The fundamental unit of grammar is a morpheme. A basic unit of written language is a grapheme. And the basic unit of sound is a phoneme. However, this is technically what Professor Crystal describes as “the smallest contrastive unit” and it is highly useful to you in explaining things - but strictly speaking may not exist in real spoken language use. That is, almost anything you say is a continuum and you rarely assemble a series of discrete sounds into a connected whole. (It is possible to do this with synthesised speech, as used by Professor Stephen Hawking - but the result is so different from naturally occurring speech that we can recognize it instantly.) And there is no perfect or single right way to say anything - which is just as well, because we can never exactly reproduce a previous performance.