🚀 go-pugleaf

lee@uhunix.uhcc.Hawaii.Edu (Greg Lee) writes:
zgilbert@titan.ucc.umass.edu (Zvi Gilbert) writes:
>>...
>>First look at the facts: "extra" is syllabified as TWO syllables in English.
>>[ek.stra]; there's no doubt that that is the case.
>
>I doubt it.  What's your evidence?

Mea culpa.  I was just informed via email that there is some phonetic research
which suggests that the /s/ may be in the first syllable; i.e., that the 
syllabification may be [eks.tra].  So I don't really know, and evidence from
phonetics would have to help us decide.  It doesn't alter the point that I
was making, which was that [extra] has TWO syllables, even though it has 
THREE sonority peaks.

>>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;...
>
>"Best stated"?  What kind of evidence is that?

We're *modeling* human language here, just like physicists model the nature
of the universe.  Ever seen an electron?  But certain facts about the physical
nature of the universe are best stated in terms of them.  Analogously, certain
facts about the distribution of segments in natural languages are best stated
by using the notion of the syllable.  This is evidence for the syllable as
a "real" element in language: it has predictive power as a model, and it's
robust. 

>Consider as an alternative Saussure's theory, according to which segments have
>a feature explosive versus implosive.  Instead of saying consonants occur
>at the beginning of a syllable, one may characterize them as explosive.

But they are only explosive before a vowel or other syllable nucleus; i.e., 
in syllable initial position.  Just saying that they are "explosive" doesn't
help, since they can shift from explosive to implosive as syllable boundaries
shift (e.g., final aspiration in Marathi).

>>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.  ...
>
>You mean, implosive obstruents are devoiced and that implosives do not
>have laryngeal constrasts?  (See?  No "notion" of a syllable.)

Greg, your theory doesn't work.  WHERE do your implosive segments occur?
HOW do you characterize their distribution?  If you look, you will 
find that they occur after syllable nuclei (like vowels, syllabic consonants,
etc.)  While you may characterize obstruents as "explosive" and "implosive," 
(and there are interesting recent attempts to do so), you still must explain
why they occur where they do.  Even if you don't call it a syllable, you're
making reference to the NOTION of the syllable if you say that "implosives
occur in position X, and explosives occur in position Y."

I've been drawing arguments for the existence of the syllable from phono-
logical literature, and so the arguments for the syllable-entity are based
on patterning facts.  

The notion of the syllable is psychologically robust too, but I don't have
evidence at hand to back that one up.  Those who know, please say so.

--Zvi