Literary Texts in Logic of Discourse

Speech act theorists have recently acknowledged that they had limited the scope of the central notion of speech acts in language sciences by emphasizing the analysis of illocutionary acts performed at a single time of literal, non-literal, or non-serious utterances at the expense of sequences of verbal actions whose performance lasts for an interval of time.^{Footnote 1} Specialists of many disciplines have initiated investigations on language competence in order to account for our ability to generate, interpret and identify different types of oral and written discourses in various contexts of utterance. Some of them have sought to develop a theory of discourse. This is the case of Daniel Vanderveken, who laid the groundwork for a logic of discourse^{Footnote 2} and proposed a discourse typology^{Footnote 3} developed on the model of illocutionary logic, which he founded with John Searle,^{Footnote 4} with the aim of contributing to linguistics and conversation analysis, cognitive and computer sciences, semiotics and literary studies. This research program was put forward despite the scepticism initially expressed by Searle.^{Footnote 5}

My intervention aims to evaluate the theoretical content of Vanderveken’s logic of discourse in literary pragmatics. More precisely, I intend to examine the possible relationships between speech acts and literary genres that Searle quickly dismissed as part of his study of fictional discourse,^{Footnote 6} thus following Jean-Marie Schaeffer and Vanderveken. The former rightly argued that a priori it is perhaps no harder (or easier) to identify a sonnet and distinguish it from an epic story than it is to identify a promise and distinguish it from a threat;^{Footnote 7} the latter claimed that the logic of discourses cannot be reduced to isolated speech acts.^{Footnote 8} Finally, I will propose a way to integrate literary genres into a general discourse typology that takes into account possible links between factual and fictional discourses in the use and understanding of verbal language.

Speech Acts and Literary Genres

In Qu’est-ce qu’un genre littéraire?, Schaeffer noted at the outset that the genre issue was raised much more acutely and with greater emphasis in literature than in other spheres of human activity where generic distinctions also have to be made. He explained the situation by drawing attention to the following fact: literary works, unlike non-verbal artistic works, take shape within a larger semiotic field, that of discourse, which includes artistic and non-artistic texts. Generic categories, considered as classes of literary genres defined in comprehension, are therefore directly related to the problem of defining literature. Having noticed that the theory of genres had become an issue in the understanding of literature,^{Footnote 9} Schaeffer later found out that the normative, essentialist-evolutionary and analytical-descriptive attitudes towards literary genres—that have dominated from Aristotle’s time to this day—cause most of the difficulties encountered by literary theorists seeking to explain generic variations and changes that have occurred in the generic system over time.^{Footnote 10}

Indeed, since Aristotle and even Plato, literary theorists postulate the existence of three generic categories: epic, drama and lyric. The epic and drama categories group fictional genres that characterize respectively the literary texts produced by the poet telling a story (epic, novel, story) and mimicking an action (drama, tragedy, comedy). Conversely, the factual genres specific to literary texts produced by a poet expressing his mental states as inspired by the facts and events of reality (poetry, sonnet, elegy) are part of the lyric category. However, literary theorists now agree that we can no longer distribute the factual and fictional genres that arise at different periods of our literary history in these mutually exclusive generic categories. Indeed, 1) the lyric category would have to include factual biographies, praises, memoirs and essays; 2) fictional stories, historical novels, science fiction and autobiographic fictions would be comprised in the epic category; and 3) the drama category would include epistolary novels, comedies, tragedies and dramas. No one will argue that Truman Capote’s non-fiction novel In Cold Blood, Wolfgang Hildesheimer’s fictional biography Marbot, Éric-Emmanuel Schmitt’s factual and counterfactual novel The Alternative Hypothesis and Edmund Morris’ fictional memoirs Dutch are works of writers who tell stories (the epic category) and express attitudes directed at entities, facts or events of reality (the lyric category), or that they imitate various discursive acts (illocutionary, reference or predicative acts) performed by characters of the play (the drama category). Those mixed literary texts are instead generated from primitive forms of factual and fictional discourse. One of the fundamental questions that arise in literary theory today is how we can account for these new language games. Can the notion of speech acts shed some light on the issue?

If the notion of speech acts is even more effective for analyzing discourse—seen as sequences of speech acts whose arrangement depends on whether one is dealing with a monological or dialogical discourse^{Footnote 11}—, and if all literary texts involve sequences of what Tuen Van Dijk calls ‘macro-speech acts’—which are set at a higher level compared to the isolated illocutionary acts that punctuate literary texts just as much as they compose most of the everyday ordinary discourse—,^{Footnote 12} then we should admit that the question of literary genres must be analyzed like any other discursive phenomenon. This is Schaeffer’s position when he argues that Maupassant’s A life has a narrative structure not because it belongs to the class of stories, but because Maupassant decided to tell a story.^{Footnote 13} We can see briefly that generic labels assigned to literary texts by members of the artistic institution contribute to the meaning of their expression in the context of utterance. The relationship between a literary text and a literary genre can be represented with the constitutive and iterative rule ‘X counts as Y in context C,’ which Searle uses to explain how we collectively construct institutional facts, and where the symbols X, Y and C represent respectively the status of the linguistic entity, its function and its context of utterance.^{Footnote 14}

Most generic labels are commonly used to express sequences of illocutionary acts at work in fictional texts (story, novel, drama, comedy, etc.). According to Schaeffer, this is the kind of macro-speech act that Maupassant has performed by telling (or pretending to tell) a story (function) through a non-serious use of language (status): A Life (X) counts as a story (Y) in the context of utterance of A Life (C). However, there are also several sequences of illocutionary acts at work in factual texts (narrative, literary essay, (auto)biography, diary, lament, praise, etc.). For example, Jean-Jacques Rousseau’s Confessions (X) count as an autobiography (Y) in the context of utterance of the Confessions (C) because the author decided to tell the story of his life (function) through a serious use of language (status). In both cases, the literary text would offer an example of sequences of illocutionary acts to members of the literary institution that assign different status and functions to literary texts they produce and receive in different contexts of utterance. It is certainly possible to create a new literary genre at the intersection of biography and novel. Can’t someone (pretend to) tell a life story (biography) while using language non-seriously (fiction), as Hildesheimer did? Conversely, can’t someone (pretend to) tell a story (novel) while using language seriously (fact), as Capote did?

Studying generic labels in this way allows us to find out that some of them denote other kinds of speech acts, such as perlocutionary acts. The term ‘comedy’ seems particularly revealing in this regard. Traditionally, we use the term ‘comedy’ to identify texts in which playwrights imitate a series of verbal acts with the intention of making the readers or spectators of the play laugh. In such a case, it is difficult to determine whether the text in question exemplifies the generic properties of a comedy only by taking into account the linguistic meaning of the text because perlocutionary acts (to amuse, intimidate, seduce, convince, etc.), contrary to illocutionary acts, cannot be expressed and communicated conventionally. Nevertheless, the use of the generic label ‘comedy’ contributes to the meaning of an utterance by conventionally appointing a sequence of illocutionary acts to the given genre. Thus, Dom Juan (X) counts as a comedy (Y) in the context of utterance of Dom Juan (C) because Molière chose, to put it simply for now, to imitate or to pretend to make a series of illocutionary acts (function) by using language non-seriously (status).

In addition to appointing sequences of illocutionary and perlocutionary acts, some generic labels also denote utterance acts, in the sense that their specifications relate to the physical properties of signs uttered, which are components of the text, rather than to their meaning in context of utterance. The sonnet, for instance, is a literary genre whose specifications are syntactic (versification) and phonetic or prosodic (sound, rhyme). Other genres refer explicitly to contexts of utterance: Greek, Roman, and Elizabethan tragedies are sub-genres of tragedy whose characteristics are socio-historical rather than linguistic. As Schaeffer does, I think that those discursive phenomena should not be dismissed when we study literary texts from a generic point of view. Therefore, I suggest extending the study of the linguistic status and the illocutionary function assigned by members of a literary institution to texts they produce and receive in different contexts of utterance. I suggest they should be analysed within a theory of discourse that is still largely unknown in literary pragmatics.

The Logic of Factual and Fictional Discourse

Vanderveken has sought to develop a logic of discourse attempting to analyze, through formalization, the structure of those types of oral or written discourses whose goal is purely linguistic.^{Footnote 15} He applied the existence of four goals in language use to discourse analysis, taking into account the limits of the human mind. Indeed, the language user can only conceive of four directions of fit between the statements that he uses to produce a certain type of discourse and the world in which he does.^{Footnote 16} Therefore, each language goal determines one of the four possible directions of fit between the words and the world. Discourses that have words-to-world direction of fit are made to achieve a descriptive goal: they are produced to describe what is happening in the world. Discourses that exhibit the world-to-words direction of fit are made to achieve a deliberative goal: they allow to deliberate about what action must be taken in the world. Discourses that have both directions of fit are made to achieve a declaratory goal: they are meant to transform the world by making statements individually or collectively. Last, discourses with a null direction of fit between words and the world are made to achieve an expressive goal: they express mental states.

According to this logic of discourse, there are only four basic types of discourse, namely: 1) description, specific to discourses whose goal is descriptive; 2) deliberation, specific to discourses whose goal is deliberative; 3) declaration, peculiar to discourses whose goal is declaratory; and 4) expression, specific to discourses whose goal is essentially expressive.^{Footnote 17} Other types of discourse are more complex. They are generated by adding or changing a finite number of new components among the following: a) the mode of achievement, b) the thematic conditions, c) the preparatory background conditions, and d) the sincerity conditions. These components determine the conditions of success and satisfaction for all types of discourse denoted by a generic label, in addition to assigning constitutive rules for their production and reception in different contexts.

Vanderveken’s logic of discourse provides a useful theoretical framework to study the possible relationships between types of discourses and the generic labels that are part of the literary sphere. Actually, it incorporates two basic theoretical elements that can hardly be neglected in literary pragmatics, namely the linguistic goal of the texts produced and received in different contexts and the direction of fit between the statements these texts include and the world in which they are used and understood. Many literary texts are in fact produced and received as written discourses whose goals are intrinsically linguistic: stories, essays and autobiographies are descriptive texts; pamphlets and literary criticism are deliberative texts; pieces that acknowledge the exceptional value of a work or that set poetic rules of writing are declaratory texts; lyric poetry, praise and elegies are expressive texts. Instead of the three traditional generic categories, one gets four generic categories deduced from the possible directions of fit between discourse and the world in which those literary genres could be produced.

Vanderveken has expressed his intention to contribute to literary studies several times, and especially he has voiced his will to “explain figures of speech and literary styles,”^{Footnote 18} but he did not analyze any literary text to validate his fundamental hypothesis about the existing links between speech acts and literary genres. Given that stories take place both in novels and in historical narratives, we can try now to define recursively a few literary genres located at the crossroads of factual and fictional discourses.

A story is a type of discourse that aims to describe what is happening in the world (linguistic goal), and is based on a series of past, present or future events (thematic condition). It is constructed from a set of evidence relating to these events (mode of achievement of the linguistic goal), assuming that they are linked to form a consistent and coherent whole (preparatory condition), and that the writer believes that they are true (sincerity condition). Following a similar pattern, a historical narrative can be defined as a story that has a descriptive goal and is based on a series of past events that occurred during a delimitated period of time (thematic condition), which are described in sources of collected information—documents, archives, testimonials—(mode of achievement of the descriptive goal), assuming they are reliable (preparatory condition). A biography is a historical narrative based on past and present events in the life of an individual (thematic condition), written on the assumption that they are the highlights of his or her personality (preparatory background condition), and that they are true events of this individual’s life (sincerity condition). An autobiography is a biography centered on the writer’s life (thematic condition). Memoirs and confessions are autobiographies that have a specific mode of achievement, preparatory condition or sincerity condition: one writes his memoirs or confessions by focusing on his reminiscences (mode of achievement of the descriptive goal), but only confessions are based on the assumption that some highlights of the writer’s life are wrong (preparatory condition) and express regret (sincerity condition).

Thanks to the logic of discourse, one can describe the formal structure underlying literary genres and explain the genesis of complex literary forms that arise at different periods of literary history. However, language games specific to various fictional types of discourse (naturalistic, fantastic and surrealist stories, science fiction, romance, drama, comedy, etc.) cannot be analyzed in this way. They can’t, at least, according to the theoretical hypothesis formulated by Searle in the context of a major controversy in which he was opposed to Jacques Derrida. Searle’s hypothesis, which is grounded in his study of the logical status of fictional discourse, implies that the existence of fiction depends on genuine language games. Now one should wonder if it is possible to characterize fictional discourses and integrate them into a general typology of discourses on the basis of Searle’s hypothesis that non-serious discourses logically depend on serious discourses.^{Footnote 19}

According to Searle, the fiction writer uses language non-seriously to intentionally pretend to make authentic illocutions―generally assertions―, as a liar does. However, Searle tends to believe that fiction is more ‘sophisticated’ than lies in virtue of extra-linguistic conventions whose function is to suspend the application of the linguistic (syntactic, semantic and textual) rules that govern the genuine use of language. Once one argues that there is only one set of rules governing the production of discourses and one set of extra-linguistic conventions suspending their application in certain contexts of utterance, one understands that it is theoretically and logically possible that all works of literary fiction be grafted on—or parasite of—all serious forms of discourse analyzed in Vanderveken’s logic of discourse. One can easily anticipate the result: every kind of factual (or serious) discourse contained in Vanderveken’s discourse typology can logically lead to a kind of fictional (or non-serious) discourse. Noting that the constitutive rules of discourse become applicable in the context of a factual discourse on fiction, such as a literary criticism of a play, one can even deduce another kind of fictional discourse whose existence is logically possible, but that may not yet have been realized, i.e., a fictional review of a work of fiction. Indeed, one can pretend to criticize a work of fiction insofar as one masters discursive rules to do it seriously, whether this criticism is sincere or insincere. From there, we may finally group literary genres into two mutually exclusive typologies of discourse: the factual ones may be classified into Vanderveken’s categories, while the fictional ones cannot for an illocutionary reason: the existence of a set of extra-linguistic conventions, described as ‘mysterious’ by Gérard Genette,^{Footnote 20} whose function is to suspend the application of the constitutive rules of factual discourses. Although Searle has admitted, after expressing doubts on the matter in his Conversation, that telling a story is a kind of illocutionary macro-speech act, the logical status of fictional discourses remains problematic in Vanderveken’s logic of discourse insofar as it leads to the development of another discourse typology characterized by the fact that all fictional genres have the same function in the use and understanding of ordinary language, i.e., a null one.

How can the fiction writer intentionally produce an apparently sophisticated and a non-defective discourse without following any genuine rule of discourse matching his macro-speech acts and the real world in which he lives? Unlike Searle, I do not think we can characterize all types of fiction (ludic behaviour) based on all types of factual discourses (non-ludic behaviour), that is to say, those involving a genuinely serious use of language. It does not seem possible, or even desirable, to argue, for example, that a science fiction story is a kind of non-serious story (fictional) that depends logically on a kind of serious discourse (factual), such as the account of scientific discoveries found in a work of epistemology, which is a specific kind of factual discourse that can be obtained by adding a thematic condition on a factual story whose goal is generally descriptive. In fact, it seems possible and even desirable to argue that science fiction, as well as naturalistic, fantastic and surrealist stories, are complex discourses generated from the primitive form of fiction. The same logic applies to stories found in fictional novels, historical narratives, (auto)biographies, memoirs and confessions. They could not be derived from factual stories that aim to describe what is happening in the world because they would be defective, as Searle rightly pointed out in his study of the logical status of fictional discourse.

Based on Searle’s analysis of performative utterances,^{Footnote 21} my theoretical hypothesis on the illocutionary force of literary fiction addresses this fundamental issue: the non-serious use of language is performative in accordance with a well-established background convention of any linguistic society. This convention states that a fiction writer has the deontic power to institute or create the new reality represented in the scope of an intensional operator such as ‘in the language game x’ by sequences of linguistic declarations.^{Footnote 22} This theoretical hypothesis has the merit of extending, by way of inversion, the logical relationships between serious and non-serious discourses about fiction observed by Keith Donnellan^{Footnote 23} as well as by Searle in “The Logical Status of Fictional Discourse.” We can try to describe how things happen (descriptive goal) on the topic of a series of past, present or future events occurring in the language game x (thematic condition) using the language seriously (status) because the author of the fictional story x has succeeded in creating those events (declaratory goal) on the same topic (thematic condition) by sequences of linguistic declarations at a higher-level of single illocutionary acts he could pretend to do with the intention of making them appear as being told (extra-linguistic goal). For example, let’s have a look at the psycho-story (Y) as exemplified by Delusions and Dreams in Jensen’s Gradiva (X), written by Sigmund Freud in the early 20^th century (C). This is a factual discourse on fiction that involves a genuinely serious use of language that commits the author to representing what is happening in Jensen’s fiction (descriptive goal); it is composed of some relatively autonomous sub-discourses of a certain type: 1) a presentation of the object and the method of investigation in psychoanalysis, 2) a narration about Jensen’s Gradiva, 3) an analysis of Norbert Hanold’s psyche, and 4) an interpretation of the character’s dreams. The conditions of success and satisfaction of Freud’s psychoanalysis logically depend on the non-serious use of language by the fiction writer, namely Jensen, in the sense that the first one is directed toward institutional facts created by the sequences of prior linguistic declarations (Y) at work in the second one (X), produced and published in 1903 (C). Most importantly, this example allows us to reveal what was missing in Searle’s theory of fiction, namely the basic function assigned to the basic status of fictional discourse in the context of its basic utterance, and from which it becomes possible to get a recursive definition of all types of complex literary fictions using the resources of formalization in logic of discourse and to integrate them within a general discourse typology.

Another Discourse Typology

Literary scholars, says Schaeffer, have been quite concerned about generic distinctions. But the theoretical interest in this issue is palpable also among analytic philosophers, especially philosophers of the ordinary language, who have tried to describe the types of functions associated with different ways to use types of sentences. Unlike Ludwig Wittgenstein, who postulated an unlimited number of language games, including fictional ones like playing on stage and inventing a story,^{Footnote 24} Searle argues that there are a limited number of uses of sentences in natural languages.^{Footnote 25} By studying real discourses, he tells us he has empirically found that language is used and understood in five ways: 1) to say how things are, 2) to engage another to do something, 3) to engage oneself to do something, 4) to express one’s attitudes and 5) to make changes in the world. His typology of illocutionary acts has been doubly important in language sciences. From a linguistic point of view, it has brought together all the elementary sentences of empirical languages. From a logical point of view, it has covered the four possible directions of fit between the types of sentences and the world in which we use them. However, it is limited to the minimal units of meaning in the use and understanding of ordinary language, i.e., to single speech acts composed of an illocutionary force and a propositional content. In his typology of factual discourses,^{Footnote 26} Vanderveken has reduced Searle’s categories to four, taking the main unit of analysis to be discursive goals that we can all meet on any topic simply by using language individually or collectively for an interval of time. In my view, the members of a linguistic society who assign a status (X) and a function (Y) to the texts they produce and receive in different contexts of utterance (C) can try to reach only one of the four linguistic goals that Vanderveken has correlated with the four directions of fit between the words and the world. Are fictional discourses more recalcitrant to typology than the factual ones that occur in one of the four discursive categories he deduced?

There is no doubt that all discourses that involve a genuine performative use of language can be integrated into a general typology of language games whose goals are purely linguistic. At first glance, one would locate the fictional story, novel, biography, memoirs, drama and so on in the descriptive category. This proposal is not uninteresting since the factual and fictional stories are generally composed of the same syntactic type of statement: the descriptive. However, this would be a category mistake in that it forces us to put on the same level literary genres that do not play the same role (function) in the use and understanding of language (status): Capote’s factual novel was produced to tell readers what was happening in the world (descriptive goal), whereas fictional novels were written rather to create new worlds (declaratory goal). Therefore, if the performative hypothesis of fiction is right, one would integrate all fictional texts in the declaratory category, which includes all discourses made in a context of master individual or collective declarations whose linguistic goal determines the double direction of fit between their constituent statements and the world in which they are used and understood in context of utterance, as in a wedding ceremony where the priest has to make master extra-linguistic declarations using language seriously. Why not? Fiction has largely been studied and understood as a discourse that linguistically resembles a descriptive discourse rather than a discourse that is logically declaratory. However, comparing factual and fictional discourses on the functional level shows how different they are even if they may look alike on the linguistic or literary level. To me, therefore, the reason seems clear enough to raise awareness about the world of differences between the factual and fictional discourses in the real use of language from an illocutionary and a linguistic or literary point of view, even though there is none from a strictly linguistic and literary point of view.

As I argued, the members of a linguistic society can succeed in producing all kinds of fiction in virtue of a collective agreement allowing them to create institutional entities, facts or events by using non-seriously all the linguistic elements contained in their language without having to name the new reality created with the prototypical performative verb ‘to declare.’ In succeeding, they produce a descriptive illusion—the same one that John Austin noticed in his analysis of performative utterances, only to claim later that the ones occurring on the stage are ‘parasitic.’^{Footnote 27} In other words, the factual discourses included in Vanderveken’s typology, and the fictional ones that I seek to integrate within the language games family whose goals are intrinsically linguistic, involve different uses of the same linguistic statements and therefore cannot be produced by making the same sequences of macro-speech acts. This is what I want to illustrate by extending Vanderveken’s discourse typology.

To integrate the fictional genres within the language games family, I retain four discursive criteria. The first criterion relates to the two universals in language use: the genuine serious and non-serious uses of language specific to factual and fictional discourses respectively (discursive status). The second criterion deals with the master illocutionary acts that must be accomplished in order to successfully produce a discourse in a context of utterance (discursive act). The third criterion applies to the four linguistic goals specific to discourses that play a role in a language society (discursive function). The fourth one refers to the various generic labels that denote their status and function within a linguistic (discursive genre) or artistic (literary genre) institution. On the grounds of those discursive phenomena, I propose a general discourse typology as a heuristic tool designed to show the logical possibilities of intersection between factual and fictional genres:

In accordance with the typology outlined, I can make a hypothesis in response to the question raised by Searle. There are at least two major ways to use and understand language in a linguistic society. Whatever the social sphere in question (artistic, literary, philosophical, scientific language games), a genuine non-serious use of language amounts to creating a new reality (declaratory goal) on various topics in a language game x (thematic conditions) by invoking the deontic power to do so within a linguistic institution (mode of achievement), assuming the ability to know how things are and how to make them in this very language game (preparatory condition), and having a network of attitudes (belief, desire and intention) double-directed toward entities, facts and events created. Conversely, a genuine serious use of language amounts to different verbal actions, including the act of creating an event which corresponds to what is represented in the semantic content of the discourse simply by uttering it (declaratory goal). These verbal actions may cover various topics (thematic conditions) and may create a new reality by invoking the power to do so within an extra-linguistic institution (mode of achievement of the goal), assuming the deontic power to do so (preparatory conditions), and expressing a complex set of attitudes composed of belief and desire or intention (sincerity conditions).

According to my analysis of the logical structure linking the genuine serious and non-serious uses of language, one must perform sequences of macro-speech acts—such as telling a story, describing objects, facts or events, referring to individuals and predicating on their properties and relations—by making narrative and descriptive utterances in order to successfully produce factual novels, biographies, memoirs and so forth, and optionally sequences of pretended verbal actions grafted to these utterances at the same time. This is what happens when the writer lies: he produces a defective novel or biography because his serious utterance does not fulfil the conditions of success determined by the components of those types of descriptive discourse. The same logic applies to fictional descriptive discourses, like traditional novels and untraditional biographies, with this fundamental difference that the writer has to perform sequences of master linguistic declarations, references and predications by making performative narrations or descriptions, otherwise these utterances would be understood for what they are not: a pack of lies or, in other words, defective discourses. Now we can extend this logic to all types of deliberative, declaratory and expressive discourse, and explain subsequently our ability to generate, interpret and identify a plethora of language games in virtue of our pragmatic competence, hence the value-free hierarchy of factual and fictional discourses in this typology containing eight generic categories in which factual and fictional genres have finally found their place.

Are There Literary Texts in These Classes?

In today’s interdisciplinary research context, I would conclude by raising the issue of the object and method of investigation in the logic of discourse. Originally founded by Austin, speech act theory was developed by Searle in the wake of research in analytic philosophy by Frege, Russell, Strawson, Grice and half a dozen other giants of analytic philosophy of language. When Searle published Speech Acts in 1969, the tendency was to oppose not only the logical and the ordinary language trends in analytic philosophy, but also two areas of investigation having language as object of inquiry: linguistic philosophy, which studied empirical facts specific to many human languages, and the analytic philosophy approach, which studied some general features of language like meaning, truth, sense, reference and reality implied in different language systems.

At the time, Searle noticed that the border between philosophy of language and linguistics was porous. In an essay about speech acts in linguistics,^{Footnote 28} he praised the linguists’ contribution to the study of an issue he considered fundamental: the links between structure and function in languages. With Vanderveken, Searle later founded illocutionary logic, whose task was to identify and formalize the properties of illocutionary forces of all types of utterances rather than the whole of linguistic forms that contribute to the meaning of utterances. On the grounds of constructive exchanges between philosophers of language, linguists, ethnomethodologists and conversation analysts, Vanderveken initiated a major shift in analytic philosophy. Unlike most analytic philosophers working within the logic and ordinary language trends, he rose from the semantic analysis of elementary and performative statements used to perform literal speech acts to the pragmatic analysis of discourses aiming to contribute to language sciences. As Vanderveken and Susumu Kubo pointed out years ago, “[i]t is quite clear nowadays that the future of speech act theory lies in the development of a general and rigorous theory of discourse.”^{Footnote 29} Today, one would wonder if scholars in language sciences, specifically in literary theory, have realized that a giant step has been made in speech act theory, an important transition that paves the way for exchanges between speech act theorists and literary theorists. Even Searle finally recognized Vanderveken’s contribution to language sciences.^{Footnote 30}

In current interdisciplinary research in language sciences, discourses of and about fiction are objects among others, of course. They are, however, quite important for the future of speech act theory as well as all theories aiming to explain how we use and understand a language in context because where there are humans who have to express and communicate their thoughts with many illocutionary forces, there are also individuals who produce and receive fictions. When he began his investigations on discourses, Vanderveken did not analyze such serious and non-serious texts about fiction. Searle did it without analyzing the natural and universal action of pretending to perform an illocutionary act. But speech act theorists cannot hope to contribute significantly to literary theory by analyzing fictional genres on the basis of the theoretical hypothesis that they are logically dependent on factual genres. Doing so is likely to draw the criticism of literary scholars and it limits the scope of a rigorous and promising theory in literary pragmatics. In response to the question raised by Stanley Fish, “Is There a Text in This Class?,”^{Footnote 31} speech act theorists can now answer ‘yes’ and reply that ‘literary texts are complex forms of life.’ This study on the relationship between speech acts and literary genres is a contribution to Vanderveken’s logic of discourse, which remains, en passant, a general theory of linguistic types of language games in which the performative hypothesis of fiction can be articulated in a systematic way.

The input of Vanderveken’s logic of discourse lies mostly, though not only, in that it reveals the logical form of literary genres and the dynamic that underlies them. In fact, one can explain the variations and changes in the use and understanding of language at work in discourses as soon as one has identified the logical form of literary genres that members of any linguistic society apply spontaneously to them. This is the path pursued by Stacie Friend in a study in analytic aesthetics devoted to the role of literary genres in the evaluation of factual and fictional texts.^{Footnote 32} Of course, the logic of discourse cannot contribute significantly and positively to identifying literature by defining recursively all types of discourses that are part of what Arthur Danto calls the ‘art world.’ However, the logician of discourse can and must contribute to the definition of literature if he cares about the empirical facts that he has to explain because, come to think, attempting to define a literary genre is already to assume that the facts that are part of its extension are literary. This study in speech act theory contributes to philosophy, language science and literature through the application of a constitutive and iterable rule of language to traces of ink on a white paper that caught the attention of its author when he tried to validate, by studying empirical facts, the fundamental hypothesis that the minimal units in the use and understanding of language are illocutionary acts: this text counts as a literary genre nowadays.