[tied] Re: Laryngeal theory as an unnatural

--- In cybalist@yahoogroups.com, "alex" <alxmoeller@...> wrote:

> Richard Wordingham wrote:
> > Did you ever try out the Sound Change Applier? You might like to
> > try out the page
> >

http://homepage.ntlworld.com/richard.wordingham/sounds/prep5d_rom.htm

> > . Input and output, I'm afraid, is still by cut and paste -

work

> > in progress, both for Javascript programming and linguistic

inputs.

> > You can add words and play around with the text. For example, I
> > haven't got the dropping of final 'r' that Miguel just cited.

What

> > I've got will only get you as far as soror > sorur . And I don;t
> > think any honest set of sound changes will give you the last

step,

> > soru > sor&, for Romanian. ( soru/sor&/#_# is dishonest.)
> >
> > On the programming front, if I can't sove the input problems, I

want

> > to make it easier to embed the words and sound changes, ideally

by

> > using <SCRIPT> tags to pull in up-to-date code.
> >
> > Richard.
>
> I did not tried it Richard. I will try it. Is that indeed 146 rules
> which you have there ?

As I said, it's not complete. You need about 200 rules (in vaguely
natural language) to get from Latin to 17th century French. As the
Sound Change Applier does not have an easy way of limiting the
application of a rule (e.g. 'before m + consonant, but not before
mn'), you need even more rules when you code them up. You may
notice that I used a lot of rules to tidy up Latin spelling, place
the Latin accent, and to convert from phonemic spelling to Romanian
spelling. Degemination can require a dozen rules or more. The C
version is limited to 200 rules (not enough!), but there's no
explicit limit, so far as I'm aware, on the number of rules for the
HTML versions.

Richard.