Re: Branching typologies

> Date:         Thu, 27 Sep 2001 19:07:31 EDT
> From: David Peterson <DigitalScream@...>
>     In my agglutinating language (is the new term polysynthetic?),
> [...]
No, the distinction between agglutinating and polysynthetic is an old
one.

An archetypical agglutinating language has a few closed classes of
morphemes, expressing simple syntactic or derivational functions, that
can be pre- or suffixed in a specific order to the base form of nouns
and verbs, base and affixes retaining their phonemic shape, to create
an inflected form (possibly of a derived word).

An archetypical polysynthetic language has both open and closed
classes of morphemes, the closed class ones often showing syncretism
of several syntactic functions, that are affixed (including infixed in
some cases) to a root, root and affixes undergoing sometimes very
complex sound changes, to create whole phrases or sentences in a
single phonological word.

But of course, languages will very rarely conform exactly to one of
these definitions, and it's sometimes a matter of judgment whether you
want to class a given language as one or the other.

Esperanto, whatever your opinion of it, is constructed to be purely
agglutinating.

Turkish is classed as an agglutinating language, but if you see some
of the extreme example of long verb forms that people have constructed
it does seem to have affixes with more complex meanings.

Inuit languages are called polysynthetic, even though the structure is
quite simple, in the sense that the meaning is built up front to back,
each suffix modifying the meaning of the stem so far, with a single
syncretic tense/number/agreement morpheme at the end. (The phonemic
rules for the joining of morphemes can be quite complex, though).

I don't know where professional linguists draw the dividing line, but
my guess is that the most important criterion is whether there is an
open class of morphemes that can only appear as affixes.

Lars Mathiesen (U of Copenhagen CS Dep) <thorinn@...> (Humour NOT marked)