Narrative Theory and Discourse Analysis

John PIER (Tours and Paris)

Narrative Theory and Discourse Analysis

Abstract: One important development in French narratology since the decline of structuralism as it was practiced in the 1960s and 70s can be found in French discourse analysis. Emphasizing discourse in social interaction over structure, discourse analysis regards narrative as one type of discourse among others. Three dimensions of discourse that relate to narrative are singled out: the text linguistic categories of text and discourse, the reformulation of langue and parole in the light of the theory of enunciation and the relations between text and context. The role in narrative analysis of enunciation (Benveniste), speech genres (Baxtin) and the scene of enunciation (Maingueneau) are examined.

Keywords: discourse, text linguistics, speech genres, enunciation

* * *

Structural narratology, now commonly referred to as classical narratology, developed in France during the 1960s and 70s but was succeeded by postclassical narratology, a highly diverse group of narrative theories that have taken form in the course of the past thirty years on an international scale. While narratology in the French-speaking countries has evolved considerably over the decades and has witnessed a renewal in recent years, it cannot be said to conform closely to the classical/postclassical model.[1] A more telling point of reference for researchers in narrative theory working in the Francophone sphere is French discourse analysis (cf. Pier 2011, 2017).

I would like to set out here a few elements of research in this field which, I believe, deserve to be more broadly known and taken account of by research in narrative theory at large. The field of discourse analysis has developed within several different national and research traditions, and it has both drawn on and been incorporated into a variety of disciplines, mainly in the social sciences. As its name suggests, discourse analysis is not the special reserve of narrative or literary scholarship, but covers all fields of study where discourse plays an influential role. Overall, it can be said that with discourse analysis the emphasis no longer falls on structure, as was the case for the structuralists, but on discourses as they occur within the framework of social interaction, thus favoring the pragmatic dimension of communication in all its forms. On this view, narrative, while it is indisputably pervasive in human societies, is ultimately one form of discourse among others.

Now, there are three areas in which French discourse analysis has made contributions that are of consequence for narratologists. The first is a close consideration of text and discourse rather than the story/discourse dichotomy that dominates much narratological research, both classical and postclassical. The second is a rethinking of the Saussurean pair langue and parole. And the third has to do with the interrelations between text and context.

Before looking at these questions more closely, however, a few words on the historical origins of French discourse analysis are in order. Research in this field began in the late 1960s, more or less at the same time as the structuralist movement and the rise of narratology. Critical of Saussurean structural linguistics, the so-called pilot science that the structuralists advocated as a model for the social sciences, the French discourse analysts did not participate in the poststructuralist movement spearheaded by Roland Barthes but were focused, at least initially, on a corpus consisting mainly of political discourses. The principal methodological concern was how to conceptualize and model linguistic units beyond the sentence, a dimension of analysis that was opened up by the American school of linguistic distributionalism starting in the 1950s but that was not fully assimilated into French structural linguistics. As it turns out, it was the accomplishment of the early French discourse analysts to have pointed the way to a trio that continues to characterize the field today: a set of concepts and analytical practices coming from pragmatics, linguistic theories of enunciation or utterance and text linguistics. Although some researchers, such as Jean-Michel Adam, began incorporating discourse analytical principles into narratology as early as the mid-1970s (cf. Pier 2011: 351–56), French discourse analysts were not by and large concerned with narrative theory or literary theory. The changes came starting in the early 1990s with a number of publications by Dominique Maingueneau, who continues to work in the field of literature and discourse analysis today. To be sure, French discourse analysts have drawn lessons from structural narratology, but few if any regard their work as narratology. It is thus important not to confuse Gérard Genette’s pioneering contribution to “discourse narratology” and the positions staked out by the French discourse analysts. The latter in fact seldom mention Genette’s Narrative Discourse, referring, instead, to his work on transtextuality and paratextuality. This explains, in part, why developments in French discourse analysis are relatively unknown among non-French-speaking narratologists.

It is against this backdrop that the three areas of study I mentioned above can now be looked at more closely.

Why have discourse analysts chosen to give priority to text and discourse rather than to story and discourse? There are three principle reasons for this.

The first is that the term “story” covers only a limited number of varieties of verbal or other communicative activities. Can a logical demonstration, a military command, a commercial contract or instructions on how to install a computer program be construed as stories? Clearly there are many forms of discourse circulating in society which can be qualified as stories only marginally, if at all, even though it is the case that such discourses might in some way be present in narratives without actually being narratives themselves.

A second reason for the choice of text and discourse over story and discourse is of a more conceptual and terminological nature. It was Tzvetan Todorov who, in his important 1966 article, “Les catégories du récit littéraire,” proposed the terminology that has become a mainstay of classical and postclassical narratology alike. However, the difficulty with Todorov’s proposal is that it superimposes principles taken from the modes of enunciation or utterance in French linguistics (histoire, characterized by the absence of enunciator and enunciatee, and discours, where their presence is marked) onto narrative content and the signifying medium, respectively; this move, in turn, is said to also replace the Russian formalists’ principles of fabula and sjužet. Taking things a step farther, in English-language narratology the French and the Russian terms are often presented as equivalent to “story” and “plot.” This unhappy mixture of concepts in which histoire, fabula and story, on the one hand, and discours, sjužet and plot, on the other, are presented as an apparently homogeneous system has sometimes been diagnosed and occasionally revised (I myself have also examined the issues from the standpoint of semiotics; cf. Pier 2003). Nevertheless, it remains widespread in narrative research.

The shift from “story” to “text” is not a mere terminological matter, but may serve to shed light on conceptual confusion. By embracing the story/discourse framework, narratologists have sometimes felt the necessity to “export” narratological concepts and principles to disciplines and to forms of expression where they may not yield the most pertinent and fruitful results. The consequence has been the rise of a sort of “pan-narrativism” in which “story” and “discourse” have been called on to do duty in areas lying beyond their conceptual origins or methodological applicability. One author, for example, considers that the story/discourse distinction forms a basis for interdisciplinary research extending up to fields such as feminist studies and international relations (Dawson 2017). These and other considerations lead me to rally fully to James Phelan’s comments on the risks of “narratological imperialism” (Phelan 2005).

Taking their distance from this situation, French discourse analysts prefer to speak of text and discourse. This move results in a significant reordering of concepts and categories. For the discourse analyst, text breaks down into a text-structure and a text-product. Text-structure, as studied by textual linguistics, bears on relations beyond the sentence at the level of microlinguistic cohesion (sometimes called ‘texture’) and at that of macrostructural coherence (also called ‘structure’). These dimensions apply to all types of text, whether narrative or not, be they verbal or in other media. As for text-product (to be distinguished from text-archive – roughly speaking, the signifier), this corresponds to the enunciate (énoncé, or utterance in English), in other words, the empirical, observable and describable trace of a discursive activity, whether written, oral or visual: a weather report, a political discourse, a short story, a film, or even an ungrammatical fragment. Moreover, text-product (or utterance) is not something that simply falls out of the sky but is the result of an act of enunciation (énonciation), defined by Émile Benveniste as “the putting into operation of language (langue) by an individual act of use” (Benveniste 1974 [1970]: 80). In the light of this definition, it is important to remember that an act of enunciation occurs within the context of interpersonal interaction, so that enunciation must actually be thought of as co-enunciation involving the participation of two enunciators or more. In French, as in other European languages, interpersonal interaction in discourse is analyzed through various linguistic markers including the use of personal pronouns and of deictic expressions such as “here,” “there,” “now,” “then.” In France, an entire school of enunciative linguistics has grown up around these notions.

Now, the problem of enunciation ties in with another feature of language that was paramount in structural narratology and that also continues to exert an influence in present-day research, particularly in its vulgarized form: langue and parole. In structural narratology, as Gerald Prince has pointed out, these concepts are often used to describe “narrative langue” as “the system of rules and norms accounting for the production and understanding of individual narratives,” the individual narrative thereby corresponding to “narrative parole” (Prince 2003 [1987]: 70). It is worth remembering, however, that Saussure’s formulation of langue and parole is itself problematic and has been the subject of various observations and criticisms. Langue, the object of linguistic study, is the abstract “system of signs,” a code governed by rules and constraints which are external to the individual but are nevertheless lodged in the minds of the members of the speech community. Parole (sometimes called the message) is the material of language and corresponds to an individual and momentary act of use. But by this very token, parole, because it escapes the rules of langue, is not the object of linguistic study: thus, while Saussure spoke of the sentence as a syntagm (a group of words ordered by syntagmatic and paradigmatic relations), he simultaneously excluded it from langue, the system of language, considering the sentence to be a manifestation of parole.

Clearly, then, there is gap between langue and parole, a situation that has drawn numerous proposals and counter-proposals for revisions by linguists and literary theorists. For present purposes, suffice it to say that Roland Barthes, in his famous article “Introduction à l’analyse structurale des récits,” postulated a relation of “homology” and even “identity” between language and literature, stating that “structurally, narrative partakes of the sentence, without ever being reduced to a sum of sentences: narrative is a long sentence, just as all constative sentences are, in a way, the outline of a small narrative” (Barthes 1966: 4). This homological relation between narrative and sentence, on the basis of which Barthes called for a “translinguistics” – a linguistics beyond the sentence – pointed in the direction of discourse analysis but fell short of a satisfactory framework for such analysis. Barthes’s proposal fails to distinguish between sentence (a grammatical construction) and enunciate, the result of an act of enunciation. Various attempts were undertaken during the 1970s and 80s to remedy this ambivalent situation, notably by enlisting Naom Chomsky’s generative-transformational grammar, with its notions of competence and performance in place of langue and parole, in order to create so-called text grammars and even narrative grammars. However, these endeavors ultimately gave way to text linguistics, one of the pillars of French discourse analysis.

The approach to the langue/parole pair adopted by the researchers in question draws on a principle quite different from that taken up by structuralist narratology: Mixail Baxtin’s principle of speech genres. Speech genres, present in all spheres of human communication, are defined as “relatively stable thematic, compositional, and stylistic types of utterances” (Baxtin 1986 [1952–53]: 64). Corresponding to neither langue nor parole, speech genres, which are both conventional and normative and thus of a distinctly pragmatic nature, play a crucial role to the extent that they serve as an interface between langue and parole. According to Baxtin, it is thanks to speech genres that verbal communication exists. It must further be noted that speech genres are not limited to the literary genres of epic, dramatic and lyric and their subgenres, since, as sociolinguistic phenomena, they exist in innumerable forms and in all social domains, both in primary or simple forms (the spontaneous exchanges occurring in everyday speech) and in secondary or complex forms (scientific studies, literary works, etc.) as well as in the rich interaction between primary and secondary forms. Within the various social spheres there exist specific functions (economic, scientific, literary, etc.) which, in conjunction with certain conditions peculiar to each sphere, give rise to different genres (stock market analysis, scientific treatise, etc., but also the sonnet, the short story, the essay, etc.). Key to the principle of speech genres is the notion of vyskazyvanie, roughly the equivalent of utterance or enunciate, that is to say, the product of an act of enunciation. The defining feature of Baxtin’s utterance as a unit of speech communication, setting it off sharply from grammatical units of language (lexical items, syntactic structures, sentences, etc.), is the change of speakers. The changing of speakers in a sequence of utterances is a significant point for many reasons. Mainly, it can be said that units of language, such as they are presented in the schematic diagrams of general linguistics, communicate nothing: as a grammatical unit, the sentence is constructed by the individual speaker, but it is directed to no one and elicits the response of no one. From the perspective of linguists such as Saussure (who is not alone in this regard), communication consists in the neutral “encoding” of a message by the sender which is then “decoded” by the passive receiver – a system evocative of the transmission of telegrams. Utterances, however, because they are defined by the changes of speaker that occur in the process of communicative exchange, function differently: an utterance, unlike a sentence, elicits an active responsive attitude in the listener, and in doing so it constitutes the initial preparatory stage of the rejoinder which is to come. The clearest example of this process is found in the dialogue, of course, where speaker and listener exchange roles in the ongoing communicative chain. But for Baxtin, the dialogical dimension of language, or dialogism, permeates all aspects of communication; it can be found even in interior monologues and other forms of mental discourse.

So the question now arises: why are Baxtin’s speech genres, a reconfiguration of the Saussurean langue/parole distinction taken over by French discourse analysis, important for the narratologist? First of all, speech genres introduce into the study of language a set of principles coming from pragmatics, thereby bringing into focus processes that are closer to actual discourse as a communicative act than, for example, the underlying system of signs that constitutes Saussure’s langue. Second, speech genres, which cover a wide spectrum, take form within a sociodiscursive space and include numerous manifestations of discourse – narrative and non-narrative, literary and non-literary and so on. In this way, speech genres tie up with the text/discourse distinction since, contrary to narratological theories that rely on the story/discourse distinction, they are not bound only to stories or narratives: speech genres do enter into the compositional dimension of narratives, of course, but also into that of legal and commercial documents, technical reports and much more, the reason why they play a crucial role in text linguistics (cf. Adam 2011 [1999]). Finally, the enunciative properties of discourse, as analyzed by Benveniste, complement the Baxtinian notion of utterance. Thus, formal features such as the use of pronouns and deictic expressions can signal the change of speakers throughout the communicative chain of utterances.

For the final step of my inquiry into the relations between French discourse analysis and narratology, I would now like to look at what Dominique Maingueneau calls scène d’énonciation, or “scene of enunciation” (literally: “theatrical stage of enunciation”). This concept brings together a set of principles that calls for reconsidering the divide between text and context. Maingueneau emphasizes that scène d’énonciation is distinct from “situation of enunciation,” which is a linguistic term, and also from “situation of communication,” which pertains to sociological conceptions of language. Both approaches view verbal activity “from the outside,” one focusing on the physical or social environment, for example, the other on the social status of the participants in an act of communication or on the circumstances of the publication of a book. As for the word scène employed in scène d’énonciation, this is clearly a metaphor: it is chosen in order to stress the idea that the act of enunciation, parole, is a mise en scène, a sort of theatrical production that takes place in the form of a self-constituting process occurring within a defined discursive space. By adopting the theatrical metaphor of scène d’énonciation, Maingueneau endeavors to enter discourse “from the inside.”

Now, Maingueneau’s system breaks down into three scenes, all of them of a pragmatic nature.

The first is the “global scene.” This category corresponds more or less to text type or discourse type: recognizing a book as a literary fiction, a collection of historical studies or a treatise on nuclear physics will trigger different kinds of presuppositions in the reader.

However, the global scene, or type, acts in close collaboration with the “generic scene,” the latter derived in large part from Baxtin’s speech genres. It is at this level that the sociohistorical conditions of enunciation for each genre come into play: for example, the finality of a given genre (the aim of a student research paper in an academic setting is not that of a talk show for late-night television viewers or that of a newspaper editorial intended for readers having a given political persuasion); the linguistic and stylistic resources employed (the lexical choices peculiar to administrative documents as compared to those of popular magazines); the medium or material support (the Internet has given rise to previously unknown genres such as blogs, chats and forums).

Finally, in order to account for the singularity of the individual discourse, Maingueneau introduces the notion of “scenography.” This term, normally used to refer to the design of the stage set, refers here to the conditions that render discourse possible and to the product of that discourse; at the same time, it adds to the theatricality of the scène (the spatio-temporal framework in which the discourse occurs) the dimension of –graphie (the process of the discourse’s occurrence). Scenography, located both upstream and downstream of the work, is thus not a mere scaffolding for contents, but a true “pivot of enunciation” of discourse (Maingueneau 2004: 192 and 201). “To enunciate,” says Maingueneau, “is not only to activate the norms of a pre-existing institution of speech; it is also to construct, on this basis, a singular mise en scène of enunciation: a scenography. […] The notion of scenography is based on the idea that the enunciator, through his enunciation, organizes the situation on the basis of which he is claiming to speak” (Maingueneau 2014: 129). In general – and this is more emphatically the case with literary works – there arises a meaningful tension between the scenography of a discourse and its generic scene.

As one of many possible examples of the scene of enunciation and the relation between generic scene and scenography, we can consider one of the early English novels, Pamela; or, Virtue Rewarded (1740) by Samuel Richardson. Published as a series of letters written by the heroine, a young servant, to her parents for the purpose of illustrating how well-bred working class ladies in eighteenth-century English society should resist the designs of their wicked masters, Pamela is consistent with the epistolary form as a speech genre. At the same time, however, the book’s scenography builds up a dramatic situation in which the protagonist’s experience is not filtered through the moralizing interventions of a narrative voice, but is witnessed by the reader as though the incidents took place on the stage. Moreover, the avowed didactic purpose of the letters is progressively sidelined as the heroine’s moral resistance culminates with the conversion of Mr. B, her social superior, to more noble ways and, in the end, with Pamela’s social ascension through marriage. The tensions produced between the epistolary form of Richardson’s novel and the scenography of enunciated speech are thus inseparable from the sociohistorical context in which the work appeared. From the discourse analytical perspective, Pamela exemplifies how text as discourse is not a formal structure to which external contextual facts and circumstances must be added, but rather a work in the course of which context is constituted.


Adam, Jean-Michel (2011 [1999]). La linguistique textuelle: Introduction à l’analyse textuelle des discours. Paris: Armand Colin.

Barthes, Roland (1966). “Introduction à l’analyse structurale des récits.” Communications 8: 1–27.

Baxtin, Mixail M. (1986 [1952–53]). “The Problem of Speech Genres.” Speech Genres and other Late Essays. Vern W. McGee (trans.), Caryl Emerson and Michael Holquist (eds.). Austin: University of Texas Press, pp. 60–102.

Benveniste, Émile (1974 [1970]). “L’appareil formel de l’énonciation.” Problèmes de linguistique générale II. Paris: Gallimard, pp. 79–88.

Dawson, Paul (2017). “How Many ‘Turns’ Does it Take to Change a Discipline? Narratology and the Interdisciplinary Rhetoric of the Narrative Turn.” In Emerging Vectors of Narratology. Peter Hühn, Jan Christoph Meister, John Pier and Wolf Schmid (eds.). Berlin/Boston: De Gruyter, pp. 405–33.

Maingueneau, Dominique (2004). Le discours littéraire. Paratopie et scène d’énonciation. Paris: Armand Colin.

Maingueneau, Dominique (2014). Discours et analyse du discours. Paris: Armand Colin.

Patron, Sylvie, ed. (2018). Introduction à la narratologie post-classique. Les nouvelles directions de la recherche en narratologie. Villeneuve d’Ascq: Presses universitaires du Septentrion.

Phelan, James (2005). “Who’s Here? Thoughts on Narratological Imperialism.” Narrative 13 (3): 205–10.

Pier, John (2003). “On the Semiotic Parameters of Narrative: A Critique of Story and Discourse.” In What is Narratology? Questions and Answers Regarding the Status of a Theory. Tom Kindt and Hans-Harald Müller (eds.). Berlin/New York: Walter de Gruyter, pp. 73–97.

Pier, John (2011). “Is There a French Postclassical Narratology?” In Current Trends in Narratology. Greta Olson (ed.). Berlin/New York: De Gruyter, pp. 336–67.

Pier, John (2017). “Von der französischen strukturalistischen Erzähltheorie zur nordamerikanischen postklassischen Narratologie.” In Grundthemen der Literaturwissenschaft: Erzählen. Martin Huber and Wolf Schmid (eds.). Berlin/ Boston: De Gruyter, pp. 59–87.

Prince, Gerald (2003 [1987]). Dictionary of Narratology. Revised Edition. Lincoln/London: University of Nebraska Press.

Todorov, Tzvetan (1966). “Les categories du récit littéraire.” Communications 8: 125–51.

* * *

John Pier

Ph.D. (New York University, 1983) and Habilitation à diriger des recherches (Besançon, 1987), is an emeritus professor at the University of Tours, where he taught 19th- and 20th-century anglophone literatures. A statutory member of the Centre de recherche sur les arts et le langage (CNRS) in Paris, he has co-directed at this center the seminar “Contemporary Research in Narratology” since 2003. His work bears on transdisciplinary approaches to narrative theory based on semiotics, complexity theory, intertextuality, intermediality and discourse analysis. Co-editor of the collection Narratologia at De Gruyter, he is also the co-founder (in 2009) of the European Narratology Network (ENN), of which he was president from 2013 to 2015. He is the author of numerous scholarly articles and book chapters and has edited or co-edited more than fifteen anthologies, including Théorie du récit. L’apport de la recherche allemande (2007), Handbook of Narratology (2nd ed. 2014), Emerging Vectors of Narratology (2017), Le formalisme russe cent ans après (2018), Jan Mukařovský. Écrits 1928-1946 (2018) and Contemporary French and Francophone Narratology (forthcoming).

[1] For a collection of essays by leading representatives of postclassical narratology translated into French, see Patron, ed. (2018).