Re: Phonemes

On Sep 8, 6:29 pm, DKleinecke <dkleine...@xxxxxxxxx> wrote:
On Sep 8, 1:35 pm, "Peter T. Daniels" <gramma...@xxxxxxxxxxx> wrote:

"Modern" phonology claims to operate on a syllabic level, but they
don't have the courage of their convictions and still break syllable
onsets and codas into "C" and "V." I suppose they would treat length
as a syllable feature rather than a segment feature (but it's easy
enough to check -- McCarthy's dissertation is a phonology of Biblical
Hebrew).

Does phonology own phonemics these days? Back in the old days
everybody took a shot at it. Some of them didn't understand what they
were doing - but we just rephonemicized it. If I remember correctly
Hockett wrote at least one article reanalyzing somebody else's
analysis.

See also Langendoen's *London School of Linguistics*, which is a
generative-phonology rewriting of several articles by Firthians
(including, IIRC, Mitchell on Arabic).

In the Arabic world I would treat emphasis, at least in the neo-Arabic
languages as a supersegmental (I think it extends beyond syllables)
feature. I could see how length could be a syllable feature but I
can't see any advantage to it. Must I subscribe to one of those odd
metric theories before I can do phonemes?

You will search contemporary phonology textbooks in vain for a mention
of phonemes, or phonological systems, let alone the "discovery
procedures," as we learned to call them, for discovering a language's
phonological system. (I'm thinking of two volumes in the Cambridge Red
series published almost simultaneously with very similar content, but
I don't think you'll get many phonemes in Goldsmith's, either.) Kinda
makes you wonder where they think the data they reanalyze comes from.

Erica Reiner's *Linguistic Analysis of Akkadian* (1966) treated length
as the symbol <colon>, which could attach either to the right or to
the left to make any adjacent segment long, thus capturing a
morphophonemic generalization (Akk. has most of the same Forms as
Arabic).

I find a colon is too tiny in most of the fonts I use so I use the
equals sign instead. If I were to make length a feature that
lengthened the following syllable I could convert the third form from
an infix to a prefix. People might prefer that. If I were to make it
lengthen the following vowel I could unify the infix action with the
8th form. But I am not all sure I want to do these things.

That's just about what Erica did (though Akk. has no equivalent of
Form III).

And the parallel version with gemination is kind of mind-boggling. I
view the gemination as the probable outcome of some feature like
nasalization in which case it is a little less frightening. The two
consonant roots (doubled verbs) I assume only acquire geminations
latterly.

So, you want a nasal infix that sometimes assimilates and sometimes
doesn't? How do you manage that?

All this is interesting but not exactly what I am asking. I am not
asking SHOULD I make such an analysis - only whether one is possible.-

Of course it's possible. Whether it's advisable depends on what it
"buys you" elsewhere in the grammar.

.

Relevant Pages
Quantcast