Gordon Tucker
A corpus-based profile of the noun representation:
Towards a lexicogrammar and socio-semantics of political representation
This research was funded by a grant from the INTUNE project (Integrated and United: A quest for Citizenship in an ever closer Europe) financed by the Sixth Framework Programme of the European Union, Priority 7, Citizens and Governance in a Knowledge Based Society (CIT3-CT-2005-513421)

1. Introduction and aims of study

This small study is based on citations of the lexical item representation elicited from the COLLINS COBUILD Bank of English (Wordbanks on-line version) (Sinclair 1987). This corpus consists of 56 million words of running text from a representative range of text-types.

The aim of the study is to produce a lexicogrammatical and socio-semantic profile of the item, on the basis that the sense of a linguistic expression is a function of its use in context. Moreover, it is through the context in all its manifestations that an expression acquires its sense(s), and by context, we understand (a) the context of culture in which the expression is used, (b) the context of situation (the situational features, e.g. the participants in the speech event and the relationship between them, the field of activity which the language accompanies, the mode of communication, written/spoken etc.), (c) the discoursal co-text, in terms of the particular discourse in which it is used, and (d) the lexicogrammatical co-text, in terms of the expression’s lexicogrammatical role and relationship to other expressions used in the same discourse.

Here we focus primarily on the narrower co-textual environment of representation, initially from a quantitative perspective, since we are examining all found citations of the word in terms of its grammatical and collocational relationship to other linguistic expressions within a span of four words either side of it.

A corpus-based approach therefore allows us to uncover probabilistic patterns of linguistic behaviour, certain aspects of which may not be readily accessible from the investigation of the expression in a small number of texts.

This is not intended to represent a full discourse perspective on the use of representation. It is not a study of complete discourses from particular sources (e.g. from the spectrum of media discourses). We are not therefore, at this point, concerned with representation as a discourse theme and the construction of perceptions and arguments involving representation within given communities (e.g. the EU or nation states). The processing of the corpus data does lead, however, to some ‘self-selection’ of genres, particularly media discourse; the salient linguistic phenomena that emerge appear, upon further investigation, to be prominent in the media sub-corpora of the overall general corpora.

What such a generalised yet lexicogrammatically restricted investigation does afford is an insight into the semantic potential associated with representation, including, in particular, those linguistic tendencies that are arguably ‘below the consciousness level’ of individual speakers. Whilst speakers would generally claim to have freedom of lexical selection, it has constantly been shown, through corpus linguistic investigation, that linguistic choices are often strongly interdependent.

2. The noun representation

The noun representation is the nominalised form of the verb represent, and therefore already metaphorical in the grammatical sense (see Halliday 2004:656). In other words, it is a process (a happening/doing/being etc.) dressed up or treated as a Thing. Clearly, representation still denotes a process, but enables the process to become a participant in another process as shown, in the invented example below:

(1) Small industry is not even represented on the forum. Industry representation is pretty poor in any case.[1]

With the process congruently expressed by a verbal form (represent) the ‘represented’ and the ‘representers’ are expressed by the subject and complement functions of the clause, with Circumstantial elements (how, when, how often etc.) being expressed by adjuncts.

(2) gay community was not adequately represented by the committee

In nominalisation, through the noun representation, these elements are expressed through constituents of the nominal group, as shown in (3) and (4):

(3)  inadequate gay community representation on the committee
(4)  the committee’s inadequate representation of the gay community

The investigation of corpus data enables us not only to ascertain how the participants and circumstances involved in ‘representing’ are expressed in the case of the noun representation, but which forms of expression are typically selected by speakers. Furthermore, the data also reveal the kinds of ‘represented’ and ‘representers’ that commonly occur in the discourse of representation.

3. Identifying the senses of representation

The observation that the context of a linguistic expression is responsible for its meaning is apparent in the case of representation and its verbal form represent.

Three senses emerge:

1. Semiotic: the ‘re-presentation’ of some phenomenon by means of an available semiotic system (language, visual, musical etc.), in order to describe, illustrate, explain, model or make accessible the phenomenon to oneself or others.

The three senses are clearly semantically related. Inner cognitive representations serve only the individual concerned. If they are to be communicated to others they must be ‘re-presented’ in some interpretable, conventional semiotic form, thus the second sense of the term. Finally, the expression of individuals and groups, in terms of ‘world views’, are ‘re-presented’ by an individual or individuals (usually in some official capacity), through the act of speaking for the group concerned or ensuring that interests are taken into consideration.

Within the context of the INTUNE Project, we are concerned with the third sense of representation.

4. Findings

4.1. Senses of representation

749 citations of representation were elicited from the corpus. Of these 485 were unambiguously identified as sense 3 of representation, giving a percentage of 65%. This suggest that sense 3 (above) is a primary sense, although the corpus contains a substantial amount of media text. A sample of 100 citations in the British National Corpus (Aston and Burnard 1998) produced 54 unambiguous sense 3 readings of representation.

Within sense 3, two sub-senses can be identified. These can be described as:

1. The existence of a ‘voice’ on some body (committee, parliament, board etc). Thus in (5), the AWU’s voice, in terms of who speak for its interests is greater than the voice of Labour Unity.

(5) On numbers, the AWU has less representation than Labor Unity.

2. The act of appealing to a body in terms of wishing to make one’s case. This sense is found infrequently and is expressed through the phrase make representation as in (6).

(6)  We as an association have made a representation to Tesco

Making (a) representation may be an individual act – i.e. putting one’s own point or position, or the act of a representative of those who which to put a given position to some body. This sub-sense, when unaccompanied by the verb make is understood from the presence of to + the body to which the representation is made, as in (7).

(7)  came in the wake of concern by media organisations and a representation to the hearing by an ABC current affairs reporter over the Speaker’s ruling.

The bulk of citations concern sub-sense 1, where the ‘voicing’ is potential rather than necessarily instantiated.

4.2 Modification of representation

Classes of denotata, such as the class denoted by the lexical item representation are susceptible of both (epithetic) description and sub-classification (Halliday 2004:320, Tucker 1998:128). Thus an attribute may be specified, e.g. an evaluation such as excellent representation, or a sub-class expressed, e.g. proportional representation. The range of specifiable attributes (e.g. colour, size, weight, age etc.) depends on the class of ‘thing’ denoted. Concrete objects, for example, are specifiable in terms of size, weight etc., which are prototypical properties of such objects. Abstract and Event things will have a different set of specifiable attributes. When prototypical attributes of one class are used with another, they are endowed with a new sense, for example, a dark feeling or a heavy punishment.

Head nouns (and the things they denote) are modified grammatically in two ways. Firstly, through pre-modification (typically adjectival structures, but also through clausal and nominal structures), and secondly through post-modification (relative clauses and prepositional expressions).
Modifying devices establish a transitivity relationship between the modifier and the modified. This is clearest in the case of relative clause post-modification, where the transitivity is explicit in the organisation of the relative clause itself, and in particular the function of the relative pronoun. The examples in (8) and (9) below show how the relative clause can express different transitivity relationships between the modified and the modifier.

(8) Women who beat their husbands (GT)
(9) Women whom their husbands beat (GT)

With pre-modification in particular, the transitivity relationship needs to be established, since its explicitness is not directly available in the structure modifier-modified. We shall see how this works in the discussion below on the premodification of representation.

4.2.1 Premodification of representation

There are essentially four types of modifier-modified relations with representation.

1. modifier = group represented
2. modifier = body on which some group is represented
3. modifier = sub-class of representation
4. modifier = degree of representation

These three types are illustrated below:

Type 1: (group represented)

Types 1 and 2 are distinguished in terms of the transitivity relationship between them, which is itself established on the basis of the semantics of the modifying terms. Thus, for example, we read nouns such as board and committee as the bodies on which groups are represented. This modification can be paraphrased as representation on the board/committee. With female, black, Palestinian etc., we understand the modifier as expressing the group represented. These can be paraphrased as the representation of females, blacks, Palestinians etc.

Representation therefore is constituted of a DECISION-MAKING OR AUTHORITATIVE BODY, a GROUP REPRESENTED and the MANNER/SYSTEM IN WHICH THEY ARE REPRESENTED.

4.2.2 Degree of representation (quantity and quality)

Representation is categorised primarily in terms of quantity. Table 1 shows the modifiers and quantifiers found with representation in its political context. The four suggested categories ultimately express quantity (since representation is typically a question of numbers of representatives), yet exploit a range of semantic and lexicogrammatical resources. Interestingly, representation is construed metaphorically in term of size, dimension and weight, suggesting, arguably, either the visual effect or impact of the representative members on a decision-making or authoritative body.

Table 1: types of modification and quantification of representation

4.2.3 Postmodification

As outlined in 4.2 above, post-modification expresses a transitivity relationship in which typically two participants are involved, either directly through a relative clause, or indirectly through a prepositional phrase.

In the case of representation, by far the most frequent form of postmodification is the prepositional phrase. And here, almost exclusively the relationship between representation and the participant expressed through the following nominal group involves either the DECISION-MAKING OR AUTHORITATIVE BODY or the GROUP REPRESENTED. These are illustrated by the concordance lines below.

The resource for these relationships is limited exclusively to of and for, in the case of the group represented, and on/in for the decision making or authoritative body.

It should be noted that there is some degree of indeterminacy in respect of the grammatical analysis of post-modifying elements of this kind, as illustrated by the first concordance line above. This arises from the potential status of preposition phrases as either direct adjunctival/circumstantial elements in the clause, or as modifiers in the nominal group. Thus, in this example, given again as (10), the prepositional phrase in parliament can arguably be analysed as either a post-modifier of representation or as a circumstance of place within the clause, the process of which is advancing. If we assume that in parliament is a post-modifier, as it would be in the case of the expression as subject of the clause, as in (11), then clearly post-modification allows for both the simultaneous expression of the group represented and the body on which it is represented.

(11) the representation of women in parliament is lamentable (GT)

4.3 representation as participant

As a nominalisation of the process represent, representation is available as a participant in clauses with other processes. That is, it may be the subject or complement of another process (verb).

Its primary function, as observed in the corpus data, is as complement of the main verb. The process types (verbs) of which representation is a complement suggest that representation is construed primarily as an issue or a problem. Indeed, this observation is partially supported by the presence of issue and question as the fourth and seventh most frequent lexical collocates of representation.

Two main semantic currents emerge. Firstly, the acquisition or loss of representation including action to obtain it. Secondly, modification in respect of some representational state of affairs, in respect of perceived under- or over-representation.

These two currents are also combined in terms of the acquisition of greater representation, as shown in the first of the concordance lines below.

It would appear, therefore, that representation associates strongly with problematicity, as a kind of semantic or discourse prosody (see Sinclair 1998, Stubbs 2001, Partington 2004). Representation becomes an issue because it is perceived as not corresponding to the needs or wishes of groups within society. The range of processes indicates a state of flux, of groups and individuals’ perception of the lack of representation and seeking, demanding, attempting to obtain representation or to improve the representation that they currently have. These categories are set out in Table 2, with illustrative corpus examples in the concordance lines below.

Table 2: Processes with representation as complement

4.4. Collocation

Collocation lists, such as the list below generated by the Wordbanks corpus software, often serve or are exploited as primary data, or at least as a point of departure for more detailed investigation of the raw linguistic data. Here, they are offered by way of an epilogue, with little further comment. Perusal of this list of the highest frequency collocates of representation, in its political sense, should now resonate with the preceding, more detailed discussion. Most of the common, all-pervasive grammatical/function words have been removed, although in a more thorough investigation, even their presence needs to be investigated. Prepositions such as in and on have been retained, however, as their clear function in post-modification was indicated in 4.2.3. Many of these collocates can now be seen in the light of their contribution to the lexicogrammatical construction of representation. Others are still in need of explanation, and this can only be achieved with recourse to their function within the clauses of the texts that constitute the corpus.

I will make two single additional observations on the basis of this, essentially unprocessed, collocation list.

Firstly, it provides some evidence particularly of socially significant categories of the represented, or rather unrepresented or poorly represented. Notably amongst these are black, women, female and Palestinian. The presence of this final category is of course partly determined by the time-span over which the corpus texts are spread, even though this would appear to be a valid significant category in the time-space before and after that included in the corpus.

Secondly, many of the lexical items included in the list constitute a lexical field for discourses on representation. There are clearly numerous items that relate to the sphere of politics and to the issue of political representation in particular. As such, these expressions, contribute to the expectations that language users develop in terms of the discourse of representation (see Tucker forthcoming).

Table 3: The most frequent collocates of representation (by t-score)

5. Conclusions

This is an exploratory study, and as such, limited in its scope. There are clearly a number of aspects of the lexicogrammatical behaviour of the noun representation that have not been investigated and analysed. Furthermore, there has been no attempt to look at the lexical item within the wider context of the full texts in which it appears in the corpus.

To a large extent, the few observations that have been made here may be considered generalisations. Yet generalisation, in respect of the linguistic resources available to language users, is a useful instrument in subsequent studies on the nature of representation (as a concept) within the various fields of discourse, such as media and political discourses, in which it discussed. A generalised lexicogrammatical profile of representation may serve as a backdrop to the particularised linguistic choices that are made in such discourses. Once we have gained some understanding of the range of options that are available to language users, we are in a stronger position to examine the choices that are made, notably in terms of competing discourses, and the competing ideologies that may respectively underpin them. Meaning, whether at the lexicogrammatical level or the wider socio-political or socio-economic level, is made through the process of selection from the full potential, a process in which what has not been selected, as well as what has, is equally of significance.

It should be noted, however, that corpus data, even when these data are substantial, provides us with evidence of a narrower set of generalisations on an item such as representation, than can be made about it in terms of its membership of the noun word class and of the subclasses of non-count noun and event noun. If there is in theory a much wider potential for its use by language users, as a member of these word classes and subclasses, this wider potential is not exploited. This suggests that what can be theoretical ‘meant’, given the organisation of the linguistic semiotic, is greater than what is ‘meant’ by the collectivity of speakers. So, in some way, the actual potential is determined by the context of culture, even with its competing ideologies.

To return briefly to the limited scope of this small study, it will be noted that no attention has been given here to the verb represent. If we accept Halliday’s notion of grammatical metaphor, and ideational grammatical metaphor in particular, this study has focussed, with its attention to the noun representation, on the metaphorical realisation of ‘representing’ rather than its congruent realisation as a verb. Clearly, a full account of ‘representing’ must include the verb as well. We have seen above how certain aspects of the use of representation reflect the underlying transitivity relations that are properties of the verb. But these are only reflections. A similar corpus-based study of the verb represent itself is likely to reveal its full transitivity, in terms of types of participant role (e.g. Agents and Affected), together with any circumstantial information that may accompany the process of representing. And from the grammatical notion of roles, such as Agent, it should be possible subsequently to throw light on the notion of ‘agents’ and ‘social actors’ as construed in the discourse of representation. But that is for another day.

References

Aston G. and Burnard L. 1998. The BNC handbook: exploring the British National Corpus with SARA. Edinburgh: Edinburgh University Press.

Halliday M.A.K. 2004. An introduction to functional grammar. (3rd edition: revised by C.M.I.M Matthiessen). London: Arnold.

Partington A. 2004. “‘Utterly content in each other’s company’. Semantic prosody and semantic preference”. International Journal of Corpus Linguistics 9.1, 131-156.

Sinclair J. 1987. Looking up: An account of the COBUILD Project in lexical computing. London and Glasgow: Collins.

Sinclair J. 1998. "The lexical item". In E. Weigand (ed.) Contrastive Lexical Semantics. Amsterdam and Philadelphia: John Benjamins. 1-24.

Stubbs M. 2001.Words and Phrases. Oxford: Blackwell.

Tucker G.H. 1998. The lexicogrammar of adjectives: A systemic functional approach to lexis. London: Cassell Academic.

Tucker G.H. forthcoming. “Exposure, Expectations and Probabilities: Implications for Language Learning”. In A. McCabe, M. O’Donnell and R. Whittaker (eds.) Advances in Language & Education. London: Continuum.