Evolution, functionalism and sociolinguistics

[Bill Croft]

I am sorry I used the terms "evolution of language" vs
"evolution of languages". I should have said "origin of
language" (which is part of biological evolution) and
"evolutionary model of language change". I won't say anything
about the former; others have thought longer and harder about
these questions, as is clear from earlier posts. But I wish
to add some things about the latter, which is central to the
relationship between functionalism and sociolinguistics. In
sum, I don't see any incompatibility between the two. One can
be a good functionalist and embrace sociolinguistics as well.

     Why an evolutionary model for language change? Evolution
is a theory of change by replication. It is applicable to any
phenomenon that changes by replication. The theory just
happened to have originated in biology, where replication
occurs through reproduction by organisms. But it also applies
to language. We replicate linguistic structures every time we
converse with one another; through altered replication of
structures, and selection of variant structures, language
changes. This is as usage-based a model of language change as
one can get - a model which most functionalists advocate, and
is opposed to the child-based model of language change
assumed by many structuralists and virtually all Chomskyans.
It is also in harmony with the sociolinguistic approach,
which is based entirely on language use.

     David Hull has developed the evolutionary model in his
"Science as a process" (U Chicago Press, 1988). Evolutionary
change is fundamentally based on variation, as Talmy Givon
notes - also a central tenet of sociolinguistics. Another
central feature of evolution is that it is a two-step
process, as Ernst Mayr has stated many times (e.g. in
"Evolution", Scientific American, 1978). The two processes
are altered replication (innovation), which produces
variation, and selection (differential replication,
propagation), where some variants are favored over others in
replication. In biology, altered replication occurs through
mutation and recombination, and selection through adaptation
(among other mechanisms). But these mechanisms are domain-
specific. Other mechanisms operate in other domains, such as
language.

      In language change, I hypothesized that altered
replication is functional, in a sense that most
functionalists would find familiar, and selection is social,
in the sense familiar to sociolinguists. This does NOT mean
that language change is "only 'culturally' and 'socially'
constrained", as Talmy puts it. It simply means that
propagation is socially constrained. But evolutionary change
is a two-step process. Altered replication is the other step.
It is functionally constrained. It is directed, leading to
the functionally motivated unidirectional processes so
familiar to us. That is, altered replication is more likely
to occur in a functionally motivated direction. Even if the
selection process is functionally neutral, as I hypothesize,
the net result will be the functionally motivated historical
changes and crosslinguistic variation that we find. There is
no incompatibility here. In fact, the two together make for a
powerful theory of language as a whole, a comprehensive
alternative to Chomsky's theory.

      Finally, there is both continuity and discontinuity
between biological evolution and language change. The
discontinuity is that the replicators in biological evolution
are genes, while the replicators in language change are
linguistic structures (what I call 'linguemes'). Two
different evolving populations, with two different sets of
mechanisms. The continuity is found in what Hull calls the
interactor, the entity whose behavior causes variation and
selection of replicators. The interactor in language change
is the speaker, that is, a human being; and a human being -
an organism - is also the interactor in biological evolution.
In both cases, as Talmy says, human behavior drives
evolutionary change, that is, change by replication.

Bill Croft