Using speech acts formulas to describe the narrative structure of affaires

Simon Smith

Institute of Sociological Studies, Charles University in Prague


Prospéro is a tool for use by sociologists interested in argumentation: it helps identify and explore the ‘argumentatively strong moments’ (Chateauraynaud 2011: 118) within large corpuses usually based around public controversies. Given, however, that the notion of a discursive configuration or discursive regime is certainly broad enough to encompass narrative as well as argumentative modes of expression or ways of knowing (Chateauraynaud 2003: 83), and that one of Prospéro’s strengths is undoubtedly its sensitivity to questions of time and temporality (identifying turning points in argument trajectories, for example), it is surprising that narrativity remains an underdeveloped area of its conceptual armoury, and that few researchers (to my knowledge) have used Prospéro to explore the narrative structure of corpuses (for an exception see Cézanne-Bert & Hériard-Dubreuil 2014).

In the project I lead at Charles University in Prague on Crisis narrativisation and institutions in politics and public policy we set ourselves two technical-methodological goals: one is to adapt Prospéro for use with Czech texts (perhaps this will be the subject of a later post). The other is to test whether Prospéro can in fact aid the search not only for the argumentatively but also for the narratively strong moments of large corpuses. In this post I want to introduce a way to use formulas in combination with a categorisation of verbs (épreuves) by type of speech act, which represents a promising first step in our effort to investigate narrativity.

The main protagonists in the Berbr affaire, Roman Berbr and Libor Kovařík (Pavel Mazáč, Deník Sport)

The connection between speech acts and narrativity

One way of introducing the theory of speech acts (developed principally by John Searle) is to start from distinction between different classes of speech acts according to how they mediate the relationship between the ‘word’ and the ‘world’. Hence a speech act, for Searle, can have one, both or neither of two ‘directions of fit’.

The so-called word-to-world direction of fit is represented by assertives. These transitive speech acts (taking a single, direct object), with an ‘epistemic’ modality (describing what is) include statements, descriptions and assertions and their “point is to represent how things are” (Searle 2011: 69). François Cooren, whose narrative reinterpretation of speech act theory is a major inspiration of my own analytical approach, adds the class of informatives as a ditransitive variation of assertives (recognising that while one can simply assert something, one can only inform somebody about something). However in practice it is usually just “an issue of emphasis” whether a verb is employed as an assertive or an informative (Cooren & Taylor 1997: 238), so here I stick to Searle’s terminology.

The opposite world-to-word direction of fit is represented by commissives and directives. These ditransitive speech acts with an essentially deontic modality (describing what should be) include promises, vows, pledges, refusals and agreements (commissives) and orders, commands, requests and arguably most questions (directives). The point of commissives “is to commit the speaker to some course of action”, and the point of directives “is to try to get other people to do things” (Searle 2011: 69). Cooren (2000) proposes a third variety of ditransitive, world-to-word speech acts, whose deontic orientation is towards what could rather than what should be. He calls them accreditives: acts that give or obtain authorisation, such as allowing, permitting, enabling, consenting, delegating or mandating.

Declaratives are a special class of speech acts which have a dual direction of fit. By using these “we make something the case by actually declaring it to be the case” (Searle 2011: 69). They act by transferring an institutional object from one actant to another, changing the recipient’s status in some way. They establish, enact, legislate, adjudicate, rule, acquit, denounce, condemn, forgive or cancel, for example.

Searle’s final class of speech acts has no direction of fit, but serves instead merely to “express the speaker’s feelings and attitudes about a state of affairs that is in most cases presupposed to exist already” (Searle 2011: 69). Expressives include apologies, thanks, congratulations, condolences, approvals, disapprovals, complaints and criticisms. However, as Cooren (2000) points out, if we are to regard an expressive as a fully-fledged act[1], and not just an attitude – in the sense that it transfers an ‘object of sanction’ between actants – then we should recognise that it does give new qualities or properties to a recipient and therefore has the same dual direction of fit as declaratives. The difference is just in the type of transformation each enacts: declaratives institute, expressives sanction, which (on a notional scale of illocutionary force) could be seen as a weaker (perhaps more resistible or reversible) form of transformation.

Cooren’s reconceptualisation of expressives is actually based on a sort of narrative insertion of speech acts, for his insight was to notice the fertility of combining speech act theory with Greimasian narrative theory[2]. The essential principle of the latter, for present purposes, is the rendition of a story as a series of exchanges of different types of object. These can include physical objects that have value in the semantic universe of the actants, institutional objects and objects of sanction, as well as modal objects – obligations, desires and commitments, or capacities and authorisations. The latter have special significance for narrativisation because they frame the temporal structure of the archetypical story. Without the transfer of modal objects it would be impossible to speak of a description of events as a story. Cooren noticed that “the speech act categories, which were linearly presented by Austin, Searle and Vanderveken, seem to be articulated to each other in a narrative form” (2001: 282); in other words that there is a correspondence between what speech acts do and the actions/transfers that occur at particular phases of Greimas’s narrative schema. The following table shows how Cooren maps speech acts onto a narrative schema.

Source: Cooren 2001: 283

If we go along with Cooren, I prefer to reposition one type of speech act that has a particular importance in the trajectory of affaires[3], and which Searle and Vanderveken categorised as a declarative (Vanderveken 1990-1) as an expressive. This is the denunciation. Once we recognise expressives as having a dual direction of fit by virtue of transferring an object of sanction, there is no difficulty in including denunciations in this category. Their categorisation as declaratives is, on the other hand, problematic since in most circumstances they have much weaker institutional backing than other classes of declarative. Almost any corpus based around an affaire would confirm that actors, regardless of the status functions they hold, are capable of and feel entitled to denounce practices they dislike. But – unless they hold the appropriate positions of authority – most actors tend to be more reticent to perform speech acts of a declarative type like decreeing, sentencing or condemning.

To describe the narrative structure of affaires (at least in the example that follows) I nevertheless found I needed to include declaratives, which are omitted from the model in the figure above. They occur frequently in my pilot corpus, and seem to articulate with the other speech acts narratively, though I remain unclear to what extent this is due to the particular features of the case. Declaratives, for Searle, represent the passage point between discourse theory and a theory of institutionalisation, so adding them to the narrative schema might indeed help us study how the embedding of narrative fragments (the fractal character of stories) articulates with the embedding of constitutive rules in texts that are built around discursive configurations that combine both narrative and argumentative constructions.

Searching for narrative fragments

To extract narrative fragments from my corpus I developed a simple formula for use in Prospéro. It is based on three suppositions derived from the above theoretical principles:

  • that the basic units of narratives are exchanges – transitive or ditransitive ‘givings’, ‘takings’ or other transfers of an object between two sets of actors
  • that most narratives are based around oppositions between a protagonist and an opponent and their respective helpers
  • that different types of exchanges point towards different phases – beginnings, endings, critical turning points – of narratives.

The Berbr affaire: intimidation and abuse of competences in the world of Czech football referees

At this exploratory stage of our work I have been working with a corpus of 35 articles from the Czech sports daily Sport from August 2015 to October 2016 which deal with alleged malpractice in and around the Football Association’s referees’ committee. The affaire began when one referee resigned in protest at these practices and made public his feelings in an interview with Sport. The newspaper then led an investigation (speaking to other referees and representatives of football) and campaigned against the vice-chairman of the FA (Berbr) and his partner (Damková), the head of the referees’ committee, who were at the centre of the allegations[4]. They, in reaction, sued most of their accusers for defamation, and the case went to court. Some verdicts are still pending, but the two decisions to date upheld the accusers’ right to voice their opinions. The corpus extends to 22,000 words: short enough to familiarise myself with thoroughly, but long enough to contain a broad, if not exhaustive vocabulary and to require considerable data reduction to ‘extract’ a concise narrative. This allowed me to apply and recalibrate Prospéro’s concepts and formulas iteratively in response to my sense of how well the formulas were capturing what I understood as key narrative moments from repeated readings of the corpus (i.e. I could see what was missing or misleading in the results and adjust formulas and categories accordingly).

In this narrative, as in many affaires, the main opposition is between critics and an establishment, who hold different evaluations of a set of past events or institutionalised procedures (recurring events). Having created collections of actors that group synonyms for ‘establishment’ and ‘critics’, as well as the names of key individuals that belong to those camps in this corpus, I asked Prospéro to look for phrases that connect representatives of the two camps using a verb (based on a formula used by Chateauraynaud & Debaz). In addition I categorised most verbs that occur in the corpus using Searle and Vanderveken’s typology of speech acts so that my results would divide into different types of exchange or illocutionary transformation that I anticipated would correspond well with narrative phases. I further extended Chateauraynaud & Debaz’s formula by looking for actions linking actors in the same camp (this was important especially in the case of commissives, where the principal ties are between actors and their helpers).

Formulas used to extract narrative fragments (Prospéro screenshot)

The following examples are designed to show the kinds of speech acts that operate the key exchanges in this narrative as well as how the formula works technically. I will take them in the order that Cooren suggests for mapping speech acts onto Greimasian narrative theory, before suggesting one small adaptation that I argue is necessary to make proper narrative sense of affaires (or at least this one).

The first example is of a narrative fragment governed by a directive:

Berbr and Dagmar Damková demand from former referee Libor Kovařík damages of 800,000 crowns and an apology in the media (24/02/2016).

It does not come from the start of the whole affaire, but initiates a new episode or a new narrative thread (initiating what Cooren would call an embedded narrative fragment). Berbr’s demand, in effect, imposes new obligations on his critics, even though they do not do as he wants and apologise. They are obliged, nevertheless, to defend their actions in court.

An example of a commissive-based narrative fragment comes from a summary that is included in several of the trial reports.

Among those who backed Berbr in court was FA chairman Miroslav Pelta (11/10/2016).

As is common in press articles covering an ongoing story (Baroni et al 2006), many later articles in the corpus contain paragraphs describing who has lined up behind whom, expressions of support for one side or the other. In effect they review/recollect the formation of alliances that logically precedes the giving and treatment of evidence.

The equation of assertives with the transfer of enablements (the competence phase of the narrative) is less self-evident, but it follows from the realisation that in many affaires or controversies the heart of the story is about evidence-gathering and knowledge-processing. Most of the verbs that introduce quotations from actors, witnesses or experts in newspaper reporting are doing nothing other than this, and at two levels: describing the accumulation of evidence in the context of a process (in this case a trial), and at the same time transferring a ‘knowing how to do’ to the reader (by providing information deemed necessary by the reporter to make a competent interpretation and form an opinion), simultaneously often giving an interpretive hint or caution through the selection of introductory verb:

Referee Libor Kovařík affirmed that it isn’t Damková but Roman Berbr who controls Czech referees (24/02/2016).

Accreditives, of course, do the same job much more directly. So when FA chairman Pelta is reported saying he will give a hearing to the two most prominent critics, he is literally transferring an enablement or an empowerment – authorising their criticisms to be heard in a forum where they might have an influence on the governance of football.

In Cooren’s model “the performance phase corresponds to the doing that is implied by the different modalities” of the previous phases (2001: 282). Hence under the heading ‘performance’ I have grouped narrative fragments governed by simple action verbs (paying particular attention to verbs that imply conflict or eventual reconciliation). These verbs don’t describe illocutionary acts (they don’t introduce or paraphrase quotations). Sometimes they refer to success and failure of actors in their respective narrative programmes; sometimes the perlocutionary effects of speech acts. In the latter cases they often refer back to the events that are problematised by the critical voices amplified in the development of the affaire, e.g.

[Tomáš] Kovařík: Referees depend on income from football. That’s why they’re afraid of Berbr (07/09/2015).

I will discuss expressives in the following section. First let me deal with declaratives, which do not feature in Cooren’s model. Most of them involve exchanges not between critics and opponents but between one of these camps and a third grouping of actors that I called judges. It may be that the preponderance of declaratives in my corpus is the result of the fact that the story includes a legal dispute and the coverage includes descriptions of a trial. The culmination of this corpus definitely institutes change, as in this example:

After Monday’s provisional acquittal of Antonín Kordula today the court overturned the verdict of the court of first appeal in the case of Tomáš Kovařík (13/10/2016).

But many other narratives will not change the status of actors in the way that a legal trial does.

Actants and speech acts Object transferred Examples
ESTABLISHMENT – DIRECTIVE – CRITICS Modal object: obligation Berbr and Dagmar Damková demand from former referee Libor Kovařík damages of 800,000 crowns and an apology in the media (24/02/2016).
COMMISSIVE – ESTABLISHMENT – ESTABLISHMENT Modal object: desire/guarantee Among those who backed Berbr in court was FA chairman Miroslav Pelta (11/10/2016).
CRITICS – ASSERTIVE –ESTABLISHMENT Epistemic object: information/know-how Referee Libor Kovařík affirmed that it isn’t Damková but Roman Berbr who controls Czech referees (24/02/2016).
ESTABLISHMENT – ACCREDITIVE – CRITICS Modal object: enablement/ empowerment Pelta reportedly wants to give a hearing to Kovařík and Kocian (06/09/2015).
CRITICS – PERFORMATIVE – ESTABLISHMENT Object of desire / object of conflict [Tomáš] Kovařík: Referees depend on income from football. That’s why they’re afraid of Berbr (07/09/2015).
CRITICS – EXPRESSIVE – ESTABLISHMENT Object of sanction: evaluation It isn’t the first time for Teplice club chairman Pavel Sedlbauer – he has been criticizing the controversial functionary for a long time (05/09/2015).
CRITICS – JUDGE – DECLARATIVE Institutional object: status function After Monday’s provisional acquittal of Antonín Kordula today the court overturned the verdict of the court of first appeal in the case of Tomáš Kovařík (13/10/2016).

Summary table of examples showing form of narrative fragments and objects transferred

The dual role of expressives in narrativising affaires

According to Cooren, “In semiotic terms, the idea of “sanction,” consists of an action used to evaluate, estimate or assess a previous action, according to a given axiology. Like commissives, expressives are clearly narratively articulated, that is, they always follow another action that has to be evaluated.” (2000: 120). In the classical narrative schema they reward or criticise the performance of the central ‘test’ in the story, thereby eliminating the initial obligation to act, restoring a balance, and closing the story: “every narrative structure functions according to this exchange principle, a principle that places directives and expressives at both ends of the narrative schema.” (Cooren 2000: 121). There are some expressives in the corpus that have this epilogue-like balance-restoring effect: after all, the very object of the legal dispute concerns whether the critics must apologise, an obligation that is eventually removed from them. Most expressives, however (and notably the denunciations, accusations and criticisms) occur earlier in the chronology of the affaire and evaluate and problematise a performance from the past. They indicate an attempt to restore balance, but are reacting to an imbalance retrospectively identified. I think it’s important, therefore, to recognise that the role of expressives in this type of narrative is not limited to the closing phase. This has a crucial implication for narrativisation: it means that expressives can play a role not only of eliminating an obligation (by evaluating it as having been either well or badly performed), but also in creating new obligations by proposing a new round of ‘rebalancing’. Indeed, assuming they succeed in mobilising political support, they necessitate it: they create what we could call mutual knowledge-testing obligations[5].

The key to appreciating this point is to recognise that the narrativity of mediatised affaires resides in the articulation of public denunciation and investigation (Quincy-Lefebvre 2011). This follows from Boltanski’s insight about the relationship between indignation and investigation in the emergence of translocal public problems. Thus the success of an accusation depends on a change of discursive regime or topos:

Pour soutenir une accusation, il faut quitter le domaine de l’intériorité pour aller vers le monde extérieur et prendre appui, plus précisément sur des objets. C’est dans des objets, c’est-à-dire des êtres objectifs, cette fois au sens où ils sont réputés impartiaux et démunis d’intérêts propres, que peut s’inscrire durablement la trace des crimes qu’il faut dénoncer. Ces preuves doivent être établies en toute généralité et, par conséquent, ne pas dépendre d’un intérêt personnel, ni même d’une perspective ou d’un point de vue. La topique de la dénonciation abandonne donc l’indignation pour se déployer dans une enquête. (Boltanski, 1993: 102)

For accusations to have a successful ‘career’ in public space they thus need to give way to an investigation or a mediation of voices with the authority to criticise institutions. For their success depends on a further mobilisation of support, not in the sense of political mobilisation but of assistance in testing/proving, either through truth tests in ‘be’ mode – establishing what really happened – or through reality tests – measuring the distance between what is and what ought to be (Boltanski & Claverie 2007: 414).[6] In an affaire it is the accusations – and the force with which they are taken up in the public sphere – that stimulate the ‘evidence-gathering’ competence phase of the story. And reciprocally, accusations depend on subsequent evidence-gathering for their illocutionary force: “Ce passage de l’indignation à l’enquête… [est] une condition pour surmonter les disqualifications dont fait parfois l’objet la dénonciation de l’injustice.” (Miliani 1997: 10).


The narrative schema in speech acts

To summarise, let me review the main functions that I suggest each class of speech act plays in the narrative structure of an affaire, and finish with a diagram showing a selection of the narrative fragments extracted from my corpus, not arranged in a strictly chronological order but in a cycle that, I suggest, helps makes narrative sense of the events.

Expressives(1) (as obligation-generating): problematising a received version of historical events and providing arguments for taking sides in a proposed ‘sequel’. Mobilisation around expressives feeds into directives (by obliging opponents to react) and articulates with assertives and accreditives (by stimulating evidence-gathering).

Directives: setting up the drama by assigning missions and obligations; or referring back to the ‘illegitimate’ directives that are the object of accusation

Commissives: formation of alliances based on expressions of support or opposition

Assertives: investigation and proving (‘testing’ what happened and whether it matches what should have happened)

Accreditives: conferring authorisations and enablements on actors or their speech acts

Performance: success and failure of competing narrative programmes; or – as performatives – actions and counter-actions in the dispute and their effects

Expressives(2) (as obligation-eliminating): evaluating the success/failure of narrative programmes

Declaratives: conferring a change of status on their recipient, instituting


  • Baroni, R., Pahud, S. & Revaz, F. (2006) ‘De l’intrigue littéraire à l’intrigue médiatique: le feuilleton Swissmetal’ A contrario 4(2): 125-143.
  • Boltanski, L. (1993) La souffrance à distance. Morale humanitaire, médias et politique. Paris: Métailié.
  • Boltanski, L. & Claverie, E. (2007) ‘Du monde social en tant que scène d’un procès’ in : Boltanski, L., Claverie, E., Offenstadt, N. & Van Damme, S. (eds.) Affaires, scandales et grandes causes. De Socrate à Pinochet. Paris: Stock: 395-452.
  • Chateauraynaud, F. (2003) Prospéro. Une technologie littéraire pour les sciences humaines. Paris: CNRS.
  • Chateauraynaud, F. (2011) Argumenter dans un champ de forces. Essai de balistique sociologique. Paris : PETRA.
  • Cooren, F. (2000) The Organizing Property of Communication. Amsterdam & Philadelphia PA: John Benjamins.
  • Cooren, F. (2001) ‘Acting and organizing. How speech acts structure organizational interactions’ Concepts and Transformation 6(3): 275-293.
  • Cooren, F. & Taylor, J. (1997) ‘Organization as an Effect of Mediation : Redefining the Link Between Organization and Communication’ Communication Theory 7(3): 219-260.
  • Greimas, A. (1984) ‘Entretien réalisé par J. Fontanille’, Langue française 61: 121-128.
  • Lemieux, C. (2007) ‘L’accusation tolérante. Remarques sur les rapports entre commérage, scandale et affaire’ in : Boltanski, L., Clavérie, E., Offenstadt, N. & Van Damme, S. (eds.) Affaires, scandales et grandes causes. De Socrate à Pinochet. Paris: Stock: 367-394.
  • Miliani, M. (1997) ‘Vous avez dit << plaisirs du sport >> ?’ Corps et culture 2. Available at :
  • Quincy-Lefebvre, P. (2011) ‘Les campagnes de presse : un creuset militant pour l’enfance. L’engagement d’Alexis Danan, reporter à Paris-Soir dans les années trente’ Revue de l’histoire de l’enfance << irrégulière >> 13. Available at :
  • Searle, J. (2011) Making the Social World. The Structure of Human Civilization. Oxford: OUP.
  • Vanderveken, D. (1990-91) Meaning and Speech Acts. Cambridge: CUP.

[1] For Cooren “speech act theory tends (ironically) not to be that interested in what is achieved in interaction and more interested in what people have in mind when they speak to each other” (personal communication, 23.1.17).

[2] Greimas, he feels, “helps us put the focus on action and performance, something that Searle fails to do” (personal communication, 23.1.17).

[3] I prefer to use the French word affaire for the type of discursive regime or public controversy I have in mind. English-language media tend to use the term scandal (or else the suffix -gate) for the equivalent thing, but I follow Lemieux (2007) in preferring to reserve the term scandal for the ‘causes that are impossible to defend’. The English term affair, however, lacks the essential connotation of indignation and accusation which I sense attaches to the French word affaire – and to the Czech equivalent kauza.

[4] The newspaper had long been a vocal critic of Berbr’s role in the governance of Czech football.

[5] Greimas incidentally acknowledged situations whereby sanctions can be put at the service of qualification rather than reward, and hence moved upstream in the narrative schema: “[la discours didactique] est un type de discours où une opération relevant des épreuves glorifiantes est « mise au service » de l’épreuve qualifiante… [ce qui] permet des jeux d’inclusion/exclusion actorielles qui dessinent en quelque sorte le chemin vers la compétence.” (Greimas 1984 : 125).

[6] One text in the corpus in which a journalist summarises the testimony of Berbr’s public critics seems to concentrate this articulation of criticism and evidence. It is the only text in the corpus where two particular categories of verb – ‘Accuse’ (a type of expressive) and ‘Investigate-calculate’ (a type of assertive) – are dominant.

2 réflexions au sujet de « Using speech acts formulas to describe the narrative structure of affaires »

  1. François Cooren

    A very smart analysis. I especially like how expressives are portrayed as a way to both close a narrative schema and open another one. Typically, negative sanctions (e.g., denunciations) are examples of expressives that open a schema where was it denounced has to be repaired in one way or another; while positive sanction (e.g., congratulations) are ways to close a schema. But note also that even closing a narrative schema can also be interpreted as the possibility to open a new one (to the extent precisely that the previous one is deemed to be closed).
    This is, by the way, why conversation analysts are wrong when they claim that the structure of conversation lies on a dyadic structure (they call it adjacency pairs). What Simon Smith brilliantly confirms and illustrates is that conversations and narrations have a triadic structures, as what is used to end an episode is also what allows us to open one.

  2. Marion Wesely

    Dear Hypotheses blogger,
    We found your article particularly interesting. To increase its visibility so the community can more easily appreciate it, we made it a headline article on the slider.
    Best regards,
    The team

Laisser un commentaire

Votre adresse de messagerie ne sera pas publiée.