On 2008-12-04 20:43, Arnaud Fournet wrote:
> What acts as "selection" in that model ?
Let's suppose that we treat words as replicators. Any word can have
slightly different realisations ("alleles"), some of which replicate
with more success than others (i.e. their relative frequency increases
in the course time). Any factor responsible for this differential
survival of variants counts as a selective pressure. For example,
analogy may work like that (a word-form consistent with the rest of its
paradigm is "fitter" than an irregular one in terms of learnability).
Even orthography may have a selective effect in a literate community
(favouring spelling-pronunciations like "of[t]en"). In fact, there are
lots of such factors, both language-internal and external.
Piotr