On 7/5/06, Mark J. Reed <[log in to unmask]> wrote:
> This concerns pronunciation in the context of English, but it is not
> YAEPT.  It concerns the way consonant clusters split up, and for once
> I'm more curious how it works in other languages than I am in
> differences between dialects within English. :)
>
> The thought began while I was listening to the song "Total Eclipse of
> the Heart", written by the great Jim Steinman (and yeah, recorded by
> Bonnie Tyler - might not have been as big a hit if sung by Steinman's
> regular partner Meat Loaf...)
>
> Anyway, it has an interesting rhyme pattern:
>
> (Turn around)
> Every now and then I get a little bit lonely
> And you're never coming 'ROUND
>
> (Turn around)
> Every now and then I get a little bit tired
> Of listening to the SOUND of my tears...
>
> And so on.  The first of each pair of stanzas ends on the rhyme, while
> the second one only rhymes in passing before continuing right on with
> nary a pause.  I'm sure there's a word for this pattern, but I don't
> know it.  In any case, one later verse in particular struck me:
>
> (Turn around)
> Every now and then I get a little bit restless
> And I dream of something WILD
>
> (Turn around)
> Every now and then I feel a little bit helpless
> And I'm lying like a CHILD in your arms . . .
>
> (Lyrics may not be exact; I didn't Google them but am relying on
> memory).  The odd thing is that, because of the en passant nature of
> the rhyme, this pair doesn't actually rhyme for me.  It sounds
> slightly off, because in connected speech and song I break "child in"
> between the L and the D, so the phrase "child in your" comes out as
> something like /'tSajl.dn=.jO`r\/.  If I intentonally lump the -d in
> with the first syllable, then it rhymes but sounds unnatural, like I'm
> delaying the onset of the "in" syllable.
>
> I'm guessing this shuffling around of syllable boundaries in the
> vicinity of consonant clusters happens quite regularly and just
> usually goes unnoticed.  I was wondering if any languages had phonemic
> distinctions based on such boundaries, where, say, "ka-ching" and
> "catching" and "cat shing" would all mean different things, and the
> distinction would lie not in the emphasis or vowel quality but in the
> location of the syllable boundary...

Speaker intuition is notoriously unreliable for determining syllable
boundaries. In a study which a colleague of mine and I conducted (and
which we're writing up now), we asked participants to divide two
syllable words in the most natural place. We didn't make the
instructions any more specific than that for fear of biasing the
results. What we found is that syllable boundaries are really
statistical rather than absolute, even for cases which seem to be
perfectly straightforward (e.g., 'attack'). Also, morpheme boundaries
seem to have the greatest effect on syllable division.

The only phonetic cues we have for syllabification are indirect and
consist of different kinds of allophonic variation which are
syllabically conditioned. For example, it is well-known that in
English voiceless stops are aspirated when they are initial in a
syllable (i.e., a syllable onset), but they are not aspirated
otherwise. This means that looking at the allophones of, say, /t/ in
contexts such as 'attack' and 'at ease' can tell us if the /t/ is
syllabified as an onset (attack [@[log in to unmask]&k]) or a coda (at ease
[&?t.i:z]).

It has long been an article of faith that languages do not distinguish
words based on syllable boundaries. But how would we know? If
syllabification can't be reliably inferred from speaker intuition and
is only apparent in allophonic distribution, it can always be claimed
that the allophones which are putatively sensitive to syllable
boundaries are in fact different phonemes because there are words
which are distinguished solely by the presence of one phone rather
than the other.

I think that it may be an undecidable question.

Dirk