II/ Chomsky's Theory: Generative Grammar

From the mid 1950's, the American linguist Noam Chomsky, established a number of
objectives which have continued to direct the cause of linguistic research to the present day.
Central to Chomsky's approach is the notion of the well-formed sentence. Any theory
capable of producing well-formed sentences and rejecting ill-formed sentences is referred to
as Generative Grammar. In his book 'syntactic structures' (1957), some principles
characterized his work from that of the structural school led by L.Bloomfield.

1. Grammar should be formalized as a system of rules that generates an infinite set of

sentences. This approach contrasted sharply with the fashionable orthodoxy that believed in
the application of appropriate procedures to a corpus of data which would yield a
grammatical description. This was rejected and the question raised was whether grammar
could generate an infinite number of grammatical sentences.

2.Chomsky lays great importance on creativity or open-endness of human language and

claims that the theory of grammar should reflect the ability that all fluent speakers of
language possess to produce and understand sentences which they have never heard before.
Example: The board behind me is made of wood. This seems to be new and understandable
to students. This creative character shows that language cannot simply be learned by
imitation, rather it is acquired by principles of abstraction which some linguists argue that
these principles are innate.

3.The number of words in the vocabulary of English is finite, and so is the number of rules
involved in the generation of the English sentences, thus grammar consists of a finite set of
rules operating upon a finite vocabulary to be able to generate an infinite set of sentences.
Recursion is a model of this.

4.Natural languages typically allow constructions that involve the repetitive occurrence of
elements. This phenomenon is known as recursion and some people call it iteration. Any
attempt to describe such growing structures by enumerating the elements in them faces the
problem that there is no upper limit to the length of such sentences with a very simple rule
system. However, the recursive character of natural language can be described.
Examples of the recursive rules
a. The first one is called coordination( S, NP)
EX: Jack saw Bill and Bill saw Mary and Mary saw Andrew and ... there is no end to this
sentence and apply the same rule again and again.
b. Adjectival iteration
EX: The wonderful, long, narrative, English ...tale
c. Prepositional phrase attachment
EX: The ball in the box under the table with the lamp ..there is no end to the number of pps
that can be added to this attachment.

-1-
5.Even though the number of sentences is infinite, not all sentences that we use in human
conversation find their way into a grammatical theory. Syntacticians define a specific level
of abstraction referred to as competence. As a native speaker of language, we are able to
make numerous intuitive judgments about our language. We do not have to consult grammar
books, or interview large groups of people of native speakers, rather by virtual of knowing a
language, we know that sentences are fine or bad. This is referred to as competence.
Consider some English sentences. Some examples:
1. The table saw the woman. The sentence is grammatically correct(s+v+o) but meaningfully
incorrect.??(odd)
2.John put the car............? sthg missing here.
ungrammatical ( it violates the function of the word put which requires presence of adverb of
place where)
3. The car put John in the garage. ungrammatical (word order).

COMPONENTS OF GRAMMAR
two components constitute the core of generative grammar and interact with each other.

1.Phrase Structure Rules PSG

Its central task is the structural description of all well-formed sentences of language. To
achieve this goal, we need a well defined rule system that combines all categories into
successively larger units. The tree is the visual manifestation of syntactic rules. This rule
system consists of phrase structure rules. A technique for analyzing sentences is referred to
as constituent analysis.
2.Lexicon
interacts with the PS component and it contains information about the lexemes of a
language. This includes phonological, morphological, syntactic and semantic aspects.
The whole system works as follows:
The phrase structure component contains a limited set of rules with which basic sentence
structures can be generated. These structures are licensed by means of lexical information i.e
the lexicon delivers the lexemes on the basis of rule information the basic sentence structures
can be generated.
Together PS component and lexicon constitute the core of any modern Generative
Grammar.

THE THREE MODELS OF GRAMMAR

1.Finite State Grammar
Finite-state grammar is known within mathematics as finite-state Markov processes, and
a model of this type is used by Hockett (1955) to model the ‘single uniqueness’ of human
beings among other animals, namely ‘the possession of speech’(1955: 3).
In 'syntactic structures' (1957), Chomsky suggests that the simplest type of grammar which,
with a finite amount of apparatus, can generate an infinite number of sentences is a finite-
state grammar.
A finite-state grammar may be viewed as a kind of machine, the machine has a finite number
of states and the capacity to change from one state to another. Among these states, one is the
initial state(So) and there is at least one final state (Sf).
-2-
It is based on the view that sentences are generated by means of a series of choices made
from left to right. After the first choice or element has been selected every subsequent choice
is determined by the immediately preceding element. In producing a sentence, the speaker
begins in the initial state by producing the first word of the sentence, thereby switching into a
second state which limits the choice of the second word. Each state through which he passes
represents the grammatical restrictions that limit the choice of the next word in the sentence.
Chomsky (1957: 19) produces the following state diagram:

By adding loops to a grammar of this kind, it can become able to produce an indefinite
number of sentences, and will thus satisfy one of the requirements grammars must meet: the
requirement that the grammar generate an infinite number of sentences from the finite
linguistic material the language provides. At the end of the sentence, the ‘final state’ will
have been reached.

Such a finite-state grammar would represent a language that permitted the sequence:
-The man comes.
-The old man comes.
-The very old man comes.
-The very very old man comes.
ECT.
-The men come.
-The old men come.
-The very old men come.
-The very very old men come.

Another requirement that grammars of natural languages must meet is that they must be able
to generate all of the possible sentences of the language, and Chomsky argues convincingly
that no finite-state grammar will be able to meet this condition, since no natural language is a
finite-state language.

-3-
2.Phrase Structure Grammar(PSG).
It is the second of Chomsky's three models for the description of language. F.S.G was
incapable of generating certain sentences of English, P.S.G is much more powerful and
satisfactory from this point of view than F.S.G. The phrase-structure component consists of
a set of phrase-structure (PS) rules which formalize some of the traditional insights of
constituent structure analysis.
As a simple example of the new form for grammars associated with constituent analysis,
consider the following:

Sentence → NP + VP
NP → T + N + Number
Number → {sing, pl}
VP → Verb + NP
Verb → Aux + V
Aux → Tense
Tense → {pres, past}
T → the
N → man, ball, etc.
V → hit, took, etc.
(NP (Noun Phrase); T (Articles etc.); VP (Verb Phrase); Aux (Auxiliary verb: for ease of
exposition the structure of Aux is radically simplified to cover only a marker of Tense);items

in curly brackets { } are alternatives, i.e., Number is either sing(ular) or p(lura)l, Tense is
either pres(ent) or past.)
Each rule is an instruction to rewrite the symbol on the left of the arrow as the symbol or
symbols on the right: informally, it can be construed as ‘the category on the left of the arrow
has the constituent(s) specified on the right of the arrow’. A derivation from this grammar
can then be represented by the tree shown in Figure 1 (adapted from Chomsky, 1957, p. 27).

-4-
 (i) S--------NP+ VP
(ii) NP-----T+N
(iii) VP-----V + NP
(iv) T--------the
(V) N--------man, ball, etc..
(VI) V-------hit, took, etc..

-3-
This set of rules is a simple P.S.G. Each of these rules is of the form x-----y where x is a
single element and y is a string consisting of one or more elements. In interpreting each rule
of x-----y of the example above as the instruction "rewrite x as y", we shall call the rules
below a derivation of sentence" the man hit the ball".

Sentence
(1) NP+VP
(2) T+NP+VP
(3) T+N+V+NP
(4) The+N+V+NP
(5) The+man+V+NP
(6) The+man+hit+NP
(7) The+man+hit+T+N
(8) The+man+hit+the+N
(9) The+man+hit+the+ball.
This derivation can be represented in an obvious way by means of the following diagram:
S

NP VP

T N V NP

T N

the man hit the ball

The diagram conveys less information than the derivation of the rules above since it does
not tell us in what order the rules were applied in .