hang-ing by a thread of phonetic similarity [was: Re: Return of the minimal pairs]

[Rich Alderson]

In response to Paul Cohen's posting regarding the status of [h] and [N] as
allophones of a single phoneme in English, both Max Wheeler and Larry Trask
bring in the notion of phonetic similarity.

As I was taught it, "phonetic similarity" was introduced into classical
phonemics precisely to counter the argument that complementary distribution
(originally the only criterion in the classical theory) required so defining
them.

								Rich Alderson