[…] The analysis of « self-constituting discourses » is not a well stabilized territory, but it gives the possibility of asking exciting questions about some important types of discourse. For quite some time I have been studying a wide range of texts, particularly religious, scientific, literary and philosophical texts, and I noticed that, if we disregard superficial differences, many descriptive concepts could be transferred easily from one to the others. So I came naturally to the assumption that in the discursive production of a society a specific area  could be delimited : that of « self-constituting discourses ». Of course, religious discourse or literature, for example, have been steadily studied for centuries by hundreds of thousands of people, but bringing those discourses together in a new discursive unit may open an interesting research program.

To clarify the notion of « self-constituting discourses », one can start from a commonplace observation. Philosophers or scientists are not supposed to appeal to the authority of journalists when they deal with philosophy or science ; on the contrary, when a debate is organized in the mass-media about important problems, particularly ethical problems, journalists request the intervention of priests, scientists, philosophers or writers. These people are perceived as not delivering just any message, but one authorized by their privileged acquaintance with « ultimate » discourses : discourses upon which others are based – that have a particular relationship with the foundations of society and with the signification of human destiny. This asymetry proceeds from the particular status of « self-constituting discourses » : discourse like others, they are also discourses which claim to be above any other type of discourse. Discourses bordering on unspeakable meanings, they must negotiate the paradoxes that such a status implies. To found other discourses without being founded by them, they must set themselves up as intimately bound with a legitimizing 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. Analysts have no access to the world beyond limits of speech, but they can analyse the textual operations by which self-constituting discourses manage their self-foundation. Self-constituting discourses take charge of what could be called the archeion of discursive production in a given society. This Greek word, the origin of Latin archivum, has, from our viewpoint, an interesting polysemy : derived from arché (« source », « principle », « order », « power »…), the archeion is the center where authority sits, a group of magistrates, and public archives too. So, this notion of archeion binds tightly founding operations in and by discourse, the determination of a place for legitimate speakers and addressees, and the management of memory.

 

[…] Referring particularly to M. Foucault (1969), I think that actually we must focus on an approach to self-constituting discourses that connects their textual properties with their status in interdiscourse, with the institutions that legitimate their speakers, with the way they circulate or they handle other discourses, the way they are transformed, stocked, and so forth.

For centuries those discourses, at least certain genres belonging to these discourses, have been intensively analysed and commented, because they were prestigious. Everyday genres, on the contrary, were ignored, because they were considered unstructured or devoid of interesting content. Nowadays the situation is quite different : in the whole world « ordinary » uses of language are privileged data for scientific investigation : for the first time in history any  discursive production may be worth being integrated in a scientific research. But that revaluation of ordinary interactions must not institute them as privileged forms of discourse, to which all forms of discourse ought to be referred. Of course, discourse is by nature interactive, but interactivity cannot be reduced to immediate, face-to-face interaction : the basic interactivity of language is distributed in a large variety of types of communication. We are always tempted to consider that a certain type of discourse is more important than others because it is supposed to give access to the ideological kernel of a society. This privilege has been attributed variously, to literature, to everyday conversation, to political discourse, or to mass media discourse. I think that we had better renounce the belief that there is such a privileged discourse area, a center that could give access to the hard core of ideological processes : what matters actually is the multitude of connections between all discursive areas, the diversity of discourse.

By nature, self-constituting discourses claim to found others and not to be founded. This property permits a definition of their status in interdiscourse, but it does not correspond necessarily to the personal convictions of their speakers. Such discourses are at once  self-  and heteroconstituents, two inseparable aspects : only a discourse that constitutes itself can found others. Therefore the name « self-constituting discourses » must not be misconstrued : ideally, we should write « (self-) constituting discourses », with « self » put in brackets. That does not mean that other forms of discourse do not act on them ; on the contrary, self-constituting discourses and other areas are always interacting, but in the case of self-constituting discourses, that interaction is ruled by specific principles, which are different for each type of discourse.

It’s not easy to establish a list of self-constituting discourses, even if we only consider our type of society, for such discourses are defined by their status in interdiscourse, not by clear-cut frontiers in social activities. Each type of society has its own self-constituting discourses or its specific way of connecting them : speaking « anachronistically », one could say that mythical discourse in traditional societies is both « philosophical », « scientifical » and « religious ». Experimental psychology is self-constituting in that it belongs to scientific discourse, but the status of psychoanalytic discourse is more fuzzy : at least in continental Europe, it behaves like a specific self-constituting discourse, similar to philosophy or religion. Unlike psychoanalytic discourse, political discourse does not seem to be a self-constituting discourse ; there is no denying that it is a very powerful idelogical frame, but it leans on scientific, religious, literary and other discourses ; it is not really autonomous. Technocratic discourse, for example, leans mainly on economics, fundamentalist discourse on religion, and other discourses. I would rather say that  political discourse is a privileged contact zone between self-constituting discourses and other areas.

In modern societies, as was already the case in classical Greece, various self-constituting discourses exist at the same time, thus competing with each other. This variety is irreductible : self-constituting discourses’ life is made of it. During many centuries philosophical discourse claimed to be prevalent : it attributed to itself the privilege of assigning boundaries to the others. Theological discourse also did so, and so did scientific discourse later…The common sense belief is that each self-constituting discourse is autonomous and has contingent relations with others ; actually their relation to others is a part of their core identity ; they must manage that impossible coexistence and the way they manage it is their very identity.

 

The two dimensions of « constitution »

 

We have still not justified the use of the expression « self-constituting discourses ».  The reason is that  « constituting » connects two interesting values :

 

• « Constituting » as action of establishing legally, of giving legal form to some juridical entity : self-constituting discourses emerge by instituting  themselves as legitimated to utter as they utter.
• « Constituting » as forming a whole, an organization : self-constituting discourses are sets of texts whose structuration modes must be legitimized by discourse itself. 

Those two values converge to the derivative name « constitution » as a body of fundamental principles according to which a State is governed, and people receive rights and duties : self-constituting discourses produce texts dedicated to embodying norms, to guaranteeing behaviors of a collectivity, to drawing the frontiers of good and evil, false and true, etc.

 

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, where groups of men and text genres are articulated. Text production delimits a space inside social space but configurates that space too. Therefore, the analysis of self-constituting discourses implies to keep a « rhetorical » conception at a distance: in fact, content and textual organization are not independent of the discursive scene which takes charge of them and through which they appear.

 

 

Position and discursive community

 

It has been said above that self-constituting discourses interact with each other and demarcate each other. So, they are not compact blocks, but form discursive fields (Maingueneau 1984) in which various positions compete : in modern societies ideological frames are steadily discussed and « discursive fields » are the space where the diversity of those « positions » is structured.

Positions are the product of an act of « positioning », in the commercial meaning of the word. The content of this notion of « position » (doctrine, school, party…) is very poor ; it only implies that no position can ocupy the whole space of a given self-constituting discourse, that the identity of each position emerges and is kept up through the interaction, often conflicting, with the others. That is a motto of many discourse analysis works : the pertinent object is not discourse in itself, but the system of relations with other discourses. Of course, most of producers of such discourses claim that their message proceeds directly from a true apprehension of God, Man, Science, Beauty, Reality, Reason, etc., but in order to understand how such discourses really work, we must refer positions to the place they hold in their field.

A position is not only a more or less systematic set of ideas, it associates a certain textual configuration and a certain way of life for a group of people. The various philosophical schools in ancient Greece, the various groups in contemporary social sciences, etc. are not only theoretical positions, they imply the existence of groups, of discursive communities (Maingueneau 1987) which may be organized in many ways. Inventing a new way of having dealings with other people and producing new discourses are two dimensions of the same phenomenon. Those communities are structured by the discourses they produce and put into circulation. So, discursive communities are paradoxically united by the texts they produce : the texts are both their product and the condition of their existence.

The key role given to such communities converges with many studies, particularly about scientific discourse (Hagstrom, 1965 ; Swales, 1990). The way people make science, practise philosophy or religion, lead an artist life, etc. is inseparable from the way they produce discourse (De Certeau, 1975 ; Debray, 1983 ; Maingueneau, 1984 and 1993). But that principle must be diversified according to the type of discourse taken into consideration : communities do not behave in the same way when they belong to scientific field or to theological field ; moreover, theology does not imply the same type of community as devotion, though both produce religious texts.

It may happen that some positions refuse any form of group. For instance, many writers claim to live in the desert ; pyrrhonian sceptics in ancient Greece refused to belong to a school, as to do so was incompatible with the principles of their scepticism. But, they cannot escape from philosophical or literary fields for all that if they produce discourse : their loneliness is integrated in their work as a condition of its textual identity.

The questions of discursive communites of self-constituting discourses is tightly bound with that of mediation.When we work on texts belonging to self-constituting discourses, we deal with highly structured discourses that speak of man, society, rationality, beauty, good and evil, etc., that have a large scope, global aims. But those discourses are produced locally, by few people set in a small sector of society. Psychoanalytic discourse, for instance, is diffused in the mass media, in many therapeutic practices, etc., but it is shaped in very limited circles belonging to a specific field.

 

The hierarchy of genres

 

Self-constituting discourses are not a genre ; they form a network of genres, which are situated in a hierarchy. It is a basic property of self-constituting discourses that some texts or genres are considered to be more prestigious because they are supposed to be nearer of their Source.

More exactly, two hierarchies must be distinguished. The first one opposes « archetexts » and « ordinary texts » ; the other one hierarchizes genres. These two hierarchies don’t coincide with each other, but they are tightly associated.

Archetexts are singular texts that are reputed to have a privileged relation to the « archeion ». Plato’s Dialogs  or Descartes’s Meditations for philosophical discourse, Newton’s Principia for physics, The Gospel for Christians, Homer’s Odyssey for literature, etc. embody, in a sense, the basic values of the discourse they partake of. But the notion of « archetext » varies according to self-constituting discourses. Religious discourse is organized around prior archetexts, whereas scientific discourse only considers archetexts  as exemplary ; in literature archetexts are « chefs-d’œuvre »…In religious discourse archetexts embody the Source of truth ; in scientific discourses such as physics or chemestry archetexts exemplify the norms of scientific activity ; but in psychoanalysis archetexts, in some respects, look a little like religious archetexts.

Although it seems contrary to common sense, it must be admitted that these archetexts, that must be commented, and the texts that comment them presuppose each other. Archetexts, when taking place in what could be called an « hermeneutic frame », receive a pragmatic status that turns them into texts worthy of interpretation and attributes to them a certain way of circulating in interdiscourse. Sometimes, the very form of the archetext depends on this status : the Gospel was not written and later commented, it was shaped through the commentaries that turned to it.

The meaning of a text that happens to be in an hermeneutic frame is supposed to exceed the abilities of its interpreters. If interpreters fail to understand it, it is not because the text is deficient, but because interprets are deficient. This failure is the consequence of their « hyperprotected » pragmatic status (Pratt, 1977). An hyperprotected text may transgress conversational maxims without losing authority. It is a « monument » (in Latin « monere » means « to remind »), whose memory lives and whose interpreters die. Rather than obscure or ambiguous, it is enigmatic. If a text is no longer interpreted, it stops being enigmatic, therefore worthy of interpretation : the more interpretations it gives rise to, the more enigmatic it appears.

Interpretation cannot be considered independently of the legitimization of interpreters. Who is authorized to interpret ? That depends at once on the discourses and on the positions in the discursive field. On discourses, because the interpretation of scientific texts, for instance, is reserved to specialists, whereas the interpretation of literary texts is claimed by scholars and by lovers of literature. That depends on positions too, because each position defines, explicitly or implicitly, who is authorized to read and interpret : some philosophers claim to write only for philosophers, others claim to write for ordinary people.

Two sorts of archetexts can be distinguished. Those that are general, which are acknowledged as archetexts by all people, and those that are only local archetexts, acknowledged by a part of the people of a given discursive field. In fact, this distinction is not clear-cut  ; for instance, « general » archetexts may be given variable values : from the viewpoint of moslems, the Bible is indirectly an archetext, as announcing the Koran ; for all philosophers Hume, Kant or Husserl’s greatests works are general archetexts, but all philosophical positions don’t attribute the same value to each of them. The definition of archetexts is always controversial : each position has its own archetexts, its proper textual pantheon, setting its own identity by modifying prevailing hierarchies. Saussure’s Cours de linguistique générale, for instance, was an archetext for most of structuralist linguists, but not for generativist linguists, who consider Chomsky’s book Syntactic structures as an archetext, the foundation of a new age of linguistics. Surrealist writers in the twenties showed that they were practicing a new literature by changing archetexts : classical works were despised and others, which had been rejected, for example Lautreamont’s or Sade’s works, became archetexts.

There may exist hierarchies of archetexts, guaranteed by institutions : handbooks of literature oppose « great » writers to others, the Catholic Church has drawn up a sophisticated scale of textual authorities.

As a rule, archetexts are considered as founders. But that notion is ambiguous : some texts claim to found a new way of speaking of God, for instance, or practising psychoanalysis or literature, others do not. But it often happens that posterity considers that texts that did not claim to be founders were actually founders, and, on the contrary, that texts that claimed to be founders did not introduce anything new. These two viewpoints are distinct and complementary.

 

The second hierarchy is more traditional : it opposes top and secondary genres. This distinction founds the possibility of popularization : on one hand are texts that are supposed to be dominated by no other text ; on the other hand one finds texts that clarify, simplify or diffuse doctrines already established. Thus the top genres of theology or basic science, for instance, are disseminated through secondary genres, such as predication in churches or TV programmes about medecine, diets, beauty. Between these boundaries (top genres and mass TV programmes) various genres may take place, such as handbooks for teachers or students and magazines dedicated to popularize science.

Those two hierarchies (archetexts vs ordinary texts, top vs secondary texts) have complex relations. Archetexts are singular texts, but they may belong to manifold genres  such as autobiography, prophecy, or commentary. A theological commentary of the Gospel addressed to theologians is a top text, because the genre it belongs to is at the top of genres hierarchy, but it is an ordinary text in the hierarchy that distinguishes archetexts and ordinary texts. The analysis of self-constituting discourses analysis must not focus only on archetexts and top genres of texts, it must take as its object the interaction of genres, from the top to the base : self-constituting discourses are basically heterogeneous and that heterogeneity must be the center of analysis. Genres and texts of high theology or of great literature are always accompanied by other genres, less prestigious, which are also necessary to the « archeion » of a society.

Generally, top genres texts are closed. By « closed » genres, as opposed to open  genres, I mean genres whose addressees are or may be producers of discourses of the same genre. On the contrary, in « open » genres the addressees, who are regularly much more numerous than the producers, belong to another sphere. The first case can be exemplified by texts of hard core science : only potential or actual writers of articles of neurobiology are expected to read neurobiology journals. The second case can be exemplified by mass-distribution newspapers, whose diffusion is very large and whose audience is not made up of journalists.

 The opposition closed / open deals with genres rules, not with reality : if millions of people happen to read a scientific article, that occurence is independent of genre.

But this distinction cannot be used too strictly :

-         Some discourses, such as physics or mathematics, by their own nature are destined to be closed. Others are in a much more unstable situation : history and sociology, for instance, even if they are directed to small communities of scholars, are regularly read by many other people.

-          In philosophy, for example, some authors choose to write open texts, addressed to a large audience (Hume’s Enquiry Concerning Human Understanding (1748), Descartes’ Discourse on Method…). The way they expose their doctrine and publish their book shows clearly their intention, which is an aspect of their philosophy. Others claim to write closed texts but actually are read by a large audience. These two circumstances must not be confounded. In the case of Descartes’ work, he himself makes a distinction between his intellectual autobiography, the Discourse on Method, written in French for ordinary people, and his much more abstract Meditations, written in Latin for specialists.

 

Inscription and medium

 

By nature, self-constituting discourses have to do with « authority » : their authors claim to deal with important matters, they speak according to strict and transcendent principles. In a word, they do not speak for themselves, they are on the track of some Other. Writing a scientific article, giving a sermon or writing a poem is more than expressing personal ideas, it is taking part in a rite, speaking under the control of something transcendent, such as Tradition, Reason, Truth. Many of those texts belong to a prestigious corpus, a textual heritage, which must be preserved from corruption, stored, copied, updated....So, rather than simple texts they are inscriptions. An « inscription » may be oral, written or recorded, depending on the society. In traditional societies oral « literature » or proverbs are inscriptions ; nowadays many inscriptions are recorded.

Inscriptions are situated in a series : resting on words already said, they are, in their turn, virtual supports for other inscriptions. The genres of self-constituting discourses genres determine precisely the ways texts may take their place in interdiscourse : scientific references are not religious quotations.

This notion of « inscription » implies further consideration of the « mediological » dimension of utterances (Debray 1991), ie the modalities of their material existence, which governs the way they circulate. Self-constituting discourses, which are ideologically dense, are highly bound with communication facilities, which is one more reason for giving a key role to genres, considered as socio-discursive devices. On that count, discourse analysis must not work with these traditional oppositions (action / representation, text / context, production / reception…) which pragmatic trends have for a long time criticized. Meaning is not locked up in a text like in a box, it implies a set of communicational conditions. The « content » of a text cannot be apprehended independently of these conditions : mediological conditions of a discourse are a dimension of its identity. It is well known that philosophical discourse could not appear in a society devoid of literacy, or that the textual organization of epic depends on the material conditions of oral performances, etc. Modifications in the way scientific texts circulate transform scientific activity. Discursive activity ties a manner of saying and a manner of transporting texts, both associated with the way specific communities live.

 

 

Scenography

 

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 (Maingueneau, 1998). To each genre of discourse is associated a « generic scene », which attributes parts to actors, prescribes the place and the moment, the medium, textual macro-structures - all conditions necessary to felicity (Austin) of a given macrospeech act. But for many genres of self-constituting discourses another type of scene  is implied : « scenography », which proceeds from the choice of discourse producer.

Roughly speaking, 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 (scenography) depends on particular discourses. Not all texts turn to scenography. As a rule, administrative genres, for instance, merely obey to the norms of their generic scenes. On the contrary, adverts have to choose scenographies according to marketing strategies : adverts for cars can use a very wide range of scenographies : a woman in her bedroom phoning a girlfriend, a engineer explaining the technical advantages of a motor, a man describing his new car to a colleague, etc.

Here « scenography » is not used in its usual way :

-         It adds to the theatrical dimension of « scene » the dimension of « graphy », of legitimizing inscription, for  scenography gives authority to discourse, it has persuasive effects on addresses.

-         Scenography is not a frame, a scenery, as if discourse occured inside of a place that is already fixed, independently 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. This paradoxical movement is the very movement of discourse. Scenography is both what discourse comes from and what discourse generates ; it legitimizes 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 a scenography are associated a certain representation of the speaker responsible for that discourse, a certain representation of the addressee, of the place (topography) and of the moment (chronography) of discourse. Those elements are tightly bound. For instance, in a contemporary religious sect a preacher may show through his speech that he is a prophet in a desert  addressing to ancient Hebrews ; the biblical period and the desert, as a place far from urban corruption, a place for fast and repentance, are essential elements of that scenography, which, by mixing two discursive scenes (the empirical one and the mythical one) gives authority to discourse.

Scenographies are selected according to ideological options of positions : Protestants give more authority to biblical scenes than do Catholics. Scenographies may be singular communicative events (for example, the Sermon on the Mount) or prototypical discourse genres (friendly conversation, sermon, lecturing, etc.).

So, in self-constituting discourses scenographies must not be considered as mere rhetorical strategies, as is the case in advertising campaign : they are consubstantial with ideological positions. When a preacher, through his discourse, shows himself as a prophetical 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.

 In Descartes’ Discourse on Method (1637) also, philosophical contents and scenography are two aspects of meaning : the scenography is not a contingent strategy, it modifies the very status of philosophy. The figure of a subject presented as simply capable of reasoning, outside of any institution, and telling his life ordinary people shows in a performative way the superiority of the cartesian « method ». The Discourse  legitimizes itself by giving to common sense (« bon sens »), independently of tradition, the part of ultimate referee, by going beyond the boundaries usually attributed to philosophical speech :

 

"ceux qui ne se servent que de leur raison naturelle toute pure jugeront mieux de mes opinions que ceux qui ne croient qu’aux livres anciens." (1988 : I, 649)

those who make use of their unprejudiced natural reason will be better judges of my opinions than those who give heed to the writings of the ancients.

 

Scenographies can be set up only in monological discourses ; if there is direct  interaction or debate, the speaker cannot really develop discursive scenes of his own : he must only assume the rules of the genre in which he is involved. If Plato were discussing with another philosopher he could not express his thought through his own scenography as he does in his Dialogs.

 

Linguistic code

 

Self-constituting discourses, mainly in top texts and archetexts, cannot have a simple relationship with linguistic diversity. When you deal with the absolute, when you  are close to the frontier of the unspeakable, you cannot consider that all languages (English, Chinese, Latin, pidgins…) are equivalent. Discourses that have such a particular status are always raising questions about the identity of the language(s) in which they invest their energy.

Therefore, for a given position language is not a mere instrument that carries information, but has the status of a specific linguistic code, which is determined from  interlanguage. By « interlanguage » we mean the space of linguistic varieties to which one can have access from a certain place. Those varieties may be internal (registers, dialects…) or external (foreign languages). But the distinction internal/external is superficial : it cannot be said out of context whether a language is « foreign » or not. For instance, for cultured people in Europe during the last century, Latin was not really a « foreign » language.

We don’t use « linguistic code » like structuralist linguists, who used that term as an equivalent to « linguistic system ». « Linguistic code » in our meaning is at once a communicational notion (it permits transmission of sense by conventional associations between sounds and interpretations) and in a sense a juridical notion (discourse uses the resources of interlanguage that it must use, in consideration of the ideological world that its own position attempts to validate).

The Gospel was written in Greek, which was not Jesus’ mother tongue nor his teaching language. Using a common language such as Greek instead of Hebrew is not independent of the Gospel’s doctrine. It indicates a breaking with the Jewish religion, implies a different diffusion area, a different practice of texts. A mass in Latin and a mass in vernacular language are not the same mass. In the case of the Discourse on Method, choosing French instead of Latin has to do with Cartesian doctrine : writing in French, challenging traditional teaching and appealing to the common sense of readers who don’t belong to academic institutions are closely connected. It is clearly said in Descartes’ quotation given above, whose complete form is :

    "Et si j’écris en français, qui est la langue de mon pays, plutôt qu’en latin qui est celle de             mes précepteurs, c’est à cause que j’espère que ceux qui ne se servent que de leur             raison naturelle toute pure jugeront mieux de mes opinions que ceux qui ne croient             qu’aux livres anciens." (1988 : I, 649).

And if I write in French, which is the language of my country, in preference to Latin, which is that of my  preceptors, it is because I expect that those who make use of their unprejudiced natural reason will be better judges of my opinions than those who give heed to the writings of the ancients.

Besides, French in the seventeenth century is involved in the same ideological process as Descartes’ thought. It is the moment when French Academy was founded (1635) ; according to its ideology of clarity, French language, which fits to the supposed natural order of reason, has to be « purified ». Therefore Descartes’ theory of « clear and distinct ideas » invests a French language that in fact is already invested by a convergent  linguistic policy. Those two movements  support each other.

Now let’s consider scientific discourse, for instance this very review Discourse studies, or others of the same type. Writing in English (in fact, articles are not really published in « English », but in that particular « dialect », academic English, most of whose speakers have not English as their mother tongue) is not merely convenient, a means to facilitate the diffusion of ideas. Such a linguistic code implies various norms, not only about the way of quoting, of choosing references, of organizing texts, but also of considering scientific activity in that field, of dealing with language (a worldwide audience excludes connivance, play on words…). The very fact of publishing papers coming from any country and destinated to any country in the world implies the possibility of bringing together texts in the same space. We could speak of a pragmatic presupposition that a worldwide field exists : scientific productions are supposed to belong to the same universe, scientists supposed to belong to the same community, independently of their ethnic or cultural origin.

 

Ethos

 

   Texts are not made to be contemplated, they are traces of discourses which try to convince subjects, to make them adhere to their universe. That adhesion is not only intellectual, in a sense it is « physical » too. One cannot study prophetic, medical or popular speech and disregard the « tone », the style of voice, the gestures associated with such characters. Discourse influence proceeds partly from that « incarnation » of speakers. Here we meet the notion of ethos, which comes from Aristotle’s Rhetoric(Aristotle 1967 : 1377b). By « ethos » Aristotle means the representation of the speaker that the addressee constructs across the production of discourse : therefore it is a representation produced by discourse, it is not what the speaker says explicitly about himself, nor the representation of the speaker that the addressee may have independently of discourse. Using pragmatic terms, it could be said that  ethos is « implied » in discourse.

In traditional rhetoric ethos is reserved to oral discourse. I proposed (Maingueneau 1984, 1993) to  use it for written texts too : even written discourse is referred to a source, it implies a « voice » that has has a specific tone. Through that determination of voice the addressee can construct a moving representation of the body of the speaker  (not, of course,  a representation of the real producer of discourse).  Readers through reading process (from heterogeneous indices given by the text) shape a more or less definite figure of the speaker’s body that corresponds to such a text. Let us name guarantor such a figure. By a constitutive paradox, it is through his very text that the guarantor must legitimate his way of speaking.

I think that such an extension of the scope of ethos is fairly consistent with the aristotelician notion, which (like its Latin equivalent, « mores ») covers not only vocal dimension of discourse, but also the psychological and physical characteristics of the orator. In a written text, the entity that I have called the « guarantor » is not immediately perceptible, like an actor on stage, it must be constructed by readers, who assign to it a temper and what could be called a corporality. « Temper » is a set of more or less precise psychological features ; as for « corporality », it is not only a corporal complexion but also a way of dressing and behaving. Of course, temper and corporality are two aspects of the same personality ; they are not real properties of real persons but positive social stereotyps on which discourse rests and that, in its turn, it contributes to reinforce. Such stereotyps circulate in diversified registers of semiotic production ; nowadays they are mainly carried by cinema and advertising, but in other periods theater, painting and novels played a much more important part in this respect.

In one of the most famous piety books in the 17° century, Introduction à la vie dévote (« Introduction to devout life ») by François de Sales (1609), the tight relationship between doctrine, scenography and ethos can be clearly observed. The scenography of this book is quite simple : a priest, a spiritual adviser, speaks in a friendly way to a woman, named « Philothée », who belongs to the gentry, explaining to her how she can lead a devout life without renouncing to the common way of life of her social class. This doctrine, which refuses rigorous devoutness and particularly calvinist discipline, is in a sense embodied in the ethos of François de Sales’ discourse : the soft, sweet  tone of the speaker makes actual the sweetness of the religious attitude that the addressee is incited to adopt. In the beginning of the book, the author stages a representation of good devoutness’ body, which is opposed to that of bad devoutness, represented by the stereotype of a melancholic character :

 

"Le monde, ma chère Philothée, diffame tant qu’il peut la sainte dévotion, dépeignant les personnes dévotes avec un visage fâcheux, triste et chagrin, et publiant  que la dévotion donne des humeurs mélancoliques et insupportables."(1969 : 34)

The people, dear Philothée, defame as much as possible holy devoutness, depicting devout persons with an unpleasant, sad and peevish face, and publishing that devoutness gives melancholic and  insufferable humours.

 

This fragment evokes a sort of « anti-guarantor », who indirectly legitimates the guarantor of the discourse : the person who speaks so softly and in such a friendly manner is the very person who lacks a « peevish face » and « insufferable humours ». According to the traditional medical categorization of the seventeenth century, this melancholic temper is opposed to the sanguine temper of the speaker.

Through ethos, by a circular process, the addressee adheres progressively to the ideological universe presented by the discourse he is reading : the way the text describes good devoutness demands the very ethos of that description. Besides, the representation of nature, society, man and God given by the texts belonging to that religious trend is consistent with that ethos. In a sense, the reader, by his reading, is already in the world whose values the text is promoting.

As ethos rests on historically determined stereotypes, an important change in culture implies important changes in the valuation of ethic representations. For instance, when occidental literature passed from the classical to the romantic regime, other types of temper and corporality were valorized : being pale and lean, looking sick, speaking in a faint voice were no longer considered  negative characteristics. Poetic meditations (1820) by the French romantic poet Alphonse de Lamartine imply a murmuring voice, as if  the speaker were only addressing to himself. The addressee perceives the voice of a speaker at death’s door and the text evokes a pale and tired man, who is wandering about the countryside or sitting, weighed down, under a tree. Such a voice is opposed to the classical ideal of « honest » conversation. One of the reasons why such poems were so successfull throughout Europe is the harmony between a way of speaking and a way of behaving. Here the poet’s morbidity (actually that of the figure of the poet that those poems imply) is not an individual fact, it depends on a global transformation of culture and, particularly, of the modes of legitimization of literary discourse.

That question of ethos, the relationship between a manner of saying and a manner of being, crosses some preoccupations of Pierre Bourdieu’s sociology. Bourdieu states that the use of speech is also  « a body technique » and that « linguistic competence, and especially phonetic competence, is a dimension of corporal hexis in which the whole relation with social world is expressed ». According to him, « the articulatory style of popular classes », for example, is « inseparable from a whole relation with body dominated by the refuse of  affected  or overpolite people and the valorization of manhood » (Bourdieu, 1977 : 31). The body makes possible a participation in social meanings : « The body believes in what he is playing…it does not represent what he is playing, it does not memorize past, he acts past…What is learned by body is not something that one owns (…) but something that one is » (Bourdieu, 1980 : 123).

These tightly articulated concepts of scenography, linguistic code and ethos must contribute to a better understanding of the efficiency of discourse and of the part it plays in ideological processes. Discourse doesn’t only persuade by the ideas it delivers : it sets the addressee in a speech scene that partakes of the semantic characteristics of the ideological universe that discourse aims to promote. But such concepts are not reserved for self-constituting discourses ; they are actually pertinent far beyond self-constituting discourses, for many sorts of texts.

 

Conclusion 

[…]  Until now discourse analysis was mainly concerned with applied research or with data that traditional academic institutions had ignored. But it is highly important for discourse analysis to show that it can approach with specific means texts that for long have been studied separately and with traditional viewpoints.

As it has been said above, the essential property of self-constituting discourses is their status inside interdiscourse : whereas « ordinary » discourses lean against them, they must lean directly against transcendent principles, against what was named « archeion ». Those discourses belong to conflicting discursive fields ; they are basically hierarchical, according to their closeness to their Source ; they are produced in small communities, although they deal with basic problems of society as a whole ; they are defined by a specific way of circulating in interdiscourse (conservation, commentary, quotation, etc.) ; what they say (« ideas », « doctrine »…) cannot be separated from the discourse scene through which they are produced and that gives them authority.

For it is only by their way of organizing their own discourse that they can show and attest their legitimacy - their conformity with the criteria of Truth. In commonplace uses of discourse, we mainly follow routines, in the hard core of self-constituting discourses speakers cannot ignore basic questions about who is authorized to be a speaker or an addressee, where and when it may be spoken, how texts must be organized, etc. Of course, such questions are implicit in any utterance, as was demonstrated by pragmatic trends, but in the case of self-constituting discourses, people are deeply committed : the answer to such questions has root consequences for their identity, their destiny and those of other people.

Unfortunately the notion of self-constituting discourse is hard to handle. Actually we are accustomed to work with texts that are brought together because they have the same function and share obviously some linguistic properties. L.S.P. studies deal with such phenomena, which are very important for applied linguistics. But what self-constituting discourses share is not obvious, it does not appear superficially. To see what a novel, a scientific article, a sermon, etc. have got in common, you need to handle very abstract operations.

Studying self-constitution discourses raises a question that cannot be dodged : the relationship between analysis and its object. For self-constituting discourses analysis is a slave to a paradox : as a scientific analysis, it belongs to self-constituting discourses, but at the same time it claims to overhang any self-constituting discourse. That paradox cannot be resolved. If discourse analysis claims to dominate self-constituting discourses, it actually assumes the old dogmatic claim of philosophy or religion, attributing to themselves the power of ruling all human discourses. Discourse analysis, on the contrary, must accept participation in a generalized interaction with other scientific disciplines and with self-constituting discourses. Discourse analysis, for instance, may question philosophical discourse, which, from its own viewpoint, may also question, discourse analysis. And so on.

 

                              

 

REFERENCES

 

Aristotle (1967), Rhétorique. Paris : Les Belles Lettres

Bourdieu, P. (1977) «L’économie des échanges linguistiques » , Langue française 34 : 17-34

Bourdieu, P. (1980) Le sens pratique. Paris : Minuit

Certeau, M. de (1974) l’Ecriture de l’histoire. Paris : Gallimard

Culioli, A. (1990), Pour une linguistique de l’énonciation. Paris : Ophrys

Debray, R. (199  ) Cours de médiologie générale. Paris : Gallimard

Descartes, R. (1988) Œuvres philosophiques. Paris : Classiques Garnier

Foucault, M. (1969) l’Archéologie du savoir.  Paris : Gallimard

Hagstrom, O. (1965) The scientific community. New York : Basic Books

Jakobson, R. (1963) Essais de linguistique générale. Paris : Minuit

Maingueneau, D. (1984), Genèses du discours. Liège-Bruxelles : Mardaga

Maingueneau, D. (1987), Nouvelles tendances en analyse du discours. Paris : Hachette

Maingueneau, D. (1993), Le contexte de l’œuvre littéraire. Enonciation, écrivain, société. Paris : Dunod

Maingueneau, D. (1998)  Analyser les textes de communication. Paris : Dunod.

Pratt, M.-L., (1977) Towards a speech act theory of literary discourse. Bloomington : Indiana University Press

Roulet E. éd. (1985)  L’Articulation du discours en français contemporain. Berne, Frankfurt/Main : Peter Lang

Schaeffer J.-M. (1989 ), Qu’est-ce qu’un genre littéraire ?. Paris : Le Seuil

Sales, F. de (1969)  Œuvres. Paris : Gallimard

Swales, J. (1990). Genre analysis. English in Academic and Research Settings. Cambridge : Cambridge University Press

Watzlavick P, Helmick Beavin, J., Jackson, Don J. (1967)  Pragmatics of Human Communication. New-York : Norton