On 4/14/06, Benct Philip Jonsson <[log in to unmask]> wrote:

> I generated a list of possible roots using John Cowan's
> everyword.pl <http://tinyurl.com/m7aba>, then ran this list
> through a self-written perl script to eliminate roots with
> unwanted phoneme co-occurrences. John's script has provision
> for excluding contiguous unwanted letter combinations like
> _yi, iy, wu, uw, kwy, gwy_ or two vowels after one another,
> and preventing certain clusters from (co)occurring, but I
> also wanted to exclude the same consonant, as well as two
> stops or fricatives at the same point of articulation or two
> nasals, from occurring twice in the same word, so I used
> regular expressions like /([$C]).*\1/ and /([pbf]).*([pbf])/
> to exclude unwanted roots.  ........

Thanks for pointing me to everyword.pl.  I made a one-line
change so it excludes regexes instead of literal
strings.  If you're interested, let me know and I'll email
you the new version.  (I reckon I should ask John Cowan
before I put the revision of everyword.pl on my site.)

everyword.pl is more flexible than the format I came
up with for my gen-redundant-morphemes.pl and
gen-all-possible-morphemes.pl, but it would be a lot
of work to rewrite gen-redundant-morphemes.pl
to use this same format, so it's a low priority.

--
Jim Henry
http://www.pobox.com/~jimhenry/conlang.htm