LINGUIST List 3.588

Sat 18 Jul 1992

Disc: Rules

Editor for this issue: <>


  1. , Rules

Message 1: Rules

Date: Fri, 17 Jul 1992 11:41 EETRules
From: <>
Subject: Rules

In a message sent on 30 May Alexis Manaster Ramer writes:
> If we are to accept that linguistics is a natural science ... we must take
> behavior ("performance") as what linguistics studies [thus getting rid of
> normativity].

This is perfectly correct (and also otherwise, Manaster Ramer has got
my meaning). It is just that we cannot get rid of normativity and study
(mere) behavior, for reasons I tried to explain in chap. 7 ('The
ineliminability of linguistic normativity') of my 1978 book (=Grammatical
theory and metascience). It follows that those who study 'language as
such' (or 'E-language') are in reality studying linguistic norms; and those
who study psychology of language or cognitive linguistics (or 'I
language') must still presuppose the existence of linguistic norms
(inherent in correct or incorrect sentences). Knowing the sentences
(norms, E-language) is a prerequisite for making hypotheses about how
one learns to understand and produce sentences. Personally, I am
working on cognitive linguistics (explaining language universals etc.),
but I make no value judgement between the study of E-language and the
study of I-language, if only because Panini was a student of E-language,
and he wipes the floor with any bunch of I-language aficionados. The
important thing is that one knows what kind of linguistics one is
practicing (which surprisingly seldom turns out to be case).

I have always held that there are (at least) two different types of
linguistics: the study of language as such (='autonomous
linguistics') and the study of the psychological mechanisms 'behind the
language as such (=traditionally, 'psycholinguistics'). I remember
vividly debating this issue with some prominent Chomskyans in the 70's
and early 80's.They repeated over and over again that just like there
is only one physics (they didn't know of Lorenzen's protophysics, and
they still don't), there is only one type of linguistics. Suddenly, almost
overnight, I discovered that they had changed their minds (read:
Chomsky had changed their minds for them). Now there were two types
of linguistics (misrepresented, to be sure, and burdened with misleading
value judgements): the study of E-laguage (i.e. autonomous linguistics)
and the study of I-language (i.e. psycholinguistics). I made another
discovery as well. These ('prominent') Chomskyans had totally forgotten
having ever denied the existence of two different types of linguistics.

Let me heal one more case of collective amnesia. In his 1955 [1975]
dissertation Chomsky was a student of E-language, as I defined this
notion above:
> An attempt to construct discovery procedures for grammar is faced
> by the difficulty that it must deal in a neutral manner with the total
> linguistic behavior of the informant, including slips, slurred speech,
> interrupted utterances, etc. A more limited approach will be satisfied
> with a grammatical description of a partially hypothetical language
> underlying actual speech in the sense that actual linguistic behavior
> can easily be characterized as a special deviation from underlying
> norms [sic] (p.149, n.21.)

This is the issue Manaster Ramer raises. As he notes, Chomsky's
solution ('competence') amounts to a de facto acceptance of linguistic
('underlying') norms.

To Rob Stainton:

The discussion has never been confined to grammatical theory. Rather,
it has included standard human sciences, physics, (philosophical) logic,
and conceptual analysis as practiced within analytical philosophy. The
last point crucial. If you compare grammatical theory only to physics,
you are bound to see similarities; but if you compare it to philosophical
logic or conceptual analysis (which you regrettably never do), you are
bound to see more significant similarities. The 'ceteris paribus' argument
was taken care of twenty years ago.

Esa Itkonen, University of Turku, Finland
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue