are striding forwards, across a field. It is dark, so the TV camera is not much use. The radar screens register a small object in the distance The information requires some careful sifting and interpretation, but in the end as you approach more closely you are in no doubt that it is a tree.

That is to say, on the basis of your perceptions of the screens and the dials you have formed the belief that the object is a tree. You do not 'perceive' the tree. All you perceive is, e.g. a fuzzy glowing shape.

Impressed by this thought experiment, a philosopher thinking about the nature of perception might reason as follows. Our situation is essentially no different from the driver of the giant mechanoid. The only difference is that we are not aware of the role that belief and interpretation plays. Inside our heads - the 'control tower' - certain data is received which the brain then interprets as a 'mountain' or a 'tree'. All we actually perceive is the data. That is the only thing that is given directly. The rest is mere inference.

To accept this picture is to accept the truth of the sense datum theory of perception. But is it necessary to accept this picture?

I would argue that the picture is fundamentally flawed. Consider what we would have to accept if this picture were true. No-one has ever, or will ever see a tree just as it is. All we actually 'see' is the raw data on which we base our judgement that 'this is a tree'.

What is 'seeing' or 'perception'? That is the question one needs to ask.

In the sense datum theory, something is given as 'data'. Seeing takes place when you assimilate the data prior to any thinking or judgement. As a result of thinking we form beliefs about the cause of the data, beliefs which might be true - or false.

But this description is surely exactly correct when applied to the things we see around us in the external world. That there is a tree in front of me now is a given. There is no place for doubt or interpretation. If I can find room for doubt, then one would have to embrace wholesale scepticism. If I don't know - just by looking - that this is a tree then what can I know?

You were asked to 'comment on the validity' of the argument for the proposition that no-one ever really perceives a tree. The premiss of the argument is correct. When you see a tree, the process involves light waves forming an image on your retina. What the argument assumes is that because a process of cause and effect is involved in the act of perception, we only 'see' the end product of the chain. The same effect COULD be produced by different causes. What is 'given' in perception must be beyond all possibility of doubt - otherwise it is not given but merely inferred.

However, this argument is valid only if we make certain restrictive assumptions about the nature of perception itself - assumptions which I would argue are unjustified.

All the best,

Geoffrey