> Thanks for the comments, BP. The *category* would still be "M", but instead
> of the features "a" and "i", you have "Green" or even just "g". Perhaps
> instead of obligatorily replacing the older features, it can be left up to
> the analyst to do so or to use the older ones. I was quite surprised to see
> that Miapimoquitch only scored 2.43; I would have thought it would be
> higher. And this tells me something about Miapimoquitch that I didn't know
> before. So I think it is essential, for the "g" measure at least, to create
> longer texts. If your language is only attested in inscriptions (for
> example), then you will have no basis for determining a "g" score. But you
> will likely also have little basis for determining "a" and "i" scores as
> well, aside from your intuitions as the creator (but even then, your
> intuitions can fail you; they did me in the case of Miapimoquitch).
>
> Dirk
>
> On Fri, Apr 4, 2008 at 8:49 AM, Benct Philip Jonsson <melroch@...>
> wrote:
>
> > I like the idea of a morph per word ratio, but I can see a couple of
> > problems with it:
> >
> > * it should still be called M rather than anything based on Greenberg,
> > please!
> > * For a conlang which lacks longer texts it may be hard to calculate
> > accurately.
> >
> > 2008/4/2, Dirk Elzinga <dirk.elzinga@...>:
> > > Thanks, Philip.
> > >
> > > A couple of things occurred to me as I reread my deathless prose from
> > 2003.
> > > In that version of the Language Code, I have:
> > >
> > > M morphology
> > > a agglutinating (+/-)
> > > i isolating (+/-)
> > >
> > > I think that these features can be replaced with a single feature, Green
> > > (short for Greenberg). This feature is the ratio of morphs to words and
> > is
> > > expressed as a real number. For example, Miapimoquitch is 2.43. That is,
> > on
> > > average there are 2.43 morphs per word. I think this is a more accurate
> > > reflection of the agglutinating/isolating dimension and isn't too hard
> > to
> > > figure out, given some amount of text. Joseph Greenberg, of typological
> > > fame, proposed this (and other ratios) as a measure of morphological
> > > typology.
> > >
> > > Also in this version of the Language Code, I claimed that English has 24
> > > consonants and 9 vowels. For the record (which is also in the archives)
> > this
> > > is not correct, and should be more like 14 or 15 for American English
> > and
> > > 19-22 for British English. Consult the archives for a nice little
> > discussion
> > > of this.
> > >
> > > Dirk
> > >
> > > On Wed, Apr 2, 2008 at 8:55 AM, Philip Newton <philip.newton@...>
> > > wrote:
> > >
> > > > On Wed, Apr 2, 2008 at 3:42 PM, Dirk Elzinga <dirk.elzinga@...>
> > > > wrote:
> > > > > Many years ago I proposed a "Language Code," which was intended to
> > > > provide a
> > > > > typological profile for a given constructed or natural language. It
> > > > should
> > > > > still be in the archives somewhere.
> > > >
> > > > Here's Take 4:
> >
http://archives.conlang.info/ge/suezhae/qhuevhunwhian.html
> > > >
> > > > Cheers,
> > > > Philip
> > > > --
> > > > Philip Newton <philip.newton@...>
> > > >
> > >
> > >
> > >
> > > --
> > > Miapimoquitch: Tcf Pt*p+++12,4(c)v(v/c) W* Mf+++h+++t*a2c*g*n4
> > Sf++++argh
> > > La----c++d++600
> > >
> >
> >
> > --
> > / BP
> >
>
>
>
> --
> Miapimoquitch: Tcf Pt*p+++12,4(c)v(v/c) W* Mf+++h+++t*a2c*g*n4 Sf++++argh
> La----c++d++600
>