INTRODUCTION
Structural grammar was the major theoretical model of language analysis in the early 1950s. The Immediate Constituent Structure provides an analysis of sentences which, in some respects, gives intuitively satisfactory account of syntactic structure. However, structural grammatical framework has been regarded in a number of ways. Chomsky’s objective in a generative grammar represents the first thorough-going attempt to come to grips with the problem of grammatical explicitness.
The innovative idea of a generative grammar in language amounted to a furthering of the ideas of European structuralism and also to an outgrowth of Harris’s ideas from his "Co-occurrence and Transformations in Linguistic Structure" (Language, 1957:33). Specifically influential in the development of the generative model of the theory of linguistic structure was Harris’s observation that morphemes which occur together may share a component, just as phonemes which occur together may. Much of Chomsky’s early work, especially his Syntactic Structures, advanced these ideas. Mathematics was at a stage to be most helpful to the linguists wishing to explore Harris’s suggestion. During the 1950s, automata theory received much attention. Many scholars saw the great breakthroughs in language translation through the use of computers. The possible theoretic models for a formalized description of language examined by Chomsky in his Syntactic Structures reflect the appeal of mathematical principles for the linguist in the 1950s.
For the linguist wishing to describe the structural relationships among morphemes in a formal way, the grammar must be finite, even though the number of possible morpheme sequences is infinite. The communication theoretic view is as follows:
Suppose that we have a machine that can be in any one of a finite number of different internal states, and suppose that this machine switches from one state to another by producing a certain symbol (let us say an English word). One of these states is an initial state; another is a final state. Suppose that the machine begins in the initial state, runs through a sequence of states (producing a word with each transition), and ends in the final state. Then we call the sequence of words that has been produced "a sentence" (Chomsky, Syntactic Structures, 1957:18-19).
This view gave rise to finite-state grammar, and subsequently to phrase structure grammar.
ENG 221 AN INTRODUCTION TO SYNTACTIC MODELS
2.0 OBJECTIVES
At the end of this unit, you should be able to:
• State what finite state grammar is and how it works.
• Explain what phrase structure grammar is and how it works.
3.0 MAIN CONTENT
3.1 Finite State Grammar
In Syntactic Structures (1957:24), 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 (or Markov-process) grammar. While finite-state grammar is not a serious candidate for an adequate grammar and has rarely been proposed as a model describing a syntactic system, it does merit some consideration.
A finite-state grammar is a grammar that is an abstract device but one that may be viewed as a kind of machine. Such a machine defines a language as the set of sentences produced from the initial state to the final state. The machine has a finite number of states and the capacity to change from one state to another as it registers different symbols. A machine defining a language this way automatically follows a sequence of operations programmed into it. The set of sentences produced by the machine defines a finite state language, and the machine producing that set of sentences is called a finite state grammar. A finite state grammar may be extended by altering the operations the machine is programmed to carry out. These are based on the view that sentences are generated by means of a series of choices made from ‘left-to-right’: that is to say, after the first, or leftmost element has been chosen, every subsequent choice is determined by the immediately preceding elements. According to this conception of syntactic structure, a sentence like:
1. This man has brought some bread.
Might be generated as follows: The word "this" would be selected for the first position from a list of all the words capable of occurring at the beginning of English sentences. Then, "man" would be selected as one of the words possible after "this"; "has" as one of the words that can occur after "this" and "man"; and so on. If we had selected "that" instead of "this" for the first position, the subsequent choices would have been unaffected:
That man has brought some bread.
Is an equally acceptable sentence. On the other hand, if we had first selected "those" or "these", we should then have to select words like "men" for the second position, followed by words like "have" for the third position – the possibilities for the fourth and subsequent positions being as before. And if we had selected "the" initially, we could continue with either "man" and "has" or "men" and "have".
One way of representing graphically what has just been said in words is by means of the state diagram shown below. (This is slightly more complicated than the one Chomsky gives on p.19 of Syntactic Structures).
Fig.1:
The diagram may be interpreted as follows: we can think of the grammar as a machine or device, which moves through a finite number of internal ‘states’ as it passes from the initial state ("start") to the final state ("stop") in the generation of sentences. When it has produced a word (from the set of words given as possible for that ‘state’), the grammar then ‘switches’ to a new state as determined by the arrows. Any sequence of words that can be generated in this way is thereby defined to be grammatical (in terms of the grammar represented by the diagram).
The grammar illustrated in the diagram above will generate, of course, only a finite number of sentences. It can be extended, however, by allowing the device to ‘loop’ back to the same or some previous state at particular points of choice. For example, we could add ‘loops’ between {"this", "that", "the", "same", "a",…} and {"man", "bread", "book",…} and between {"these", "those", "the", "some",…} and {"men", "books",…} making possible the selection of one or more elements from the set {"awful", "fat", "big",…}, and thus the generation of sentences beginning:
That awful man …
That big fat man …
Some big fat awful men …
The grammar could also be extended in an obvious way to allow for the generation of compound sentences like:
2. That man has brought us some bread and this beautiful girl has eaten the cheese.
Sentences such as this are still very simple in structure; and it would clearly be a complicated matter, even if it were possible, to construct a finite state grammar capable of generating a large and representative sample of the sentences of English. It would be observed, for example, that we had to put both with "this", "that", etc., and with "these", "those", etc. We should also have to put {"awful", "fat", "big", etc.} in several different places because "this awful man" and "these awful men" but not *"these awful man" and *"this awful men" are acceptable. Problems of this kind would multiply very quickly if we seriously set about the task of writing finite state grammar for English; and the conception of syntactic structure that underlies this model of description has little to recommend it other than its formal simplicity. But Chomsky proved that rejection of finite state grammar as a satisfactory model for description of a natural language is more solidly based than it would be if it rested solely upon considerations of practical complexity and our intuitions as to how certain grammatical phenomena ought to be described. He demonstrated the inadequacy of finite state grammars by pointing out that there are certain regular processes of sentence formation in English that cannot be accounted for at all, no matter how clumsy or counter-intuitive an analysis we are prepared to write, the framework of finite state grammar.
SELF ASSESSMENT EXERCISE 1
1. Define ‘finite state’ grammar.
2. To what extent is finite state grammar adequate in accounting for all and only the sentences of English?
3.2 Phrase Structure Grammar
This is a type of grammar discussed by Chomsky in Syntactic Structures (1957). Phrase structure grammars contain rules which are capable, not only of generating strings of linguistic elements, but also of providing a constituent analysis of the strings, and hence more information than finite state grammars. They are not, however, as powerful as transformational grammars, as the latter are more capable of displaying certain types of intuitive relationship between sentences.
Phrase structure grammars are currently looked upon as a bit old-fashioned by some linguists, but they remain a practical tool for the teacher and are a substantial part of the available literature, so they cannot be ignored.
A phrase structure grammar is a series of rewrite rules. These rules break down sentences, establishing their basic structures, regardless of the final form the sentences may take after transformational rules have applied. Each rule in a phrase structure grammar has the form:
X → Y + Z
which means ‘X consists of Y and Z’. The first phrase structure rule, for instance, is:
S → NP + Aux + VP
Which means that a sentence consists of a noun phrase (NP) followed by an auxiliary element (Aux) like tense that in turn is followed by a verb phrase (VP). Successive phrase structure rules indicate what NP, Aux, VP, and other constituents consist of, until there are no more constituents in the sentence to account for.
No complete phrase structure grammar for English has yet been fully accepted. It is possible that none ever will be; for the phrase structure rules depend in part on the theories that underlie them, and these change with the linguist and the year.
The phrase structure grammar (PSG) has two specific functions to fulfill. First, it must indicate what chunks can be combined in a given language to form constituents of one or more such chunks. For example, if a native speaker of English is asked to look at sentence (3) below and make some kind of logical division of that sentence into parts, he or she will come up with something like the arrangement in number (4). Every speaker of English, including the most linguistically unsophisticated, will reject number (5) as a possible division:
3. The baby kicked over the lantern
4. The/baby
The baby/kicked over
kicked over/the lantern
the/lantern
the baby/kicked over the lantern
5. The baby kicked
baby kicked over the/lantern
the/elephant kicked
That is, the native speaker has some kind of intuitive feeling that the baby is a constituent, and that the individual words are constituents, and that kicked over the lantern is a constituent. He has no such feelings about possible combinations like kicked over the or baby kicked.
The second function of the PSG is to tell us in what basic order the constituents are to be arranged. Native speakers of English will accept number (6) below, but not any of the other examples:
6. The elephant went stumbling through the orchard.
7. *Stumbling elephant the through the orchard went.
8. *Through stumbling elephant the orchard went
9. *The orchard went through stumbling the elephant
It is customary in Transformational grammar to put an asterisk in front of sentences that would not be considered grammatical by native speakers.
Having explained the two functions of PSG, it is time to bring in the actual rules:
1. S → NP + VP
2. NP → (Det) + N
3. VP → V + (NP)
This is a partial set of PSG rules for English. The first one tells us that S (sentence) is composed of NP (Noun Phrase) and a VP (Verb Phrase) in that order. The second rule says that a Noun Phrase consists of a Noun (N) which may be preceded by a Determiner (Det). The parenthesis around an element indicates that it is optional. The third rule says that a Verb Phrase is composed of a Verb (V), which may optionally be followed by a Noun Phrase (NP).
This set of rules will permit all of the sequences in examples (10)-(13), but will not allow any of (14)-(17):
10. Ngozi screamed
11. The boy left
12. The girl drank a milkshake
13. Some boys saw Elizabeth
14. *Screamed Ngozi
15. *Boy the left
16. *Milkshake a drank girl the
17. *Boys some Elizabeth saw
The acceptable sentences above are made up of words arranged in a particular order. We shall refer to words out of which sentences are composed as its ultimate constituents (they are not further analyzable at the syntactic level). The order in which the ultimate constituents occur relative to one another may be described as the linear structure of the sentence.
But how does this system assign to sentences the appropriate phrase structure? The answer to this question is given by a convention associated with the operation of ‘rewriting’. We can use the partial phrase structure grammar above to generate the deep structure of some sentences. To illustrate how the deep structure is generated, we can use a tree diagram. The relationship between the phrase structure rules and a tree diagram is quite strict; a phrase structure rule of the form X → Y + Z matches the tree:
X
|
Y Z
Let us generate the deep structure underlying sentences (10)-(13):
10. Ngozi screamed
S
/ \
NP VP
| |
N V
Ngozi Screamed
11. The boy left
S
/ \
NP VP
| |
Det V
| |
N (none)
The boy left
12. The girl drank a milkshake
S
/ \
NP VP
| |
Det VP
| / \
N V NP
| |
| Det
| |
| N
The girl drank a milkshake
13. Some boys saw Elizabeth
S
/ \
NP VP
| |
Det VP
| / \
N V NP
| |
| Det
| |
| N
Some boys saw Elizabeth
It is evident that the phrase markers given above convey the following information: in (10), the terminal elements "Ngozi screamed" form an S which consists of two constituents, NP (Ngozi) and a VP (screamed). In (11), the terminal elements "the boy left" form an S which consists of NP (the boy) and a VP (left); the NP consists of two constituents, Det (the) and N (boy). In (12), the terminal elements "the girl drank a milkshake" form an S which consists of NP (the girl) and a VP (drank a milkshake). The NP to the left of VP consists of two constituents, Det (the) and N (girl); the VP consists of V (drank) and an NP (a milkshake), and the NP to the right of VP consists of two constituents, Det (a) and N (milkshake). We can account for (13) in the same way.
We thus see that the phrase structure grammar is much more satisfactory than the finite state grammar. Any set of sentences that can be generated by a finite state grammar can be generated by a phrase structure grammar, but the converse does not hold: there are sets of sentences that can be generated by a phrase structure grammar, but not by a finite state grammar. Phrase structure grammars are intrinsically more powerful than finite state grammars – they can do everything that finite state grammars can do – and more. Phrase structure grammars contain rules which are capable not only of generating strings of linguistic elements but also of providing a constituent analysis of strings, and hence more information than finite state grammars.
SELF ASSESSMENT EXERCISE 2
1. Define ‘phrase structure grammar’
2. State two functions of phrase structure grammars
3. Using the partial PSG rules discussed above, generate the deep structure of the following sentences with the aid of tree diagrams:
(a) The student passed the exam.
(b) Some scholars received their pensions.
(c) Students drive flashy cars.
4.0 CONCLUSION
In this unit, we have tried to present in an outline, what finite state grammars and phrase structure grammars are. Neither the traditionalists nor the structuralists provided an adequate account of the syntax of natural languages. Although the grammars we have discussed are not without problems in syntactic analysis, they recognize that language is creative, if not infinite. Therefore, grammars must be generative.
5.0 SUMMARY
You have learned in this unit:
• What finite grammars are;
• What phrase structure grammars are; and
• How to use tree diagrams to generate deep structures of sentences.
6.0 TUTOR-MARKED ASSIGNMENT
1. What do you understand by a tree diagram?
2. Compare and contrast finite state grammar and phrase structure grammar.
3. Draw the deep structure trees for the following sentences:
(a) The Vice Chancellor attended the party
(b) She wept
(c) Most lawyers are Senior Advocates
(d) The musician performed
Leave a comment behind