🚀 go-pugleaf

goer@midway.uchicago.edu writes:
>hahn@csd4.csd.uwm.edu writes:
>>zgilbert@titan.ucc.umass.edu writes: 
>>> This is one of those facts about syllabification in English and many other 
>>> languages; /s/ can do many things that are not permitted to any other 
>>> consonants.  It violates the sonority hierarchy all the time, in words like
>>> "extra" and "straw" (three sonority peaks, but clearly only two syllables.
>>
>>Some consider these to be one sound, i.e. /st/ or /sp/,  and some  
>>say  that they are extrasyllabic, thus not part of the delinquent sonority
>>peaks.
>
>I've always love the terms "extrametrical" and "extrasyllabic."  Let me
>translate:  We don't know what the hell to do with them, but we'll tack
>them on somehow or other.  

Well.  I think that your characterization is a mite unfair.  I'm not going
to talk about extrametricality in this post, but it is a widely attested 
characterization of certain phonotactic facts.  Let's look at how "extra"
gets syllabified again, and talk about syllabification.

First look at the facts: "extra" is syllabified as TWO syllables in English.
[ek.stra]; there's no doubt that that is the case.

The question is how to explain this fact within the theory of phonology. 
One of the explanations for why things syllabify as they do is the sonority
hierarchy, as mentioned in a previous post.  However, the problem with the
sonority hierarchy is that it doesn't get the facts about /s/ in English (there
are other problems to, like "table" in French: one syllable [tabl], but two
sonority peaks [a] and [l]).

The sonority hierarchy gets 99% of the case. There are a few special consonants
in various languages that are a problem; /s/ is one of them.

I think that the use of "extrasyllabic" above wrt /s/ is incorrect; the /s/ 
is clearly syllabified into the second syllable in "extra."  What we must 
explain is why this segment violates the sonority hierarchy, since it is the
putative EXPLANATION for syllabification.  At present, there's no really 
good answer; but an approach to a solution may be modular.  That is, there
are various componants of the phonology that contribute to syllabification:
the phonotactics of the language, the sonority hierarchy, the Avoid V 
principle, etc.  There may be different constraint rankings; i.e., you can
never violate the phonotactics of the languages (no *tl onset clusters in 
English), you can violate the sonority hierarchy in particular places, and
you can violate "Avoid V" (in English, anyway), all over the place.

So lets be clear what we're talking about.  The FACTS of syllabification
are not in doubt.  The explanation for why segments syllabify the way that
they do is more complicated than just saying, "The sonority hierarchy gets
everything"; there are other factors (phonotactics and the like) that are 
involved.

>I'd really hoped that somebody would discuss the objective evidence that
>syllables, in fact, exist before giving everything names.  Despite the
>rush of work in syllabic and metrical phonology, there still appear to be
>a lot of phonologists who don't accept the notion of a deep-structure
>syllable.  

I'm not exactly sure what you mean by "deep structure syllable."  Deep
structure is a syntactic term... the generative phonology tradition in 
which I work doesn't have syllables at the underlying level in the phonology
(at least in English); syllabification happens at the level of lexical 
phonology.

Quick summary of evidence for the syllable (drawn from Selkirk, 1982):

1) phonotactic evidence: many, many genralizations about the distribution of
segments in languages are best stated in terms of the syllable; e.g., /ng/
(the velar nasal) only occurs syllable-finally in English, whereas /h/ in 
English only occurs syllable-intially.  

2) phonological evidence: many rules in languages can be stated only with
reference to the notion of a syllable; e.g., in Polish, all syllable-final
obstruents are devoiced.  In Thai, there are no laryngeal contrasts in
syllable-final position.  

3) metrical and stress evidence: phenomena such as tone & stress can only
be adequately treated with reference to a theory of the syllable (various
stress rules in various languages; tone changing rules, etc.)

(See the article for more info.)

Hope this helps!

--Zvi
Phonologist in Training at UMass, Amherst