Journal of Language and Literature Volume 2 Number 1 2003 ISSN 1478 - 9116

1. Introduction

This paper is an exploration, from a language ecology perspective, of aspects of the construction of meaning in language. These are exemplified through analysis of a poem, but the primary focus of the paper is on linguistic rather than literary issues. Linguistic expression is taken as in essence poetic (Goodman, 1973) and, although not all writers would agree (Bradford, 1993: xiii), good poetry can be said to provide an excellent illustration of basic linguistic processes employed in a skilful manner.

The nascent field of language ecology owes its beginnings to a proposal by the linguist Einar Haugen (1971) that biological ecology would provide a helpful metaphor for studying the interaction of language and its social environment. Recent developments, however, have begun to point the way to a deeper and more all-encompassing relationship between ecological thought and language (Garner, forthcoming, 2003). Ecological thinking has begun to challenge many fundamental philosophical assumptions in a diverse range of disciplines (Hayward, 1995: 8). Although there is no clear consensus on what constitutes an ecological way of thinking (Laferrière and Stoett, 1999: 24-5), four common characteristics can be identified in the literature. Ecological thinking focusses on phenomena as holistic, dynamic, interactive, and situated. From this perspective, language is part of the large, complex whole of human life, in constant interaction with other aspects of human behaviour, and dependent for its significance on the situations in which it occurs.

Every instance of language use is intended to express something and to communicate that something to another human being. A particular element of language, such as a word or an utterance, never occurs alone, but always in the context of other language and other communicative behaviour, in a particular physical and cultural environment, and so on. A holistic view of the object of linguistic study goes beyond the traditional concerns with constructing abstract grammars for decontextualized sentences (Linell, 2001).

Ecologically, language is a manifestation of human sociality. Sociality can be defined as the "capacity for complex social behaviour" (Carrithers, 1992: 34); in humans, it takes the form of "social practices ordered across space and time" (Giddens, 1984: 2). In particular, it is realized in systems of meaningful behaviour that constitute community, communication, and culture. An ecological linguistic description shows how the meaning of any given linguistic item is interdependent with those systems.

There is no space here to attempt a complete ecological description of meaning (see Garner, forthcoming, 2003, for a more extensive treatment). Discussion is limited to two specific aspects of the construction of meaning. One is the dynamic balance between countervailing imperatives in language: for stability and predictability on the one hand, and for change and creativity on the other. The second is the construction of meaning through word association, a process which in poetry is characterized by the controlled interplay of predictability and creativity.

The first two sections of the paper give a brief theoretical outline of predictability and creativity, whilst the third examines the associative networks of words. This is followed by an analysis of aspects of language meaning in a poem in the light of the theoretical concepts.

2. Language Use and the Predictability Imperative

It is a basic supposition of traditional linguistics that "language is rule-governed, i.e., that speakers obey an internalized set of instructions in the way they construct and use sentences" (Finch, 2000: 1). Researchers in socially-oriented fields, however, such as sociolinguistics (Linell, 2001) and pragmatics (Wilson, 2001: 339), have argued that focussing on abstract systems of rules gives a misleading picture of what it means to learn how to use language (e.g., Hymes, 1964; Narasimhan, 1998). Corpus studies, too, reveal a very large proportion of prefabricated strings in normal usage that are not constructed from rules but are used as wholes (Wray and Perkins, 2000). From an ecological perspective, language consists of a vast repertoire of patterns that the language-user learns, along with appropriate non-linguistic behaviour, through constant communicative use.

Patterning in language results from the imperative (apparently common to all sentient creatures) for iterative behaviour. The repetitiveness of linguistic patterns gives it the essential element of predictability, which is communicatively highly functional:

In normal interaction, the default setting is formulaicity, both for production and for comprehension … [T]his enables the individual to focus his/her analytic faculties away from the linguistic "packaging" and onto the production and evaluation of propositions, the updating of contextual information, and the making of predictions about what is going to happen next. (Wray and Perkins, 2000: 19)

Beyond the individual, sharing a repertoire of communicative behaviour expresses communal membership and locates the individual within the collective. The community serves to keep uncertainty and misunderstanding, potentially ever-present in human interactions, to an acceptable level, and thus enables normal living. Predictability is a safeguard against breakdown in communication and hence communality. It is a hedge against chaos.

An important way in which a literate speech community maintains predictability in language is through universally accepted, formalized standards for the language. These are enshrined in recognized authorities such as dictionaries, grammars, and normative texts (e.g., Shakespeare, the Bible), which play a vital sociolinguistic role.

Predictability in communicative behaviour is an ecological imperative, but it has its dangers. Its extreme leads to the iteration of totally and invariably predictable patterns. Stagnation is as great a threat as chaos: when no new meanings are possible, communication is as impracticable as when all meanings are new.

This brings us to the counterbalancing ecological factor: the capacity for creativity.

3. Language Use and the Creativity Imperative

Language users are, by and large, unaware of the overwhelmingly repetitive nature of language usage, and find the established patterns adequate to almost all of the meanings that they require. This is because there is always the possibility of variation, the potential for creativity. Every instantiation of a pattern reinforces the predictability of the code, but it also represents a meaningful choice. Although creativity is less evident in language than most speakers realize, it is more than simply a potential. Life is changing continuously, and new, often subtly different, meanings are required. With occasional exceptions, most adapt their language use to the new circumstances without even being aware of what they are doing. And of course from time to time significant and striking new language forms are created - especially in artistic works, and most obviously in poetry. Creativity is, in Giddens' (1984) term, "agency": the capacity of the individual to contribute to the resources of the community, whether ephemerally or in a more lasting way.

Total creativity, however, is as inimical to the ecology of language as total predictability. Transformational grammar over-emphasized creativity, making it the central characteristic of an individual's language competence:

The language generated by the grammar is infinite. Putting aside irrelevant limitations of time, patience, and memory, people can in principle understand and use sentences of arbitrary length and complexity. … Language serves as an instrument for free expression of thought, unbounded in scope, uncontrolled by stimulus conditions, though appropriate to situations, available for use in whatever contingencies our thought processes can comprehend. This "creative aspect of language use" is a characteristic species property of humans. (Chomsky, 1980: 220-2)

If, however, every linguistic string had to be generated afresh and nothing could be taken for granted, interaction would be impossibly difficult. The counterbalance is provided by community's language competence. A creative act that is found to be valuable - the coining of an expressive phrase by an advertiser, for example - is adopted by a wider group. It then ceases to be a creative pattern, and becomes as predictable as any other phrase in the repertoire. The particular form of any specific linguistic element is the outcome of an ever-present interplay between predictability and creativity. Goodman (1973: 47-48) writes:

I would suggest that the subject matter of linguistics is something actual and universally common, namely: the relation of the tension that exists between a speaker's power to speak and the code of the speech-community; and the relation, often the tension, between a speaker's need to say his say and the need to be clear to the hearer. These relations and tensions produce language.

The predictability of a pattern allows it to bring its history of assumptions along with it into every exchange; creativity makes that history modifiable.

Amongst those who lead the community in modifying the patterns of meaning are the poets, and central elements of those patterns are words.


4. Word Meaning and Associative Networks

Focussing on individual words in context does not undermine the claim that language comprises a repertoire of patterns. Separating words out from the larger patterns in which they occur helps to illustrate the relationships that take place between any chunk of language and its meaning. The next step, in a more extended analysis than is possible here, would be to set each word into the patterns of which it is part, so as to provide a fuller understanding of the concepts of word and of meaning within the ecology of English. The ensuing discussion takes concept of word for granted, but problematizes the concept of meaning. Meaning is not a property of language but a concomitant of its use; it is brought into being through human interactions with one another and with the environment (in its broadest sense, incorporating all aspects of the community and its culture).

It is widely taken for granted that words are, at least in principle, definable in terms of denotation and connotation:

"Denotation" tends to be described as the definitional, "literal", "obvious" or "commonsense" meaning of a sign. … [It] is what the dictionary attempts to provide. … The term "connotation" is used to refer to the socio-cultural and "personal" associations (ideological, emotional, personal, etc.) of the sign. These are typically related to the interpreter's class, age, gender, ethnicity and so on. Signs are more "polysemic" - more open to interpretation - in their connotations than in their denotations. (Chandler, 2002: 140)

This satisfying bifurcation of word meanings is, however, far from robust. Definitions, even those found in high-quality dictionaries, are inherently inconclusive, and will not bear close interrogation. It is possible to argue endlessly about such questions as whether, for example, to describe something as a table necessarily implies that it has one or more legs. Metaphorical use of a word makes its denotation more problematic: for example, is it a metaphor to refer to an outward-jutting ledge of rock in a cliff as a table? Whatever the answer, where does the denotation of the word end and metaphorical extension begin?

Nor is the distinction between denotation and connotation clear, since there are no independent grounds for allocating part of the meaning to one or the other. If aroma is always used only for pleasant smells, is pleasantness part of the denotation or the connotation (an emotional evaluation)? If the former, does that imply that aroma has no connotations? Do all words (or at least all content words) have connotations, or are some purely denotative in meaning? If the latter, how do we decide between those which have connotational meaning and those which do not? Contextual meaning would appear to be connotation, yet it crucially influences the definition of words with significantly different meanings in different contexts: for example, phone as a unit of sound and as a communication device.
There are cases in which contextual determination is not a case of "either-or" but a sort of sliding scale. Neurotic, for example, has a clear technical definition in psychology and psychiatry, and a related but far broader one in everyday conversation - including conversations involving psychologists and psychiatrists. A psychiatrist, who in a casual conversation accused a friend of being neurotic because she locked the house whenever she went into the garden, apparently did not intend the technical meaning. The context determined the interpretation in that particular instance, but a similar level of contextual determination of meaning occurs almost continuously. The question that arises is: to what extent does context determine the denotation of a word?

This is not mere sophistry: word meanings are inherently highly fluid. Whilst it is useful for some purposes to treat them as if they were clearly specifiable, it is inadequate for a serious analysis of how meanings are created and interpreted in context. For this purpose, a flexible approach is necessary, one which is predicated on the ecology of language in use.

5. Word Meaning and Associations

When a word is used, it is meaningful not because its definition is known to the user, but because it has some sort of articulation with his or her communicative experience. It has, in other words, previously been used, in any of the four language modes, in a context in which it successfully communicated. A word without this history is a word that does not mean for the user, for meaning in use is personal. That is not the same thing as idiosyncratic: if a word's meaning is not substantially shared, it does not communicate. What is shared, however, is not an abstract, conceptual definition, but similar historical communicative experiences.

Thus connotation is at the core of the process of meaning: "the socio-cultural and personal associations" referred to by Chandler. Indeed, from now on, I prefer the term "association" to "connotation" in order to avoid the implied contrast, arising from their history of use, with its alter ego "denotation". (Saussure (n.d. [1916]: 125-7) included associative meaning as a part of the signification of language, but it has received relatively little attention from linguists; see Barthes (1967: 90)). Communication is constantly occurring, and members of a speech community are continually developing and reinforcing the shared associations that allow them to make meaning to one another.

The more people who communicatively use a particular word, the more widely is its web of associations known. All competent speakers of English more or less agree on the meanings of a large portion of the word-stock of the language in common: in other words, they use these words with common associations. In order to preserve the agreement - the predictability of the code - these common associations are summed up in a series of abstract formulations that we call a dictionary. (Barthes (1990: 9) refers to "denotation, the old deity, … foreordained to represent the collective innocence of language"). The actual meanings, however, are in the use. Definitions do not communicate, except in a very derivative sense. Their role is socio-ecological: to safeguard commonality through predictability. That is why there are many examples of "words" in dictionaries that no longer mean:

What keeps a word alive is the fact that large numbers of people agree to using it in a certain connotation. A word dies … when there is no longer a sufficient number of people for whom it carries a connotation. … A word is archaic, therefore, not because it connotes something out-of-date, something no longer seen or believed in, … but because it has a connotation which is generally not understood … (Blamires, 1981: 73)

These connotations (associative meanings) arise from a number of factors (Barthes, 1967: 75-86), but can be grouped for convenience into two broad and overlapping categories. The first contains collocations, words that occur together with predictable frequency in larger patterns of language. Throughout a discussion of the weather, for example, words like hot, windy, sun, humid, and clouds are likely to occur, and so are words such as forecast, expressions of time like season, yesterday, and weekend, and expressions of evaluation such as too (hot), (don't) like, typical, and lovely. Most of these words also occur in other forms: hotter, humidity, and typically. There are, in addition, some words which are more or less interchangeable, such as hot, baking, tropical. Words hunt in packs, and an essential part of being able to engage meaningfully in this sort of discussion - in other words, of knowing the language - is a knowledge of which words belong in the pack. Associations of this type are highly sensitive to cultural developments; phrases gain currency for many different reasons, and form associations that become a part of the community's language knowledge either for a limited period or permanently. Until recently, for example, English speakers' network of associations with web would have included spider but not surf. The history of a community's culture leaves its traces in the associative meanings of the community's words.

The associative nets of different words interact in a significant way. Hot is associated with lukewarm as well as weather, but the associations of weather do not (typically) include lukewarm. As the topic of a communication is established, some of the associative nets are brought into prominence and others recede.

The other broad category of associations comprises those that are highly individual. A colleague, for example, told me that for him there is a strong negative association between teacher and red-haired, a legacy of early traumatic experiences with a red-haired teacher. Everybody has such associations that are not shared by large groups of users of the language; they inevitably play some part in the construction of meaning. Each word sets off a chain of associations in the mind of both speaker and hearer, and the meaning which each takes from the interaction is determined by those associations that existed prior to the exchange and by those that are created by the exchange itself. In the interests of establishing common understanding, they are usually bracketed out, but more often than we are aware they may contribute to unexpected misunderstandings. These idiosyncratic associations, nonetheless, are a potential source of creativity, and through communicative interactions - especially, in fact, as the next section demonstrates - they may become shared; the individual may contribute to the meanings of the community.

It appears that words exist in the minds (and perhaps the brains) of language users as the nodal points in a sort of net or web. A simple demonstration is a word association test: starting with any particular word, a competent speaker can readily mention a dozen more that are in some way associated with it. Such associations are highly context-dependent: when discussing domestic furniture, the associative net around table will include one set of words; in a discussion of a technical report, the associative meanings of table will be quite different.

During communication, associations function to make meaning. They set up a cloud of expectations in the mind of a collocutor which enable him or her to make predictions of the language that is to follow, and hence to construct and adjust the meaning as the interaction progresses. Similarly in the written language: at the end of a page, the reader predicts the next word or two while turning over. Associative meanings have to be predictable: enough of the expectations triggered by the net of associations need to be confirmed for the participants to feel that they are making sense of and to one another. But they must also be creative: unpredicted words help to establish new associations and hence new meanings.

The language of poetry provides a highly crafted exemplar of these ecological processes at work. The following section traces the way in which a poet plays upon associations, balancing the predictability of the shared with the creativity of the new.

6. Analysis of Associations in a Poem

The particular poem, by Robert Graves, was chosen because its theme conveniently reinforces the argument. Owing to limitations of space, the analysis has been abbreviated to a few indicative examples; conducting this sort of analysis in class typically takes up to two hours. Note that the poem is introduced consecutively, word by word (occasionally in longer chunks), and it is recommended not to read the entire poem (reproduced at the end) until the analysis has been worked through. Citations are given in bold; the punctuation is included as a guide, but where it seems helpful to indicate incompleteness, dots have been added.

The title begins with The Cool …, which sets up two sets of associations. Negative ones include, e.g., cold, chilly, unfriendly; positive ones include, e.g., refreshing, as well as recently acquired ones that would almost certainly not have been shared by the author, e.g., nice, good, fine, funky, laid back. One can predict, therefore, a very wide range of words that could follow in the title; the one that does, Web, comes as a surprise, associated as it is on the one hand with complexity and interrelatedness, and on the other with entanglement (spider's web; web of deceit). The associative nets of these two words do not normally intersect; here, the emphasis is on the creative rather than the predictable; interpretation has to be suspended. The Cool Web leaves us puzzled (except perhaps as a term of approval for the Internet).

The first stanza opens with Children …, the associations of which are rich: e.g., naive, inexperienced, vulnerable, immature, lively, uncontrollable. We predict that the poet will say what they do, or what they are like, and we expect at some stage perhaps a contrast, stated or implied, positive or negative, with adults, also part of the associative net. … are dumb resonates with children, but activates two new sets of association: silent, tongue-tied, deaf; stupid, unknowing. Interpretation continues to be suspended; the ambiguity is well balanced, sustained by what follows: … to say. This is then followed by … how hot, which has rich associations, e.g., burning, fierce, extreme, and collocations with weather, passion, fire, water, tip, property. One of these sets is picked up immediately by … the day is. The complete line is:

Children are dumb to say how hot the day is,

It has ended predictably: the pattern is thoroughly familiar, but the dual possibilities of dumb remain unresolved.

The next line is:

How hot the scent is of the summer rose,

bringing together predictable nets of associations: scent and rose, summer and rose, as well as the creative intersection of hot and scent (for some readers, this association may have already been established through cosmetics advertisements, but it is still creative in this context). This line has resolved nothing: dumb is still ambivalent, and there is as yet no way of telling whether, to put it simply, it is a good thing or a bad thing - hence, whether to use the positive or the negative associations of cool in the title in constructing the meaning. We need to read on.

How dreadful … introduces the third line. The associations here are rich and obvious: the immediate ones are strongly negative, but for the thoughtful reader there are also more distant, parallel ones (what we might call overtones) such as reverence and awe. The connexion of the two nets intensifies the developing meaning. The next words are: … the black …. The opening two lines have set up a web of meaning that activates negative associations such as night, despair, death. These are confirmed by the next word: … wastes ..., a creative collocation, followed by a perhaps more unexpected phrase: … of the evening sky, …. The complete line is:

How dreadful the black wastes of evening sky,

The fourth line repeats the pattern of the third:

How dreadful the tall soldiers drumming by.

Here, the more literary associations of dreadful are drawn to the fore by linking the word with tall and soldiers. So the poet brings us to the end of the first stanza with the interpretations to be made of cool and dumb - and hence of the whole poem - still a puzzle.

The second stanza introduces a significant change in the meaning process. But we … unambiguously although implicitly delivers the contrast with children. The predicted contrast with adult is finally confirmed. The interpretative process begins to proceed rapidly here; the next two words … have speech … resolve the continuing uncertainty around dumb. We can assign meaning to the title as the first line continues: …, to chill the angry day,, which harks back to the first line of the poem by picking up the associations of hot and angry. The complete line is:

But we have speech, to chill the angry day,

The meaning at this point is quite transparent, and no unconfirmed predictions remain. The association of vulnerable with children is retrospectively transferred to dumb, for lack of speech leaves them exposed to the fearsome elements.

The next line is:

And speech, to dull the rose's cruel scent,

This underlines the interpretation, which is rounded off in the remaining two lines of the stanza:

We spell away the overhanging night,
We spell away the soldiers and the fright.

There is no need to comment on all of the associations being drawn upon here, except to note the juxtaposing of the different but mutually reinforcing networks associated with spell.

The opening of the third stanza, There's a cool web of language …, confirms the interpretation of the title. The poem is about language and how it means; its role in helping human beings to relate to the world. But there follows an immediate hint that this comforting interpretation is not the end of what the poet has to say; that there is a new twist in the meaning. The associations of web with entanglement, which have been left hanging, so to speak, since the title, are reactivated in the next few words: … winds us in, which suggests spiders, or fishermen, catching their prey. The full line is:

There's a cool web of language winds us in,

The next line begins with a word whose associations pick up those of wind: Retreat …. Or are the consequences not so undesirable, since what follows is from too much joy and too much fear? The complete line is:

Retreat from too much joy or too much fear:

The collocation of retreat with strategic is potentially appropriate here.

The next line begins We grow sea-green …, a colour that, in conjunction with people, is associated with, e.g., bilious and drowned; it also, of course, affirms earlier associations with fishing and killing. Death, finality, drowning are all reinforced by the remainder of the stanza: … at last and coldly die / In brininess and volubility. The decisive judgement can now be made: language is damned by association with talkative, excess, volume. The final two lines of this stanza are:

We grow sea-green at last and coldly die
In brininess and volubility.

So we are trapped in a web of words; unable to engage with the world, we drown in our own verbosity. How fortunate (certainly not stupid) the children are, to be unable to speak!

The final stanza, like the second, introduces a turning-point: But …, and this is followed by a series of confirmations of predicted associative connexions:

But if we let our tongues lose self-possession,
Throwing off language and its watery clasp,
Before our death, instead of when death comes,

Then we are taken back to the opening stanza, but this time instead of retreating, we pick up the implications of its opposite:

Facing the wide glare of the children's day,
Facing the rose, the dark sky and the drums,

And now for the final line. At this point, the process of constructing the meaning through the linking of associations, at times confirming our predictions, at times making new and creative links, has been followed through to a conclusion. There are no loose ends, no associations left hanging, no unresolved tensions between possible interpretations. But now one word (a compound) relativizes the meaning of the whole through associations that cut right across the poet's carefully woven web:

We shall go mad, no doubt, and die that way.

The associations of many of the key words have to be revisited and realigned: cool, web, dumb, winds in, and so on.

7. Conclusion

This paper has attempted to illustrate important aspects of the ecological processes of creating meaning in language. The possibility of creating meanings - of communicating - depends on the tension between two contradictory forces: predictability, a characteristic of the community, and creativity, a characteristic of the individual.

Word meaning was used as an example of this possibility. The common distinction between denotation and connotation was turned on its head, with the latter seen as the determiner of meaning, and the former a derivative concept with a socio-ecological function. The sets of associations that cluster round each word give rise to predictions and hence make possible a continuous process of interpretation. The conjunction of the predictability-creativity tension with associative word meaning was illustrated through an outline analysis of Graves' poem "The Cool Web".

Appendix: the complete text

The Cool Web
Robert Graves

Children are dumb to say how hot the day is,
How hot the scent is of the summer rose,
How dreadful the black wastes of evening sky,
How dreadful the tall soldiers drumming by,

But we have speech, to chill the angry day,
And speech, to dull the rose's cruel scent,
We spell away the overhanging night,
We spell away the soldiers and the fright.

There's a cool web of language winds us in,
Retreat from too much joy or too much fear:
We grow sea-green at last and coldly die
In brininess and volubility.

But if we let our tongues lose self-possession,
Throwing off language and its watery clasp
Before our death, instead of when death comes,
Facing the wide glare of the children's day,
Facing the rose, the dark sky and the drums,
We shall go mad, no doubt, and die that way.

About the Author

Dr Mark Garner is Programme Director of the Master of Arts in Applied Linguistics in the English Language Centre at Northumbria University.

Email: m.garner@northumbria.ac.uk

References

Barthes, Roland (1967) Elements of Semiology London: Jonathan Cape

Barthes, Roland (1990) S/Z Oxford: Blackwell

Blamires, Harry (1981) The Secularist Heresy London: SPCK

Bradford, Richard (1993) A Linguistic History of English Poetry London: Routledge

Carrithers, Michael (1992) Why Humans Have Cultures: Explaining Anthropology and Social Diversity Oxford: Oxford University Press

Chandler, Daniel (2002) Semiotics: the Basics London: Routledge

Chomsky, Noam (1980) Rules and Representations Oxford: Blackwell

Coupland, Nikolas, Sarangi, Srikant, and Candlin, Christopher N. (Eds.) (2001) Sociolinguistics and Social Theory Harlow: Longman

Finch, Geoffrey (2000) Linguistic Terms and Concepts Basingstoke: Macmillan

Garner, Mark (forthcoming, 2003) Language - an Ecological View Oxford: Peter Lang

Giddens, Anthony (1984) The Constitution of Society Cambridge: Polity Press

Goodman, Paul (1973) Speaking and Language London: Wildwood House

Haugen, Einar (1971) "The ecology of language", in Dil, Anwar S. (Ed.) (1972) The Ecology of Language - Essays by Einar Haugen Stanford: Stanford University Press

Hayward, Tim (1995) Ecological Thought: An Introduction London: Polity Press

Hymes, Dell (1964) "Introduction: toward ethnographies of communication", in American Anthropologist, 66 (6)

Linell, Per (2001) "Dynamics of discourse or stability of structure: Sociolinguistics and the legacy form linguistics", in Coupland, et al. (Eds.) (2001)

Narasimhan, R. (1998) Language Behaviour: Acquisition and Evolutionary History New Delhi: SAGE

Saussure, Ferdinand de (n.d. [1916]) Course in General Linguistics London: Peter Owen

Wilson, John (2001) "Who needs social theory anyway?", in Coupland et al. (Eds.) (2001)

Wray, Alison and Perkins, Michael R. (2000) "The functions of formulaic language: an integrated model", in Language and Communication, 20