Re: The Language Code, take 2 (or 3)

On Wednesday, June 11, 2003, at 05:00  PM, Christian Thalmann wrote:

> --- In conlang@yahoogroups.com, Dirk Elzinga <dirk_elzinga@B...> wrote:
>
>
> > P       phonology
> >         .
> >         .
> >         s       syllable template {c,v}
>
> Maybe more categories would be useful, e.g. {c} consonant,
> {v} vowel, {l} consonant cluster, {t} stop, {f} fricative,
> {n} nasal, {s} spirant, {g} glide, {d} diphthong.
>
> English could then be roughly described as (l)[v|d](l),
> while Obrenje would be (t)(f|s|g)v(c).
Perhaps this is getting too specific. Here are some thoughts. Instead
of stating specific consonant types, indicate only whether syllables
have onsets and codas, and whether they can be complex or not. English
would be P ... s(c+)v(c+).

> I'd also suggest adding a category:
>   P       phonology
>           .
>           .
>           m       mutations/sandhi {c,v}
>                   i       initial
>                   f       final
As I understand the terms, mutation and sandhi are different phenomena;
mutation is the alternation among consonants as a marker for
grammatical categories, while sandhi is the phonologically conditioned
alternation of consonants across morpheme or word boundaries. The types
of formal marking of morphological categories is indicated in the Code
only by the agglutinating/inflecting/isolating attributes; more
specific discussion is best saved for the language documentation, I
think. Sandhi is part of the allophony attribute.

> > W       writing system
> >         .
> >         .
> >         r       +/-     regularity/irregularity
>
> I'd suggest this instead:
>
>   W       writing system
>           .
>           .
>           s       spelling
>                   r       regularity (+/-)
>                   p       phoneticness (+/-)
>                   m       mutation/sandhi
>
> Where r+++ means "if I learn the rules, I can pronounce
> every written word correctly" (e.g. Spanish) and p+++ means
> "if I learn the rules, I can spell every spoken word
> correctly" (e.g. Finnish, Esperanto).
This reminds me of the biuniqueness problem in American Structuralist
phonemics. There are two directions: text-to-speech, and
speech-to-text. If I understand you correctly, you are proposing that
"r" refer to the degree to which the text-to-speech mapping is regular
and predictable. This implies that speech-to-text might not be regular
or predictable, and your example of Spanish is a good one. "p" then
means speech-to-text. Your examples show this kind of predictability as
well, but also show a high degree of text-to-speech as as result. That
is, "If I learn the rules [of Finnish, Esperanto, etc] I can pronounce
every written word correctly." I don't know what it would mean for a
language to be r---p+++; that is, to be able to reduce the language
correctly to writing, but then not be able to predict the correct
pronunciation from the transcript. Oh, hang on; isn't this how
shorthand works?

> The tag m tells us
> whether or not mutations/sandhi are written (e.g. Welsh) or
> not (Jovian).
This would be a property of either "r" or "p", I think.

> This would allow us to distinguish etabnannery (r+++p---)
> from maggelitude (r---p---).
You would have this four-way distinction:

r+p+  "Finnish"
r+p-  etabnannery
r-p-  maggelitude
r-p+  tachygraphy (?)

Again, I'm not sure that all of the detail is warranted. I suppose the
true test is if people find it useful.

Dirk
--
Dirk Elzinga
Dirk_Elzinga@byu.edu

"I believe that phonology is superior to music. It is more variable and
its pecuniary possibilities are far greater." - Erik Satie