Newsgroups: sci.lang
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!zombie.ncsc.mil!news.duke.edu!news-feed-1.peachnet.edu!insosf1.infonet.net!internet.spss.com!markrose
From: markrose@spss.com (Mark Rosenfelder)
Subject: Re: Critique : "An introduction to language", by Fromkin/Rodman
Message-ID: <D5GHJp.DrF@spss.com>
Sender: news@spss.com
Organization: SPSS Inc
References: <jon-1403951433390001@hfmac323.uio.no>
Date: Wed, 15 Mar 1995 00:33:24 GMT
Lines: 118

In article <jon-1403951433390001@hfmac323.uio.no>,
Jon Hareide Aarbakke <jon@babel.ifl.uib.no> wrote:
>This is a well-intended criticism of "An Introduction to Language" by
>Fromkin and Rodman. The intention is that linguists stick the finger in
>the ground, as we say, and make sure they don't speculate too wildly about
>the human language faculty and its relation to their own grammars. This
>might have positive consequences for theory-building.

You make many excellent and justified criticisms of this book.  However,
in many places I think you are simply showing your own pedagogical or
theoretical biases.

The pedagogical issue raised here is how to introduce students to the
field of linguistics.  There's an interesting essay by Geoffrey Pullum
on this in _The Great Eskimo Vocabulary Hoax_, contrasting American and
English approaches to the problem.  The English (according to Pullum)
preferred to give a sort of Cook's Tour of the field, presenting many
theories and viewpoints, with their positive and negative points, but not 
going into any great detail on any of them.  The Americans preferred to give 
an in-depth account of a single theoretical viewpoint.  Evidently Fromkin 
and Rodman take this approach, but you seem to be out of sympathy with it:

>It was also a surprise to discover how imperialistic the book is. The
>title must be said to be misleading - this is much more an introduction to
>a certain way of looking at language, loosely characterized as the
>generative "paradigm", than it is to language as such.  As far as I can
>see, other approaches to looking at language are not mentioned, and I find
>this monopolizing of the field a bit disturbing. 

I don't propose to say who's right or wrong, but I think it's worth noting
that Pullum ends up preferring the American system; he thinks it's better
for students to know enough about one school to understand it fairly well
than to have a superficial acquaintance with many.  James McCawley, though
his own introductory works are more careful to note controversies than
Fromkin and Rodman are, has similar concerns; he thinks it silly to pretend
that introductory students are going to be in any kind of position to
intelligently criticize entire schools of linguistics.  That will come, 
but only with a good deal more education.

>Fusing the object under study with the descriptive device, coupled with
>the difficulty of getting access to the elements assumed to exist by the
>theory except _through_ the theory,  makes the theory completely
>waterproof and hence unscientific, according to Popper; it cannot be
>falsified. The first step to clarification is to distinguish between
>formalism and object of study.

So we must not let Chomsky dominate our linguistics textbooks; but they
must take their philosophy of science from Popper?

Popper's warnings on falsification are useful, but after Kuhn and Feyerabend,
to believe that theory can be so cleanly divided from observation
is either naive, or privileging of one particular school of epistemology.

>"....the rules of grammar written by linguists or posited as being in the
>grammar should explicitly characterize the actual rules known
>unconsciously by speakers of the language.
>How can we state these rules in the most explicit and simple way ? With
>the development of linguistic theory, technical notations began to be used
>as in other sciences, both to simplify the theoretical statements and to
>reveal the "laws" of language. Every physicist knows that  E= mc2 means
>'Energy equals mass times the square of the velocity of light'. "
>
>It is unclear what is meant by unconscious knowledge, and how it is
>possible to characterize the unconscious rules if they are indeed
>inaccessible to consciousness. Surely the best we can do is to
>characterize the output of these unconscious mechanisms ?

The unclearness I would take to be your problem, not Fromkin and Rodman's;
the kind of unconscious knowledge referred to is simply the ability of 
all speakers to form grammatical sentences even if few of them can explicitly
verbalize what rules they are following.

Their book certainly implies a certain epistemology about such knowledge--
but so do you (in the above-quoted paragraph).  Frankly I'm more skeptical 
of yours than of theirs.

>The reference to physics here seems inspired by Chomsky's assertion that
>the methods of physics are the appropriate ones for linguistic inquiry.
>The methodological statement follows from the conception of linguistics as
>a branch of biology, presumably in itself reducible to physics (according
>to the same line of reasoning). In my view, this is not a very fruitful
>way of conceiving of language.

So?  Should your view be imposed in introductory linguistics classes 
instead?  You can do whatever you want in your own classes, of course.
But it's not a very compelling criticism of someone else's textbook
that it doesn't happen to match your views.

>A final quote shows the fusion of the formalism of the grammarian and the
>object s/he is studying:
>" All speakers of English have as part of their linguistic competence -
>their mental grammars- the ability to put NPs in PPs in NPs ad infinitum".
>
>Here three elements are fused into one :
>1) competence
>2) mental grammars
>3) an arbitrarily chosen formalism
>
>This cannot lead to fruitful scientific insights. It may sound appealing,
>but unfortunately it is unproven as it stands. If you wanted to find the
>NPs in people's heads, you'd need to leave the desk and get out your
>brain-scanners. You'd need a different scientific method. As it stands, it
>ranks with flat earths and geocentric astronomy.

And you are just confusing a textbook with a theory.  That Fromkin and
Rodman (allegedly) confuse fact and formalism does not mean that Chomsky
or other linguists do, and it certainly doesn't show that Chomskyan
linguistics is wrong (which you'd need to do before a comparison with
outmoded astronomical theories is even relevant).  

>I am a bit saddened to see that new students of linguistics are treated to
>such muddled thinking as I have pointed to in my discussion. Linguistics
>needs a re-orientation, and a willingness on the part of its researchers
>to face the epistemological problems that beset it. 

At best you've shown that a particular introductory textbook contains
some ill-considered and even ill-informed statements.  This hardly shows
that "linguistics" faces any epistemological problems.  
