Bob Port (welcome, Bob) has posted a comment on my post about the article he wrote with Adam Leary for Language. It is good that Bob is here now to clear things up. (I somehow do not manage to put my reaction below his, so I will put it here instead.)
In my view, a strong argument in favour of the assumption that languages contain an alphabet is what Nick Clements calls Feature Economy (I do not find any paper online, but here are a few abstracts.) If we have a language L1 which has a sound [p], and a language L2 which does not have that sound, the chances that language L1 has also a sound [b] are significantly bigger than the chances that L2 has [b].
Port and Leary discuss a similar argument in their paper, but they conclude (I am citing from the preprint):
Note, however, that what these phenomena support is only phonological categorization on a language-specific basis. This evidence offers no support at all for universal phonetic sound classes.
It seems to me that we can agree on this: languages all seem to use categorisation – that is what phonology is about –, but the relation between this and its phonetic implementation is (i) complex and (ii) possibly language-specific (it probably also interacts with sociolinguistic factors such as age, group identification, etc.).
I do not see where the specific problem is for formal phonology. If we acknowledge that there are two levels of reality, one of abstract categorisation and one of phonetic continuity, it does not make sense to deny the existence of one of these two just because it is difficult to establish the relation. Of course, if we are only willing to take phonetic measurements as evidence, we may not find the kinds of evidence we are looking for as phonologists. The evidence for phonology comes from phonological facts, such as the structure of sound systems – Feature Economy is a specific prediction made by this.
If Port and Leary’s arguments in their article are right, it does indeed imply that the strong position defended in SPE – that there is a universal one-to-one relation between phonological categories and the phonetics is too strong. (As a matter of fact, I do not think that Chomsky and Halle were seriously taking that position: binary phonological features are mapped into multivalued phonetic features in probably a complicated and even language-dependent way in SPE.)
I want to come back to this sentence which I apparently do not understand:
There is only one route left to justify doing traditional generative phonology or for studying only the abstract sound structures of a language and deny the relevance of articulatory, acoustic and auditory details. It is to claim: We don’t care about linguistic behavior, only about linguistic knowledge.
This probably describes the difference between phonology and phonetics in most textbooks. (I do not really like the phrase we don’t care, and would replace it with we abstract away from.) The phonological literature is filled with examples of generalisations on abstract sound structures; there are predictions of the Feature Economy type, and they seem confirmed. There is no ‘utter lack of evidence’ for the assumptions on which formal analysis of phonologies are based; there is plenty. Maybe we are not going to find it in the phonetics. But then, if we only would take phonological facts as evidence, there would not be a lot of evidence for many phonetic details: it would be a bit funny to conclude from this that these phonological facts put an unbearable empirical burden on phonetics.