Re: THEORY: Active case-marking natlangs

One more thing regarding the active-case-agreement debate. This
is a summary of what Johanna Nichols says of the matter. There
are some quite interesting points she makes, imho.

The active language type strongly prefer head-marking morphology.
This is consistent with the fact that the verb is the favoured
part of speech for showing active marking. This does of course
not say that it is impossible for a dependent-marking language to
have active alignment, just that it is more rare. The table below
summarizes the number of head-, double- and dependent-marking
active languages in a survey by Nichols.

Table 1. Dominant marking of 21 languages with active alignment.

   Dominant alignment    Head    Double    Dependent    Total
        Active            14       4           3         21

Of the three active languages in Nichols' sample (Georgian, Eastern
Pomo and Tonkawa) plus Batsbi (referred to as Ts'ova-Tush by me)
that are dependent-marking, three are classified as fluid-S
(Ts'ova-Tush, Eastern Pomo and Tonkawa) and only one (Georgian)
as split-S. The fluid-S language type is very uncommon among active
languages. Thus we can see a clear tendency for fluid-S languages
to be case-marking. Head-marking languages with active alignment
are split-S with the exception of Acehnese.

This interesting correlation is explained thus: Languages with
active alignment prefer head-marking since they grammaticalize
lexical categories of verbs. However, dependent-marking active
languages are generally fluid-S, as we saw, which means that
they rather grammaticalize nominal semantic roles and not verb
categorization. In conclusion it can be put like this: alignments
that favour nominal marking (i.e. fluid-S) are associated with
grammaticalization of nominal semantic functions, and those that
favour verbal marking are associated with the grammaticalization
of verbal semantics.

This explanation is as you can see very functional, which may
please some, and upset some. I know what it does to me. :-)

And please excuse the somewhat stiff language. I've snipped large
portions of this mail from my thesis.

So do you have any comments on this?

daniel

--
<> QHEIL RYNENYA ALANDEA <> daniel.andreasson@telia.com <>
<>           RINYA LAWEA <> Daniel Andreasson           <>