Re: [SHOEBOX] "Morphophonemic form gives cyclical pattern..."

In a message dated 7/23/00 5:53:01 AM Central Daylight Time,
dbell@GRAYWIZARD.NET writes:

> Yes, this is on my wish list as well, but I can't see how this alone would
>  lead to your cyclical error.  Unless our Master Shoeboxer, Jeff, can come
up
>  with something, however, I think this one will remain on the wish list.  I
>  have however been toying with turning this problem into a "feature"  I have
>  begun to envision a development of the language in which this kind of
>  ambiguity is reduced to a minimum.  I could use Shoebox interliearization
to
>  identify the ambiguities and "evolve" them away in the new form of the
>  language.
>
Unfortunately, parsing never takes into account the part of speech, it just
shows it to you when available, and when ambiguities occur.  SIL could fix
this by adding another type of rule file, similar to Generate and Rearrange
rules, but instead describing what kind of affixes can be applied to roots,
and describing word order, etc.  Probably won't happen soon, and it still
wouldn't eliminate all ambiguities, either, at least in a naturalistic
language.  Somebody ought to try interlinearizing a good logical language to
see how well they can be parsed by a linguistic parser.  If you are
attempting a morphophonetic change via the Generate step, then you could work
up an ugly hack to sort of fix the problem, but it would be rather ugly.

Jeff Sheets

"On two occasions I have been asked by members of Parliament, `Pray, Mr.
Babbage, if you put into the machine wrong figures, will the right answers
come out?'  I am not able rightly to apprehend the kind of confusion of ideas
that could provoke such a question." -- Charles Babbage