Actional and Relational verbs in newspaper editorials.

1. Introduction

In this paper, we will analyse 60 editorials from the British newspapers The Sun and The Times, using the lateral verb model outlined in Kress and Hodge's (1979) "Language as Ideology", a seminal work that advocates closer ties between linguistics, sociology and psychology.

A study of newspaper discourse should require no justification because it is probably both the most read of all text types and that of which the greatest volume is printed. Moreover, the two papers in question, both owned by Rupert Murdoch, represent the two ends of the press spectrum: The Sun is Britain's top-selling tabloid daily, known for its sensationalist scandalmongering; while The Times has long been considered a venerable organ of the establishment.

Our starting hypothesis is that the differing styles of these two papers owe something of their distinctiveness to the writers' choices of verb types as carriers of new information. Thus, we will determine the relative proportions of the various Relational and Actional verbs, in the above model, used in each paper. Where statistics for The Sun and The Times differ, we will attempt to discern rhetorical motivation for this.

2. Theoretical background

Kress and Hodge (1979) explain that perception is not purely psychological, but dependent on language, which reconstitutes the world and provides the systematic organising assumptions through which we see "reality". As the practical consciousness of society, language is:

inevitably a partial and false consciousness. We can call it ideology, defining ideology as a systematic body of ideas, organized from a particular point of view (Kress -- Hodge 1979:6).

More practically, for our purposes, language is:

an instrument of control as well as of communication. Linguistic forms allow significance to be conveyed and to be distorted. In this way hearers can be both manipulated and informed, preferably manipulated while they suppose they are being informed (Kress - Hodge 1979:6).

Whereas Theme generally gives prominence to certain actors and sets the agenda for a message, some schemata - which according to Kress and Hodge, classify events in the world - are principally carried by the Rheme. Among these is one they call the Lateral model. Focused on the verb, it portrays "reality" in terms of two sub-models, the Actional and the Relational, and these provide the theoretical framework for the present study. The Actional submodel, firstly, relates one entity - Actor - to another - Affected - by a process (Transactive) or, alternatively, a single entity and a process (Non-Transactive). The Relational submodel, secondly, involves writer comments on entities or actions. Again, there are two types of Relationals: Attributives describe and are either Possessive (eg. "Bill has courage") or Qualitative (eg. "Paula is brilliant"); Equatives equate two concepts, usually entities, (eg. "John is President").

Kress and Hodge claim for their model, which they call essentially semantic, a higher degree of rigour and delicacy than was possible with the traditional distinction between transitive and intransitive. Unlike the latter, for instance, their lateral model would not confuse actions and non-actions (eg. "the parcel weighs ten pounds"). Unfortunately, a semantic analysis is at least partially dependent on reader interpretation. This means that even their model fails to clarify, definitively, the philosophical question of what constitutes an action in a world subject to perpetual change (Halliday 1993:23). So as not to become embroiled in the Kantian dilemma of subjective versus objective reality, therefore, we opted to employ the following simplified model:

A. Kress and Hodge's lateral model, reduced to:

B. Analysis was by form only, though we noted instances in which form and function diverged.

C. Ignoring Kress and Hodge's mental versus physical process dichotomy, we nevertheless aimed at a measure of delicacy by including the following subcategories:

a) Within Active Transactives -

abstract object

indirect object

abstract subject AND object

abstract subject

reflexive subject/object

b) Within Passive Transactives -

implied subject

explicit subject

no subject

D. Analysis was limited to the main verb only of each clause.

3. Findings (1): form-function divergence

Though it is possible to envision written discourse as a form of dialogue between author and projected audience (Hoey 1983:28), in crucial ways the dialogue is dictated by the writer and is non-negotiable. Parts of the syntagm - or certain classificatory acts -- may be obligatory, such as tense and modality, but others are optional. And these the writer controls.

Now, if the language of a newspaper represents a vision of the world offered to the reader, it is not offered directly. The writer has the power to delete, substitute, combine and reorder many elements of the syntagm by what Kress and Hodge refer to (confusingly, in view of the Chomskyan associations) as "transformations", for example active [right arrow] passive. Such transformations tamper with the discourse, contrarily to modals, which clarify:

Transformations serve two functions, economy and distortion, often so inextricably mixed that even the speaker cannot separate them (Kress - Hodge 1979:9).

A brief analysis was, therefore, conducted of form- function divergence, to gauge its relative importance in the two papers, based on the hypothesis that this divergence must indicate one of the transformation areas that is more accessible to the researcher. Though it would be difficult to prove that a given transformation was primarily a rhetorical strategy, as opposed to a stylistic preference, on the part of the writer, the choice of say Relational form to convey the semantics of an Actional is something that can be documented.

Overall, results showed a form-function divergence for 12.1% of laterals in The Sun; 6.3% of those in The Times. By far the greatest divergence occurred when the chosen form was Attributive Relational. In The Sun, 29% of these verbs had a function that was different from what the form suggested, for instance:

We have never been more delighted to wash dirty linen in public.

In this clause, the form (BE) + ADJECTIVAL COMPLEMENT makes it an Attributive Relational, but the function is Actional, in the sense of "we are washing ... with pleasure." The additional presence of a Discourse Participant Theme (We) makes it all the more likely that the divergence is rhetorically significant.

In The Times, meanwhile, this divergence was only found in 11.4% of Attributive Relational verbs, suggesting a less hortatory rhetoric than The Sun's, but was still greater than for other verb forms in The Times.

Form-function divergence was present but less prominent in Non-Transactive

Actionals and Equative Relationals, while in Transactive Actionals it was negligible. In each case the divergence was greater in The Sun than The Times and for Equative Relationals the difference was eight fold. The rates for each category of lateral verb can be seen in table 1, below.

In summary, form and function are more likely to diverge in The Sun than The Times, possibly for reasons relating to ideology, since there must be a good reason for not choosing the most congruent form available (Halliday 1985:321). However, divergence is not likely to occur in Relationals, which represent only a small minority of all laterals, as we shall see below. in Actionals, a much larger group, form and function almost always coincide, especially in Transactives, the largest group of all. Thus, while interesting in the case of Relationals, divergence is only of marginal significance when counted overall.

4. Findings (2): overall levels of relationals and actinals

An analysis of our two subcorpuses against Kress and Hodge's lateral model produced many strikingly similar statistics and a few notable differences. The ratio of Attributive Relationals to Equative Relationals, for example, is very similar for both papers: Relationals are approximately two thirds Attributive, one third Equative. The ratio of Relationals to Actionals, on the other hand, differs: The Times employs, proportionally, 60% more Relationals than The Sun. This tentatively suggests that the former may be more analytical; the latter more narrative-like, in style. Below, we examine these ratios in more detail.

Relationals

Out of a total of 884 main predicates in The Times, 262 (or 29.6%) were Relationals; the remainder Actionals. in The Sun, 35 (18.4%) Out of 190 were Relationals, a significantly lesser proportion than in The Times, as mentioned above. But, within these totals, the ratio of Attributives to Equatives is similar:

Taking Attributives, to begin with, The Sun's 22 constitute 11.6% of all predicates, while The Times' 178 represent 20.1%. This option is, therefore, relatively more important in The Times.

A Theme-Rheme breakdown reveals that about four fifths of Attributives in each paper -- 77.3% in The Sun in; 80.3% in The Times -- co-occur with unmarked Themes (or those fronting a subject participant) only. The balance -- 22.7% in The Sun; 19.7% in The Times -- co-occur with marked (non-subject) Themes and/or Modal Verbs or Ellipsis. Of "marked" (ie. not with subject Theme only) Attributives in The Sun, 60% have the structure Modal Verb + Attributive, while of those in The Times, 62.9% have Modal Verb + Attributive and a significant minority (17.1%) have the structure B Clause Theme (ie. the whole subordinate clause as Theme) + Attributive. If we take the unique occurrences of various possible combinations of elements into consideration, it transpires that B Clause Themes co-occur with 22.9% of "marked" Attributives in The Times, and Interrogative Themes, Verb Group Themes and Ellipsis each with 8.6%.

Thus, the major differences between Sun and Times with regard to Attributives are, firstly, that these latter are employed nearly twice as often in The Times; secondly, that The Sun's "marked" structures are relatively simple -- in no instance do they combine even two elements with Attributives, in contradistinction to The Times. We note with Kress -- Hodge 1979:22:

Reducing the complexity of an argument and limiting the terms which it can contain is a drastic intervention. Showing less means someone else seeing less. And seeing less means thinking less.

This observation holds true both at a micro and a macro level. The absence of even the smallest element means the loss of a nuance that could be rhetorically vital.

As for Equatives, meanwhile, they are the least common of the four predicate types under investigation. The Sun's 13 make up only 6.8%; The Times' 84 only 9.5% of all laterals and, in each paper, they are only half as common as Attributives.

Again, the proportions employed in clauses with unmarked Themes are similar in both papers: 76.9% in The Sun; 78.6% in The Times. Again, the balance -- occurring with marked Themes or elements, considered more significant for ideological evaluation -- is slightly higher in The Sun.

Of "marked" Equatives, The Sun has only unique occurrences, with Modal or Ellipsis or Interrogative Theme. In The Times, 55.6% have the structure Modal Verb + Equative, a majority again. If we count co-occurrence alone and ignore discrete instances of combinations, 61.1% of Times "marked" Equatives co-occur with Modals; 22.2% with B Clause Themes; 16.7% with Verb Group Themes and 11.1% with Ellipsis. As a whole, then, the figures for Equatives are similar to those for Attributives.

Actionals

Actionals are easily the predominant laterals in both papers, accounting for 81.6% of all predicates in The Sun; 70.4% in The Times. They arc especially prominent in The Sun, suggesting that this paper is either, relatively, more interested in events than The Times, or that it tends the more to adopt a narrative/populist discourse style.

Transactives, firstly, form the largest sub-category. In The Sun, they constitute a majority (50.5%) of all predicates and in The Times (49.4%) this is almost the case too. Table 3, below, shows them as a percentage of Actionals.

Relative to Non-Ts (Non-Transactives) then, Ts (Transactives) are more important in The Times than The Sun. A Theme-Rheme analysis, meanwhile, finds that 74% (71) in The Sun; 68% (297) in The Times, co-occur with unmarked Themes. Thus, 26% in the former; 32% in the latter, co- occur with marked elements, so, in this instance, it is The Times that is possibly more evaluative.

Of The Sun's "marked" Ts, 56% consist of Modal Verb + T. If all co-occurrence, in various combinations, is counted, 72% of Ts take Modal Verbs; 16% take Interrogative Themes; 12% Passive Voice and 8% each Ellipsis and B Clause Themes.

In The Times, an almost identical majority of "marked" Ts -- 56.4% -- have the structure Modal Verb + T. In terms of co- occurrence, 85.7% of TimesTimes Ts take Modal Verbs; 20.7% take Passive Voice; 13.6% B Clause Themes and 6.4% Verb Group Themes.

Thus, around 70% of Ts generally co-occur with unmarked Themes. In the remainder, Modals, are by far the most significant element, followed by Passives. Then trends diverge, The Sun having, for example more Interrogative Themes, The Times more B Clause Themes, with "Marked" Ts. Verb Group Themes sometimes co-occur with Ts in The Times but not in The Sun. These Themes denote a marked word order not found in The Sun (except in Exclamatives) and here, as in other respects, our observation again holds true about The Times having the more complex structures, with the ensuing implications of a simplified discourse in The Sun.

For insights into who is doing what to whom, we analysed the entities of Actor and Affected, which are explicit and show causality in Transactives, unlike Non-Ts which obscure the Affected. As Kress and Hodge do not describe the different types of causality at work in language, we are thrown back onto the old-fashioned notions of subject and object to shed more light upon the processes presented in these newspapers as "reality", as in Table 4, below.

It should not be forgotten that a majority of all predicates in both papers present happenings in terms of concrete, direct, explicit subjects and objects. These we consider to be the unmarked cases. The marked cases often involve two or more of the elements above simultaneously (such that marked cases are less common than one might assume from Table 4).

Indirect Objects and Reflexive Subject/Objects have almost identical occurrence ratios in Times and Sun Both reveal something about the process but little about Actor or Affected. What Table 4 does reveal is that The Times has a greater tendency than The Sun to depict abstract entities, overwhelmingly so in the case of Abstract Subject and Abstract Subject + Object. For Abstract Subject, The Times' rate of occurrence is nearly fivefold that of The Sun, suggesting that The Times is more likely to discuss concepts, while The Sun, broadly speaking, sticks to entities.

As for Passive Voice, it provides writers with the option of deleting the agent, or Actor, responsible for an action - another way of reducing the complexity of an argument. Kress and Hodge exemplify with an extract from press coverage of the 1970's miners' strikes: ... essential maintenance in the pits is done overtime (Kress - Hodge 1979:25).

The people who have to work in unpleasant conditions are deleted and perhaps taken for granted by focusing on the process alone. This strategy has clear political possibilities and we, accordingly, compared Passives in Sun and TimesTimes for their relative proportions of implicit and explicit subjects.

In The Sun, the ratio of implicit to explicit subject was 82.4% to 17.6%, while that in The Times was 74.6% to 25.4%. This is interesting because Kress and Hodge claim that middle class language is more likely than working class language to employ agentless Passives. Our own belief is that The Sun simulates working class; The Times middle class discourse, up to a point. So more research is needed here. Perhaps The Sun's hortatory side, or desire to persuade (as opposed to merely informing), is simply stronger than any need to be consistent in imitating working class language. At any event, our findings regarding The Times' use of more subordinations and more complex verbal structures than The Sun are in line with those of Bernstein (1971:109) on elaborated and restricted code. The contention of Kress and Hodge that abstracts would be more prominent in middle class than working class speech, then, is also fully supported in our data.

Turning, finally, to Non-Transactives, they account for 31.1% of all verbs in The Sun; 20.9% in The Times. They also represent a similarly higher proportion of Actionals in The Sun than The Times. Non-Ts either obscure the Affected or portray "actions" for which there is no Affected, often using congnitive or affective verbs (THINK, FEEL ...) with THAT ... or paralinguistic verbs such as SAY.

A survey of the occurrence of Non-Ts vis-a-vis various elements reveals that 62.7% in The Sun and 68.1% in The Times co-occur with unmarked themes. Conversely, 37.3% in The Sun; 31.9% in The Times co-occur with marked elements. Therefore, with Non-Ts - as with both types of Relationals - it is The Sun which appears to be the more evaluative paper.

Of "Marked" Non-Ts in The Sun, 54.6% have the structure Modal Verb + Non-T. The same is true of 61% of Non-Ts in The Times. Focussing on simple co-occurrence, 68.2% of Non- Ts in The Sun and 72.9% in The Times involved Modal Verbs. Each paper's Non-Ts this time, frequently involved B Clause Themes or Verb Group Themes. The major difference between them was that 22.7% of Non-Ts in The Sun took Interrogative Themes, while the figure for The Times was only 1.7%. On the whole, similarities again outweigh differences from one paper to the other.

5. Conclusion

Form-function divergence, which possibly indicates rhetorical tricks, was found to be more prominent in The Sun than The Times, but occurred most often in Relationals. Among Actionals, the main category, it was so infrequent as to be of only marginal significance.

From a survey of the overall levels of Relationals and Actionals, it is clear that similarities, from paper to paper, predominate. Some ratios were surprisingly close. Among Relationals, for instance, two thirds were Attributives; one third Equatives in each paper. 80% of Attributives, in each, occurred with unmarked Themes and, of those that did not, a similar majority in each paper employed Modal Verbs. The picture was similar for Equatives. Transactive Actionals, the largest subcategory, account for half of all predicates in each.

Such differences as there were, and our interpretation of them, were as follows. The Sun's structures were simpler than those of The Times, which could relate to our hypothesised simulation of working class discourse by the former and might, conceivably, denote a hortatory manipulation of the "facts". The relatively greater proportion of Actionals in The Sun could betray some measure of a resort to populist, narrative-like discourse. On the other hand, The Times appcared more ready to discuss abstract concepts than The Sun. Non-Ts, which can obscure the Affected or depict verbal actions (eg. SAY) with no Affected, were more in evidence in The Sun, possibly reflecting its habit of citing the spoken reactions of individuals to events. Last of all, more Non- Ts in Sun than Times co-occurred with marked thematic elements, which may reveal a more overtly evaluative rhetoric in the former than the latter.

Table 1

Form-Funcion Divergence by Lateral Verb Type

                            SUN   TIMES

Attributive Relationals    29.0%  11.4%
Equative Relationals       18.8%   2.3%
Transactive Actionals       2.0%   1.8%
Non-Transactive Actionals  13.2%  11.1%
Table 2

Attributive vs. Equative Relationals

                SUN         TIMES

Attributive  22 (62.9%)  178 (67.9%)
Equative     13 (37.1%)   84 (32.1%)
Table 3

Transactives Vs Non-Transactive Actionals

                    SUN         TIMES

Transactive      96 (61.9%)  437 (70.3%)
Non-Transactive  59 (38.1%)  185 (29.7%)
Table 4

"Marked" Subject and Object Types in Transactives

                      SUN   TIMES

Abstract Object      31.3%   35%
Indirect Object      11.5%  11.4%
Abstract Subject      2.1%  10.1%
Abstract Subj + Obj   3.1%   9.6%
Reflexive Subj/Obj     1%     1%

REFERENCES

Bernstein, Basil

1971 Class, codes and control. London: Routledge and Kegan Paul. Davies, Florence

Bernstein, Basil

1971 Class, codes and control. London: Routledge and Kegan Paul.

Davies, Florence

1993 Theme analysis as a heuristic for analysing genres and text-types. [Unpublishcd MS, University of Liverpool]

Fairciough, Norman

1987 "Discourse representation in media discourse", Sociolinguistics 17: 125-139.

Gosden, Hugh

1993 "Discourse functions of subject in scientific research articles", Applied Linguistics 14: 56-75.

Halliday, Michael A.K.

1985 An introduction to functional grammar. London: Arnold.

1993 "Language in a changing world". Occasional Paper W013, Applied Linguistics Association of Australia.

Hawes, Thomas P.

1990 Language and bias against women in The Sun and The Daily Mirror. [Unpublished MA thesis. University of Birmingham]

Kress, Gunther

1986 "Language in the media", Culture and Society 8: 395-419.

Kress, Gunther - Robert Hodge

1979 Language as ideology. London: Routledge and Kegan Paul.

McQuail, Denis 

1987 Mass communication theory An introduction, SAGE.

Redd, Teresa M.

1991 "The voice of Time: The style of narration in a newsmagazine". Written Communication 8, 2: 240-258.

Thomas, Sarah

1991 A merging of voices: An investigation of the way discourse is reported in medical research articles. Uupublished Ph.D. dissertation, University of Birmingham]

Thomas, Sarah - Thomas P. Hawes

1994 "Thematic options in reports of previous research", Australian Review of Applied Linguistics 17: 45-72.

Thomas, Sarah - Thomas P. Hawes

1994 "Reporting verbs in medical journal articles", ESP Journal 13: 129-148.

Weizman, E.

1983 "Some register characteristics of journalistic language: Are they universals?", Applied Linguistics 5: 54-65.