THE CAMBRIDGE ENCYCLOPEDIA OF LANGUAGE

Date: 2015-10-07; view: 491.

DAVID CRYSTAL

13 LINGUISTIC LEVELS

There is too much going on in a piece of speech, writing, or signing to permit us to describe its characteristics in a single, simple statement. Even in a short spoken sentence such as Hello there!, several things are taking place at once. Each word conveys a particular meaning. There is a likely order in which the words may appear – we would not say There hello! Each word is composed of a specific sequence of sounds. The sentence as a whole uttered in a particular tone of voice (poorly signalled in writing through the exclamation mark). And the choice of this sentence immediately constrains the occasions when it might be used – on a first meeting (and not, for example, upon leave-taking). While we say or hear the sentence, we are not consciously aware of all these facets of its structure, but once our attention is drawn to them, we easily recognize their existence. We could even concentrate on the study of one of these facets largely to the exclusion of the others – something that takes place routinely in language teaching, for instance, where someone may learn about aspects of 'pronunciation' one day, and of 'vocabulary' or 'grammar' the next.

Selective focusing of this kind in fact takes place in all linguistic studies, as part of the business of discovering how language works, and of simplifying the task of description. The different facets are usually referred to as levels of linguistic organization. Each level is studied using its own terms and techniques, enabling us to obtain information about one aspect of language structure, while temporarily disregarding the involvement of others. The field of pronunciation, for example, is basically analysed at the level of phonetics, using procedures that are quite distinct from anything encountered at other linguistic levels. When we do phonetic research, we try to disassociate ourselves from the problems and practices we would encounter if we were carrying out a study at the level of, say, grammar. Similarly, grammatical study takes place using approaches that are in principle independent of what goes on in phonetics. And other levels, likewise, provide us with their own independent 'slant' on the workings of language structure.

The notion of levels is widely applicable, especially when we engage in the analysis of a range of languages, as it enables us to see and state patterns of organization more clearly and succinctly than any other way that has so far been devised. Levels appear to have a certain empirical validity in psychological and neurological contexts also. At the same time, we must never forget that, when we isolate a level for independent study, we are introducing an artificial element into our enquiry, whose consequences must be anticipated. The sounds of speech that we study via phonetics are, after all, the substance through which the patterns of grammar are conveyed. There will therefore be interrelationships between levels that need to be taken into account if we wish to understand the way language as a whole is organized. As with any structure, the whole cannot be broken down into its constituent parts without loss; and we must therefore always recollect the need to place our work on individual levels within a more general structural perspective.

HOW MANY LEVELS?

It is not difficult to sense the complexity of language structure, but it is not so easy to say how many levels should be set up in order to explain the way this structure is organized. Some simple models of language recognize only two basic levels: the set of physical forms (sounds, letters, signs, words) contained in a language, and the range of abstract meanings conveyed by these forms. More commonly, the notion of forms is sub-divided, to distinguish different kinds of abstractness. In speech, for example, the physical facts of pronunciation, as defined by the processes of articulation, acoustic transmission and audition, are considered to be the subject matter of phonetics. The way different languages organize sounds to convey differences of meaning is the province of phonology. And the study of the way meaningful units are brought into sequence to convey wider and more varied patterns of meaning is the province of grammar. The term semantics is then used for the study of the patterns of meaning themselves.

Four-level models of language (phonetics / phonology / grammar / semantics) are among the most widely used, but further divisions within and between these levels are often made. For example, within the level of grammar, it is common to recognize a distinction between the study of word structure (morphology) and the study of word sequence within sentences (syntax). Within phonology, the study of vowels, consonants, and syllables (segmental phonology) is usually distinguished from the study of prosody and other tones of voice (suprasegmental phonology). Within semantics, the study of vocabulary (or lexicon) is sometimes taken separately from the study of larger patterns of meaning (under such headings as text or discourse). All of these are regularly referred to as 'levels' of structure.

We could continue, making divisions within divisions, and recognizing more subtle kinds of structural organization within language. We could extend the notion to include other aspects of language functioning apart from structure (as when some scholars talk of a pragmatic 'level'. But there comes a point when the notion ceases to be helpful. When a theory sets up a large number of levels, it becomes difficult to plot the relationships between them, and to retain a sense of how they integrate into a single system. At that point, alternative models need to be devised.

WHICH LEVEL FIRST?

Is there a 'best' direction for the study of a language, using the framework of levels? The American linguist Leonard Bloomfield (1887 – 1949) recommended an approach in which one worked through the various levels in a particular order, beginning with a phonetic description, proceeding through phonology, morphology, and syntax, and concluding with semantics. In this view, the analysis at each level apart from the first is dependent on what has gone before. Workers in the Bloomfieldian tradition would talk about starting at the 'bottom', with the phonetics, and working 'up' to the semantics – though in view of the complexity of the task facing phoneticians, phonologists, and grammarians, it is a moot point whether anyone who strictly followed this approach would ever arrive there. Apocryphal stories abound of informants in field studies who have died before the investigating linguist got around to studying the meaning of the speech patterns that had been so painstakingly transcribed!

In any case, it is now recognized that it is possible to carry out an analysis at one level only if we make certain assumptions about other levels. Our choice of sounds to describe phonetically depends to some extent on our awareness of which sounds play an important role in a language (phonology), which in turn depends on our awareness of the way sounds distinguish words (grammar) enabling them to convey differences in meaning (semantics). Similarly, when we study grammatical patterns, such as sentence structure, we need to be aware of both semantic factors (such as the relationships of meaning that bring the patterns together) and phonological factors (such as the features of intonation that help to identify sentence units in speech). In a sense, when we work with levels, we need to be able to move in all directions at once. The British linguist J.R. Firth (1890 – 1960) once likened the business to a lift that moves freely from one level to another, in either direction, without giving priority to any one level. The simile makes its point, but the two-dimensional analogy is still misleading. To capture the notion of levels, multi-dimensional geometries are required.