Williams: "Synaesthetic Adjectives" (1976)

Based on an inspection of English and Japanese adjectives, this paper claims that the etymological development of meaning by metaphorical extension always follows a particular pattern, shown in the figure below.

Figure 1 from Williams' paper (p. 463).

For instance, sound can be conceptualized as touch (soft music), but the opposite is not the case (loud touch). Williams emphasizes, though, that the law is quantitative, not logical:

Even in English, the generalization is not exceptionless. But its regularity varies between 83% and 99%, depending on how we compute what counts as an observation of it. (p. 463)

These numbers refer to the number of dictionary entries inspected (not, for instance, token frequencies in running text).

Towards the end of the paper, he wonders whether "these sequences [of consecutive word meanings] might be reflected in any physical basis of sensation" (p. 472) like its evolutionary history, or the developmental history of the individual. In the conclusion, he writes:

Obviously it is presumptuous, to say the least, to seek a biological foundation for a phenomenon that may not universally exist, in an aspect of human cognition about which very little is known. But the parallels that do exist … indicate … a point of interaction between mind an brain. (p. 473)

However, it is important to keep in mind that he explicitly presents his hypothesis as one about "rule-governed semantic change" (p. 473), not about cognition.