20.412 being of the plain style

               Humanist Discussion Group, Vol. 20, No. 412.
       Centre for Computing in the Humanities, King's College London
  www.kcl.ac.uk/schools/humanities/cch/research/publications/humanist.html
                        www.princeton.edu/humanist/
                     Submit to: humanist_at_princeton.edu

         Date: Thu, 25 Jan 2007 07:18:20 +0000
         From: "Clai Rice" <cxr1086_at_louisiana.edu>
         Subject: being of the plain style

Uh-oh, Willard. Mucking about with colorless green ideas might generate
an avalanche of replies. You are correct that Miller is talking about a
mode of interpretation rather than, or as much as, a restriction on
usage. I don't know about Miller, but Chomsky clearly is aware of the
type of examples you describe. He said in _Aspects_, of "deviant"
strings breaking subcategorization or selectional rules, that "it is
necessary to impose an interpretation on them somehow--this being a task
that varies in difficulty or challenge from case to case--whereas there
is no question of imposing an interpretation in the case of ... strictly
well-formed sentences" (149). So Chomsky sees two (or more) types of
interpretation at work, one that involves the universal grammar and one
that involves something else (later on 149 he mentions "analogy to
well-formed sentences" for selectional violations, but throws up his
hands for subcategorization violations). This problem at the
syntax-semantics interface has dogged him throughout all later
permutations of his theories, and is still largely ignored by many
mainstream (American) linguists.

But other approaches to language do not agree that there is one sort of
dominant mode of interpretation for language (called "grammar") and
other modes for poetic usage. In some brands of construction grammar,
for example, your example from Miller would be handled by general
inferencing at the level of the comparative construction (A is x-er than
B). Constructions specify particular relations among the items involved.
Just as there is a relation that specifies two items cannot be identical
(try "John is taller than John"), other constructions specify two items
must be hyponymic (X is a kind of Y). So if by chance you use nouns in
the two slots that aren't usually understood as hyponyms, you find a way
to adjust the understanding of the nouns to make them hyponyms anyway
("an idea is a type of virus").

So you can have a science of language I believe without backing yourself
into this kind of corner. Whether the plain style is what gives rise to
Chomsky and Miller's semantic blinders, I can't say, though it's a
provocative idea. The usual whipping boy is "truth-conditional
semantics", but that solution just begs that same question you ask.

--Clai Rice
Received on Thu Jan 25 2007 - 02:54:55 EST