Functional and social factors in language (change)

[Bill Croft]

      This message picks up a thread begun by Martin
Haspelmath some months ago, on a hypothesis about factors in
language change proposed in my book "Explaining Language
Change: An Evolutionary Approach" (Longman, 2000). The
hypothesis is:

Mechanisms of innovation are functional, mechanisms of
propagation are social.

[By 'functional' I mean 'pertaining to the mapping between
morphosyntactic form and semantic-pragmatic substance, and
between phonological form and phonetic substance'. By
'social' I mean 'pertaining to language in social interaction
and social organization'.]

      The chief objection to this hypothesis that I have
encountered (in the thread and elsewhere) is the belief that
mechanisms of propagation can be functional as well as
social. I have just come across a reference to another line
of evidence supporting my hypothesis, having to do with
borrowing patterns. The reference is Cecil Brown's "Lexical
Acculturation in Native American Languages" (Oxford, 1999),
p. 9:

"The findings of these studies [he cites Brown 1987b, 1989b,
1994; Voegelin & Hymes 1953; Bright 1960a, Dozier 1956] and
of others (e.g. Diebold 1962, Spencer 1947; Scotton & Okeju
1973; Mixco 1977) challenge the long-held assumption of
linguists that structural features of some languages make
them more inclined than others to adopt loanwords for
introduced items (e.g., Sapir 1921:205-10; Herzog 1941:74;
Haugen 1956:66; Bonvillain 1978:32). For example, Haugen
writes that 'loanwords are easily accepted by languages with
unified, unanalyzed words, but not by languages with active
methods of word compounding' (p. 66). Thomason and Kaufman
(1988) assemble considerable evidence that indicates -
reminiscent of Bright (1960a) - that sociolinguistic factors,
such as the degree of bilingualism, rather than language
structure, significantly affect the extent of borowing,
especially when grammatical features are involved."

      What Brown calls "structural features" falls under the
definition of 'functional' given above, since a feature such
as "unanalyzed words" refers to the form-meaning mapping in
the language. The adoption of loanwords by a language is the
propagation of the novel (borrowed) form in the speech
community. Hence the research Brown summarizes indicates that
propagation of borrowed words and grammatical features is
driven by social, not functional, factors.

      I am reviving this thread not simply because of this
further evidence for the hypothesis I proposed. There is a
bigger issue here which I think functionalists must address.
A truly comprehensive alternative theory of language to the
Chomskyan one must integrate functionalist and
sociolinguistic theories and empirical results. After all,
like functionalism, sociolinguistics is fundamentally usage-
based, variationist, and in fact functional, in that the
function of language is to facilitate social interaction in
social groups.

Bill Croft