Skip to navigation – Site map

Modality in debate

On CAN as a mediation marker
Laurent Rouveyrol
p. 187-201


Cette étude poursuit deux objectifs. Le premier est de montrer comment l’analyse de discours peut remettre en question certaines visions traditionnelles concernant la modalité. Le second est d’explorer la façon dont l’auxiliaire modal CAN est utilisé par différents locuteurs dans un corpus de débats télévisés britanniques : le débat-panel Question Time. La fréquence d’emploi de l’auxililaire modal CAN dans le corpus est le double (0,46 %) de ce que l’on trouve dans un corpus non marqué (0,23-0,25 %). Comment donc expliquer cette fréquence surprenante dans le débat télévisé considéré ?
Le genre du débat politique télévisé impose tout d’abord une approche intégrative et pluridimensionnelle de la modalité : si les personalités invitées sur le plateau doivent exprimer leur opinion sur l’actualité de la semaine, tout énoncé renvoie de près ou de loin à une forme de subjectivité quels que soient les marqueurs apparaissant. Il faut donc distinguer le niveau des modalités, concernant les choix de marqueurs tels que les déterminants, de celui des modalisations, correspondant aux marqueurs explicites de subjectivité : les adjectifs axiologiques, les complétives de type « I think that » et les auxiliaires modaux dont CAN fait partie. L’analyse statistique laisse apparaître que les énoncés contenant CAN concernent les ordres de discours politique, ainsi que celui de la vie quotidienne. Les gloses méta-énonciatives ayant pour fonction de demander la parole sont nombreuses. Par ailleurs, les notions d’argumentation, d’opposition et la valeur d’actualité  les caractérisent principalement.
L’analyse qualitative montre que dans le cadre d’une argumentation, CAN permet de faire référence à un savoir implicite, présupposé, venant appuyer la thèse défendue. Ce type d’emploi correspond à l’étymologie de CAN : Cunnan signifiait Connaître en Vieil et Moyen Anglais. On défendra donc l’idée que l’auxiliaire modal CAN incarne l’activité cognitivo-discusive du débat argumenté, tout d’abord parce qu’il correspond au jeu interactionnel à l’œuvre et ensuite parce qu’il opère une médiation entre un savoir implicite possédé par un locuteur donné, et un discours médiatisé necessairement polémique, dans une situation de communication mettant des locuteurs/idéologies face à face. De la médiation du savoir à la médiatisation du discours, tel est le processus marqué par CAN dans le cadre du débat Question-Time.

Top of page

Full text

1. Introduction

1This research aims both at illustrating how the study of modality in specific corpora may challenge traditional theoretical boundaries and at exploring the way the marker CAN, as a modal auxiliary, is used by different speakers. The corpus scrutinized will be transcripts of the British panel-debate Question Time.

2In this BBC weekly programme, speakers belonging to the panel are invited to answer questions asked by a sample audience representative of the British nation. The panel is generally composed of two politicians who are members of opposed parties as well as experts, academics, high civil servants or leading personalities; in general four persons and the journalist sit around the table. The persons invited are then expected to voice their opinions on matters and topics drawn from the current affairs or current events of the week.

3Why is CAN so typical of such a situation of communication? It appears twice as frequent in Question Time as in unmarked corpora from a very basic statistical point of view: after including negative utterances in the counts, we find 310 occurrences in a corpus composed of six programmes, which corresponds to a percentage of 0.46 of the whole. A simple request on the B.N.C shows that utterances containing CAN only form 0.23% of the entire corpus. Another percentage calculated taking into account Biber et al’s figures (1999) on all four registers of discourse taken together (news, fiction, academic discourse and conversation) is equivalent to 0.25%.

4How can we account for such a surprising frequency? What is the link between the marker and the discourse genre the programme belongs to? Such are the questions that we shall address in this article.

2. The panel-debate genre and the notion of modality

  • 1 The notion of “order of discourse” is adapted from Foucault (1969) and has become a key-concept o (...)

5The communication genre is that of the political interview/debate; and following up some of the precepts constituting the Critical Discourse Analysis approach (Fairclough 1995,1998) and the Bakhtinian theory (1984), it is possible to determine that the programme corresponds to the “generic complex” of the “mediatized political order” of discourse1:

A generic complex articulates orders of discourse (genres, discourses) together in particular ways, and effects particular positionings of agents in relation to them. For instance, contemporary political interview might be looked at as a device for articulating together the orders of discourse of the political system, the media and the ordinary lifeworld of the audience. Political interviews typically mix their genres and their discourses. In complex ways, politicians characteristically shift into conversational genre, and draw upon lifeworld discourses, in finding ways to address mass audiences who are listening or watching in mainly domestic environments. (Fairclough, 1998: 149)

6Four sorts of agents are identifiable: politicians and experts, who both belong to the panel; ordinary citizens, who are part of the sample audience; the journalist, who plays the role of an “interaction-conductor” and confronts agents.

7As a result the programme is crossed by three main orders of discourse: the political, media and ordinary life ones. As suggested in the quotation above, the link between agents and orders of discourse is flexible: a politician, for instance, can resort to the ordinary life order of discourse to meet electoral goals, which implies that the corpus under study is necessarily heterogeneous and evolves dynamically.

8The debate participants are present to voice and discuss their own views; they were invited or volunteered to take part in the programme to say what they think. This specific situation of communication necessarily demands a broadening of the notion of modality. The strict relation traditionally established between markers and semantic domains needs to be challenged.

2.1. Modality and markers

9Modality is conventionally defined as the “expression of speakers” and writers’ attitudes towards themselves, towards their interlocutors, towards their subject-matter”. (Fairclough, 1995: 27). Or in Lyons’ terms it is concerned with “the opinion and attitude of the speaker” (1977: 452).

10If on the one hand a consensus is easily reached concerning a definition of the term, it cannot be denied, on the other hand, that the listing of linguistic markers encoding modality is problematic.

2.1.1. Towards an integrative approach

11Linguistic approaches can be grouped under two main perspectives. The first traditional one is formal: the link between modality and markers is a strict one, implying that only certain markers are modal. In the second type of perspective, the link is flexible and opens the door to a notion of degree rather than to a dichotomy. In such a broader integrative perspective, any linguistic form encodes a certain degree of subjectivity. Thus apparently “neutral” utterances may also be viewed as the result of choices, whether conscious or unconscious.

  • 2 The integrative approach described here is inspired from Ducrot (1980): linguistic forms are cons (...)

12In the case of Question Time, participants are invited in order to voice their own opinions or that of their party on recent burning issues. Therefore any utterance pronounced during the show is nothing but “the expression of a speaker’s vision towards the world”, whatever form it may take. The perspective we need to adopt is necessarily integrative2, which avoids the epistemological paradox that an utterance which does not contain a “subjective marker” is an “objective” one.

13Similarly, labelling epistemic modality as “subjective” and root modality as “objective” does not seem to be relevant in the televised debate genre: expressing one’s doubts would be “subjective” whereas voicing one’s knowledge for the sake of argumentation would be “objective”. Language activity is always speaking-subject dependent; if some objectivity is felt, it is rather the result of an objectivization strategy. The objective/subjective dichotomy is not coherent with an interactional phenomenological conception of language in which speakers are replaced in the sociological environment they shape. Objectivity is therefore not an option, and the problem seems to be one of marking, which can be direct or indirect.

14Adopting an integrative approach, we can take any component of language activity into account: pronouns, modal auxiliaries, choice of nouns, speech acts, activity types such as narration, argumentation, description, along with the fact of anchoring one’s discourse in a specific order, intonation etc. Modality is then broadened to speakers’ stance. Adopting this view, we meet Fairclough’s (1995: 27):

The concept of “modality” is used in a very general way [...]. Choices of pronouns, modal auxiliaries, speech acts, and many others, are included within modality.

15French researchers such as Lapaire and Rotgé have also tried to pave the way towards such a perspective (1991: 323):

L’on s’interrogera enfin sur la possibilité d’étendre à d’autres chapitres de la grammaire le concept de modalité. Tout bien considéré, des outils comme les déterminants indiquent différentes manières de présenter et de travailler une idée nominale (assimilable à un dictum).

16Any linguistic element (segmental or not) may then be brought into play, but it does not entail that all elements bear the same value and encode the same degree of modality. At this point, a reconception of the traditional notions of “modus” and “dictum” is needed to bring the picture together.

2.1.2. “Modus” and “dictum” reconsidered

17The distinction between “modus” and “dictum” originates in the writings of the Scholastics and has been widely discussed since then. It is generally believed that the utterance containing a modal marker can be split into two parts: (i) the part containing the modal marking is called the “modus” and (ii) the other remaining part is designed as the “dictum”. In this view the “dictum” is considered neutral, disconnected from the speaker’s subjectivity, which does not fit in with the parameters defining the debate genre previously analysed.

18It is suggested that the notion of modality, broadened to speaker’s stance, encompasses the level of the “dictum” and that of the “modus”. The diagram below offers a representation of the theoretical frame:

Figure 1. Modality, modus and dictum

Figure 1. Modality, modus and dictum

19As for the “dictum”, it is viewed as the result of an indirect marking process and concerns the level of “modalities”. The “modus” is linked to markers directly encoding subjectivity, at the level of “modalisations” constituted by two sub-categories, axiological lexical choices (good, bad) and grammatical choices (modals, “I think that”…). The direct markers of stance retained are the ones listed by Biber et al. (1999, Chapter 12). The adoption of such taxonomy renders cross-register comparisons possibe.

20Modality is thus conceived as an integrative notion operational on the panel-debate genre. At least indirectly, it is always possible to relate the “dictum” to a subjective source, having structured the utterance and selected the different items containing it, whether in a conscious or unconscious way. However strong the appearance of neutrality may be in an utterance, the utterance can never be considered as disconnectable from its locutionary source, from a speaker’s subjectivity.

21In this overall frame, the analysis of the marker CAN will focus on the auxiliary and the relation it establishes with other discourse levels, in particular discourse orders and activity types.

2.1.3. A multi-dimensional model

22In order to analyse the corpus, a muti-dimensional model was designed to be used as a scheme for Mick O’Donnell’s software Systemic Coder. Computer statistical treatment of data prevents the linguist from focusing on non-significant phenomena. In the environment provided by Systemic Coder, the researcher makes the analysis, attributing the relevant features to segments, and the software elaborates counts and statistics. The approach remains qualitative but allows quantification of data.

  • 3 For an extensive presentation of the approach and the “star model”, see Rouveyrol et al, 2005.

23The scheme combines four different levels of reality and essentially bears on Fairclough’s Critical Discourse Analysis as well as on Vion’s interactional enunciative theory3:

1) Pragmatic parameters

24This level deals with external data corresponding to speaker types (agents): gender, political party, function (journalist, politician, academic).

2) Discourse and genre parameters

25Interaction genres (debate, fierce debate, joke), orders of discourse (media, political, ordinary life), and activity types (narration, description, argumentation) are taken into account. This part of the model is inspired from Fairclough (1995), who draws a distinction between genres and discourse types and in so doing allows possibilities of variation and heterogeneity.

3) Enunciative staging

26The third level concerns the notion of enunciative staging designed by Vion (1998). The different voices woven in a speaker’s discourse are distinguished. Does the speaker give the impression of speaking alone or does he seem effaced? Questions such as these are examined. Five modes are set: unicity (the speaker seems to speak alone), duality (two points of view are constructed), parallelism (the speaker quotes or speaks along with other speakers), opposition (the speaker goes against another speaker), and effacement (the speaker seems absent form his/her own speech).

4) Modalisations

27The last level is that of modalisations and lists segmental units encoding direct subjectivity according to Biber et al.

28Different levels of analysis enable the linguist to transcend the limits of a strictly formal approach which also limits the fuzziness of an analysis which would not take segmental markers into account. These stages are interconnected and all levels are hierarchically equivalent.

29During the tagging process, utterances are determined by the features applied depending on their relevance; but it is also true that each feature is characterized by the combination of features also applied with it. Discourse is seen as an interactional reality, unstable, made of breaks, interaction, overlaps at all discursive levels. The holisitic approach presented here is an attempt at formalizing and modelling this dynamic aspect.

3. Descriptive statistics

30The corpus is constituted of six programmes broadcast in 1992 for the first one and 1997 for the last five. Interaction 6 provides a relevant sample since 45 utterances contain the modal CAN, which represents 0.43% of the whole interaction (the overal average was 0.46 for the entire corpus). All 45 segments received the relevant features of the scheme and they can be characterized as follows.

3.1 The pragmatic and generic level

31At the pragmatic level, it appears that they are mainly expressed by politicians (60%) and are produced in the heart of the debate (not during opening or closing sessions). They were embedded in an argumentation process for 66.7% of them; those that were not had a meta-interactional function serving to take a turn or asking for the permission to do so, as in “Can I just answer ?” pronounced by a politician. Consequently, the orders of discourse at work are the political order:

[1] Any Prime Minister can do that. (politician)

[2] What I find difficult is that an entire party can tell us all that every single thing they believed in … was rubbish, (politician)

and the media order:

[3] It can’t be undone because it’s been done? (journalist)

[4] Nor can I bring in the last question (journalist)

32The identification of a discourse order is not an easy task. Several parameters are in combination: types of speaker and lexical choice are the essential ones. In the previous examples, it is clear that the expressions Prime Minister and party, point towards the political order of discourse. Likewise the meta-term question is typical of the journalist, whose main function is to put questions to the panel.

3.2. The Enunciative level

33It has to be noted first, that utterances can be multi-coded; an utterance may belong to several modes at the same time; only features belonging to the same mode/system are exclusive. When carrying out larger scale analyses, it was noticed that utterances belonging to the political order of discourse belonged to the mode of parallelism, as well as to that of opposition. This apparently paradoxical marriage is, however, easy to account for: politicians speak mainly in the same line as the other politicians or citizens sharing their ideology and, in doing so, they also go against those who favour diverging opinions.

3.2.1. Unicity

34A relatively high proportion of utterances (51%) correspond to the mode of unicity (the speaker speaks alone), where explicit or implicit markers are used.

Explicit unicity
[5] Er I think it can be exaggerated (politician)
[6] And I think the least you can expect of a competent Minister (politician)

Implicit unicity
[7] And it’s only the bigger ones which can produce it today.

3.2.2 Opposition

35Almost two thirds of examples (64%) concern the opposition mode. Polyphonic utterances, which seem to contradict unidentified virtual speakers, form the most important sub-category (47%); whereas diaphonic utterances, contradicting co-participants, represent 18% of oppositional segments.

Polyphonic opposition
[8] It can’t be undone because it’s been done? (Journalist)
[9] And so we shall be quite different, I can asure you of that.
[10] Isn’t it true that Liz Lynn can commit the Liberal Democrats to whatever she wants because they’re not going to be in Government? (Audience)

Diaphonic opposition
[11] No but what about the voter who can’t say what they want as MPs? (Journalist)
[12] Technically it can’t be undone. Yes it can. (Politicians)

36Other enunciative modes such as effacement, duality or parallelism are unmarked when utterances containing the modal CAN are considered.

3.3 Linguistic stance

37When CAN is used, axiological terms are generally not found in the utterance. Only 40% of the sub-corpus contains subjective lexical markers:

[13] I think the great advantage of having the party system as it is now, is that you can see policy developping at conferences…

38As for the type of stance, the feature “style of speaking” characterizes 24% of utterances. In such examples, a comment is made on the utterance itself or on discourse in general, most of the time by the journalist:

[14] And the question is how can voters make a decision when the two parties appear to share the same manifesto?

[15] Now if I can put it...

  • 4 Epistemic stance concerns semantic notions such as the axis certainty/doubt, limitation, precisio (...)

39What Biber et al. refer to as epistemic stance4 concerns 69% of all examples. The semantic value labelled actuality represents 67% of the utterances containing CAN. In this type of stance, markers comment on the “reality or actuality of a proposition as real-life fact” (Biber et al., 1999: 854).

[16] It’s got to be discussed and debated in the House and then we can express any views or reservations which we’ve got.

40The utterances not present in the category either contain markers such as if, or are semantically (contextually) governed by modals such as should or would, or are interrogative sentences.

41It now seems possible to state that the most relevant features to characterize utterances modalized by CAN are: political discourse, opposition, debate, argumentation and actuality.

4. CAN brought into Question

42The statistic description above prompts us to take a close look at the predominant notions of argumentation, opposition and actuality to further the qualitative analysis of the modal auxiliary CAN in the panel-debate Question Time.

43Interactional opposition is directly linked to the situation of communication and notably to the presence of politicians of diverging opinions in the studio. The journalist also, using diaphonic opposition, can put questions to the panel in a more aggressive way and therefore stimulates exciting answers. Opposition is clearly variational, 30% of all the utterances in Interaction 6 were pronounced by a single politician: Edward Heath; other panel members oscillate between 10 and 18%. It is therefore suggested that oppositional segments should be treated in relation with argumentation or actuality.

4.1. Actuality

44Actuality is identifiable by the presence of markers which “comment on the “reality” of a proposition”. Biber et al (1999) provide a certain number of examples:

[17] Actually I’m not really fussy at all.

[18] If I like the music enough, I really couldn’t care less who the people are.

Such examples are present in the corpus:

[19] And the really successive rulers have killed millions

[20] And that’s the actual specific person.

45Actuality is mainly marked by adverbials; we intend to suggest that CAN is another marker, especially in an argumentative process. The notion of actuality is not far from what Palmer (1990: 105-109) calls Existential modality or rational modality. Palmer (1990: 7) notes that:

The existential mode is […] more concerned in ordinary language with some, any, all, than the expressions of modality. But it is of interest for two reasons. First, CAN is used in an existential sense to mean “some” (though more commonly “sometimes’) eg “Lions can be dangerous’.

  • 5 Palmer (1990: 105) explains that in such examples “there is no reference to what is strictly impo (...)

46In the same way, rational modality5 could be a sub-type corresponding to the more general term actuality. The super-ordinate concept of actuality would then include rational and existential modality, which seems consistent with the debate genre: the expression of existence is rationalized, part of an argument. It is interesting to note that in the rational examples given by Palmer (21-22), the mediatized political order of discourse prevails.

Rational modality
[21] These are terms we cannot accept. No British government should, no labour governement would. These terms are unacceptable. (Palmer, 1990: 105)
[22] The societies can expect to face difficult times. (Palmer, 1990: 106)
[23] It will be a voice of love saying this is not how our society can go on. (Q.T)

Existential modality
[24] Lions can be dangerous (Palmer, 1990: 107)
[25] Roses can be mauve (Palmer, 1990: 107)

47Examples pertaining to this semantic category are to be found in the Question Time corpus:

[26] It is only they who can actually make sure that these thugs are rejected.

48It appears that the previous utterances, whether positive or negative, are part of an argumentation process and their role in relation to that cognitive task has to be clarified.

4.2. Argumentation

49The notion shall be conceived as the linguistic action consisting in orienting an utterance towards a conclusion, following Plantin’s work (1990). Let’s take two examples:

[27] I can’t remember the referendum I’m too young/ but as a student at the moment studying agriculture it seems to me that a lot of small farming businesses are going out of business and they can’t stay as they are and it’s because of Europe. (Interaction 6)

[28] Audience: I work as an educational psychologist in a London Borough, and I work every day with a lot of extremely hard-working highly competent teachers who are doing their best to work with children from the whole ability range, from a very great mix of backgrounds (…)
Journalist: No but tell us from your experience, why children attack their teachers?
Audience: I can’t tell you why individual cases occur without knowing about the cases, what I can tell you is that teachers feel extremely overstressed (…)

50In both examples, the speaker introduces himself by his professional activity which induces a form of comptetence based on acquired knowledge or know-how. In the second example, there is a very interesting game between power (linked to knowledge) and non-power linked to ignorance. Here, CAN is a factivity marker in close relation to what is known out of experience resulting from a cognitive process. In the previous examples, CAN functions as a mediation marker between some knowledge and a mediatized utterance. The general formula to comment on such examples could be:

I know (x) and therefore I think and say that (x) or (y). Element (x) is presupposed in some way or other.

51In the co-text surrounding CAN, lexemes such as know, clear, true, because, or the expression I think are very easily identified by a software like Wordsmith Tools. Generic determiners such as Ø article or pluralised nouns in the position of subject are also frequent. Why is it so?

4.2.1 Diachronic incursions

  • 6 For Visser (1969: 384) Cunnan belongs to the category corresponding to verbs of knowledge, unders (...)

52Diachronic linguistics provides some answers and, as often, sheds light on current discourse use. Visser, in his History of the English Syntax notes that in Old and Middle English the verb Cunnan, from which CAN is derived, meant know as a full lexical verb6.

53In the sequence CAN + Direct object, paraphrases would be: know, be acquainted with, have practical knowledge. The expression CAN skill, found in the 14th or 15th centuries meant to have discrimination on, knowledge in a certain matter (1969: 500). According to Visser (1969: 497):

The development from full to auxiliary verb is a slow and intricate process which in a few cases is still going on, and in which the basic meaning, though often considerably obscure, is never completely lost.

54The author gives two examples: Children can be a great nuisance, It can be bitter cold in here and provides the following comment (1969: 1741):

  • 7 Contemporary linguists generally see CAN as mainly related to “action”; few have commented on oth (...)

Here, CAN refers to something which the person or thing functioning as the subject of the sentence occasionally -sometimes even by way of exceptions- is or does7. The notion of capability is only very vaguely present. The actions or states referred to are mostly felt as unpleasant, disagreeable, disappointing censurable, or reprehensible.

55The over-presence of CAN in a debate is therefore not surprising, neither is the fact that the semantic notion of knowledge prevails in the context.

4.2.2. Presupposition, actuality and argumentation

56In our corpus, we can determine four cases when the paraphrase I know that something exists and this is why I think and I say (x) or (y) is relevant, but the nature of the presupposed element is different in each case.

a) Presupposition corresponds to a property of the syntactic subject:

[29] Schools like any society can only operate with authority and discipline
I know something: schools only operate with authority and discipline
          (School = [+discipline / + authority])

b) Presupposition corresponds to the actual knowledge of the subject/predicate validity or non validity:

[30] Nobody can impose a European social model on Britain.
I know that there is nobody to impose a European social model on Britain.

c) Presupposition corresponds to an exterior state, a fact (which is very often a premiss) or a cause justifying the utterance:

[31] I think in a vast country like Russia, you can’t introduce democracy overnight.
I know (X) [Russia is a vast country] and this is why I know and I say:
the introduction of democracy overnight is not a valid proposal.

[32] Lots of children welcome the fact that there is authority and discipline in the school, you can’t have education without it.
I know (x) [education works with authority and discipline], some pupils even agree,
this is why I know and say education without it does not exist.

57This example is an incomplete syllogism; the missing premiss exactly corresponds to the presupposed element. It is almost possible to formalize the previous example using equations:

Premiss: Education = [+authority +discipline]: a = x + y
Parallelism: pupils agree
Conclusion: Education without it = Ø : a – (x + y) = Ø

d) Presupposition corresponds to the evocation of a cognitive (knowledge-acquiring) process:

[33] You can know John Major is going to give in when he starts talking tough.
I’ve already seen / heard that, I know there’s a way of knowing

58In the preceding example, CAN does not point towards the validity of the subject-predicate relation (encoded by the verb know), but strengthens the main verb evoking that a cognitive process (encoded by the temporal clause) has been at work to generate experience.

59Even if all the previous examples are undoubtedly linked to actuality, existential and/or rational modality, we believe that the auxiliary is still dynamic in the sense that it is destined to weigh on the adressee’s values and integrated representations in order to change them. To argue is to perform some sort of action oriented towards perlocutionary goals, whether they be tangible or not. A politician on T.V. is always (more or less desperately) trying to win votes and convince citizens.

5. Conclusion

60In an interaction like Question Time, the study of a modal marker imposes an integrative modelisation of modality and an extension of the concept to the notion of stance or subjectivity conceived dynamically. In a televised debate more particularly, we can argue with Lyons (1994: 14) that:

[…] locutionary subjectivity is really interlocutionary subjectivity and that, in consequence, subjectivity in so far as it is manifest in language if not more generally, is really intersubjectivity.

61The marker CAN incarnates interactional features (such as the fact of asking for permission to take a turn) as well as pragmatic ones which intrinsically define the communication genre in which they appear: a speaker on the programme and more precisely a politician has to argue to convince that their own vision (what they know and think of the world) is the relevant one. The marker CAN in our corpus may be seen as a mediation operator between some upstream knowledge (which is coherent with the original meaning of the marker in Old and Middle English, Visser: 1969) and a necessarily polemical mediatized discourse in a situation confronting different speakers/ideologies. From the mediation of polemical knowledge to the mediatization of discourse is the process that CAN marks.

62In the genre studied, utterances containing CAN mainly comment on the actuality of propositions in an argumentation process. In this corpus, and perhaps more generally, it does not seem possible to dissociate the marker from validated/acquired knowledge i.e a degree of factivity. From this respect, CAN, in oral communication, is very different from MAY: only 8 occurrences are identifiable in our sample interaction. The example below shows that CAN is never equivalent to MAY in the situation of a mediatized debate:

[34] In Particular in Albania, but also earlier in former Yugoslavia, in events that can and may happen in a lot of other parts of Europe.

63All these elements may explain why Question Time is half way between the registers of conversation and academic prose according to Biber’s statistics, when the marker CAN is analysed. This is perhaps the way it should be; we are dealing with experts or authorized speakers taking part in a “topical debate on matters of interest and concern to a representative cross section of the British public”, as Peter Sissons would have it in the opening sequence of the programme.

Top of page


Bakhtine, M., 1984. Esthétique de la création verbale, Paris : Gallimard.

Biber, D. et al., 1999. Longman Grammar of Spoken and written English, London: Longman.

Ducrot, O., 1980. Les mots du discours, Paris : Editions de Minuit.

Fairclough, N., 1995. Media Discourse, London: Arnold.

Fairclough, N., 1998. “Political discourse in the media: an anlytical framework”. In A.Bell and Garrett P. (eds.), Approaches to Media Discourse. Oxford: Blackwell.

Faucoult, M., 1969. L’archéologie du savoir, Paris : Editions de Minuit.

Lyons, J., 1977. Semantics, vol 2, Cambridge: C.U.P

Lyons, J., 1994. “Subjecthood and subjectivity”. In Subjecthood and Subjectivity. The Status of the Subject in Llinguistic Theory, Gap : Ophrys.

Lapaire, J.R., & Rotgé, W., 1991. Linguistique et grammaire de l’anglais, Toulouse : Presses Universitaires du Mirail.

O’Donnel, M., 2002. Systemic Coder. U.R.L :

Palmer, F.R., 1990. Modality and the English modals, London: Longman.

Plantin, C., 1990. Essais sur l’argumentation, Paris : Kimé.

Quirk, R., & Greenbaum, S., 1973. A University grammar of English, London: Longman.

Rouveyrol, L., 2003. « Etude pragmatique de la variation linguistique dans le débat politique médiatisé en anglais ». Thèse de doctorat N.R, F. Dubois-Charlier (Dir), Aix-en-Provence.

Rouveyrol, L. et al, 2005 “A linguistic toolbox for discourse analysis: towards a multidimensional handling of verbal interactions”. In Discourse Studies, N° 7, 3: 289-313. London: Sage.

Scott, M., 1998. Wordsmith Tools. U.R.L:, Oxford: O.U.P.

Vion, R., 1995. « La gestion pluridimensionnelle du dialogue. » In Cahiers de linguistique française, Université de Genève, N° 17.

Vion, R., 1998. « La mise en scène énonciative du discours. » In B. Caron (ed), Proceedings of the 16th international Congress of Linguists, [CD-ROM], Oxford: Elsevier Sciences.

Visser, F. Th., 1969. An Historical Syntax of the English Language. Volume II, First half, syntactical units with one verb. Volume III, First half, syntactical units with two verbs, Leiden: E.J. Brill.

Top of page


1 The notion of “order of discourse” is adapted from Foucault (1969) and has become a key-concept of the Critical Discourse Analysis approach which Fairclough (1995: 55) defines as follows:
The Critical Discourse Analysis approach thinks of the discursive practices of a community –its normal ways of using language- in terms of networks which I shall call “orders of discourse”. The order of discourse of a social institution or social domain is constituted by all the discursive types which are used there. The point of the concept of “order of discourse” is to highlight the relationships between different types in such a set.

2 The integrative approach described here is inspired from Ducrot (1980): linguistic forms are considered in their pragmatic function and meaning cannot be dissociated from its conditions of use.

3 For an extensive presentation of the approach and the “star model”, see Rouveyrol et al, 2005.

4 Epistemic stance concerns semantic notions such as the axis certainty/doubt, limitation, precision, actuality and knowledge perspective.

5 Palmer (1990: 105) explains that in such examples “there is no reference to what is strictly impossible but only to what is unreasonable and unacceptable”.

6 For Visser (1969: 384) Cunnan belongs to the category corresponding to verbs of knowledge, understanding, belief, doubt, trust and mistake. All verbs denote mental processes. Cunnan as an independent verb also corresponds to the category of perception with verbs like See, Hear, Feel (1969: 553).

7 Contemporary linguists generally see CAN as mainly related to “action”; few have commented on other uses. Quirk and Greenbaum (1973: 53) belong to the exception when they mention the notion of “theoretical possibility” next to “permission” or “ability”. The examples given are: anybody can make mistakes or the road can be blocked.

Top of page

List of illustrations

Title Figure 1. Modality, modus and dictum
File image/png, 58k
Top of page


Bibliographical reference

Laurent Rouveyrol, « Modality in debate », Anglophonia/Sigma, 15 (30) | 2011, 187-201.

Electronic reference

Laurent Rouveyrol, « Modality in debate », Anglophonia/Sigma [Online], 15 (30) | 2011, Online since 20 April 2015, connection on 28 June 2017. URL : ; DOI : 10.4000/anglophonia.437

Top of page

About the author

Laurent Rouveyrol

Université de Nice Sophia Antipolis – UMR 6039 – Bases Corpus, Langage.

Top of page


Licence Creative Commons
Anglophonia – French Journal of English Linguistics est mis à disposition selon les termes de la licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International.

Top of page
  • Logo Presses universitaires du Midi