Page 39 - Communication and the Evolution of Society

16 Communication and Evolution of Society

indistinguishable from that between theory and reality in other
nomological sciences, in the explicative version the linguistic
character of the object necessitates a relation that can hold only
between different linguistic expressions: the relation between
explication and explicandum, where the language of explication
(that is, the construct language of linguistic science, which is a
standardized version of ordinary language) belongs in principle
to the same level as the natural language to be explicated. (Nei-
ther in the descriptive nor in the explicative case of theory forma-
tion can the relation of linguistic theory to its object domain be
conceived as that of metalanguage to object language. )*°
Theory and Everyday Knowledge. There is yet another pecu-
liarity arising from these differently oriented conceptualizations.
An empirical-analytic theory in the narrow sense can (and as a
rule will) refute the everyday knowledge of an object domain that
we possess prior to science and replace it with a correct theoretical
knowledge regarded provisionally as true. A proposal for recon-
struction, by contrast, can represent pretheoretical knowledge
more or less explicitly and adequately, but it can never falsify it.
At most, the report of a speaker’s intuition can prove to be false,
but not the intuition itself.3* The latter belongs to the data, and
data can be explained but not criticized. At most, data can be
criticized as being unsuitable, that is, either erroneously gathered
or wrongly selected for a specific theoretical purpose.
To a certain extent, reconstructions make an essentialist claim.
Of course, one can say that theoretical descriptions correspond
(if true) to certain structures of reality in the same sense as
reconstructions bear a likeness (if correct) to the deep structures
explicated. On the other hand, the asserted correspondence be-
tween a descriptive theory and an object allows of many epis-
temological interpretations other than the realistic (e.g., instru-
mentalist or conventionalist). Rational reconstructions, on the
contrary, can reproduce the pretheoretical knowledge that they
explicate only in an essentialist sense; if they are true, they have
to correspond precisely to the rules that are operatively effective
in the object domain—that is, to the rules that actually determine
the production of surface structures.*7 Thus Chomsky’s correla-
tion assumption, according to which linguistic grammar is rep-