SOME PUZZLES

It is possible that S is an icon of O, S* is a symbol perceptually indiscriminable from S and that S is not an icon of O or indeed of anything else. I might stumble upon a strange symbol that looks to me for all the world like the portrait of a pear but which turns out to be, upon anthropological investigation, an abstract symbol of the local native godhead. What distinguishes this symbol from an icon of a pear? Obviously not just natural generativity, for in both cases one might naturally generate the interpretation ‘It's a pear. Of course in one case, the iconic case, this naturally generated interpretation is the correct one while in the other, the totemic case, it is not. So what makes the naturally generated interpretation the correct one when it is the correct one?

As it happens, this problem is formally identical with the problem of the colour-reversed icon. A colour-reversed icon is one in which some or all of the colours have exchanged places with other colours so that, for example, red pigment on the canvas would represent an object as green while green pigment on the canvas would represent it as red. Wollheim posed the crucial question: ‘What makes a colour-reversed icon less natural than the normal icon?’ He rightly points out that many accounts of depiction cannot answer it.