(toc) #title= (Table of Content)
What are Rules?
Rules form an important aspect of Transformational Generative Grammar (TGG). Rules are used in linguistic analysis to capture the way native speakers acquire and use their language. A competent native speaker of a language does not simply string words together randomly because he recognizes those words as the words of his language, but adheres to specific rules of the language that make his utterances fall in line with how other speakers use them. These rules not only make it possible for him to produce the correct sentences of his language, but also make his sentences acceptable by other native speakers. These rules are, therefore, generalized descriptive statements that summarize the way acceptable sentences are produced in a language. They predict and reflect the native speaker’s competence of his language.
Phrase Structure Rules
The basic idea in Chomsky’s Transformational Generative grammar is that sentences of a language can be divided into constituent parts such that each word in a sentence is a constituent of some phrase that, in turn, is ultimately a constituent of the sentence. The constituent structure of sentences is determined in part by phrase structure rules of the grammar.
In TGG, finite (i.e. limited) rules are used to generate infinite (i.e. unlimited) sentences in a language. By ‘generative’ we mean ‘to specify’ or ‘to account for’ all the grammatical sentences in a language. Phrase Structure (PS) rules generate constituent structure representations for sentences in a language. Thus, a set of PS rules can generate many different sentences in a language. To see how PS rules can do this, consider the following set of PS rules needed to generate sentences like:
The man likes the car
1. (a) S -> NP VP
(b) VP -> V NP
(c) NP -> Det N
Phrase structure rules are a subclass of rewrite rules where a single syntactic category is ‘rewritten’ or ‘expanded’ as a string of categories (including the string consisting of a single category). The arrow should be interpreted ‘rewrite as’ or ‘expand as’. Thus, rule 1(a) above should be read as ‘the category S rewrites as the string consisting of an NP followed by a VP”. The rules apply as follows. Beginning with an initial symbol S, the rules apply one at a time until each phrasal category (e.g. S, NP, VP) has been rewritten as a string of lexical categories (e.g. Det, N, V).
The rules is (1) allow several distinct but equivalent derivations – two of which are given in (2).
2. (a) S 2. (b) S NP VP (by {1a}) NP VP (by {1a})
NP V NP (by {1b}) Det N VP (by {1c})
Det N V NP (by {1c}) Det N V NP (by {1b})
Det N V Det N (by {1c}) Det N V Det N (by {1c})
The derivations 2(a) and 2(b) are equivalent in that the same phrase structure rules are applied to the same categories in each case. They differ in the order in which the rules are applied: rule (1b) precedes rule (1c) in (2a), whereas rule (1c) precedes rule (1b) in (2b). As a result, derivation (2a) contains a string NP-V-NP that is not contained in derivation (2b); and conversely, (2b) contains a string Det-N-VP that is not contained in derivation (2a). The set of strings from all equivalent derivations is called a phrase marker. The phrase marker is a set of strings that assigns a phrase structure interpretation to a terminal string (the string of lexical categories that terminates the phrase marker). The phrase marker of the terminal string Det-N-V-Det-N is the set of strings S, NP-VP, NP-V-NP, Det-N-VP, Det-N-V-NP, NP-V-Det-N, Det-N-V-Det-N. This phrase marker can be represented as a tree diagram:
S
VP
NP
likes
Transformational Rules
Phrase Structure rules are not powerful enough to provide an adequate description of the syntax of English, and the inadequacies can be overcome by introducing rules of a new kind, called transformations. A phrase structure grammar consists exclusively of PSG rules and assigns to each sentence a syntactic structure in the form of a single marker (except in the case of syntactically ambiguous sentences), whereas a Transformational Grammar consists of a set of PS rules plus a set of transformational rules and assigns to each sentence a series of phrase markers, varying in the level of abstraction involved.
A context-free PSG could account for the differences between transitive and intransitive verbs but could not at the same time express the significant likeness between them. For example, consider this active/passive pair:
1. The dog frightened the boy
2. The boy was frightened by the dog
There is a systematic syntactic relation between (1) and (2) that cannot be expressed by means of PS rules. The highly elementary apparatus available in PSG does not suffice to handle this type of relation. Since PS rules do not allow for operations involving the movement of an element from one part of the structure to another: to formalize this kind of operation we shall need transformational rules.
There are various types of Transformational Rules. There are inversion rules, movement rules, deletion rules, insertion rules, and so on. (Detailed discussion of these rules will be taken up in Module 4). Phrase structure rules and transformational rules will generate an infinite number of sentences in any language.
Transformations can perform many functions as mentioned above:
(i) they can move elements from one position to another in a sentence. Consider examples (1) and (2) above. Notice that in (2), the subject NP has been moved to the object position and vice versa.
(ii) Transformations can insert new elements into a structure. In example (2) above, by has been inserted into the sentence to generate the passive sentence.
(iii) Transformations can substitute an element in a structure. Consider examples (3) and (4) below.
3. Sam loves Sam
4. Sam loves himself
The second Sam in (3) is replaced by himself in (4). We call reflexivization transformation.
(iv) Transformations can invert the order of elements in a structure. This is evident when a declarative sentence is changed into a Yes-No question as in (5) and (6).
5. Jane can sing
6. Can Jane sing?
In the above, (6) is derived by moving the modal auxiliary to the initial position in the sentence.
(v) Transformations can also delete elements in a structure. This is also called gapping. Consider examples (7) and (8):
7. Ngozi peeled yam and Ngozi cooked it
8. Ngozi peeled yam and cooked it
Example (7) represents the deep structure. The noun Ngozi in the two conjoined clauses refers to the same person, i.e. the two noun phrases are coreferential. In this case, the second Ngozi becomes redundant, and it is deleted by a transformation on the basis of coreferentiality.
In syntax, it is usually the case that rules are ordered.
By rule ordering, we mean the logical and effective application of rules in the derivation of structure. For example, a rule can serve as an input or can produce an input for other rules to apply. In other words, the first rule must apply to produce the structure for other rules to apply. So, the rules are ordered.
Although it is possible for more than one rule to apply to a structure, rules can also be recursive. This is possible if the symbol to the left of the arrow also appears to the right of the same arrow as in (9) below:
9. NP -> NP S'
In this case, if this is the only rule we have, then, any time we expand the NP to the left of the arrow, we will get another NP S' to its right. When the first NP after the arrow is expanded, we will get the same result. Hence we have circularity in the application of the rule. This rule is, therefore, recursive since it produces an indefinite number of rules in a circular manner.
We thus see that transformational rules are powerful, and when used with phrase structure rules, can account for an indefinite number of sentences in a language.
Notational Conventions
In Generative Grammar, various notational conventions are used. These notations refer to the set of symbols which are introduced into the analysis to facilitate the formulation of a statement, especially in the form of a rule.
The syntactic rules describe the sentences of a language in terms of syntactic categories. A syntactic category is a group of words in a given language that can replace one another in any sentence of the language whatsoever without affecting grammaticality. We assume that each word of English will be assigned to its correct syntactic category in the lexicon, thus, boy, cat, and toy are nouns, collect, destroy, and disturb are verbs, purple, funny, and good are adjectives, a, an, and the are determiners, and so on. We can abbreviate the categories noun, verb, adjective, and determiners by N, V, Adj, and Det, respectively. Observe that it is now possible to talk about categories rather than about individual words. The fact that the boy destroyed a toy is a grammatical sentence of English could be stated in our grammatical description of English; but this would be an extremely cumbersome and unrevealing approach to take in general, owing to the fact that we would have to list all of the other grammatical sentences of English in precisely the same way. The list would have to include sentences like The girl destroyed the toy, A girl destroyed a toy, The man met the woman, A cow ate the hay, and so on, all of which display the identical pattern Det, N, V, Det, N. Without having access to these categories, however, it would be impossible for us to explicitly represent the fact that sentences that display this pattern are grammatical sentences of English. The problem is further compounded by the fact that the number of grammatical sentences of any language is infinite.
It is clearly impossible to construct an exhaustive list of the grammatical sentences of any language when there are an infinite number of them. The only hope of even beginning to deal with the question of what constitutes a grammatical sentence of the language is to state the conditions under which strings of words are grammatical in terms of the categories of the words in the string. We can thus state our observations in terms of rules. We will use notations in which such rules can be expressed in precise ways.
The symbol S represents the notion ‘sentence of a language’. The symbol -> means ‘may be of the form’ or ‘rewrite as’. In the grammar of English, then, the fact that a grammatical sentence may be of the form Det N V Det N is expressed as the following formula:
10. S -> Det N V Det N
Such a formula is called a phrase structure rule. It is not a rule in the sense that it says what a grammatical sentence must look like, but in the sense that it says what a grammatical sentence can look like. The rules are thus descriptive and not prescriptive. There will be other rules specifying what other sentences can look like. All of these rules together specify what sentences look like.
Other notations employed include the following:
* asterisk: indicates a deviant sentence
( ) parentheses: enclose optional elements in a rule { } braces: enclose alternative elements in a rule [ ] square brackets: enclose syntactic features / oblique: is used in context-sensitive rules double arrow: is used in transformational rules zero: indicates the absence of any element
CONCLUSION
In this unit, we have tried to provide some introductory information that will prepare you for syntactic analysis. We have discussed what rules are and provided illustrations of what Phrase Structure and Transformational rules are, and how they operate. We have also given some notational conventions that are used in current syntactic analysis.
Leave a comment behind