LINGUIST List 5.331

Tue 22 Mar 1994

Disc: Mainstream Linguistics

Editor for this issue: <>


  1. Esa Itkonen, Mainstream Linguistics
  2. , mainstream linguistics
  3. , Re: 5.322 Mainstream Linguistics

Message 1: Mainstream Linguistics

Date: Tue, 22 Mar 1994 12:41:32 Mainstream Linguistics
From: Esa Itkonen <>
Subject: Mainstream Linguistics

 About one week ago somebofy expressed the wish that
the discussion on mainstream linguistics should
concentrate on the point raised by Martin Haspelmath,
namely the POWER exercised in the hiring policy. This
is quite right, but there is also power of another
kind that should not be left unmentioned, namely the
power of publicity. Paul Deane assumes that the worst
thing that can happen in this area is to be ignored by
representatives of mainstream linguistics. I admit
that this is bad enough, but there can be worse. I
give one example.
 A couple of years ago I published a book with the
title Universal History of Linguistics: India, China,
Arabia, Europe (Benjamins 1991). The first review of
the book appeared in HISTORIOGRAPHIA LINGUISTICA 1992
2/3, and it contained e.g. the following remark: "To
read this book brings more interest and pleasure than
this review can convey". At about the same time John
E. Joseph noted in a rather similar vein in LANGUAGE
1992/4: "Finally [Malmberg 1991] is of course not at
all comparable to those recent works like Itkonen
(1991) and especially Auroux (1990) which aim at far
greater comprehensiveness and a much more advanced
level." Exactly one year later, however, he gave a
totally negative review of my book in LANGUAGE 1993/4.
In it he makes twelve critical remarks, eleven of
which are totally baseless. The remaining point of
criticism is that he does not like the title of my
book, and this may well be true. (I have justified
elsewhere in greater detail the claims I am making
 What had brought about Joseph's change of mind? At
the World Congress of Linguists in Qu b c, August
1992, he offered a truly Chomskyan solution to the
normativity of language: we just have to postulate
separate modules for normativity and for language. I
pointed out that treated in this way, modularity
degenerates into a deus ex machina (if it hasn't done
so already), from which Joseph inferred that I must be
one of the bad guys. And when dealing with the bad
guys, Chomskyans shy away from (practically) nothing.

 This attitude explains the interesting fact noted by
David Gil in his original posting which started the
whole mainstream-linguistics discussion: people wish
to remain anonymous when there is any danger that they
could be interpreted as offending the mainstream.
(Notice that the wish not to be singled out is
equivalent to the FEAR of being singled out.)
 Everyone could cite cases similar to mine. Just
recall Johnson-Laird reviewed by Stephen Anderson and
Bates & McWhinney reviewed by Edward Gibson. All this
makes me wonder if there is much point in letting
Chomskyans review either anti-Chomskyans or other
Chomskyans, because in both cases the result is
equally predictable. If you point out that free press
should not be tampered with, I reply that you are
right as far as FREE press is concerned.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: mainstream linguistics

Date: Tue, 22 Mar 1994 09:22:07 mainstream linguistics
From: <>
Subject: mainstream linguistics


Professor Vicki Fromkin writes:

>Glad to see the answers exposing the myth that the so-called mainstream
>GB/P&P/formal (whatever) linguists do not deal with 'real languages'...

>Another point in response to H. Samuel Wang who asks if some of the
>proposed principles in GB theory are "god-given truths" and if not
>"why are these 'generalizations' so important"? Of course they are not
>God-Given -- but as Einstein put it for physics " the empirical contents
>and their mutual relations must find their representation in the conclusions
>of the theory"

>So, if one is interested in explanation as well as description, the
>generalizations are indeed important.

I don't think any linguist would deny that "generalizations are important".
But generalizations is not the issue here. I believe a distinction needs to be
made between "linguistic generalizations" and "what is important in
GB/P&P/formal etc. linguistics".

There is something in the very nature of "mainstream linguistics" that aids in
widening the gap between the two notions. Linguistics in the Chomskian
tradition has chosen an axiomatic approach that is characteristic of pure
mathematics. Although physics is often cited as a parallel, from what I see,
"mainstream linguistics" is much more madly axiomatic than physics (there is as
yet no unified theory in physics, but "mainstream linguistics" starts out by
assuming the existence of universals). As an ex-mathematician, I'd like to
remind all linguists out there that mathematics is the description of an ideal
world, but linguistics is not. In (pure) mathematics, there is theory, theory &
nothing but theory, but linguistics has to deal with real data, real languages,
and that I think is the essence of the many messages on this topic that
stressed the importance of "real languages". Although Professor Fromkin also
mentioned that recently mainstream linguists have also started to value real
data, however good their intentions may be, an inherent danger of such a
strictly axiomatic approach is that what is "interesting" is ultimately
determined by the theory, the axioms; the data has very little say in it when
you start out by wanting to build a completely coherent world from scratch,
rather rather than assuming the more modest task of describing phenomena a
little bit at a time. Another often taken-for-granted feature of the
"mainstream" approach is the adoption of Boolian
logic under all circumstances. This has led to the denial of external
evidence, and who know what other important "real data" that would have been
important had it not been rendered unimportant by the theoretical approach.

Each theory is a kind of lens that distorts the phenomena in some way or other,
magnifying certain aspects and shrinking others. And the point I've been
trying to make is that the "mainstream" approach, due to its axiomatic and
binary nature which divorces it from the real world, is particularly prone to
data distortion. That is why we need the theory-free study of real languages.
That is why we need non-mainstream approaches, to allow us to look at the
phenomena from other angles. And that is why the domination of the present
"mainstream" is particularly unhealthy to the study of language as a whole.

Wenchao Li
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 3: Re: 5.322 Mainstream Linguistics

Date: Mon, 21 Mar 1994 23:11:36 Re: 5.322 Mainstream Linguistics
From: <>
Subject: Re: 5.322 Mainstream Linguistics

Mike Maxwell wrote:

> As for the "real" reason for inversion in Italian psych-verb
> constructions (another issue Connol[l]y and his interlocuters refer to),
> in many such issues there are really two questions: why inversion (or
> whatever construction) occurs (a pragmatic issue, perhaps, having to
> do with the speaker's intent), and the exact constraints on the
> construction--why it sometimes doesn't occur, or why it affects the
> particular part of the sentence it does (probably a syntactic issue).

Maxwell's general answer fails to address the particular problem: that
psych verbs in many languages show word order peculiarities that are due
*entirely* to the fact that the surface subject is not the Experiencer,
which typically has the form of an indirect object, but rather the
lower-ranking Theme (aka Patient). But in accusative languages, the
highest-ranking argument of the verb "ought" to be the subject. What we
then find is that the Experiencer shows various properties which are
typical of subjects in that language, while the actual subject lacks these
properties, though it has others. Very often, as in Italian, the
Experiencer has the *positional* subject properties, while the
Theme-subject controls verb agreement, reflexivization, etc. But languages
vary in how these properties are split: in Icelandic, the Experiencer
possesses most syntactic subject properties; the subject controls only verb
agreement and nominative case. At the other extreme, English pays
attention only to the surface subject, regardless of how it ranks in any

It should be emphasized that those psych verbs which do make the
Experiencer the subject show none of this unusual syntax: in accusative
languages, the Experiencer-subject possesses all alleged subject traits.
(See Belletti & Rizzi, who demonstrate that this is true for Italian.)

Pragmatics has nothing to do with such phenomena. Nor can they be
explained by syntax as the term is commonly understood: it is not a
"syntactic issue", but rather a semantic one (well, on the interface of
semantics and syntax). And while it is perhaps possible to treat such
things convincingly in a GB or Barriers framework, the explanation is
presyntactic, so that all the structural trees, nodes, projections etc. are
merely something the GB person feels duty-bound to produce: they add
nothing whatsoever to the explanation or understanding of the phenomenon.

Please note: syntax and structure can explain much; so can pragmatics.
It's just that they can't explain everything.

--Leo Connolly
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue