[OJFord]

You're just introducing more and compound things that we don't know if we have the same experience of.

Consider an enum type: we agree on labels (colour names) for different stimuli; as those stimuli combine they correspond to some other enum variant, and we interpret them as the same label. But we don't know, and I agree with the top-level comment that we can't know, whether our perception of the stimuli with that label is the same.

This is distinct from colourblindness, which is inability to distinguish stimuli for some of the enum variants, and so the labels get merged/boundaries between labels move. (Ok fine we also need a hashmap from stimuli to the enum..)

We agree to call the rose pink, and it's the same pink we see when we look at a colour card and agree it's the colour of the rose, but we don't know that what our brains reconstruct from the light when we look at 'pink'-labelled things is the same.