A theoretical mechanism used in linguistics
, and particularly in phonology
, to describe the method by which an infinite variety of physical events (speech sounds) are perceived as belonging more or less ambiguously to a finite number of phonemes
When someone thinks you have a thick accent, it's because your phonemic grid and theirs are different. You might be pronouncing something unacceptably because your phonemic grid can't tell the difference between what you said and what you should have said.
In American English, unvoiced stops are aspirated (or not) based on their position within an utterance. (If followed by a vocalic segment and not preceded by an 's', a stop is aspirated.) In Korean, unvoiced stops belong to distinct phonemes depending on the amount of aspiration; the word 'kil' means entirely different things base on whether the 'k' is unaspirated (as in English "skill" minus the 's'), lightly aspirated, or heavily aspirated (as in English "kill").