SETTING OUT THE PROBLEM

Imagine that Churchill and the other Allied leaders are going to get together for a meeting. It is crucial to get the seating right so the conference table is being arranged ahead of time by ‘seating’ dots around a rectangular piece of paper. It would be perfectly natural, given the convention, to point to these dots and say That's Churchill on Roosevelt's right’ or ‘Churchill should be moved closer to Uncle Joe’ and so on. The ‘advance men’ have anticipated and arranged the meeting by constructing a model of it. Yet the pieces of the model are not icons of the participants or anyone else and the whole model is not an icon of the conference. The model is, in some sense, a natural way of simulating the conference and for this purpose it is probably more useful than a series of drawings. Of course if one is staging a spectacle – a play, a masque or a film – one becomes interested in visual detail. But for the purposes of the organisers this model is what they want.

But the mere addition of information concerning visual detail would not in itself transform our model into an icon. It depends upon how the information is expressed. Suppose we use dots. That one is Churchill's nose, that one is his mouth, these are his eyes, and these are his ears. This will be his cigar, this one will be his furrowed brow and that one a frown.