strings and serialization

[Jess Tauber]

Hi all. In cataloguing some less usual typological features of language I
started noticing some interesting things. Languages with large
manner-encoding ideophones (apparently made up serially) oppose as a type
those which have large pathway/locational terms (such as Siouan and
Athabaskan-Eyak)- such terms also possibly serial in origin.

This raises questions about whether such languages are merely extreme
variants of those with more balanced manner/pathway terms, often
morphologized (or even lexicalized) onto stem structure, as found in
Delancey's "bipartite stem belt" and containing languages from Penutian,
Hokan, and Uto-Aztecan.

Working re-editing a dictionary of Yahgan (isolate, Tierra del Fuego) it
became clear that the language when set down (mid 19th C.) was in its eastern
dialect in the early stages of entering into "bipartite" (or in this case
even tripartite) habits. And yet many of the manner/bodypart prefixes, bases,
and pathway/locational suffixes apparently themselves were built up from
smaller parts. Thus I hypothesized that the ancestor was an isolating,
analytical language (perhaps with remnant, lexicalized traces of morphology-
as in Sinitic, or many other Tibeto-Burman). Serialization and creation of
standardized "freezes" created the basic morphemes then fed into the higher
level of bipartite structure.

Now, many Penutian languages appear to have similar serialized strings
underlying many of the base morphemes, as do a number of Hokan languages.
Others in the two stocks, especially those with much more simplification of
the bipartite scheme, use expressive roots to a much greater extent (such as
in Pomo).

But getting back to spatial terms- Siouan has elaborated pathway/locational
sets, this in contradistinction to the very simple i-, a-, o- differentiation
on the verb stem. Might we think of such pathway terms as spatial
"ideophones"? There appears to be some sort of complementary relationship
between manner and pathway terminology in terms of access to the base- even
in those languages possessing both glommed onto the base one layer appears to
be older than the other, and perhaps there has been an alternation in
historical terms. Perhaps we are witnessing evidence of an alternation
between nominal and verbal sources for the verbal equivalent of case marking,
which may also be taking place in the serializing languages of the
Tibeto-Burman stock.

Thoughts??


Best,
Jess Tauber
zylogy at aol.com