CHAT: the language code (longish)

Hey.

I thought I'd have a go at a revision of the language code, based on
the outline which Boudewijn posted and incorporating some comments from
Carlos. I haven't quoted either of them in order to keep the post a
little shorter.

T       type
        f       fictional
        l       logical
        x       auxiliary
        p       personal

P       phonology
        t       tonal
                c       contour tones
                        r       register
                l       level tones
                        !       downdrift/step
                #       number of tones

If you don't know what tonal register or downdrift/step are, you
probably don't have them. Tepa is t* -- no tones (using the '*'
convention from Taliesin's original post).

        p       phonemes
                +/-     allophony
                        p++++   loads of allophonic variation; a
                                phonologist's dream come true
                        p++     some interesting variation
                        p       some allophony, but nothing worth
                                mentioning (so I won't)
                        p--     some speakers are lazy, and I don't
                                approve of their sloppy habits
                        p----   WYSIWYG
                #       consonant phonemes
                #       vowel phonemes

Here, I thought it'd be nice to give some indication of how complex
sound patterns are in your language with respect to allophony, as well
as the number of consonant and vowel phonemes. If you don't know your
language's score WRT to p, it's probably between p and p----. Tepa is
p+++11/4.

        s       syllable template: fill in with c,v,g

I liked Boudewijn's solution here; provide the possible syllable shapes
using parentheses for optional items; Tepa is s(c)v(v/c).

M       morphology
        a       agglutinative
        i       isolating
        f       inflecting
        h       head-marking
        d       dependent-marking

Any of these can be scaled using +'s or -'s. Head/dependent marking
indicates whether the relevant morphology is marked on the head of a
phrase or on the surrounding bits. For example, case in German is
largely a dependent marking deal, since nouns themselves don't take case
endings (well, some do, but under well-understood conditions), so German
would be Mf+d-; that is, fusional morphology which, in the case of noun
phrases, is marked more consistently on the dependents (article,s
adjectives) than it is on the head noun. Tepa's score here is Mf++h+++;
that is, pretty solid fusional morphology rather than anything else, and
any inflection present is found on the head (with the exception of some
modal and pronominal clitics).

        t#      number of distinct tenses
        a#      number of distinct aspects

It might be desireable to combine tense and aspect into a single
variable t/a, since in some languages it's not easy to tease apart the
two in a verbal inflectional system. Also, lots of languages might show
tense distinctions in some aspects which are not available in others, or
some aspect distinctions in some tenses but not in others. For instance,
a project I'm working on right now has a perfective/imperfective
distinction only in the past tense. So it has three t/a categories:
present, past perfect, and past imperfect, rather than two tenses and
two aspects. Thus for that project: t/a3 instead of t2a2.

        c#      number of distinct cases

A thought just occured to me. How do we mark pronominal categories? Do
we want to? It's common for pronouns to make more distinctions than
nouns, so is the total number of cases based on the count from pronouns
as well as from noun phrases, or just from noun phrases? For instance,
English would only be c2 when looking at nouns but c3 when taking
pronouns into account. There are other pronominal categories which
aren't marked on nouns like inclusivity/exclusivity of non-singular
first persons.

We also don't have a way to indicate morphological class behavior, such
as verbal conjugation patterns (think -er, -ir, -re verbs in French) or
morphological gender in nouns.

S       syntax
        b       basic word order; s,v,o; f 'free'
        arg     argument marking
                n       nominative/accusative
                e       ergative/absolutive
                a       active/stative
                t       topic/focus
                s       split/mixed system

As I understand them, Philipino languages could be considered an example
of an 'argt' type of language, where argument marking has to do with the
status of the argument as topic. A language where word order is
dependent on topic/comment status of arguments would thus be Sbf argt. I
think. This may need to be adjusted. Tepa is Sbvosarga.

        L       lexicon
                c       compounding (add +/- as needed)
                d       derivation (add +/- as needed)
                #       number of words so far

I moved compounding to the lexicon and adopted Carlos' suggestion for a
+/- scalar annotation. I've also included derivation so that you can
indicate if you create lots of vocabulary by adding morphologically
bound elements or not. Nowa, for example, would be c++++d++++ (given
its design principle of oligosynthesis). Tepa is pretty boring here:
Lcd+600.

The complete code for Tepa is:

Tepa: Tf Pt*p+++11/4s(c)v(v/c) Mf++h+++t*a2c* Sbvosarga Lcd+600

The observant reader might notice that the phonology part is much more
articulated than the others, and from that draw the conclusion that I'm
a phonologist. That would be a safe bet :-). I would like to see more
articulation of the syntax and morphology sections as well, but don't
really feel up to the task. Thoughts?

Dirk

--
Dirk Elzinga
dirk.elzinga@m.cc.utah.edu                   "All grammars leak."
http://www.u.arizona.edu/~elzinga/                 -Edward Sapir