Volume 8, No. 2, Art. 21 – May 2007
Discourse Analysis in France: A Conversation
Dominique Maingueneau & Johannes Angermüller
Abstract: In this interview, the linguist Dominique MAINGUENEAU sketches the history of discourse analysis in France. After discussing the intellectual background in which it emerged, he turns to certain key figures of French discourse analysis like Michel FOUCAULT and Michel PÊCHEUX. Special attention is given to the role of Michel FOUCAULT, who has crucially influenced his research on discursive scenography and self-constituting discourses. Concerning FOUCAULT's methodological impact MAINGUENEAU emphasises the problem of "enunciation". A French version of discourse pragmatics, enunciative linguistics focuses on the way texts mobilise their contexts.
Key words: (French) Discourse analysis, Michel Foucault, Michel Pêcheux, linguistics, social sciences, pragmatics, Critical Discourse Analysis
Table of Contents
1. Discourse analysis in France: The impact of Michel FOUCAULT and Michel PÊCHEUX
2. Current Tendencies: Enunciation, Scenography, Self-constituting Discourses
3. Discourse Analysis as an Interdisciplinary Field in an International Context
About the Interview
This interview is the product of the collaboration between Dominique MAINGUENEAU and Johannes ANGERMÜLLER, who are members of CEDITEC (Centre d'étude des discours, images, textes, écrits, communications) at the University of Paris 12. The main part was done in March 2007 in Paris. The interview was conducted in French and translated by both authors. 
About Dominique MAINGUENEAU
Dominique MAINGUENEAU is professor of linguistics at the University of Paris-XII (Créteil), where he takes part in the activities of the CEDITEC ("Centre d'étude des discours, images, textes écrits, communications"), and member of Institut Universitaire de France. Since the 1970s, his research has mainly focused on discourse analysis. He has published many books in this area of study: from Initiation aux méthodes de l'analyse du discours (Paris, Hachette, 1976) to the co-edition of a Dictionnaire d'analyse du discours (Paris, Seuil, 2002). His research associates a pragmatic outlook on discourse with linguistic "enunciation" theories and Michel FOUCAULT's Archaeology. For twelve years he has been working in the field of "self-constituting discourses" (philosophical, religious, scientific, literary etc.), that legitimate the entire discursive production. 
His numerous publications on discourse analysis include:
Maingueneau, Dominique (2004). Le discours littéraire. Paris: Armand Colin.
— (2002). An academic genre. Discourse Studies, 4(3), 319-342.
— (1995). L'énonciation philosophique comme institution discursive. Langages, 119, 40-62.
— (1994). Die "französische Schule" der Diskursanalyse. In Konrad Ehlich (Ed.), Diskursanalyse in Europa (pp.187-195). Frankfurt/M.: Peter Lang.
— (1993). Le contexte de l'œuvre littéraire. Énonciation, écrivain, société. Paris: Dunod.
— (1991). L'Analyse du discours. Introduction aux lectures de l'archive. Paris: Hachette.
— (1987). Nouvelles tendances en analyses du discours. Paris: Hachette.
— (1984). Genèses du discours. Bruxelles: Pierre Mardaga.
— (1983). Sémantique de la polémique. Lausanne: L'Age d'Homme.
— (1981). Approches de l'énonciation en linguistique française. Paris: Hachette.
— (1976). Initiation aux méthodes de l'analyse de discours. Problèmes et perspectives. Paris: Hachette.
Charaudeau, Patrick & Maingueneau, Dominique (2002). Dictionnaire d'analyse du discours. Paris: Seuil.
Maingueneau, Dominique & Cossutta, Frédéric (1995). L'Analyse des discours constituants. Langages, 117, 112-125. 
1. Discourse analysis in France: The impact of Michel FOUCAULT and Michel PÊCHEUX
J.A.: Dominique, you have been working on discourse analysis since the 1970s. In this conversation, we want to know more about the history of discourse analysis in France and especially about Michel FOUCAULT, who is an important reference in your own work. The pioneering works in French discourse analysis appeared in the late 1960s (PECHEUX, 1969; FOUCAULT, 1969; Langages's special issue, no.13). What were, in your view, the conditions at the time that helped establish this new field? 
D.M.: Discourse analysis developed in France in the late sixties because the conditions were favourable. Various factors came together: a scientific tradition, a school practice and an intellectual climate.
The scientific tradition is European philology, which always associated historical studies and text analysis. The philology of the 19th century defined itself as an auxiliary discipline of history. With very different theoretical backgrounds, discourse analysis occupied a good part of the territory that traditional philology had neglected.
The school practice is a kind of "close reading" (explication de texte) taught to pupils in secondary schools and to students of the humanities. Of course, French discourse analysis has never been a continuation of literary commentary, but it does find its roots in that practice. It must be recalled that since the late 19th century the French way of practising stylistics has been rather specific: while the inspiration of German stylistics (which was much more prestigious) was mainly psychological and hermeneutic, French stylistics was mainly based on grammatical analysis: the analysis of linguistic phenomena was considered to be an obligatory step to the interpretation of texts (CRESSOT, 1947; MAROUZEAU, 1941). In this way, French discourse analysis claimed to found itself upon linguistics.
The intellectual climate was French structuralism, particularly literary structuralism, in its heyday in the sixties. By deciding to apprehend texts "immanently", rather than with reference to the intentions of their author, this trend opened new ways of studying texts. 
J.A: In the 1970s, Michel PÊCHEUX gathered a group of discourse analysts who became what was sometimes called the "French school of discourse analysis". Who were they and how did they do their research? Is it still possible to speak of a "French school"? 
D.M.: French discourse analysis is heterogeneous. As you said, two main trends can be emphasised: Michel FOUCAULTs "archaeology" and Michel PÊCHEUXs discourse analysis. Only the latter claimed to be a discourse analyst; the influence of the former was much more indirect. In its narrow meaning "the French school of discourse analysis" refers only to PÊCHEUX's project. This "French school" emerged amongst Marxist researchers influenced by ALTHUSSER's thought. They belonged to various disciplines, especially philosophy (PÊCHEUX), history (Régine ROBIN, Jacques GUILHAUMOU and others); but most of them were linguists (Denise MALDIDIER, Jean-Jacques COURTINE, Jacqueline AUTHIER-REVUZ etc.). 
ALTHUSSER did not deal with discourse analysis but his doctrine implied a project that aimed to study the way in which ideology is invested by language. Language was considered to be relatively autonomous from "infrastructure" and it ought to be considered in its "materiality" and not as a simple vehicle for representations previously elaborated. 
This trend of discourse analysis developed chiefly in three places: the linguistic department of Paris X-Nanterre University, the "Centre for Political Lexicometry" (Ecole Normale Supérieure de Saint-Cloud), where computers were used to study political discourse; a laboratory of social psychology in the CNRS (Centre National de la Recherche Scientifique), where Michel PÊCHEUX developed a computer program, "Automatic Discourse Analysis", to identify ideological processes in corpora. 
In 1969, discourse analysis was discovered by a larger audience. Langages, which was the most important journal of linguistics in France, published an issue entitled "Discourse analysis" (DUBOIS & SUMPF, 1969); at the same time Michel PÊCHEUX published a book, Analyse automatique du discours, in which he presented his project. These two publications officially mark the birth of discourse analysis in France. Discourse analysis claimed to be a true analysis, in the psychoanalytic sense of the word, aiming at decomposition in order to reach an unconscious content. The text was considered to be a deceptive totality: its continuity had to be broken up to uncover the "other", hidden discourse. The analyst detected invisible gaps in texts, cut-up fragments (words, syntactic schemata etc.) and connected them to build an interpretation in terms of class struggle. 
In such a conception of discourse analysis, scientific investigation and militant preoccupations could not be disassociated: studying ideological processes in texts contributed to transforming society; from that viewpoint, the analysis of any discourse was politically orientated. This approach shares a few features with contemporary Critical Discourse Analysis. The French school, however, was deeply influenced by psychoanalysis and Marxism whereas most research in Critical Discourse Analysis has recourse to socio-cognitive theories and is concerned with gender or ethnic prejudice. 
J.A.: While FOUCAULT's Archaeology proffered a strong theoretical program, it is less explicit about his analytical method. What was his impact in French discourse analysis? How has FOUCAULT informed your linguistic work? 
D.M.: PÊCHEUX published his book Analyse automatique du discours at the same time when Michel FOUCAULT published his Archéologie du savoir, which in many respects holds a conception of discourse quite opposite to that of the ALTHUSSERian analysts. Later on, FOUCAULT's problematic took advantage of the development of pragmatic preoccupations in the field of social sciences and linguistics, particularly in France with the development of theories of "enunciation". Some notions elaborated by FOUCAULT were quite positive for discourse analysis: discourse must no longer be considered as signs referring to representations, but as "practices that systematically shape the objects of which they speak" (1969, p.67 [English translation: FOUCAULT, 1989, p.54]). FOUCAULT states that the forms of subjectivity cannot be separated from the institutions of speech that intervene in their construction. For him, discourse is "a set in which the dispersion of the subject and his discontinuity with himself can be determined", "a network of distinct sites" (1969, p.74 [English translation: FOUCAULT, 1989, p.60]). FOUCAULT, unlike the "French school", rejected the hermeneutic attitude that compelled analysts to look for concealed meanings: instead of searching "below what is manifest, the half silent chattering of another discourse", let us "show why it could not be other than it is" (1969, p.40 [English translation: FOUCAULT, 1989, p.31]). 
For the historian of ideas, the concurrence of the two publications is an interesting case in point. The trend represented by PÊCHEUX is influenced by ALTHUSSER' s Marxism and LACANian psychoanalysis. Far from following PÊCHEUX' s line of thought, FOUCAULT' s book opened an entirely different approach to the discursive. His was not just a succession of brilliant intuitions, but a dense network of concepts articulating a powerful and coherent conception of discourse. Such a book, which was followed by L'Ordre du discours [The Discourse on Language] (1971) could not but exert a great deal of attraction on discourse analysts. However, you should not fall a victim to the retrospective illusion by making FOUCAULT the initiator of discourse analysis. For a long time, his influence was diffuse, without opening an alternative space capable of producing empirical work in a precise methodological framework. However, in the course of time, the Archaeology has benefited from the success of the pragmatic currents in the entire social sciences and linguistics, particularly the theories of enunciation. 
Undoubtedly, FOUCAULT's position raises a number of difficulties for a discourse analyst who wants to keep his moorings in linguistics. By limiting the field of linguistics, which he reduces to the science of "langue" (SAUSSURE), to the study of words and sentences, FOUCAULT excludes it from his archaeology. Given such a weak notion of linguistics, FOUCAULT claims for his archaeology the exclusive right to "discourse" whereas in the current conjuncture discourse analysts seem more prone to having recourse to linguistics. 
Though FOUCAULT talks about "discourse" or "enunciative function", he manipulates elements situated on a prelinguistic level, even a pretextual level, as it were, which is revealed by these lines from the Archaeology of Knowledge:
"What are described as 'systems of formation' do not constitute the terminal stage of discourse, if by that term one means the texts (or words) as they appear, with their vocabulary, syntax, logical structure, or rhetorical organisation. Analysis remains anterior to this manifest level, which is that of the completed construction […]. If analysis studies the modalities of enunciation, it questions neither the style nor the succession of the sentences; in short, it leaves the final placing of the text in dotted outline." (1969, p.100 [English translation: FOUCAULT, 1989, p.84]) 
This type of affirmation is hardly compatible with the postulates of discourse analysis, which cannot but refuse the idea that textual organisation is just a surface phenomenon, that interaction strategies are accessory, i.e. "style", "rhetorics" etc.. As is shown by the same recourse to the notion of "archaeology", there is in FOUCAULT a certain tension between a clearly structuralist inspiration and a movement of thinking which gives special emphasis to the "enunciative function" and, more generally, to problems familiar to current trends in pragmatics. 
However, I think that the Archaeology elaborates a number of important ideas for discourse analysis:
the assertion of what founds, in my view, any real discourse analysis, namely the opacity of discourse, neither reducible to "langue" nor to social or psychological instances. On this, FOUCAULT offers a few well-known phrases, e.g.:
"But what we are concerned with here is not to neutralise discourse, to make it the sign of something else, and to pierce through its density in order to reach what remains silently anterior to it, but on the contrary to maintain it in its consistency, to make it emerge in its own complexity. […] I would like to show with precise examples that in analysing discourses themselves, one sees the loosening of the embrace, apparently so tight, of words and things, and the emergence of a group of rules proper to discursive practice. […] A task that consists of not – of no longer – treating discourses as groups of signs (signifying elements referring to contents of representations) but as practices that systematically form the objects of which they speak." (1969, pp.65-67 [English translation: FOUCAULT, 1989, pp.52-54])
the recourse to an enunciative subjectivity irreducible to traditional frames of reference in philosophy or psychology. By closely linking discourse and institution in an apparatus (dispositif) of enunciation, which allows the occurrence of enunciative events at the same time that these events are constituted by their very existence, FOUCAULT calls into question traditional cleavages.
a distancing vis-à-vis spontaneous hermeneutics as a model for the analysis of texts:
"The analysis of thought is always allegorical in relation to the discourse that it employs. Its question is unfailingly: what was being said in what was said? The analysis of the discursive field is orientated in a quite different way; we must grasp the statement in the exact specificity of its occurrence; determine its conditions of existence, fix at least its limits, establish its correlations with other statements that may be connected with it, and show what other forms of statements it excludes. We do not seek below what is manifest the half silent murmur of another discourse; we must show why it could not be other than it was." (1969, p.40 [English translation: FOUCAULT, 1989, p.31]) 
J.A.: There are many strands of "enunciation linguistics", which offers a great deal of analytical instruments for the analysis of texts. What is the problem of "enunciation" about and why has it been picked up by many discourse analysts since the late 1970s? Would you say that FOUCAULT was a precursor of enunciation theory? 
D.M.: As I said, FOUCAULT's approach has largely benefited from the pragmatic and enunciative turn which has taken place in linguistics and the entire social and human sciences. The structuralist conception of language and texts has turned out to be of limited compatibility regarding the reflection on the social inscriptions of discourse. Originally, the work on enunciation was done by linguistics (Charles BALLY, Emile BENVENISTE, Roman JAKOBSON, Antoine CULIOLI and others), who were not interested in discourse analysis at all, but just wanted to improve the understanding of the way natural languages work. They studied phenomena like deictical elements (particularly linguistic persons: "I", "you", "we" …), modalities, verbal tenses, reported speech, etc. Pragmatic currents (John L. AUSTIN, John SEARLE, Paul GRICE etc.) by contrast, were mainly interested in the philosophy of language. But both of them have contributed to the elaboration of powerful instruments for the analysis of discursive practices. As FOUCAULT underlines, "the order of discourse", is situated on the unstable border between language as a system (langue) and social practices. Whatever can give consistency to the study of this order of discourse is, therefore, useful, and particularly notions like speech acts, genres of discourse, enunciation, polyphony, textual coherence, etc., which imply a notion of language as cooperative activity rather than monological representation, as was the case in the dominant tradition of the Western philosophy of language, in which rhetoric was separated from logic. 
J.A.: As you say, discourse analysis in France has been closely connected to linguistics in a broad sense (sciences du langage). Given the interest many discourse analysts have in sociological and historical questions, how would you describe what they do? 
D.M.: I think that the specific interest of discourse analysis is to apprehend discourse as articulating texts and social places. Consequently, its object is not textual organisation, nor communicative situation (as it can be described by traditional sociology), but what knots them together is a certain genre of discourse practice. This notion of "social place", however, must not be considered from a literal view point: this "place" may be a position in a symbolic field (political, religious, etc.). Genres must not be considered as text classes, in a taxonomical way, but as communication devices, at once social and linguistic: in a word, discursive. 
Although discourse analysis, by its very nature, is at a crossroads between linguistics and social sciences, I think that discourse analysis must keep in privileged contact with linguistics. This does not mean that linguistics must provide all concepts and methodological tools. On the contrary, discourse analysis has always created its own network of concepts. This is the reason why many scholars speak of "discourse linguistics" for research not belonging to traditional areas of linguistics (lexicology, syntax …). 
2. Current Tendencies: Enunciation, Scenography, Self-constituting Discourses
J.A.: If enunciative linguistics offers a range of methodical instruments for discourse analysis, how would you describe the methodological differences between linguistic and sociological approaches? Would you say that discourse analysis privileges qualitative methods? 
D.M.: The first question is not that easy to answer. It depends on what sort of sociological approach you are thinking of. If you are a sociologist who practises conversation analysis, or if your theoretical background is the sociology of knowledge of LUCKMANN and BERGER, you are very different from people who work out criteria to define social classes or analyse types of consumer behaviour. Roughly speaking, a traditional sociologist considers words as the "expression" of realities beyond language; from this point of view, discourse is passive, it is only a place where social realities are projected. For me, discourse analysis can be distinguished from such approaches by three main characteristics; 1) it apprehends discursive practices as playing a crucial role, beyond any elementary distinction between words and social realities : discourse too is a social activity, and its manifold genres play a key role in the construction of social reality; 2) discourse analysts do not study communication settings but the interaction of communication settings and of the "enunciative scenes" that are implied by communicative events. 3) discourse analysis researchers takes into account at the same time concepts and methods from systemic linguistics and from discourse linguistics and the conditions that make discourse practices possible (for example, for printed texts a wide range of phenomena: typography, circulation networks, authorship etc.). Therefore, the distance between sociology and discourse analysis approaches is very variable. It is a matter of degree. Some sociologists can be true discourse analysts, whereas others do not worry at all about discourse. 
There is no doubt that the opposition between "qualitative" and "quantitative" methods is difficult to draw. The French tradition of discourse analysis obviously includes both qualitative and quantitative methods, with the help of computers for large corpuses. From their viewpoint, close reading and computers have to be associated. Roughly speaking, one could say that discourse analysis is mainly qualitative because, as a rule, it always takes into account communicative, textual and linguistic properties of discourses even when they turn to quantitative methods. 
J.A.: Your first book on discourse analysis was Initiation aux méthodes de l'analyse de discours (MAINGUENEAU, 1976). How did you come to be interested in this topic? What were the major models and inspirations for your work then? 
D.M.: I do not really know why I came to be interested in discourse analysis. When you were a student in that structuralist period you had to study texts; but first of all I wanted to be a linguist. I guess that discourse analysis was a strategy to reconcile texts and linguistics. In the sixties and the seventies analysing texts with new concepts and new methods was very exciting: it was at the same time a political and a philosophical decision. Being a discourse analyst, defending a young discipline was a way of being young … 
I was in a paradoxical situation. On one hand I was in the "hard core" of French school (the handbook that you have mentioned belongs to that trend), on the other hand my own research was very far from the point of view of the ALTHUSSERian French school. I was working on religious discourse. Little by little I mixed my own ideas with FOUCAULT's inspiration and I elaborated my own conception of discourse analysis in the eighties (MAINGUENEAU, 1984). 
The Archaeology was of great importance for me, for it radically criticises the philological approach: when it rejects notions like "world vision" (Weltanschauung), "work" (œuvre), "author", "document", "influence", "context", etc., it frees a space for an approach to discourse analysis which is centered on what I call "discursive institutions", the reciprocal entanglement of the use of language and a place, through genres of discourse which are dispositifs of enunciation. By drawing upon the Archaeology, on linguistic and pragmatic theories of enunciation, we can rethink a range of entrenched practices and notions which still determine our approach to texts. 
Currently, I am working on what I call "self-constituting discourses" (MAINGUENEAU, 1999). For quite some time I have been studying a wide range of texts, particularly religious, scientific, literary and philosophical, and I noticed that if we disregard superficial differences, many descriptive concepts could be transferred easily from one to the others. So I came to the assumption that in the discursive production of a society a specific area could be delimited: that of "self-constituting discourses". Bringing those discourses together may open an interesting research program. These discourses claim to be above any other type of discourses. They must set themselves up as intimately bound with a legitimising source and show that they are in accordance with it, owing to the operations by which they structure their texts and legitimate their own context, the way they emerge and develop. Discourse analysts can study the textual operations by which self-constituting discourses manage their self-foundation. 
As the analysis of self-constituting discourses is discourse analysis, it aims at showing the connectedness of textuality and action, of "intradiscursive" and "extradiscursive" dimensions. Discourse as text and discourse as activity are tightly knotted in discursive institutions, which articulate groups of men and text genres. Text production delimits a space inside social space but configures that space, too. Therefore, the analysis of self-constituting discourses implies keeping a "rhetorical" conception at a distance: in fact, content and textual organisation is not independent of the discursive scene which takes charge of them and through which they appear. 
The concept of "enunciation scene" is very important for me. The meaning of an utterance cannot be separated from its pragmatic frame; even a doctrine must stage its own discourse. But a distinction must be made between generic scene and scenography. Roughly speaking, the generic scene is part of the context, it is the very scene that the genre prescribes, whereas scenography is produced by the text. So, two texts belonging to the same generic scene may stage different scenographies. A preaching in a church, for instance, can be staged through a prophetic scenography, a conversational scenography, and so forth. In the former case the speaker will speak in the way prophets do in the Bible and will give the corresponding role to his addressees; in the latter case he will speak in a friendly way to the audience. As a result, addressees interpret discourses through the association of two scenes, two contexts: one (generic scene) is imposed by the genre, the other one (i.e. the scenography) depends on particular discourses. Not all texts turn to scenography. As a rule, administrative genres, for instance, merely obey the norms of their generic scenes. Here "scenography" is not used in its usual way: It adds to the theatrical dimension of "scene" the "graphic" dimension, of legitimising inscription, for scenography gives authority to discourse, it has persuasive effects on addresses. 
Scenography is not a frame, a scenery, as if discourse occurred within an already fixed place, independent of discourse. On the contrary, discourse puts progressively into place its own communicational device. So, -graphy must be apprehended simultaneously as frame and process. Scenographies are determined according to the content of discourse: speaking through a prophetic scenography implies that only prophetic speech is convenient for the very world that the particular discourse is referring to. Discourse implies a given scenography (a speaker and an addressee, a place and a moment, a given use of language) through which a certain world is shaped, and that world must validate the scenography through which it is shaped. Scenography is both what discourse comes from and what discourse generates; it legitimises a text that, in return, must show that this scenography from which speech is proceeding is the pertinent scene for speaking of what it is speaking of. 
In self–constituting discourses scenographies must not be considered as mere rhetorical strategies, as is the case in a advertising campaign: they are consubstantial with ideological positions. When a preacher, through his discourse, shows himself as a prophetic figure, somebody who speaks directly, roughly, who denounces sinners and demands intense repentance, it defines implicitly what legitimate religious discourse has to be and, correlatively, the nature of illegitimate religious discourse: he is reaffirming his own act of positioning his identity inside the field. 
J.A.: Today's situation looks in many ways different from that of 30 years ago. Is discourse analysis on the way to becoming a new discipline, as you seem to say in your seminal Dictionnaire d'analyse du discourse (CHARAUDEAU & MAINGUENEAU, 2002), or does it remain a field for researchers from various disciplinary backgrounds? Moreover, where would you locate the French tradition in the European field of discourse analysis? In what way would you say that FOUCAULT represents this tradition? 
D.M.: In France like in any other country, discourse analysis—situated at the crossroad of human and social sciences—cannot appear at all as a homogeneous field of investigation. Various factors contribute to its diversification: diversity of reference disciplines (sociology, psychology, anthropology, history etc.), diversity of schools, diversity of data; one must also take into account the specific aspect of discursive activity under investigation (discourse production, discourse circulation, reception etc.), the applied or not applied character of the analysis … Such diversity may trigger local conflicts, but does not necessarily call into question the existence and the legitimacy of a "globalised" field of scientific and intellectual production. 
In the sixties the situation was quite different. The various trends of discourse analysis were taking place with fairly protected national frontiers, such as Great Britain (HALLIDAY), France (the "French School" with its two outstanding figures: ALTHUSSER and FOUCAULT), the USA (ethnomethodology, ethnography of communication). Although these trends spread out to many countries, they were deeply rooted in national intellectual traditions. Since the eighties, discourse analysis has been subject to a process of globalisation: nowadays more and more researchers throughout the world are exchanging ideas about discourse, and, thanks to modern information technologies, more and more quickly. Trends are defined more by international networks than by countries. 
The "French school", in the narrow sense of the expression, was marginalised from the seventies onwards by the development of pragmatic and "enunciative" trends but also by the return of Marxism and psychoanalysis in France. As a result, nowadays there are no longer any active groups of researchers working in the field outlined by the French school of PÊCHEUX. Discourse analysis is less "charismatic", and theoretical controversies are much less severe than before. It could be said that the way of researching is more professional. Yet, it cannot be denied that there is still, in much research done in France, an identifiable family likeness, some specific "French tendencies". We cannot speak of a "school", only of some features (which are not shared by all the members of the community) that are visible in many articles or books. In particular:
The interest in "constrained" corpora that are non-conversational. These "constrained" corpora are not necessarily written, but they are bound to an institutional frame, they are controlled by the speakers and are associated with the memory of other texts.
A preoccupation for "linguistic materiality": the emphasis is put on the study of linguistic forms (morphology, syntax, enunciation phenomena), and not directly on their social function. But, of course, if such-and-such linguistic phenomenon is studied, it is because of explicit assumptions based at once on the specificity of the corpus and on the properties of the linguistic units taken into account.
A close relationship with the linguistic theories of "enunciation": this European trend of reflecting on language (Charles BALLY, Emile BENVENISTE, Roman JAKOBSON, Antoine CULIOLI, Oswald DUCROT and others) focuses on the activity of enunciation, which cannot be reduced to the act of production considered independently of language. Due to the essential reflexivity of language, the enunciation is the reference point of the utterance, of which it bears many traces: person, time, place marks, modality, determination …
The primacy of interdiscourse: this "dialogical" principle rejects any approach that would consider discursive identities to be closed domains. The identity of a discourse is not given; it is a constant process of preservation of a determined configuration. Discourse is always criss-crossed by the manifold forms of presence of other discourses, virtual or effective. Meaning is not a mere projection of communicative intention; it is on the contrary a negotiation inside a radically conflicting space.
A certain conception of subjectivity: the subject of discourse is not the subject as it is ordinarily understood by sociologists or psychologists. The subjectivity implied by discourse questions the very distinction between these two forms of subjectivity. Such a problematic interferes with that of interdiscourse: "heterogeneity" (AUTHIER-REVUZ, 1995) and "polyphony" (DUCROT, 1984), from two distinct viewpoints, systematise ideas developed by M. BAKHTIN.
Discourse analytic research in France is easily associated with some "philosophical" reflection on language or discourse: theoretical backgrounds often play a key role in determining corpora and methods, or at least in legitimising them. The style of research in not empiricist: "facts", as a rule, are considered to be the product of a construction and the emphasis is put on the conceptual coherence of the investigation. 
3. Discourse Analysis as an Interdisciplinary Field in an International Context
J.A.: Where do you see the contribution of the social sciences in French discourse analysis? What analytical methods does linguistics provide for discourse analysis? 
D.M.: Concerning the social sciences, I think they have a crucial role to play in discourse analysis. In the U.S., the research on conversation was initiated by sociologists in ethnomethodology and one of the tendencies which have strongly influenced discourse analysis, the ethnology of communication, was clearly anchored in anthropology. In France, discourse analysis did not emerge in the departments of literature, but in the human and social sciences. PÊCHEUX, for example, was attached to a laboratory of social psychology. Interdisciplinarity was the norm: the French school at the time was situated between Marxism and structuralism, neither one locked in a discipline. 
But the conception of interdisciplinarity has seen a change which has strongly affected its meaning and practices. "Discourse" has from the beginning been thought as an interface between different disciplines. Today, we tend to grant primacy to the idea of interdisciplinarity: having several research teams work together in a bigger framework does not just mean putting distinct disciplines together, but must lead a joint production of knowledge. Even categories like "text", e.g., which one may consider reserved to linguists are not their property; a considerable part of the theory of textuality has a cognitive or a social dimension. This is not without consequences for the entire disciplinary spectrum. The historian or the sociologist, working on "discourse" or interactions, is bound to change their relationship to their discipline. It will still take years for this process to come to its end. 
In France as well as in many other countries, the social and human sciences are under the influence of a "linguistic turn", which might be better termed a "discursive turn", but this does not mean that a great number of sociologists, historians or psychologists engage in a methodical study of statements (énoncés) in a discursive theoretical framework. It seems that we can get out of this ambiguous situation only by the means of teamwork, which combines diverse disciplines. 
J.A.: Where do you see the major changes going on in the field of discourse analysis? Would you say that discourse analysis in France continues a tradition that goes back to FOUCAULT's and PÊCHEUX's pioneering works? 
D.M.: It is very difficult to respond to this question. Before speaking about FOUCAULT and PÊCHEUX, I will point out three major evolutions which characterise the current situation in discourse analysis: a) the growing interest in the material support, b) in large corpuses and computer-aided processing of data, c) in interaction and action.
We have been increasingly accustomed to the fact that the transmission of the text does not follow its production that the way it institutes itself materially is inextricably bound up with its meaning. Two different approaches which are especially interested in this question come to my mind: a) the socio-cognitive research by Jack GOODY (1977), for who writing opens a new regime of thinking. These investigations converge with those of historians of the book, in France, Roger CHARTIER (1987) for instance; b) the work on mass communication between audiovisual media and information technology, which is backed up by "mediology" (DEBRAY, 1991).
In France, discourse analysis has always been in a symbiotic relationship with information technology. Today, discourse analysts dispose of a whole array of programs in order to scan corpuses, test hypotheses or construct new ones. The constant growth of computer performance and storage capacity, the existence of enormous data bases of all sorts (data, bibliographies, software) on the Internet considerably changes the way research is done. Those working on archives can claim exhaustiveness. Thus, notions like "representative sample" get profoundly transformed, and discourse analysis participates in the remarkable development of "corpus linguistics", which investigates enormous masses of data.
Important shifts have taken place—in the methodologies, the objects as well as in theories. There is a particular interest in the activity of subjects in interaction, in their active cooperation in the construction of interactions and dialogues, in their co-action. It is true that Francophone discourse analysis, preferring institutional corpuses, has long remained skeptical with respect to interactionist corpuses. But social activity rests on the mix of written genres and oral interactions and we see today symbolic practices multiply which can no longer be classified with the terms oral and written. In particular, I think of all the forms of electronic writing. 
The discovery of the interactionist dimension is closely linked to the praxeological turn in the social and human sciences. Certainly, this question, coming out of the philosophical tradition, and then renewed by linguistic pragmatics and HABERMAS's philosophy, is nothing new in linguistics where the praxeological dimension of language use has been intensively discussed. The very notion of discursive genre presupposes a shift of the text toward the activity of which the text is the trace. 
Concerning the relation of today's discourse analysis with the works of FOUCAULT and PÊCHEUX, it seems clear that the current evolution has vindicated FOUCAULT rather than PÊCHEUX. This having said, the "analytical" tendency represented by PÊCHEUX is far from obsolete if you do not reduce it to the LACANo-ALTHUSSERian conjuncture with which it is closely associated. The points PÊCHEUX made are characteristic of certain ways of doing discourse analysis today. I will insist on two points:
The focus on the critical dimension, seeking to reveal political and philosophical presuppositions, turns up in the strands of Critical Discourse Analysis and in a number of works in the field of Cultural Studies.
The recourse to methodologies calling into question the continuity of texts. Thus, the postulate that a text is constructed on a constitutive unsaid which it masks and reveals at the same time, is not going to disappear. We can even say that this is a constant preoccupation in discourse analysis even though it remains a minority position. 
Thus, we may say—both in a simplifying and paradoxical way—that discourse analysis, as a full-fledged disciplinary field within the social and human sciences, would not exist if the study of the discursive and textual workings, based on a reflection on discursive genres, was not dominant. But discourse analysis would not exist either if critical, even paraphilosophical approaches were not possible. This is probably a constitutive tension in the ongoing trends in discourse analysis. 
Authier-Revuz, Jacqueline (1995). Ces Mots qui ne vont pas de soi. Boucles réflexives et non–coïncidences du dire. Paris: Larousse.
Charaudeau, Patrick & Maingueneau, Dominique (2002). Dictionnaire d'analyse du discours. Paris: Seuil.
Chartier, Roger (1987). Lectures et lecteurs dans la France d'Ancien Régime. Paris: Seuil. [English translation: The cultural uses of print in early modern France. Princeton: Princeton University Press]
Cressot, Marcel (1947). Le style et ses techniques. Précis d'analyse stylistique. Paris: PUF.
Debray, Régis (1991). Cours de médiologie générale. Paris: Gallimard.
Dubois Jean & Sumpf Joseph (Eds.) (1969). L'Analyse du discours. Langages, 13.
Ducrot, Oswald (1984). Le dire et le dit. Paris: Minuit.
Foucault, Michel (1969). L'archéologie du savoir. Paris: Gallimard. [English translation: The archaeology of knowledge. London: Routledge, 1989]
Foucault, Michel (1971). L'Ordre du discours. Paris: Gallimard. [English translation: The archaeology of knowledge and the discourse on language. New York: Pantheon, 1972, pp.215-37]
Goody, Jack (1977). The domestication of the savage. Cambridge, London: Cambridge University Press.
Maingueneau, Dominique (1976). Initiation aux méthodes de l'analyse de discours. Problèmes et perspectives. Paris: Hachette.
Maingueneau, Dominique (1984). Genèses du discours. Liège-Bruxelles: Mardaga.
Maingueneau, Dominique (1999). Analysing self-constituting discourses. Discourse Studies, 1(2), 175-199.
Marouzeau, Jules (1941). Précis de stylistique française. Paris: Masson.
Pêcheux, Michel (1969). Analyse automatique du discours. Paris: Dunod.
Dominique MAINGUENEAU is professor of linguistics at the University of Paris-XII (Créteil). His research associates a pragmatic and "enunciative" outlook on discourse and Michel FOUCAULT's Archaeology. He is working out a theory of "self-constituting discourses".
Tel.: 331 451 71 127
Johannes ANGERMÜLLER is Assistant Professor (Wissenschaftlicher Assistent) at the Department of Sociology at the University of Magdeburg. His fields of research include theories and methods of discourse analysis, sociology of intellectuals and science, political and cultural theory.
Institut für Soziologie
PSF 4120, 39016 Magdeburg, Germany
Tel.: (00)49 160 97623314
Maingueneau, Dominique & Angermüller, Johannes (2007). Discourse Analysis in France. A Conversation [48 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 8(2), Art. 21, http://nbn-resolving.de/urn:nbn:de:0114-fqs0702218.