It's often put like this: the last syllable
in Latin is "extrametrical" (with lexical exceptions), i.e. invisible to the
stress rule, no matter what its internal structure. Then we look at what remains
and stress the rightmost syllable if it is heavy (CV: or CVC) or if there is no
other syllable to stress; otherwise stress is placed on the next syllable to the
left.
Moras (defined as units of phonological
weight) do play a role here, because in order to capture the notion of a "heavy
syllable" in an elegant way you have to think in terms of rhyme complexity
(otherwise you end up with a disjunctive statement: "a closed syllable or a
syllable containing a long vowel or diphthong"). If we assume that a consonant
in a rhyme position (i.e. in the syllable coda) contributes one mora
of phonological weight (this is a language-specific option), CV: and CVC
syllables each carry the same weight (two moras) and both count as "heavy". Only
CV syllables count as "light", i.e. unimoraic.
Words like <remitto:> or
<de:ductum> are syllabified thus: re.mit.(to:), de:.duc.(tum)
(extrametrical material parenthesised). The syllable-final obstruent in each
penult is moraic by virtue of the position it occupies in the syllable
structure. This may, but need not necessarily entail a different
pronunciation. Whether <solvit> is realised as disyllabic sol.(wit) or
trisyllabic so.lu.(it), the stress rule will make sure that it receives initial
stress.
Latin avoids stress domains ("metrical
feet") with a light stressed syllable at the end. This is a common phenomenon,
since a metrical foot preferentially ends in a falling contour of auditory
prominence (so that a change of pitch can be used to signal the location of
stress), and it is hard to squeeze such a contour into a single mora.
But stress (unlike tone) is a property of
syllables, not of moras, and the role of the latter is restricted to assuring
that there is enough "space" for stress to be placed. Actually, in disyllabic
words with a light penult the final syllable was attracted into the stress foot
(to prevent the creation of a suboptimal structure) and was no longer
extrametrical. The phenomenon of "iambic shortening" (LH > LL, L= light, H =
heavy) can be explained as the phonological reduction of a foot-final unstressed
syllable.
There is some metrical evidence that Latin
stress assignment was iterative in longer words (that is, there were secondary
stresses preceding the primary one), but the details are not quite
clear.
Piotr
----- Original Message -----
Sent: Monday, January 28, 2002 9:17 PM
Subject: Re: [tied] the rhythm and stress in Latin
>If we redefine these rules in terms of moras, the rule is
simpler: "Stress
falls always in the >penultimate mora anteceding last
vowel (conditionlas
are removed from this rule!)"
Some
problems:
(a) Why should the moraic length of the last vowel be ignored, if
the accent
rule is moraic? A syllabic rule does not have that
problem.
(b) You also have to handle words such as illuc, with accent on
the last
syllable (!)
(c) In words like deductum you apparently would
count "c" as moraic, and in
words like remitto, the first "t" as moraic,
which is not particularly nice.
The traditional rule which looks at open and
closed syllables avoids that
problem.
(d) You would have to add a rule
that a resonant (indeed any consonant!) is
"moraic" before another consonant,
but non-moraic before another vowel.
There is no evidence of such different
pronunciation in these contexts.
(e) In words such as "solvit" your
analysis suggests so.l.vit, but we know
that in fact it was the "v", not the
"l" which was liable to pick up extra
syllabicity: sol-uit.
And
what do you mean by removing "conditionals" from your
rule?
Peter