RE: Deconstruction of Deep Structure and other Improbabilities

From: Micheal Palmer (
Date: Mon May 26 1997 - 00:39:26 EDT

At 2:59 PM +0100 5/25/97, Jim Beale wrote:
>At 11:04 AM +0000 5/25/97, Clayton Bartholomew wrote:
>>This is exactly what I was driving at when I said that invoking the
>>term *deep structure* involves accepting a whole set of
>>presuppositions. There are more presuppositions than the one stated
>>above, but one is enough to illustrate the point. My reference to
>>Augustine's Logos Doctrine was not being flippant. I think there is
>>some kind of rationalism lurking in the term *deep structure.* Jim
>>Beale has made this rationalism explicit in his statement quoted
>>above. I'm sure he will respond by saying I haven't used the word
>>rationalism correctly.
>>If you don't accept Jim's restatement of Chomsky (early Chomsky?) then
>>the term *deep structure* should be avoided.
>I am not aware of any substantial changes in his position. In
>fact, it has been restated numerous times since _Aspects_ was
>written. It is definitely anti-empirical. Though Chomsky may
>himself be a rationalist, not all who reject empiricism are to
>be classified as rationalists; except perhaps in the narrow sense
>of accepting the possession of innate knowledge of universals.

I seem to have missed Jim's restatement of Chomsky (If you're reading, Jim,
would you mind sending me a copy?), but if he's talking about Chomsky's
position on deep structure, it has very definitely changed. In his work in
the 1990s he has increasingly rejected deep structure as a usable concept.
The most thorough statement of his new positing is _The Minimalist Program_
(MIT, 1995) in which he thoroughly rejects deep structure and builds a new
research paradigm without it. He remains, however, very much a rationalist.

>>I avoid it for that reason.
>One can avoid the term, but the concept itself seems unavoidable.
>It is the basis of the possibility of translation. If one affirms
>the possibility of translation, then he also affirms the existence
>of a deep structure, call it whatever he will.

I disagree. Deep structure is a theoretical assumption which is not at all
necessary for translation. We could just as well talk of a set of logical
or semantic relationships which arise out of the surface structure (roughly
what we have in the text), which we can then restate in a number of
different ways in another language. There is no a priori reason to see the
relationships commonly attributed to deep structure as lying behind, or
beneath the text as we have it. They can just as well be seen as arising
out of that text.

>Consider a pair of examples given by Katz:
>(1a) the doctor's arrival, the arrival of the doctor, ...
>(1b) the doctor's house, the house of the doctor, ...
>(2a) John is easy to leave
>(2b) John is eager to leave
>The surface structures of each of these two are identical. But they
>possess quite different underlying syntactic structures. Since this
>is so obvious, why should one oppose the term "deep structure"?

This is *not* so obvious. Why do we have to see the differences as
'underlying'? They can also be seen as semantic differences which arise out
of the properties of words like 'arrival' and 'house' or 'easy' and
'eager'. In other words, the differences may be lexical in nature, and
require no assumption about any underlying structure. Clearly there is an
abstract level at which the implied structural relationships in these
sentences are quite different, even though their grammatical structure
(surface structure) is quite similar, but there is no reason to see this
abstract level as 'underlying' or in any sense prior to the text as we have

Deep structure can be, however, a very useful way of talking about the
logical relationships which obviously are present in virtually any verbal
or written communication. It can also be a very useful theoretical tool for
translators. It is not, however, essential.

Micheal W. Palmer
Religion & Philosophy
Meredith College

Visit the Greek Language and Linguistics Gateway at

This archive was generated by hypermail 2.1.4 : Sat Apr 20 2002 - 15:38:17 EDT