accusative and ergative languages

[Larry Trask]

On Sat, 22 May 1999, Patrick C. Ryan wrote:

> However, there is also the typological angle. It is my impression
> that most typologists believe that nominative-type languages
> developed from ergative-type languages.

No; certainly not.  It is definitely not the view of most typologists
that accusative languages generally descend from ergative ancestors.
Indeed, I can't name a single linguist who holds such a view, apart from
one or two Russian linguists, notably G. Klimov.

This is not surprising, since such unidirectional development would be a
form a stadialism -- the view that every language must proceed through a
predictable series of stages.  And stadialism, we know, is wrong.

It is now perfectly clear that a language can change in any direction at
all.  For example, an ergative language can lose its ergativity and
become accusative, while an accusative language can acquire ergativity.
Likewise, a language can acquire or lose tones, or noun-cases, or
object-agreement, or any of a zillion other things.

All that we can say is that certain changes are far more likely than
others, because obvious pathways exist.  For example, an agglutinating
language can easily become fusional, and a fusional language can easily
become isolating, and an isolating language can easily become
agglutinating.  But it's not so easy for a language to change directly
from agglutinating to isolating (say), because there exists no plausible
pathway.

Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larryt at cogs.susx.ac.uk