ORIGIN OF STRUCTURAL GRAMMAR AND ITS VARIANTS

Also Read

Origin of Structural Grammar

A combination of factors provided the impulse away from traditional grammar toward something both more objective and more accurately descriptive. We mentioned in the last unit of the last module that the techniques of traditional grammar were inadequate in handling, not only some syntactic constructions, but also the analysis of most non-European languages. Exploration and expansion of trade had brought about an increased awareness of the great variety in languages, and tentative comparisons of English and other languages were made. Throughout the late 18th and 19th centuries, scholars became interested in comparative linguistics, and the study of lexical and syntactic relationships among languages. Sir. William Jones had established a connection between Latin and Greek with Sanskrit, a language spoken in India. This observation prepared the way for the development of comparative linguistics as we know it today. The implicit or explicit identification of linguistics with the methods used in the historical and comparative study of the earliest stages of Indo-European languages was in part due to the success of this discipline (during the 19th century) in providing proof of the genetic relationships between Indo-European languages and in identifying precise rules which govern linguistic change. Comparative grammar attained such a degree of certainty and exactness that it came to be regarded as the linguistic study which is scientific par excellence. This ‘scientific’ appeal inspired scholars into finding ways of analyzing language scientifically.

With the wide appeal empiricist philosophy enjoyed in the intellectual community at this time, there was greater respect for the methods and results of science. Social scientists and philosophers, envious of the dramatic achievements in the 19th and early 20th century natural science, asked: ‘How can we be scientific too?’ The quest for this led to the advent of structural linguistics.

What is Structural Linguistics?

The exact time when the label ‘structural linguistics’ was first used is uncertain. But today, it designates those trends of linguistic which deliberately and explicitly try to gain insight into the systematic and structural character of language. In other words, any approach to the analysis of language that pays explicit attention to the way in which linguistic features can be described in terms of structures and systems is referred to as structural linguistics. In structural linguistics, the relationship between the corpus of data and the theoretical description had to be a direct one. Any linguistic statements about universals, or in fact any theoretical terms at all which were not directly labl o phenomena were ruled out; the only useful generalizations about language were inductive generalizations.

The structuralist theory of language was the first major new approach to linguistics in the 20th century. Introduced probably by the Swiss linguist, Ferdinand de Saussure and pioneered in the United States by Leonard Bloomfield, structuralism flourished for about thirty years until early 1960’s. In its emphasis on the investigation of concrete linguistic data, structuralism logically followed the late 19th century neo-grammarian school. But structuralism was geared toward descriptive linguistics.

Structural linguists analyse language in accordance with external form. This typically involves isolating, classifying, and segmenting the observed language data. For example, a structuralist encountering English for the first time would begin by collecting samples of English utterances and would then record the strings of sounds that constituted those utterances. The structuralist would find that sounds like [p] and [b] in English contrast because they distinguish words like pit and ith. Whether the [p] sound is produced with accompanying aspiration would be considered rather insignificant because it would not affect the meaning of what was being said.

After isolating and classifying significant sounds of a language, the structuralist would notice that certain sounds appear together in recurring segments that signal some form of meaning. Thus, the sounds p,r,o would often be found in succession (pro-) to signal “in favour of”. The structuralist would then classify the larger units. For example, a word like child would be labelled a noun because it has inflections for the possessive and the plural: child, child’s, children, children’s. By contrast, I with the three additional forms: me, mine, my is a pronoun because it has forms characteristic of this morphological class. Book and table would be placed in the same class because they could both be found in utterances like: The costs a lot of money, but rare in utterances like: The man the house. In summary, structuralism began with concrete language data and organized these data into levels of structure for the purpose of description.

Structural linguists view languages as consisting of various layers or strata. Each layer is treated separately; you don’t mix levels. The first layer is the phonological (analysis of the sound system of a language). There upon the forms are analyzed in a morphological stratum. Subsequently, sentences are treated in a syntactic stratum in which nominals (so-called) to distinguish them from the morphological class of nouns), verbals, and so on, are identified. Relationships between these layers are described. For example, the representation for the morphological affix indicating the genitive is labeled {Z} at the morphological level. This corresponds to three variants at the phonological level: /s, z, lz/. The precise distribution of these and other such phonological variants are determined. The study of relationships between layers was known by compound terms such as morphophonemics.

Structural grammars aimed to be highly rigorous. The elements of any one language, whether at the phonological, morphological, or other levels, were solely identified by their form. Definitions based in part on meaning were rejected. Unlike the practice of traditional grammar, nouns were not to be identified as ‘names of persons, places, or things’ but rather as elements identified by inflections. This approach was held to be greatly superior to that of traditional grammar, by which words like fire, beauty, possibility did not fit the traditional definition, for none of these is the name of a person, place or thing.

Structural grammars were especially successful in treating exotic languages at the phonological and morphological levels. In contrast with earlier grammars for exotic languages, these were not presented as languages similar in some respects to Latin, different in others. Instead, the elements were identified by their role in the various layers.

There are many variants of structural grammar, but here we shall discuss only two: Tagmemics and Scale and Category grammars.

SELF ASSESSMENT EXERCISE 1

1. What was the immediate cause of the emergence of Structural Grammar? 2. Explain what you understand by ‘Structural Grammar’.

3.2 Tagmemic Theory

Kenneth L. Pike produced in his book Language in Relation to a Unified Theory of the Structure of Human Behaviour, published in 1950, a wide-ranging discussion of current methods in linguistics and an attempt to lay the foundation of a comprehensive interpretation of language within its cultural setting. Pike elaborated a theory which he called gram(m)emics, and later tagmemics, based on the notion that utterances can be analyzed simultaneously according to three hierarchies: a lexical one (in which the minimum unit is the morpheme), a phonological one (in which the minimum unit is the phoneme, or the distinctive feature), and a grammatical one (in which the minimum unit is the tagmeme or gram(m)eme).

The tagmemic theory views language as a particular kind of behaviour. In this view, language data are patterned behaviour within a patterned context. The patterning extends to human behaviour in general, and the same kinds of pattern characterize both aspects of human behaviour – verbal and non-verbal. Such behaviour types (verbal and non-verbal) are equally meaningful.

Let us consider some patterns that are discoverable within the verbal portion of behaviour. These patterns will show sameness and recurrence in various ways. Consider the following examples:

1. Our boss asked us to leave.

2. We asked John to leave.

3. John asked us to leave.

4. John was asked to leave by our boss.

5. We elected John boss.

6. We elected John to do the dirty work.

7. We went.

A patterning may be discovered in the above examples in that identities of sound features recur. In "boss" and "asked," we find two tokens of an identity of sound pattern; in "to" and "asked," there recurs a second identity pattern. In the above examples, certain words may be identified as the same or similar in terms of sound pattern and as the same in terms of meaning, e.g., "boss" in (1), (4), and (5). To express this second type of patterning, we must provide a framework for expressing this identity of items. We see that some items once identified may occur in the same positions with respect to some sequences of items. For example, "our boss," "we," and "John" may occur initially, and in that position, they are the ‘subject’ of the sentence.

Classes of words exist to perform functions. Some classes perform only one. For example, the class of correlative conjunctions "and" and "or" (perhaps including "either," "neither," "nor," and "whether") serves just one purpose: to join together elements of equal rank. Usually, the two things joined are themselves members of a single class: "Mary and John" (two nouns), "to be or not to be" (two verbs), "it’s slimy and wet" (two adjectives).

More often, a class performs two or more functions. Then it becomes necessary to name the syntactic operations both in terms of the class that performs the function and the function that it performs. In sentences (1) and (2) above, "our boss" and "we" respectively exhibit a grammatical pattern: ‘Subject-as-Actor’. Description of this pattern consists in (i) identifying a position or slot; (ii) associating a structural meaning with the slot; and (iii) correlating this slot with a morpheme class within the lexical level.

The combination of class and function is sometimes called a ‘slot-class correlation’, and the term for it is tagmeme, and the class of items grammatically acceptable in each slot is called fillers. Tagmemes are the particles of syntax. The adjective in: "one sure thing," "one thing sure," and "the thing is sure" occurs in three different tagmemes, since each of those positions represents a different function – the meanings are not the same. The adverb "clearly" in "clearly he can’t see" and "He can’t see clearly" occurs in two tagmemes: adverb-as-sentence-modifier and adverb-as-verb-modifier.

As with other levels, particles are ranged in strings. A typical string in syntax is ‘noun-as-subject plus verb plus noun-as-object’: "Monkeys love bananas." Syntactic strings are called syntagmemes, that is, tagmemes taken together. A number of other simple sentence syntagmemes can be mentioned: noun-as-subject plus linking verb plus adjective-as-complement (Lead is soft); interrogative-pronoun-as-complement plus linking verb plus noun-as-subject (who is that man?).

Syntagmemes are the different syntactic patterns that a language provides for. Besides sentence syntagmemes, there are subordinate syntagmemes, like noun phrases and prepositional phrases. In the prepositional phrase "by the author," we find an additional noun tagmeme: noun-as-preposition-object. In the noun phrase: "the visible stars," we find adjective-as-premodifier, whereas in the noun phrase: "the stars visible," we find adjective-as-postmodifier: the functions are different because normally the first means ‘stars whose magnitude is great enough to make them visible’ while the second means ‘stars that can be seen because conditions (the weather, for example) are favourable’.

Identical tagmemes can be arranged in different syntagmemes. The result then ‘means the same’ but the ‘style’ is different. "I didn’t see John" and "John I didn’t see" contain the same noun-as-direct-object, but it occurs at the end of the first sentence and at the beginning of the second. "The gate is straight" and "straight is the gate" are different syntagmemes with identical tagmemes in reverse order.

We have tried to give a summary of the kinds of patterning that contribute to language in this theory and examined briefly some of the formal ways of accounting for this patterning.

Tagmemic grammar is well-suited for describing morphology, and it is adequate for describing syntactic patterns that have been identified in the data the linguist has collected. But the scheme is incapable of going beyond the immediate corpus. Assume, for example, that in a language the subject slot is filled by a noun. If an expanded corpus shows that this slot can also be filled by a pronoun, we must either revise the original statement, formulate an additional statement, or somehow redefine nouns and pronouns so that both are members of some larger class that can be taken as filler for the subject-slot.

Because of the great variety of alternative patterns available in most languages, any system of tagmemic notation involves more complexity. The question of course is not really how complex the descriptive system as such. If a language has complexities, a description of it will also have complexities. The problem, therefore, is to find a way to describe syntactic structures so that systematic patterns are discernible in the maze of variety that engulfs the investigator.

Although many structuralists have approached syntax in terms of tagmemic description, it has often seemed just as reasonable to offer a simple description of a few typical sentence patterns. A few examples, often in the form of a sample text, have served to illustrate the most common sentence types. The goal has been to exemplify the typical rather than to provide an exhaustive account of all possible sentence types. The typical grammar has usually contained a listing of phonemes, an extensive discussion of morphology, and a mere hint of syntax.

SELF ASSESSMENT EXERCISE 2

(a) What are the aims of Tagmemic Grammar?

(b) How useful is it in the analysis of language?

3.3 Scale and Category Grammar (Systemic Grammar)

National boundaries are usually insignificant in the development and propagation of scientific theories and methodologies. But clearly the first effects of new development will be felt in the immediate circle of a scholar’s colleagues and pupils. Transformational-generative theory started in America but has now spread its influence to almost all centres of Linguistics in the world. Tagmemics, likewise American in origin, has found adherents in Europe and elsewhere. In Great Britain, the work of J.R. Firth and the theories of the context of situation and of prosodic analysis associated with him represent a variant of structural linguistics. Since Firth’s death in 1960, work in these fields has continued, along with work on other lines, but a body of linguistic theory stemming from Firth’s teaching has been developed by Michael Halliday. It represents, in fact, an attempt to do what Firth never did, namely, to work out an explicit theory of language and of linguistic description on the basis of Firth’s teaching and his published writings. For this reason, it was first known as ‘Neo-Firthian Linguistics’. To what extent Halliday’s theory has actually developed what Firth would have worked out for himself or would acknowledge as his own, and how far Halliday has projected his own thinking onto Firth’s is open to question.

Syntactic analysis in the London style is commonly called ‘systemic grammar’ (other, less significant terms have also been used). The fundamental ‘categories’ of Halliday’s linguistic description, introduced and explained in Categories of the Theory of Grammar, are four in number: unit, structure, class, and system; additionally, these categories are related to each other and to the phonic substance along three scales: rank, delicacy, and exponence. The prominence of the two terms, scale, and category, also led to the use of scale and category linguistics by some commentators to refer to this development.

The general conception of language assumed in the theory contrasts linguistic form with substance (phonic or graphic representation) on the one side, and with situations (in relation to which linguistic form has meaning) on the other. Central in linguistic form are grammar and lexis (lexis being the vocabulary of a language, consisting of its individual lexical items). Grammar and lexis are related to their phonic representation through phonology (i.e., the distinctive sound units and sound features of the language), and to their graphic representation through orthography (the alphabet and spelling rules of the language). On the other side, the semantic functions or meanings of grammatical and lexical elements are stated in terms of contexts of situation abstracted by the linguist as descriptive and analytical frames within which to summarize the multiple relationships between linguistic forms and the world of human experience wherein they are meaningfully used.

Phonology serves to link grammar and lexis to their phonic representation. This envisages phonology very much as Firth envisaged. Within this framework, important work has been done relating tonation of English sentences to their grammatical structures and to the various semantic distinctions.

Units in grammar are such entities as sentences, clauses, phrases, words, and morphemes. They belong to the same level and are related to each other in terms of size or inclusion: sentences include, or are made up of clauses; clauses are made up of phrases (or groups); phrases are made up of words; and words are made up of morphemes. Likewise, at the phonological level, tone-groups are made up of feet, feet of syllables, and syllables of phonemes.

The interrelations of units in size or inclusion are referred to as the scale of rank, and in moving within one level (grammatical or phonological) up or down in size in a description, one is moving up or down the rank scale. Thus, broadly, Halliday’s rank corresponds to tagmemic level.

Along the rank-scale, units, except the smallest at each level, exhibit structures, that is to say, an ordered internal composition of units next below on the rank-scale within the level; sentence structures consist of clauses, clause structures of phrases or groups, and so on. It is a tenet of systemic grammar that structures must always be regarded as comprising the units next below in rank. If a clause contains only one word, it must be analyzed as a clause containing only one word. Tagmemic does not impose such a requirement, and clauses can be described directly as composed of word tagmemes.

Part of the recursive possibilities of linguistic structures is treated in terms of what is called rank-shift: a unit is shifted in rank when it occupies a structural place, not in the structure of the unit next above, but in the structure of a unit in the same rank-size as itself or below it. For example, adverbial phrases can occur as part of other phrases (by the house in our garden), and in English, relative clause constructions shift sentences to the position of modifiers in nominal group structures (I admitted (him) to my house; the man admitted to my house was an exconvict).

In conformity to what has been said, units other than the highest in rank (the largest or most inclusive) at each level are grouped into classes according to the functions they can fulfill in the structure of units next above them. For example, in English clause structure, nominal groups or phrases form a class in that they can (in general) function in positions of subject and complement; and nouns form a class because they constitute the head, with or without a modifier, of noun groups.

Each class is either a closed class, to which new members cannot readily be added, like the class of English prepositions, or an open class, whose membership is freely extensible, like the English nouns and verbs. Classes consist of the units that comprise their membership; in the case of word classes, the members are words.

Along the scale of ‘delicacy’, the subclasses of a class are said to form systems, and when the practicable limits of subclassifying have been reached in lexis, the individual lexical forms themselves constitute the terms of systems.

The scale of exponence relates form to substance, that is, the abstractions of grammar, lexis, and phonology to the actual phonic (or graphic) data, the exponents. Any descriptive move nearer the data is a move along the exponence scale. Thus, in passing from predicate as an element of clause structure to verb group, one is passing nearer to the actual data, and in passing from verb to enjoy as a lexical member of the subclass of transitive verbs, one comes nearer still to the data; and the limit on the scale of exponence is reached when the verb enjoy is finally referred to as an uttered sound sequence, or is narrowly transcribed as [en i] (or written in an actual letter sequence ‘enjoy’). Logically, these three basic scales, rank, delicacy, and exponence are independent in the theory, since they concern different sorts of relationships.

In order to grasp the rationale of systemic grammar, it is important to appreciate that its advocates do not normally suggest that it is more successful than transformational grammar at carrying out the task for which the latter was designed – namely, defining the range of grammatical sentences in a language. Systemic grammar aims rather to provide a taxonomy for sentences, a means of descriptively classifying particular sentences.

Systemic grammarians claim, with some justice, that their sort of theory is much more relevant than the generative approach to the needs of various groups of people who deal with language: while Chomskyan linguistics appeals to the psychologist, systemic linguistics appeals to the sociologist. The psychologist wants a theory that describes languages so that he can see what kinds of languages human beings are capable of using; the sociologist wants to be able to describe any patterns that emerge in the particular choices that given types of individuals make in given circumstances from the overall range provided by their language. Other purposes for which systemic grammar is held to be more relevant than transformational grammar are literary criticism and language teaching.

At the same time, there are problems about the assumptions underlying systemic theory. Alongside the notion of system, Halliday introduces into syntax the notions rank and delicacy. ‘Rank’ refers to a scale of sizes of grammatical units, roughly speaking; the lowest-ranking unit is the morpheme, the highest-ranking is the sentence, and for any given language, there will be a fixed number of intermediate ranks (English is said to have five ranks in all). Any grammatical system will operate at a specific rank. If we think in terms of Chomskyan hierarchical phraseograms, Halliday is saying, as it were, that sentences can be represented not merely as trees but as trees that are regimented in such a way that along any branch, there are the same number of intermediate nodes between the ‘root’ and the ‘leaf’. For Chomskyan grammars, this is untrue; some morphemes are dominated immediately or almost immediately by the root S node, other morphemes are reached only via a long chain of intermediate nodes and branches representing the application of many rules. Halliday appears, with his notion of rank, to be putting forward a new universal of syntactic structure. It is likely that he did not appreciate what he was committing himself to in introducing the term.

As for delicacy, this is a scale of relative preciseness of grammatical statement. Thus, car will be distinguished from shiny at a very gross syntactic level, since there are few verbal contexts in which one of the words could be substituted for the other in a syntactically well-formed sentence; on the other hand, car and hovercraft will be distinguished only at a more delicate level – the two are largely interchangeable but hovercraft does not take –s in the plural. The notion of ‘delicacy’ might be harmless, except that Halliday’s motive for introducing it is to argue that there is, in principle, no end to the process of increasing the delicacy of grammar: at a delicate enough level even the words boy and girl, for instance, would be syntactically distinct. Some scholars have refuted this, maintaining that boy and girl are syntactically equivalent at the most delicate level. What Halliday has in mind is that, for instance, the utterance This girl is pregnant is more probable than This boy is pregnant; but this is because of sociology and because people do not often utter patent untruths, not because the latter sentence is in any way un-English.

4.0 CONCLUSION

In this unit, we have tried to present, in general outline, what structural grammar is and how it works. We have also discussed two variants of structural grammar – Tagmemics and Scale and Category (systemic grammar). Although there are several variants of structural grammar, the above represents, in general, the aims and objectives of the theory.