syncretism w/o paradigms

[Heidi Harley]

Hi all! Jonathan writes --

 Meta-syncretisms (as I used the term) do not mean that a contrast is
 systematically absent from the language. I (and I think Williams,
 implicitly) use the idea only for cases of syncretism in which the
 relevant contrasts are indeendently attested in the language.

This is the way I was understanding it, too.

 The following paper in particular makes the case that some of the
 "diagonal" syncretisms (say, neuter plural nominative = masculine
 singular genitive) we try to explain in Indo-European occur only in
 Indo-European, a result that questions just how much syncretism should
 fall out of the system rather than being listed as homophony / language
 particular historical accidents.

But are the kind of meta-syncretism facts that you & Williams discuss
facts only about indo-european lgs like latin & russian, or does the
meta-syncretism effect show up in other lgs too? I expect those surrey
people know; I'll have to take a look.

I think the point is particularly well-made by the impoverishment
analysis of the various English verbal inflection paradigms that you
present early in the Syncretism w/o Paradigms paper. It's of course
very easy to do each of those syncretic patterns entirely with
undespecified vocab items, as you note. And, as Martha says, on grounds
of minimalist aesthetics, that might be the nicer way to go if
possible. But in order to capture the *meta*syncretism effect across
those paradigms (e.g. the 'to be' paradigm and the regular paradigm,
which involve different vocab items), noted by Williams, you had to go
to the Impoverishment story even for those English facts ? vocab
underspecification just doesn't cut it.

*Within a single paradigm*, that is, vocab underspecification is a fine
way to go. But when the same syncretisms show up again in another
paradigm of the lg that is spelled out with different vocab items, it'd
be a remarkable coincidence if it just so happened that the
undespecification of the vocab items conditioned by features relevant
to paradigm A created the same syncretic effects as the
underspecification of the vocab items that are conditioned by features
relevant to paradigm B. The point, which I think you prove admirably
well, is that the only way to get that effect is with Impoverishment of
terminal nodes (post-syntax, pre vocab insertion, as John Frampton
notes).

So the question really is, how robust are metasyncretic effects? If
they're just accidents of a single language family, great. But if
they're a very widespread phenom, such that we are moved to propose
Impoverishment analyses in a wide range of cases of syncretism, not so
great (though if that's what you gotta do, that's what you gotta do, of
course -- I still would think a postsyntactic morphology would be the
way to go for independent reasons).

from a learnability perspective, I imagine there'd have to be a
simultaneous process of building up the feature structure relevant to a
language from the contrasts visible in the morphology (e.g. a 3-way
person, 2-way number, 3-way gender split in English based on the
pronominal system), and also positing Impoverishment rules to account
for syncretisms they hear in other places where they would *expect* to
get distinct forms. So, e.g., the english acquirer knows that there's
18 different possible feature combinations available, based on the
feature they've had to posit to distinguish the pronouns from one
another. Plus the english acquirer knows there's verbal agreement with
those features (based on, say, the different forms of 'to have'). Each
time the learner hears a verb form that's identical with a form they
have remembered hearing with a featurally distinct subject, they say,
aha! that's a place where I have to invent an Impoverishment rule.
Then, the syncretisms they observe in, e.g. the 'to have' paradigm will
be predictive of syncretisms they expect to hear in other paradigms
(though if they're not, as in the 'to be' paradigm, that's no problem
either, they just go back and annotate their initial Impoverishment
rule with 'not for "to be"', or 'only for verbs of class X'). The
overall effect will be a tendency to metasyncretism, without requiring
a Basic Paradigm in Williams' sense.

But of course, there's no place in that kind of system for
underspecification of vocab items to play a role; it would all be taken
care of by Impoverishment.

hmm!

yrs thoughtfully, hh