Page 362 - DENG504_LINGUISTICS

Linguistics

Notes PS Rules vs. T Rules
Transformational rules map deep structures into intermediate or surface structures. They operate on
the output of the underlying string as well as on their own output. They need the derivational history
of underlying string. They do not apply to single, unexpanded nodes one at a time to generate deep
structures, but rather to whole trees. They are not the expansion rules which apply on single nodes,
but are used to perform the operation of permutation (rearrangement of the parts of a tree), deletion
(removing parts of a tree), and adjunction (attaching elements to nodes of a tree), etc. All of them are
not obligatory; some are obligatory while others are optional. They signify grammatical distinctions;
they simplify grammars. They capture the LSG (linguistically significant generalizations).
Transformations can be singular (operating on a single Phrase-marker) or generalized (operating on
two or more Phrase-markers) or produce a single new Phrase-marker by a process of embedding or
conjoining). The following pairs of sentences represent singular transformations :
(i) (1) Mohan kissed Radha —> Radha was kissed by Mohan
(2) He will go —> Will he go ?
(3) She was good —> She wasn’t good.
The following sentence represents generalized transformations :
(ii) (1) The linguist wrote this book) The linguist who wrote this book
The linguist bought a house) bought a house.
(2) He said it) He said that he was waiting.
He was waiting
(3) They have to work hard Their having to work hard is a nusiance
It is nuisance

Transformational rules cannot convert one P-marker into another P-marker. They are applied in a
definite order. They account for structural index, structural description, and structural change. They
can disambiguate the ambiguities better than PS rules. They are more powerful than PS rules, as they
account for more facts or account for them more ‘correctly’. Recursion is the property of PS rules, and
not of T rules.
PS rules rewrite one symbol into another. Only one symbol at a time is expanded by them. They do
not allow addition, deletion and permutation. They specify the categorical component of the base
and define basic grammatical relations that function in the deep structure. They determine the ordering
(organization) of elements in the deep structure. PS rules are both biniary and sequential. They do
not generate a sentence but only an underlying string : we cannot always generate a full grammatical
sentence without applying some transformational rule to the underlying string. In the base component,
we cannot expand a symbol into zero; we cannot expand a symbol into the same symbol (* A—> A).
At a given time there should only be one symbol (*BC—> DEF).
We can summarize the difference between PS rules and T rules in the following manner :
(1) PS rules do not require any derivational history; they are somewhat independent. T rules require
derivational history; they are applied to the output of the PS rules (underlying string); and thus
are somewhat dependent, or supplementary.
(2) PS rules apply to single, unexpanded nodes, one at a time, and generate deep structures. T rules
apply to the whole string (deep structure) to produce intermediate and surface structures.
(3) PS rules account for the relationships among constituents and sub-categories; T rules account
for various syntactic relations between one sentence and other sentence/s. That is to say T rules
show inter-tree relation, PS rules indicate ‘intratree’ relations.
(4) PS rules cannot delete, or adjunct, or permute; T rules can.
(5) All PS rules are (though a rule may contain optionality within it) obligatory; some T rules are
obligatory and some optional.

356 LOVELY PROFESSIONAL UNIVERSITY