Re: Stress question

Eric Christopherson wrote:

> > Your first secondary stress rule is, well, weird. It's unlikely
> > that you'd locate primary stress with reference to the right
> > edge, then "double back" to locate secondary stresses.
>
> Really? I'm quite surprised, as that's how I tend to analyze my own accent
> patterns in English.
I  believe he was talking about predictable systems, where, unlike
English, stress is not phonemic (cf. "incite" [In.'saIt] vs. "insight" ['In.saIt]).

> > Here's an example. Start on the left edge and stress every other
> > syllable (shown by an "x" above the vowel of the syllable; the
> > bracket shows the reference edge):
>
> Hmm, I hadn't ever heard of an accent system counting syllables from the
> *left* before, although I have heard of ones that always stress the very
> first syllable. But I don't see why it shouldn't be possible.
Well there are three distinct issues here:  what type of feet the language has,
the way in which foot structure is assigned to syllables, and where the primary
stress (if any) is placed.

(1) The first can be further divided into two questions:  is the
foot binary (must have two moras or syllables), or can it be degenerate
(having one mora or syllable) or ternary, quaternary, etc.; and, where in the
foot does stress go? Most languages like having binary feet, as seen in Dirk's
example.  That variable predicts the rate of iteration -- there'll be y number of
stresses for every x number of syllables or moras.  Next, if these feet are binary,
then does the stress go on the first or second syllable internally? If the first,
it's a trochee; if the second, an iamb (from poetics, obviously).
If the feet are not binary, that's because, in Optimality-Theoretic terms,
some constraint like Parse-syl (see below) outranks Ft-Bin.
(2) Second, languages can simply assign which edge of the word it wants
the stress to start from.  In OT, these are known as the All-Feet-Right
and All-Feet-Left constraints (which are mutually exclusive).  If these
constraints are not dominated by any other constraints, then you end
up having a noniterative stress system, i.e., one where there is only one stress
per word.  If, on the other hand, the higher ranked one of these (either AFL
or AFR) is dominated by a third constraint, Parse-syl, which says that all syllables
should be parsed into feet, then you'll have an iterative system, where whatever
footype you have, that will be repeated to the other edge of the word unless
it runs into yet another constraint (like Non-Finality, which makes the last
syllable of a word extrametrical for purposes of stress).
(3) Primary stress, in OT, is determined by yet another set of constraints:
Main-Foot-Left, or Main-Foot-Right.  Note that because these are separate
constraints from AFL and AFR, you base the foot structure from edge X of
the word, while assigning primary stress from edge Y, the opposite edge.
(Obviously, the primary stress constraint has to outrank AFL or AFR.)

Thus, in a roundabout response to your comment,  from which edge you start
assigning feet is actually integrally  linked to the rest of the foot structure.  This
is one of the neat things about OT, that, despite some of the faults I've mentioned
before, it can tie seemingly different phenomena together.

===================================
Thomas Wier | AIM: trwier

"Aspidi men Saiôn tis agalletai, hên para thamnôi
  entos amômêton kallipon ouk ethelôn;
autos d' exephugon thanatou telos: aspis ekeinê
  erretô; exautês ktêsomai ou kakiô" - Arkhilokhos