[Lexicog] polysynthetic languages and dictionaries

[Mike Maxwell]

Sean M. Burke wrote:

>In my experience, Kimmo's formalisms (last I looked) were way too
>simplistic to handle anything with actually complex morphophonology.
>
>
Just a quick note on this.  As one who learned generative phonology
during the '70s and '80s, I would agree that two level phonology would
not be my favorite formalism to develop s.t. in, although I believe it's
more a question of how messy it would be, rather than whether it's
possible.  (It might be a question of taste, too, although I don't
really believe that :-).)

Bill Poser mentioned the Xerox transducers, which are indeed a finite
state technology.  (Transducers, meaning both generation and parsing.)

While the Xerox tools do include 'twolc', an implementation of two-level
phonology (basically the same as Kimmo's formalism, so far as I know),
the formalism that is featured in the Beesley/ Karttunen book in fact
implements a phonology with linearly ordered phonological rules.  This
is similar to the notion of phonological rule ordering in older theories
of generative phonology.  (Phonological features are not
implemented--rules are written in terms of phonemes or characters.)  It
is also possible to implement stratal phonology (where a rule set
applies only to a subset of the items in the lexicon.

As for morphology, the Xerox tools allow non-concatenative affixation.
(If you're familiar with the limits of finite state languages, you may
question whether it can do full reduplication.  It's basically done
through a lexicon compilation process.)  I've not implemented the
polysynthetic part of a polysynthetic language (which is to say, I'm
doing a Nahuatl parser, but we haven't gotten to incorporation).  But I
don't see any problem in principle to using the Xerox formalism to
describe polysynthetic languages.

In addition, Lauri Karttunen has a paper on implementing (a version of)
Optimality Theory using the Xerox tools.

Some years ago, I wrote a morphological/ phonological parser/ generator
which did not use finite state technology.  This was Hermit Crab, which
was distributed as part of LinguaLinks (from SIL).  I don't know of
anyone who is currently using this.  Hermit Crab implemented an SPE-like
phonology, i.e. it used phonological features, and handled
non-concatenative morphology.

    Mike Maxwell



------------------------ Yahoo! Groups Sponsor --------------------~-->
Make a clean sweep of pop-up ads. Yahoo! Companion Toolbar.
Now with Pop-Up Blocker. Get it for free!
http://us.click.yahoo.com/L5YrjA/eSIIAA/yQLSAA/HKE4lB/TM
--------------------------------------------------------------------~->


Yahoo! Groups Links

<*> To visit your group on the web, go to:
     http://groups.yahoo.com/group/lexicographylist/

<*> To unsubscribe from this group, send an email to:
     lexicographylist-unsubscribe at yahoogroups.com

<*> Your use of Yahoo! Groups is subject to:
     http://docs.yahoo.com/info/terms/