>>13982138Yes, but they admit degrees of identity. There is no clear set of necessary and sufficient acoustical conditions that entail a particular phoneme has been pronounced. Due to slight differences of dialect, physiology, environmental, sociological+psychological, and pragmatic conditions, two different utterances of the same phoneme will not sound the same, and a lot of the perceived auditory regularity we hear is the result of the structure imposed upon our auditory perception by our cognitive system.
If you wanted to model particular phonetic categories, you would probably need to use some like fuzzy sets, and even that would definitely have terrible real world performance, but it's a start. More accurate models of phonetic categories (and other linguistic categories) probably require a combination of more empirical data on how the human brain processes and structures sound and other sensory functions, and of course a deeper and more accurate formal theory of grammar and the representational/computational processes involved.
