Evolution and Grammaticalization

[Rob Freeman]

Hi Östen,

On Wednesday 01 March 2006 08:48, Östen Dahl wrote:
> ...
> (The two meanings are "a nonfunctional property or byproduct" and "an
> effect which by itself has no effects in the physical world whatever".)

I don't think we need be surprised, all words have different and even 
contradictory meanings. If they do it is probably because we need them all.

In particular it is not unusual for causes and effects to get mixed up.

To me it is your first meaning of "byproduct", in the sense of "not being a 
direct effect", or "having no direct causes" which is the most interesting 
and relevant for epiphenomena in the context of language.

Of course an epiphenomenal model of language does not mean there are no 
causes, only that the causes are not direct. We would still have grammar, it 
is just identification as an epiphenomenon would mean we would not seek to 
directly describe the grammar in terms of rules. What rules we have would 
only describe indirect causes, like ways of generalizing usage.

> The term "emergence", by the way, is arguably even more ambiguously used by
> linguists and others. I discuss the two terms "emergence" and
> "epiphenomena" in my book "The growth and maintenance of linguistic
> complexity", Benjamins 2004.

Can you outline the main issues for us? I would like to hear what you have to 
say about "emergence" (especially in the epiphenomenal sense of "having no 
direct causes" rather than the evolutionary sense of gradual change!!)

-Rob