If you ask two people the colour of a certain object (under the standardised lighting conditions) you usually get the same answer (or variations on the same ans

If you ask two people the colour of a certain object (under the standardised lighting conditions) you usually get the same answer (or variations on the same answer: "turquoise" vs. "blue", for example, but there's a fairly well understood vector space in which these fall). Ask two synaesthetes what colour a smell is, though, and I suspect the answers you'd get are much less correlated (that's certainly the case for the colour of numbers) though there are cultural aspects (apparently the colour synaesthetes see in letters are very often the exact colours those letters had in the poster or book they learnt the alphabet from).