Re: Reversible sound change applier

Alex Fink writes:

 > Since asserting that I've realized that all my difficult examples
 > rely on the non-directional application that rsca does.  I assume
 > you apply the transformations left-to-right or right-to-left to any
 > given word?  In this case there's probably no problem, since you
 > can unapply in the opposite direction.

I don't, though. :)

I haven't done extensive tests on the output, but those I have done
have worked fine. If you could send me your difficult examples, that
would be a great help to me.

[Specifying additional phonemes added by a reverse rule]
 > This approach needs a bit of elaboration to do the right thing with
 > persistent processes, though; perhaps it's too patchworkish for
 > your liking anyway.

Not so much that it's patchwork, as that it places a burden on the
user to add such elements throughout the rules file (potentially). I'd
rather such a list were kept in one place - which means it can't be by
rule, obviously. I really don't have any data on which to determine
whether it would be a problem to have a single list of all possible
phonemes over the entire ruleset. It would doubtless generate some
incorrect results; the question is how many, and at what computational
cost?

Again, if anyone has any example rules, lexicons and reversed output
files (in whatever format; I should be able to convert to IPA Zounds
syntax in many cases) for testing, I'd love to see them so I can
refine the process.

Jamie
--
Artefact Publishing: http://www.artefact.org.nz/
GnuPG Public Key: http://www.artefact.org.nz/people/jamie.html

Reply