LINGUIST List 8.680

Wed May 7 1997

Disc: Functionalism

Editor for this issue: Anthony Rodrigues Aristar <aristarlinguistlist.org>

Directory

1. reiss, Question on functionalism

Recent discussion on the Linguist List and the OT List about
functionalism has got me wondering: It appears to be widely accepted
that two of the forces that play a role in grammars are

 1) ease of articulation is good and
 2) maintenance of contrast is good.

(1) reflects the speaker's inherent laziness and self-interest whereas
(2) can be interpreted as "be nice to the hearer". An appreciation of
the interaction of these "competing" factors is supposed to give
insight into the patterns we see in language: sometimes underlying
contrasts merge, presumably for the sake of simplifying some
articulatory challenge; sometimes contrasts are maintained, which makes
things easier for the hearer, since there is less ambiguity. No language
could be so dominated by one factor or the other such that the language
was unpronounceable or that everything came out as maximally simple
'ba', because such a language would not be very useful for
communication. I tried to see what I could come up with if I changed my
initial assumptions.

Here are two alternatives based on different views of human nature.

 A. Contrarianism
The " forces" necessary to understand language are that
 1a)difficult articulation is good and
 2a)merger of contrast is good.

(1a) means something like "the speaker likes a challenge" and (2a) says
"the heck with the hearer". As far as I can tell this set of assumptions
leads to exactly the same outcome and provides the same level of
understanding as the "standard" functionalist view sketched above.

 B. Anarchy
Assume that the effects of grammar are not motivated by any principles
of what is *good* or *bad*. Each language has a basically random
sampling of 'processes' that allow difficult articulation and merge
contrasts. Since it is probabilistically unlikely that a language would
have all of one kind of 'process', all languages have a more or less
balanced distribution, and we find both 'markedness' and 'merger' to
varying degrees in all languages. Again, this seems to lead to the same
result as the "standard" theory.

 It seems to me that all three of these sets of initial assumptions are
equally valid, since they lead to the same result. Evidence from sound
change doesn't seem to help since, as Bloomfield (Language p.371) says,
concerning the causes of sound change, "The greater simplicity of the
favored variants is a permanent factor; it can offer no possibilities of
correlation."

Can someone help me to see where I have gone wrong?



 Charles Reiss
reissalcor.concordia.ca