Knowledge of ‘the external world’ can only take those forms of which the bodymind is capable, and the finer points of that knowledge are learned by each individual, not generically ‘given’. But the need to interact symbolically forces us to forget all that, for we must assume a world external to all of us, and a common language referring to it, in order to anchor our meanings in a common ground. We could hardly converse at all if we had to take into account each person’s semantic idiosyncrasies. So we cannot help feeling that we ‘share experience with one another symbolically’ (Tomasello 1999, 42). When you can’t make that assumption, you can’t communicate in the fully human, symbolic sense (though some semiotic exchange may still be possible).
You can’t perceive your perception, but you can investigate your perceptual process by making inferences from what you do perceive. Investigation, being dialogic, has a social context: you can and do assume that others perceive the same world you do, and that other perceptions of that same world differ from yours – for each individual perceiver is unique in some respect. And on top of that, polyversity virtually guarantees that perceptual differences may be either masked or exaggerated by the idiosyncratic use of common symbol systems to report those perceptions.