Semiotics and Language

AutorGreimas, Algirdas Julien, 1917-1992
PealkiriSemiotics and language : an analytical dictionary / A. J. Greimas, J. Courtés ; translated by Larry Crist ... [et al.]
IlmunudBloomington : Indiana University Press, c1982
ViideGreimas, Algirdas Julien 1982. Semiotics and Language: An Analytical Dictionary. Translated by J. Courtés. Bloomington: Indiana University Press.

In the course on the Schools of Semiotics [semiootika koolkonnad] we are taught by this - Greimas's - dictionary. I comprehend little to nothing in the Estonian translation and even the method of reading the text out loud did nothing to better my apprehension. Then I came across this raggedy-looking paberback in the library and right away decided that the best way to go about learning this jargon is to type in all the entries we are supposed to learn. Thus, I am reduced to mere mediator in learning this, and since other formatting is already taken within the text, I have marked the notions and passages I find to be relevant for my own work in underline. Nevertheless this is the most "semiotic" post in this blog and will undoubtedly direct or conduct my understanding of semiotics in general.

The term semiotics is used in different meanings, depending on whether it designates (A) any manifested entity under study; (B) an object of knowledge, as it appears during and after its description; and (C) the set of ways that make knowledge about this object possible.
(A) Semiotics as object
1. Clearly, the standard definition of semiotics as a "system of signs" is not appropriate for meaning (A), for it already presupposes a recognition of the signs. By using the expression "system of significations," we would introduce the broader concept of "signification," and, further, by replacing "system" - which is limiative, precise, and theoretic notion - by "set," we can proposed to define, at least temporarily, a given semiotics as a signifying set that we suspect, at least hypothetically, possesses an orgnaization, i.e., an autonomous internal articulation. It may also be said that any signifying set, as soon as one proposes submitting it to analysis, can be designated as an object-semiotics; this definition is tentative, since it is valid only within the framework of a descriptive project and therefore presupposes a metasemiotics that, theoretically, encompasses it. Furthermore, the concepts of signifying set and object-semiotics are not coextensive: the results of the analysis sometimes show that only one part of the signifying set is embraced by the constructed semiotic system, or, on the contrary, the latter sometimes accounts for more entities than those initially thought to be part of the signifying set (see semantic field).
2. These preliminary remarks, apparantly idle, assume their importance when we have to deal with the status of the so-called natural semiotic systems and with the pertinence of the dichotomy between what is "natural" and what is "constructed": moreover, such a problem involves semiotic theory as a whole.
By Natural semiotic systems we mean two vast signifyng sets: on the one hand, natural languages, and on the other, "extra-linguistic contexts" that we consider as semiotics of the natural world. They are called "natural" because they impose themselves upon human beings rather than being constructed by them - people are immersed in their mother tongue and are projected, from birth, into the world of "common sense." However, the boundary between what is given "naturally" and what is constructed, is blurry: literary discourse uses a given natural language, and logic finds its origin in natural langues; yet unquestionably, they are genuine cosntructions. The semiotics of space experiences the same difficulty in distinguishing between "built" space and "natural" space: a "natural" countryside is obviously a cultural concept and has meaning only with respect to space formed by humans. Contrary then to F. de Saussure and L. Hjemslev, for whom natura languages are semiotic systems among others, natural languages and the natural world appear to us as vast reservoirs of signs, as the place where numerous semiotic systems are manifested. Furthermore, the conept of construction must also be revised and reasserted from this point of view: inasmuch as construction implies the existence of a constructing subject, room must be made for collective subjects, alongside of individual subjects (ethno-literary or ethno-musical discourses, for example, are constructed discourses, whatever may be the status the genetic anthropology may attribute to subjects producing such discourses). Consequently, it would seem desirable to substitute for the opposition naturallconstructed (or "artificial") the opposition scientific semiotic systems/non-scientific semiotic systems; here, by scientific semiotic systems - in the broad sense of "scientific" - we understand an object-semiotics treated within the framework of a semiotic theory, explicit or implicit (the construction of a documentary language, for example, is built on a theory, even if the latter is only barely scientific).
3. It consequently becomes indispensable to define the status of these macro-semiotic systems, in which particular semiotic systems are organized. These macro-systems, i.e., natural languages and natural worlds, are "natural" only in the sense that "nature" is perceived through "culture," which makes them relative and permits the use of the plural. In the first place, it is necessary to record the correlations that exist between the two sets: thus, the affirmation that the natural world is translatable into natural language hs to be interpreted as the correspondence that can be established between units stemming from two types of semiotic systems (phemes of the natural world correspond, on the figurative plane, to semes in the natural languages; somatic activities - behavior - are "described" as linguistic processes, etc.). The result of this is a certain interpenetration of segments semming from the two semiotic systems, recognizable on the syntagmatic plane: linguistic deictics refer to the natural context, gestural segments replace verbal syntagms, etc. In the second place, the affirmation that natural languages are the only languages into which the other semiotics systems are translatable (which the reverse is impossible) can be explained two ways: first, because figures of the natural world are coded semantically in natural languages; secondly (and especially), because only natural languages can lexicalize and manifest abstract (or universal) semantic categories that generally remain implicit in other semiotic systems.
4. For us, the macro-semiotic systems - natural languages and natural worlds - are the loci where all other semiotic systems are manifested.
(B) Semiotic typology
1. While, in meaning (A), the term "semiotic" serves to designate a signifying set prior to its description, in a different acceptation it is used to refer to any object of knowledge that is being or is already constituted; we are then dealing with an object-semiotics considered either as a project of description, or as already having undergone analysis, or finally as a constructed object. In other works, we can speak of semiotics only if there is a point of encounter between the object-semiotics and the semiotic theory that apprehends it, gives form, and articulates it.
2. Following the tradition of L. Hjemslev, who was the first to propose a coherent semiotic theory, we can accpet the definition that he gives for semiotics: he considers it to be a hierarchy (i.e., a network of relations, hierarchically organized) endowed with a double mode of existence, paradigmatic and syntagmatic (and therefore which can be grasped as a semiotic system or process), and provided with at least two articulation planes - expression and content - the union of which constitutes semiosis. present research, focusing on analyses of discourse and of semiotic practices, seem to favor the syntagmatic axis and semiotic processes; but not at all modify Hjemslev's definition. One can envision a later phase of research being devoted to the systematization of established results.
3. To these common characteristics, let us try to add some other, more specific, features, in order to open the way for a typology of semiotic systems. At the present time, two types of classifications are implicitly or tacitly accepted: a distribution of semiotic systems, based on channels of communication, and another, based on the nature of the recognized signs. neither of them, however, correspond to our definition of semiotics. The classification according to the channels of sign transmission (or according to the orders of the senses) depends on our taking into consideration the expression substance: but the latter is not pertinent for a definition of semiotics (which is, first and foremost, a form). Moreover, a distribution according to the nature of the signs is based on relations that these signs (symbols, icons, indices, etc.) maintain with respect ot the referent. Such a criterion, because it infringes upon the principle of autonomy (or of immanence) of semiotic organizations, established by F. de Saussure, cannot be accpeted, for it also is not pertinent. In any case, one can wonder whether, given the present state of semiotic research, any such classification is not premature.
4. The typology of semiotic systems as proposed by L. Hjemslev in his Prolegomena is of a very different nature. In order to avoid confusion, we shall first present it concisely, adding our own remarks afterwards. This typology is based on two criteria of classification: (a) scientificness (a semiotics is called scientific when it is a description conforming to empirical principles), and (b) the number of (language) planes that constitute a given semiotics. It is, therefore, possible to differentiate monoplanar semiotic systems (or systems of symbols, in Hjemslev's terminology) that are also scientific or not, and pluriplanar semiotic systems which are biplanar semiotic systems of which at least one of the planes constitutes a semiotics (called an object-semiotics) - the case in which only one of the two planes is an object-semiotics is by far the most frequent. Pluriplanar semiotic systems are subdivided according to whether (a) they are scientific or not, and (b) their object-semiotics is scientific or not. The following schema represents this distribution:
  • pluriplanal semiotic system
    • connotative semiotic systems (non-scientific)
    • meta-semiotic systems (scientific)
      • scientific meta-semiotic systems (the object-semiotics of which is a scientific semiotics)
      • semiologies (the object-semiotics of which is non-scientific)
To this classification are added two other categories: a meta-semiology and a meta-semiotic of connotative semiotics, that examine, respectively, semiologies and connotative semiotics.
5. To understand, interpret, and evaluate such a typology, several remarks are necessary: (a) With respect to the classifications given above (in par. 3), that of Hjemslev is differentiated first by the introduction of the criterion of scientificness, that is, by the absolute necessity of having an explicit theory (when speaking of semiotics) that is to account for it; in addition, Hjemslev's classification differs from the previous ones, by the utilization, as a criterion, of the planes of language (signifier and signified taken as a whole), a criterion already part of the definition of semiotics and, therefore, homogeneous with it (whereas substance of referent introduce suplementary and heterogeneous variables). This forces us to consider the proposed typology as being part of a theory taken as a whole; it is possible to reject the entire theory, but not the classification itself; (b) In order to obtain a homologenous terminology, we should note that our definition of semiotics corresponds, in Hjemslev's typology, to the meta-semiotics called semiology: every signifying set treated by semiotics (semiotic theory) becomes a semiotic system; (c) Scientific meta-semiotic systems are linked to the issues concerning metalanguages, issues common to logic, mathematics, linguistics, and semiotic theory; (d) The creation of a separate category for mono-planar semiotic systems, which Hjemslev considers as systems of symbols not having the status of "semiotics," does pose a problem. The definition that he gives for mono-planar semiotic systems - they should be identifiable by the conformity of the two planes, by the isomorphism and their isotopism, by the one-to-one correspondence of their units - does not necessarily mean that they involve only a single plane of language, bu that they are manifested as a signifying form (in the Saussurian, and not Hjemslevian, sense). Moreover, a distinction could be established between such monoplanar semiotic systems, depending on the type of conformity recognized. Formal languages (or systems of symbols) would be, in this sense, "elementary." This is because each element, taken separately, is recognizable either on the expression plane, or on the content plane (in this latter case it is called "interpretable"), since the differentiation is based on nothing other than discrimination (which allows us to identify these languages on the expression plane alone). By contrast with formal languages, then, there are "molar" or semi-symbolic languages, characterized not by the conformity of isolated elements, but by the conformity to categories. Prosodic and gestural categories, for example, are signifying forms - "yes" and "no" correspond, in our cultural context, to the opposition verticality/horizontality (nodding vs. shaking the head) - as are the categories recognized in abstract painting or in certain musical forms. Thus, what is at stake in the distinction between interpretable monoplanar semiotic systems and those that are signifying is considerable; (e) The handling of connotative semiotic systems (linked with that of denotation), left outside the field of scientificness, is problematic. it is easy to see that the difficulty of a rigorous description of these languages of connotation resides in the fact that by proceeding from their expression plane, it is impossible to predict connotations (which have as signifier a peculiarity of pronounciation, or the choice of a lexeme, or a syntactic turn of phrase, etc.). It is even more difficult to propose a hierarchical distribution of these connotations (i.e., a connotative semiotic system). For these reasons, R. Barthes' Mythologies, however ingenious and refined they may be, are but connotative fragments and do not succeed in even suggesting an underlying system.
From this we conclude that a directly opposite approach to languages of connotation must be attempted, one that would begin by establishing a theory of connotation, from which a description of connotative systems could be undertaken, based on the content plane. We have barely outlined this theory, by dealing with social connotations that are presented in the form of connotative taxonomies ("profane" and "sacred" languages, "internal" and "external" languages, etc.). in ethno-semiotics, or in the form of connotative syntaxes (corresponding to a typology of discourses) in socio-semiotics. Research in this field has barely begun; in addition to social connotations, Hjemslev suggests tha there are also individual connotations (more or less corresponding to ancient and modern characterology), connotations of which we have scarcely a vague idea; (f) Present usage tends to differentiate between linguistic semiotic systems and non-linguistic semiotic systems, with reference to the two privileged loci of manifestation of semiotics that we call (perhaps improperly) macro-semiotic systems: natural languages and natural worlds. This can be done only by postulating a separate, specific status for macro-semiotic systems - contrary to Hjemslev for whom a natural language is one semiotic system like another (having, however, a privileged character). These macro-semiotic systems must be considered capable of containing and developing autonomous semiotic systems (for example, a number of recent analyses concerning religion and legal discourse justify this last point).
Immediately, however, another problem comes to mind, that of the transgression of the boundary that we have just established, in the form of syncretic semiotic systems, the existence of which is obvious at once. The expression plane of these systems is constituted by elements belonging to several heterogenous semiotic systems. If oepra and film are presented at the outset as preemptory examples of syncretic discourses, one might wonder whether natural language - and oral discourse in particular - constitute only one, although essential, element of communication, beside other paralinguistic or proxemic elements; in this case, communication itself is syncretic; (g) Other distinctions may also be proposed, taking into account the generative trajectory of discourse. Thus, we may oppose figurative and non-figurative (or abstract) discourse, and, at the same time, figurative and non-figurative semiotic systems, depending on the level of depth that is textualized and manifested. The difference between discourse and semiotic system lies in the fact that the former is nothing else than the apprehension of the latter as process.
All these distinctions and reorganizations, even if they sometimes are a source of confusion within the semiotic field, should be considered as a sign of health and vitality for a semiotics that proposes itself both as a project for research and as already on-going research.
(c) Semiotic theory
1. In definition (B), semiotics was conceived as the adequate superimposition of an object-semiotics and a descriptive language. Here we view it not only as (a) the locus for the elaboration of procedures, for the construction of models, and the choice of systems of representation, governing the descriptive level (i.e., the methodological metalinguistic level), but also as (b) a locus for the verification of the homogeneity and coherence of these procedures and models, as well as (c) the locus for the explicitation (in an axiomatic form) of undefinables and of the foundation of this entire theoretical construction (this is the epistemological level, strictly speaking). From this point of view, semiotics will be taken either as general semiotics, thus insisting on the requirement imposed upon it, that it account for the existence and functioning of all the particular semiotic systems, or as semiotic theory, inasmuch as it is called upon to satisfy conditions of scientificness proper to any theory, and inasmuch as it is defined, therefore, as a metalanguage (both as scientific metasemiotics and metasemiology, in Hjemslev's terminology).
2. In principle, several semiotic theories - just as several generative grammars - may be elaborated: only their formalization would possibly allow them to be compared and to be evaluated with respect to one another. Such a comparative approach is at the present time absolutely impossible, since no semiotic theory worthy of its name yet exists; on the other hand, we find intuitive theories having no operational procedures (often replaced by peremptory "professions of faith") and, on the other hand, procedures that are sometimes formalized, but which are not based on any explicit theory. This authorizes us to limit ourselves here to a brief account of what we consider to be the general conditions of a semiotic theory, while, at the same time, referring to our theoretical project.
3. Semiotic theory must be presented, first, for what it is, i.e., as a theory of signification. Its first concern, therefore, is to render explict the conditions for the apprehension and production of meaning; this is to be done in the form of a conceptual construction.Thus, by being situated in the Saussurian and Hjemslevian tradition (according to which signification is the creation and/or apprehension of "differences"), it will have to bring together all those concepts that, while being themselves undefinable, are necessary in order to establish the definition of the elementary structure of signification. This conceptual explication then gives rise to a formal expression of the concepts retianed by the theory. Considering structure as a network of relations, semiotic theory will have to formulate a semiotic axiomatics that will be presented essentially as a typology of relations (presuppositions, contradictions, etc.). This axiomatics will permit the constitution of a stock of formal definitions, such as, for example, semantic category (minimal unit) and semiotics itself (maximal unit). The latter includes, following Hjemslev, the logical definitions of system (the "either ... or" relation) and of process ("both ... and"), or content and expression, of form and substance, etc.
The next step consists in setting up a minimal formal language. The distinction between relations-as-states (for example, contradiction) and relations-as-operations (for example, negation) allows us to postulate symbol-terms and operator-terms, thus opening the way for the calculation of utterances. It is only then that the theory will have to be concerned with the choice - or free choice - of the representation systems in which it will have to formulate its procedures and models (for example, the semiotic square or elementary utterance).
These few remarks are meant to give only a general idea of the approach that appears to be necessary for the construction of a semiotic theory. The elements of our semiotic project are scattered throughout this work.
4. To these general features of a semiotic theory, we must also add other, more specific, options upon which the articulation of its total economy will, nevertheless, depend. The first of these options is the generative form which, we believe, is suitable for its unfolding. By that, we mean, in a very broad sense, seeking out the definition of the semiotic object, viewed from the angle of its mode of production. This approach, leading from the most simple to the most complex and from the most abstract to the most concrete, has the advantage of allowing us to introduce, at appropriate intervals, a certain number of established items from linguistic theory, such as issues relative to "language" (Benveniste) or "competence" (Chomsky), but also the articulation of structures into levels according to their modes of existence: virtual, actual, or realized. Thus, the semiotic generation of a discourse will be represented in the form of the mapping of a generative trajectory involving a good number of levels and components, distinctions that are perhaps only temporary and operationa, but that allow us to situate the different fields of exercise of semiotic activity with respect to one another.
5. The second of our options consists in introducing into semiotic theory the questions of enunciation; the putting of language into discourse (Benveniste) and of the specific, definable conditions that surround it, a problem treated, altogether differently, by American pragmatics. To the deep semiotic structures, located "in language" and in which "competence" finds its soure, we have added discourse structures that are less deep, in such a way that they are constructed by passing through the filter constituted by the phenomenon of enunciation.
Semiotic theory must be more than a theory of the utterance - as is the case for generative grammat - and more than a semiotics of enunciation; it must reconcile what appears at first to be irreconcilable, by integrating them into a general semiotic theory.

1. The term semiology is used, concurrently with that of semiotics, to designate the theory of language and its applications to different signifying sets. It goes back to F. de Saussure, who used this label when he called for constituting the general study of "systems of signs." As for the fields of knowledge (or wanting-to-know) that these two terms cover, they were first constituted in France in the 1960s, within the framework of what is called French structuralism (Merleau-Ponty, Lévi-Strauss, Dumézil, Lacan, etc.), influenced in the area of linguistics by the heirs of Saussure: L. Hjemslev and, to a lesser degree, R. Jakobson. Of the two terms (used interchangeably for a rather long time), "semiotics" was finally favored: hence, the name of the International Association for Semiotic Studies (in French, l'Assiciation Internationale de Sémiotique). Despite this instutitonalization, the term "semiology," solidly in place in France (among the disciplines of R. Barthes and, partially, those of A. Martinet) and in other Latin countries, continues to be widely used, and it was not until the 1970s that the methodological content of semiology and of semiotics was progressively differentiated, making the distinction between the two designations significant.
2. The semiological project was quickly reduced to almost nothing since it was developed within the limited framework of Saussure's definition (and outside of any contact with the epistemology of the human sciences of the period). The concept of "system," in this project, excludes semiotic process and, at the same time, the most diverse signifying practices; the study of "signs," inscribed in the theory of communication, consists in the almost mechanical application of the model of the "linguistic sign," etc. Semiology was thus reduced to the analysis of a few artificial, supplementary, codes (cf. the analysis of Prieto and Mounin), making semiology appear as a discipline appended to linguistics.
3. In order to discover the reasons for the decisive impact of F. de Saussure on the development of semiological studies, we should turn, not to the aforementioned narrow formulation, but to the theory of language (of which Saussure outlined the basic dimensions), considered in its entirety. Thus, it is in its Hjemslevian formulation (cf. R. Barthes' Elements of Semiology and A. J. Greimas' Sémantique structurale) that Saussurianism entered for good into French semiology. However, Hjemslev, while keeping Saussure's term, endows it with a precise definition: by semiology he means the scientific meta-semiotics, the object-semiotics of which is not scientific. Thus, he excludes from the domain of semiology, on the on ehand, connotative semiotic systems (i.e., languages of connotation) and, on the other hand, meta-semiotic systems that have scientific semiotic systems (logical languages, for example) as their object-semiotic systems.
4. These terminological subtleties may appear futile, but to us they seem necessary, as they serve as a reference point, allowing us to locate basic options that were present during the progressive differentiation between semiology and semiotics. Thus, with respect to the Hjemslevian definition of semiology, R. Barthes' first "disloyalty," even before Elements of Semiology, was his interest in the connotative dimension of language (cf. his Mythologies), a domain excluded by Hjemslev from the definition of semiology but which we place within socio-semiotics (for social connotations) and psycho-semiotics (for connotations at the level of the individual). Obviously, this was not an act of disloyalty, but a fundamental attitude regarding signs and languages, and one may remember the shock effect produced by the originality of this approach, as well as the almost immediate result - the recognition of the establishment of smeiology. However, this oblique approach to language left too much to the intuition of the describer (or Barthes' "scriptor"): since the signifier of connotative languages is disseminated along the whole lenght of discourse, inaccessible for any direct structuration, it could only be approached by a preliminary and arbitrary postulation of the signified. Inasmuch as it was no longer supported by an imagination subjected to a rigorous conceptual discipline, semiological analysis of a connotative nature could only succeed in producing a redundance of commonplaces, unless it were to seek a foundation elsewhere, either in a certain form of psychology (at which time the object-semiotic system, not analyzed, becomes the "signifier" for the psychoanalyst) or in a certain sociology (at which time semiology becomes the post facto justification of a theory of ideologies). As soon as we let signifieds choose their own signifiers (a consistent connotative approach could proceed in no other way), we abandon the basic postulate of semiotics - the reciprocal presupposition of the signifier and the signified, which constitute the force and specificity of semiotics.
5. The reverse "disloyalty" - still with respect to the Hjemslevian definition of semiology - consisted in an interest in meta-semiotics whose object-semiotic systems were already scientific semiotic systems (i.e., scientific discourse and formal langugaes), a domain that Hjemslev had left to logicians and mathematicians. It was obviously not a question of taking a place of these two groups (although the difference between the semiotic and local points of view could possibly reveal their complementarity), but to see how they treated a particularly difficult problem, that of descriptive metalanguages. This encounter with the Viennese School of logic and its Anglo-American continuation (which defines semiotics as the union of two components: syntax and semantics), as well as with the Polish School of mathematicians (which developed the issue of the hierarchy of metalanguages), only confirms Hjemslev's requirement of a "scientific" metalinguistic description. It must be said - from this point of view - that semiology (in the restricted sense that we are beginning to give to this term) was never very interested in the problems of semantics; for it, the description of the signified was a mere problem of paraphrase. However, to avoid an unverifiable subjectivity, the paraphrase must be regulated and the paraphrastic description of the signified plane (of the semiotic system) must be submitted to analysis; if the description is recognized as a construction, the analysis must be coherent and adequate. It is not, as some would claim, a question of an unwarranted domination of linguistics over semiology, but of general conditions under which any activity having a scientific calling is exercised. A gap therefore divides semiology (for which natural languages serve as instruments of paraphrase in the description of semiotic objects) and semiotics (the main task of which is the construction of an appropriate metalanguage).
6. Finally, the last point of contention lies in the evaluation of the relations between linguistics and semiology/semiotics. Semiology seems to challenge the primacy of linguistics, by insisting on the specificity of signs and organizations recognizable within non-linguistic semiotic systems, semiotics is considered tightly bound to the methods of linguistics [isn't it the other way around?]. In rality - and this is particularly clear in the field of visual semiotics (see planar semiotics) - semiotics more or less explicitly postulates the mediation of the natural languages in the process of reading signifieds belonging to non-linguistic semiotic systems (pictures, painting, architecture, etc.), whereas semiotics challenges this mediation. Beginning with Systéme de la mode, the most Hjemslevian of Barthes' works, in which, in order to describe the semiotics of clothing, he makes use of the mediation of "written fashion" (considering, however, that it was simply for convenience sake, and not required by methodology), the semiology of painting came to be conceived as the analysis of discourse on painting. The misunderstanding goes back to the period when linguistic theoreticians such as R. Jakobson, fighting against the psychologism of "thought," expressed by the "tool" of language, openly affirmed the indissoluble interconnection of these two "entities." Recognizing that there is no language without thought, nor thought without language, does not imply that we have to consider natural languages as the only receptacle of "thought": the other, non-linguistic, semiotic systems are also languages, that is, signifying forms. Consequently, the felt and the experienced - terms by which we designate, for example, the hold that architectural forms have over us - are but the signifieds of these forms, that a constructed, more or less adequate, but arbitrary, metalanguage is supposed to account for.

1. A term is said to be denotative when it encompasses a definition which aims at exhausting a concept as far as its extension is concerned (cf. J. S. Mill). Thus, for instance, a linguistic unit is denotative when it subsumes all its occurrences.
2. More broadly speaking, a denotative character is attributed to semiotic systems, complex objects that they are, insofar as these systems fulfill the demands of the principle of empiricism (and, more specifically, the requirement of exhaustivity). A denotative semiotic system is, for L. Hjemslev - and merely in a preliminary attempt of definition - a system none of the planes of which is itself a semiotic system. When one of the two planes is itself a content plane, such a semiotic system can no longer be viewed as denotative.
3. Such a definition adds nothing to the definition of a bi-planar semiotic system (that is, a "semiotic" properly speaking in the terminology of the famous danish linguist). Thus Hjemslev completely abandons this definition of denotative semiotic system so as to propose a new distinction between scientific and non-scientific semiotic systems. He himself explains this shift of terminology by noting that in order to establish his definition of a semiotic system he had presupposed an ideal text and postulated that it was characterized by structural homogeneity. But no such text exists. Any text, as product, belong sot several different systems. Consequently, (a) a text cannot be viewed a priori as a homogeneous entity. On the contrary, it is constructed in terms of the level of pertinence chosen, as the analysis progresses; (b) a natural language is not a denotative semiotic system, and the manifested discourse belongs to several systems (a semiotic system, a number of connotative semiotic systems, of non-scientific meta-semiotic systems, etc.); (c) "everyday language" is not a semiotic concept. Consequently, it can in no case be identified with the concept of denotative semiotic system which, taken as a whole to be a "signifier," would have a signified, which would make of it a connotative semiotic system (or language of connotation).

1. A term is said to be connotative if, when one of the features of the concept considered in terms of its comprehension is named, that term points to the concept as a whole (cf. J. S. Mill). Since the fature(s) taken into consideration has(have) been selected either through a subjective choice, or because of social convention, connotation is a process which is difficult to define with precision. This explains why it has given rise to such a diversity of definitions and why the use of this term is often confusing.
2. From a semantic perspective, connotation could be interpreted as the establishment of relations among one or more semes located on a surface level and the semene to which they belong, which must be read at a deeper level (cf. deep structure). Their connotation is akin to metonymy, the well-known rhetorical figure. The relation estbalished by connotation could be either hypotactic of hypnonymic. It would then be a case of the so-called oblique definition, studied elsewhere.
3. In his typology of semiotic systems, Hjemslev made room for a specific class of connotative semiotics. The only common point between the connotation of concepts (located at the lexematic level) and the languages of connotation (which overdetermine discourses) is found in the rather intuitive recognition of a gap or of an oblique relation which would exist between a primary, "denotative," signified, and a secondary, "connotative," signified. However, in order to postulate the existence of a plane of connotative content, one needs to introduce the semiotic function (or semiosis) which would connect it to an expression plane. The latter cannot be considered as identical to the expression plane which is correlated with the denotative signified, for that would leave only one content plane. Thus Hjemslev postulates an expression plane which is itself a semiotic system (for example, a natural language). The connotative semiotic system would then be a particular kind of meta-semiotic system.
4. The main difficulties appear when an attempt is made at analyzing such a connotative semiotic system. In order to identify the units of the connotative signifier, one must begin with the description of the semiotic system in question, viewed as "denotative": the units which will be found here can, alone, eventually be bi-valent and belong simultaneously to both semiotic systems. Account must then also be taken of the fact that the role of the connotative signifiers can be fulfilled by the signs of the semiotic system under analysis as well as by the figures of its expression and/or content planes, and also by the two substances which constitute its form (for example, a Southern accent is recognized through its phonetic, and not phonological particularities). Moreover, not all the units of these various planes have connotative properties. As a result, the listing of the signifier-connotators can be performed only by extraction - a process which, according to Hjemslev, is not scientific. Thus he classifies connotative semiotic systems among non-scientific semiotic systems.
5. Yet, the existence of semiotic connotations cannot be doubted, and their importance has been clearly shown in the work of R. Barthes. For him, "ideology would be, after all, the form of the connotative signifieds, while rhetoric would be the form of the connotators." The urgent need for a theory of connotative semiotic systems is thus apparent.
6. While, in Hjelmslev's perspective, the description of a connotative semiotic system must begin by the study of its expression plane (a study which, because of the principle of commutation, would be expected progressively to elucidate the form of the connotative content), one wonders whether theoretical research should not first be focused upon the substance of this content. This would then allow us to begin by the identification of the main topical spaces where connotative activity takes place. A socio-semiotic approach (cf. the "external linguistics" of F. de Saussure) which would elaborate tentative models of the possible loci of the manifestations of connotation, would help better to circumscribe the connotative phenomenon and to show, in part, the interrelations among social connotations. Such research would have to deal with and integrate problems concerning the categorization of the world (the hypothesis of Humboldt, Sapir, and Whorf), concerning the functions of language according to Jakobson, concerning the genre typology, etc. Simultaneously, psycho-semiotics sould need to deal with individual connotations, as Hjelmslev himself suggested. The inverse approach, elucidation of the connotative forms, would be taken up only at a later stage of research.

1. As a philosophical term, concept carries numerous and varied definitions, which after all refer, more or less, to entities of the signified (= ideas) that can organize the data of experience.
2. F. de Saussure used the term to designate - in a first approximation - the signified, the sole determination of which is to participate in the constitution of the sign (the concept of a tree and the acoustic image of a tree constitute the sign "tree"). Later, he eliminated this notion in favor of signifying form.
3. In semiotic theory, the term concept can be kept in the sense of name (the signification of which is clarified by its definition). The clarification of concepts by successive definitions then becomes the major preoccupation of all the metalinguistic constructions of the theoretician. For indeed, it appears that linguistic or semiotic theories contain many non-defined concepts. These are borrowed from natural languages, and, more particularly, from often implicit philosophical doctrines. While these concepts are frequently strongly suggestive, and can cover crucial sets of issues, they are not thereby integrated into the overall theory. The construction of a theory should then involve a conceptual phase, where the concepts are called upon to be replaced by definitions and interdefinitions whose constituent elements are more abstract and more general than the original concepts. It is only at the top of such a conceptual hierarchy that we find non-definable concepts (such as "relation," "object," "description," etc.) that form an inventory which helps to establish an axiomatic system.
4. In such a perspective, inspired by Hjelmslev, the term concept, as an element of the metalanguage, helps to name classes of objects (the semiotic units) as well as procedures and models. It is in this sense that within a theory, a distinction is drawn between concepts that are "real," i.e., integrated into scientific metasemiotics, and operational concepts, which are bases for procedures or models that seem effective but which, because they are not integrated within the theories, can only be considered provisiona.

1. The diverse and varied uses of the word form reflect practically the whole history of Western thought. Thus the status attributed to this concept in one or the other semiotic theory (or, in a more limited way, in linguistic theories) lets its epistemological foundations be recognized easily. Indeed, the notion of form has inherited its eminent position in the theory of knowledge from the Aristotelian tradition. Opposed to the matter which it "informs" (to which it gives form) at the same time it "forms" the knowable object, form is what guarantees the permanence and the identity of the knowable object. Taken in this fundamental sense, form is close to our conception of structure (cf. Gestalt).
2. When the concept of form is applied to "thought objects," the matter it informs is interpreted progressively, via semantic drift, as the "meaning," the "content," giving rise thus to dichotomies established by daily use. From this point of view, the word form comes close to and becomes almost synonymous with expression. The "base meaning," consireded as unvarying, is submitted to variations on the phonetic, syntactic, or stylistic planes. Contrawise, when meaning is considered as something which nothing can be said (Bloomfield), form is valorized. It alone can be submitted to linguistic analysis (cf. American structuralism).
3. F. de Saussure's affirmation that language is a form resulting from the union of two substances must be understood in this context. Being neither "physical" substance nor "psychic" substance, but the locus of their convergence, form is a signifying structure (cf. Merleau-Ponty). The ontological independence of semiotic form which is thus affirmed confers a status of autonomy on linguistics (which has an object the coherent and exhaustive description of this form).
4. L. Hjelmslev's interpretation of the Saussurian conception of form permits a refining of the mechanism of semiotics, a mechanism which is both epistemological and methodological. While the monist formulation of signifying form (which applies, stricto sensu, only to the prosodic categories of natural languages) is not called into question, it is broadened by postulating the existence of a form proper to each of the two substances: the expression form and the content form must be recognized and analyzed separately, prior to their joining, which produces the semiotic schema.
5. Recognition of two forms proper to each of the two planes of language has, within a general theoretical framework, allowed the resituation of both phonology, the expression form, and of phonetics, the study of substance (cf. -emic, -etic). It has also authorized the transportation of the same distinction to the content plane, thereby opening the way to the elaboration of a formal semantics.

1. Semiosis is that operation which, by setting up a relationship of reciprocal presupposition between the expression form and the content form (in Hjelmslev's terminology) - or the signifier and the signified (Saussure) - produces signs: in this sense, any language act implies a semiosis. The term is synonymous with semiotic function.
2. By semiosis can also be meant the semic category of which the two constituent terms are the expression form and the content form (or, the signifier and the signified).

1. Signification being the key concept around which all semiotic theory is organized, it is not surprising to see that it has a place in the different positions of the area of study that semiotic theory hopes to delimit. The term signification became progressively excluded from its initial delimitation because of the definitions and names that were set up incorporating it. Nevertheless, it kept its parasynonymous uses in everyday language. We shall note some of these.
2. Like all nouns of this subclass (cf. description, operation, etc.), signification may sometimes designate a doing (signification as a process), sometimes a state (that which is signified), thus revealing a dynamic or a static coneption of the underlying theory. From this point of view, signification may be paraphrased either as a "production of meaning" or as "meaning already produced."
3. We arrive at first delimitation of the semantic field embraced by "signification," by opposing it to "meaning," i.e., by saving the latter term for that which is anterior to semiotic production. We will thus define signification as articulated meaning. That means that the term "signification" is sometimes used to designate the "purport" in the Hjelmslevian sense, but this acceptation ought to be excluded from semiotic metalanguage.
4. Conjointly with the term meaning, signification is still used to denote the content substance. As the content substance has already been selected with signification in mind and as it presupposes the existence of the content form, use of the term "signification" is not incorrect; it is superfluous. We have the same situation when signification is employed as synonymous with the signified of the sign or with the content plane in general.
5. Signification is also used as a synonym for semiosis (or the act of signifying) and is then interpreted either as the union of the signifier and the signified (constituent of the sign), or as the relation of reciprocal presupposition that defines the cosntituent sign.
6. Since these uses of the term are already provided with particularizing semantic labels, we reserve the term signification for what appears essential to us, i.e., for "difference" - the production and reception of gaps - that defines, according to Saussure, the very nature of language. Understood in this sense as the construction and perception of relationships, signification is established as "articulated meaning" in the meaning/signification dichotomy and, as a general concept, subsumes at the same time all the acceptations presented above.
7. To this axiomatizing defintion of signification, we must add another one, of an empirical character, bearing no longer on its "nature," but on the means of apprehending it as a knowable object. We then realize that signification can be grasped only during its manipulation, when, pondering over it in a given language and text, the enunciator is led to operate transpositions and translations from one text to another text, from one level of language to another, and, finally, from one semiotic system to another semiotic system. This paraphrasing activity may be considered as the representation of signification as a producing act, uniting in one domain the interpreter-enunciatee (as signification is not an ex nihilo production) and the producer-enunciator. As a programmed cognitive activity, signification is then supported and sustained by intentionality, which is another manner of paraphrasing signification.
- MEANING; STRUCTURE (B: Elementary structure of signification); CONTENT

1. Ferdinand de Saussure introduced the dichotomy synchrony/diachrony in order to designate two distinct ways of approaching linguistic phenomena. In actual fact, only the notion of synchrony was of consequence to Saussure, for it allowed him to establish linguistics as the study of coherent systems. The term diachrony came then to designate the area of studies in historical grammar. Thus, the opposition between synchrony and diachrony, which Saussure set forth as two interrelated temporal dimensions for research, has nonetheless been taken for a long time to be an opposition between a structural attitude and an atomistic approach with regard to language facts.
2. The initially categorical opposition between the two terms of the Saussurian dichotomy has been progressively softened. Given the fact that a semiotic system is not defined by the synchronization of the elements constituting it but by their internal, logical coherence, it was possible to interpret diachrony as a set of transformations which are located and recognizable between two systems taken as wholes (or between two given states of a natural language considered as the loci wherein two distinct systems are set). Such a conception, which likens the distance between two related languages, in fact does away with diachrony and allows for the practice of an achronic comparativism.
3. Instead of making use of the questionable procedure which consists in postulating a priori the existence of two states of language before knowing the transformations which alone could define these states, we can conceive of diachrony in terms of transformations located within a semiotic system (or natural language) provided that we then would identify the original states and the results of these transformations as semiotic (or linguistic) states. This approach can be illustrated by two examples.
4. While associated with the Prague School, Roman Jakobson offered an interpretation, which according to him stems from diachronic philology, of changes in the expression form of grammatical categories. These changes would be due to the redundant overdetermination of the morphemes manifesting the grammatical categories. So, for example, the disappearance of Latin declension endings can be explained by the redundant and prolonged coexistence of superfluous morphemes which denoted the same grammatical categories (such as determinants, prepositions, etc.). It can be said that the establishment of this secondary, emphatic, system resulted in the withering of these now useless flexional morphemes.
5. Other linguists (Martinet, Haudricourt) who begin with the postulate of equilibrium (which every semiotic system must maintain in order to be able to function) conceive of the diachronic process as chain-reaction transformations. These transformations are induced by the intrusion of a foreign element within a system (the vowel system, for example); these are transformations which seek to reestablish the lost equilibrium and succeed in constructing a new system based on a new equilibrium. This is a particularly interesting approach, for, instead of starting with states of language seeking to undergo possible transformations, it first describes the transformations which alone can define these states.
6. If we agree to consider such transformations as diachronic transformation, there is no reason not to use the same nomenclature for the transformations that we recognize in the unfolding of the narrative discourse (admittedly at the level of the content form). This discourse, which locates its performances between two structural states (initial and final), is comparable, due allowances being made, to the linguistic process that a linguistic community accomplishes between two states of a natural language.

1. The term synchrony was proposed by F. de Saussure, in opposition to diachrony, in order to designate simultaneity as a criterion for gathering together a set of linguistic phenomena which thus constitute a state of language; this for the purpose of a systemic study of language.
2. Synchrony was an operational concept, inasmuch as it made possible the creation of the concept of linguistic system (conceived as a relational hierarchy, the functioning of which is insured by its own internal organization). Although the concept of synchrony was useful in order to conceive the idea of system, it is no longer useful for the analysis of system. Indeed, this notion is as imprecise as that of "present," for example. Is a metaphor, invented by the speaking subject at the very moment in which s/he is speaking, a synchronic or diachronic phenomenon? A state of language - therefore, a synchrony - lasts several hundred years and embodies many various internal transformations (called conversions by L. Hjelmslev). Linguistics today operates in achrony, the concept of synchrony no longer being operational.

1. Although it is a property common to all the different semiotic systems, the concept of meaning is undefinable. Intuitively or naively, two approaches to the problem of meaning are possible: it may be considered either as that which permits the operation of paraphrasing or transcoding, or as that which grounds human activity as intentionality. Before its manifestation in the form of articulated signification, nothing can be said about meaning, unless metaphysical presuppositions full of implications are introduced.
2. L. Hjelmslev proposes an operational definition of meaning, by identifying it with the primary "material" or "support" by which any semiotic system, as form, is manifested. Meaning thus becomes synonymous with "matter" (the term "purport" covers both words): "meaning" and "matter" are used intechangeably when speaking of the two "manfiestants" of the expression plane and of the content plane. The term substance is then used to designate meaning as it is assumed by a semiotic system; this allows us to distinguish the content substance from the expression substance.

1. An actant can be thought of as that which accomplishes or undergoes an act, independently of all other determinants. Thus, to quote L. Tesniére, from whom this term is borrowed, "actants are being or things that participate in processes in any form whasoever, be it only a walk-on part and in the most passive way." From this point of view, "actant" designates a type of syntactic unit, properly formal in character, which precedes any semantic and/or ideological investment.
2. The term "actant" is linked with a particular conception of the syntax which interrelates the functions of the elementary utterance. These functions, such as subject, object, predicate, are defined independently of their realization in syntagmatic units (for example, nominal and verbal syntagms). This syntax also poses the predicate as the nucleus of the utterance. All this is to say that the actants are to be considered as the resultant terms of the relation known as the function. This concept of actant is likewise to be interpreted in the framework of case grammar (Fillmore), where each case can be considered as the representation of an actantial position. In this respect, actantial grammar, which is semiotic in nature, is seen as a more abstract formulation of case grammar. At a deeper level, actantial grammar, not subject to phrase linguistic form, is able to account for the organization of narrative discourse at the level of narrative syntax (called the surface level), thanks to functional, syntactic categories - subject, object, predicate, etc. - which it makes explicit in view of its own construction. From this point of view, it is distinguished from categorical grammars, which operate with morphological classes, and from syntagmatic grammars, which rely on distributional classes.
3. The conept of actant has the advantage of replacing, especially in literary semiotics, the term of character as well as that of "dramatis persona" (V. Propp), since it applies not only to human beings but also to animals, objects, or concepts. Furthermore, the term character remains ambiguous since it also corresponds in part to the concept of actor (where syncretism of actants may occur), which is defined as the figure and/or the empty locus wherein are invested syntactic and semantic forms.
4. Typologically, the following are distinguished within the uttered discourse: (2) The actants of communication (or of the enunciation), which are not only the narrator and narratee, but also the interlocutor and interlocutee (which participate in the structure of second degree in interlocution, the dialogue). (b) The actants of narration (or of the utterance): subject/object, sender/receiver. From the grammatical point of view, the syntactic actants (inscribed in a given narrative program) such as the subject of state and the subject of doing, are opposed here to the functional (or syntagmatic) actants, which susbume the actantial roles in a determined narrative trajectory. Bearing in mind the two dimensions recognized in discourse, we distinguish between, for example, pragmatic subjects and cognitive subjects. The latter appear either in syncretism with the pragmatic subjects, or as autonomous actors, in the case of the informant, for example; or, in the case of the observer actant, they are recognizable at least as implicit positions. Since at the level of discoursive semantics, the actant is taken in charge by the procedure of figurativization, it is then termed either individual, dual, or collective.
5. Any actant may be projected onto the semiotic square and thus articulated in at least four actantial positions (actant, anactant, negactant, negantactant). When aritculated in this way, the actant is known as the proto-actant and is transformed into an actantial category.
6. As the narrative discourse progresses, the actant may assume a certain number of actantial roles, defined both by the position of the actant in the logical sequence of the narration (its syntactic definition) and by its modal investment (its morphological definition). Thus the hero will be the hero only in certain parts of the narrative - s/he was not the hero before and s/he may well not be the hero afterwards.

1. Historically the term actor has gradually replaced character (and dramatis persona), indicating thereby a greater desire for precision and generalization - a magic carpet or a business firm, for example, are acors - thus extending its use outside the purely literary.
2. Obtained by procedures of engagement and disengagement (which belong to the domain of enunciation), the actor is a lexical unit, nominal in type, which, once incorporated into the discourse, may receive, at the moment of its manifestation, investments of surface narrative syntax and discoursive semantics. Its semantic content proper seems to consist essentially in the presence of the seme of individuation, which gives it the appearance of an autonomous figure in the semiotic universe. An actor may be individual, (for example, Peter), or collective (for example, a crowd), figurative (anthropomorphic or zoomorphic), or non-figurative (for example, fate). The individuation of an actor is often marked by the attribution of a proper noun, though that does not in any way constitute a sine qua non of its existence - a vague thematic role, "father," for example, may often be used to denote the actor. Thus, onomastics, included under discoursive semantics, is complementary to actorialization, which is one of the procedures of discoursive syntax.
3. At first, the term actor was linked and opposed to the term actant. From a comparative point of view, when dealing with a corpus of tale variants, it can be noted that a single subject-actant, for example, can be manifested by several occurrence-actors. Nevertheless, distributional analysis, used in this manner, underlines particularly the invariant nature of the actant, while not giving thereby any information on the nature of the actor. It should also be borne in mind that the actor goes beyond the limits of the sentence and, thanks to anaphora, is maintained throughout the discourse - or at least throughout a discoursive sequence - in keeping with the principle od ifentity. It ceases, at that point, to be the variable of a single, invariant actant, and successively assumes various actantial roles. In the same way, since discourse is the unfolding of semantic values, the actor may receive one or more different thematic roles.
4. Thus a more precise definition may be established by viewing the actor as the point of convergence and investment of both the syntactic and semantic components. In order to be designated actor, a lexeme should have at least one actantial role and at least one thematic role. It should be further noted that the actor is not only the point of investment of these roles, but also of their transformations, since discourse consists essentially of the interplay of successive acquisition and loss of values.
5. At the surface level of the text, then, is seen an actorial structure which is nothing else than a topological structure. The various actors are built up into a network of loci which, while empty by nature, are the loci where narrative and discoursive structures are manifested.
6. From the point of view of the production of discourse, a distinction can be made between the subject of the enunciation (which is an implicit actant logically presupposed by the utterance) and the actor of enunciation. An example of the latter case would be, for example, "Baudelaire" as defined by the totality of his discourses.

1. In a very general sense, contract can be understood as the fact of establishing, of "contracting" an intersubjective relationship which has as its effect the modifiction of status (being and/or seeming) of each of the subjects involved. Even though this intuitive notion cannot be defined rigorously, we need nevertheless to posit the term contract so as to determine progressively what are the minimal conditions under which the establishment of "entering into a contract" between two subjects takes place. These conditions can be viewed as being presupposed by the establishment of the structure of semiotic communication. It is indeed necessary to recognize, hidden under the contract, this "phatic communication which constitutes the necessary and preliminary undergirding for any communication and which seems to involve both a tension (a well-disposed or a mistrustful expectation) and a relaxing (as a kind of response to the expectation). Indeed, the establishment of the intersubjective structure is at once, on the one hand, an opening toward the future and toward possibilities for action, and, on the other hand, a constraint which somehow limits the freedom of each of the subjects. We propose to use the term implicit contract to designate this set of preliminary components on which the intersubjective structure is based.
2. In the semiotic perspective, it does not seem appropriate to opt for either one of the two opposed ideological attitudes which consider social life respectively as made up of confrontations and struggles or as based upon "charity" and "well-intentioned" conventions. On the contrary, the structural approach demands that both the positive and the negative terms of a category be considered together. In other words, the polemical structures (given a priori or resulting from breaches) must be viewed as constituting the opposite pole of the contractual structures (stricto sensu). Actually, these two types of structures belong to the same contractual organization of intersubjectivity.
3. At first glance, two kinds of contracts can be distinguished. A contract is said to be unilateral when one of the subjects makes a "proposal" and the other makes a "commitment" to that proposal. A contract is bilateral or reciprocal when "proposals" and "commitments" are interwoven. Yet such a definition, borrowed from common dictionaries, shows the modal nature of the contractual structure. The "proposal" can be interpreted as the wanting of subject S1 that the subject S2 do (or be) something. "Commitment" is nothing else than the wanting or the having-to of S2 taing upon itself the suggested doing. In this perspective, the contract appears as an orgnaization of reciprocal cognitive activities which bring about the transformation of the modal competence of the subjects involved.
4. The preceding remarks may seem to be inspired by philosophical or sociological preoccupations. But this is not the case. They are developed exclusively and above all on the basis of growing number of concrete analyses of discourses, and, more specifically, of narrative discourses. Such discourses involve numerous descriptions of contractual structures and thus they constitute the main source that semioticians can exploit in their effort to establish a typology of contractual structure. Thus, for instance, the canonic narrative schema derived from V. Propp's descriptions appears, in one of its aspects, as the syntagmatic projection of the contractual structure. In this narrative schema, the contract, established from the beginning between the Sender and the subject-Receiver, governs the overall narrative development. What follows it in the narrative appears as its execution by the two contracting parties: the trajectory of the subject, which is the Receiver's contribution, is followed by the sanction, both pragmatic (retribution) and cognitive (recognition) by the Sender. It is clear that this syntagmatic organization, based upon the articulation of the contract, can be broken up into a series of contractual units such as the establishment, the breaking, the re-establishing, and the execution of the contract.
5. The concept of contract must be related to that of exchange, the theory of which has been elaborated by M. Mauss. In such a case, the contract appears, at first, as a delayed exchange: the distance which separates its conclusion from its execution is filled up by a tension which is both like a credit and a debit, like a confidence and an obligation. Yet a closer look shows that a simple operation involving the exchange of two objects of value is not a mere pragmatic activity. Rather, it taks place essentially on the cognitive level. For indeed, in order that the exchange might take place, it is necessary that the two parties be assured of the "value" of the value of the object to be received as counterpart for the object given. In other words, it is necessary that a fiduciary contract (often preceded by a persuasive and by an interpretive doing of both the subjects) be established prior to the actual pragmatic operation.
6. Such a fiduciary contract can be called utterative in so far as it is inscribed within the utterance-discourse and as it involves pragmatic values. Yet it is manifeted as well on the level of the structure of enunciation where it is seen to be an enunciative contract (a term proposed by F. Nef), or a veridiction contract, since it aims at establishing a fiduciary convention between the enunciator and the enunciatee involving the veridictory status (on saying-the-truth) of the utterance-discourse. The difuciary contract which is thus established can be based upon evidence (i.e., upon a self-evident certainty) or it can be preceded by a persuasive doing (causing-to-believe) of the enunciator to which the enunciatee responds by way of an interpretive doing (a believing).

1. Parallel and closely related to information theory, a schema of linguistic communication has been developed. It remains bound to a theory quote mechanistic in character, even though it claims to be more respectful of intersubjective verbal exchanges. According to the psychologist Bühler, linguistic activity can be defined by its three functions: expression (from the addresser's viepwoing), appeal (from the addressee's viewpoing), and representation (which has to do with the referent or the context). Roman Jakobson has taken over and completed this triadic schema, using a new terminology. For him, verbal communication is based on six factors: the addresser and the addressee, the message transmitted from one to the other, the context (or referent) - verbal or verbalizable - with which the message is concerned, the code (more or less common to the actants of the communication) thanks to which the message is communicated, and finally the contact, which is based on both a physical channel and a psychological connection. To each of these different elements there corresponds a particular linguistic function, respectively: emotive (or expressive), conative, poetic, referential, metalinguistic, phatic.
2. It goes without saying that the Jakobsonian functions of language do not exhaust their object, and that such an articulation, however suggestive it may be, does not lay the basis of a methodology for discourse analysis. The six function schema is both too general to allow for an appropriate taxonomy and an appropriate syntax and at the same time too specific in that it deals only with verbal communication - and even then it does not take into account - to the exclusion of all other semiotic systems. Thus, for example, this schema seems to deal only with informative doing, which can be articulated, according to the addresser/addressee relationship, as emissive doing/receptive doing. Now, there are other ways in which the transmission of knowing may be conceived, particularly when it is modalized: such is the case with persuasive doing and interpretive doing, which depend more on manipulation than on "communication."
3. It is clear, on the other hand, that if language is communication, it is also production of meaning. It cannot be reduced to a simple transmission of knowing on the "I"/"you" axis, as a certain brand of cuntionalism would have it. Furthermore, it develops, so to speak, for itself, for what is, with its own internal organization that the theory of communication alone - taking a somewhat external viewpoint - does not seem able to account for.
4. Although independent from Bühler, Jakobson, Martiner and all the functionalist trend, English philosophy of language, with J. L. Austin, shares - beyond differences in terminology and preoccupations - the same concern, that of accounting for language as an intersubjective operation, even though it attempts to integrate a great part of human activity. J. R. Searle's notion of "speech act," which has been progressively elaborated, and, beyond it, the theory of pragmatics (in the American sense), transcend the limits of simple "communication" by focusing upon the conditions of its exercise; despite a certain lack of coherence in terminology arising from a philosophico-linguistic amalgam, their constribution to the study of linguistic activity should not be ignored.
5. To avoid a conception of communication that would be too mechanistic (which goes back to the information model), or one that would be too restrictive by confining itself to "extralinguistic" parameters, it is essential to place this key notion in a larger context. Human activities as a whole are generally considered as taking place along two main axes: that of action on things, by which human beings transform nature - the axis of productions - and that of action on other persons, which creates intersubjective relations, the foundations of society - the axis of communication. The concept of exchange in the French anthropological tradition which, especially since M. Mauss, covers this second sphere of activities, can be interpreted in two different ways, either as the transfer of objects of value, or as communication between subjects. The transfer of objects that are manifested as acquisitions and deprivations can only affect the subjects and constitute, insofar as they take canonical forms, certain systems of interhuman relations, regulating the desire of human beings. Lévi-Strauss has proposed that three fundamental dimensions of these communication-transfers be distinguish: to the exchanges of women, considered as a process, there correspond kinship structures, which have the form of systems; to the exchange of goods and services there correspond economic structures; and to the exchange of messages correspond linguistic structures. This very general schema can of course be modified or refined: in place of linguistic structures particularly, it is advantageous to register wider semiotic organizations. From the concept of exchange, on the other hand, should be taken away the euphoric connotations which allude to a universal good will among humans in their mutual relations. For indeed, it is difficult, if not impossible, to establish clear distinctions between contractual structures and polemical structures which govern communication. Nevertheless, it remains true that such a conception of communication enables us to establish a strictly semiotic approach to the problem, which is quite different from those of economic theories, on the one hand, and from the theory of communication, on the other.
6. Insofar as communication takes place between subjects and insofar as value invested in the objects put into circulation (pragmatic or cognitive values) are considered as constituents of the being of the subject (constantly undergoing increases or decreases of its being), it is clear that the sender and the receiver can no longer be treated as abstractions, as the empty positions of source and destination. They are, on the contrary, competent subjects, considered at a particular moment of their becoming, each inscribed in its own discourse. It can also be understood why a dialogue that appears inside a narrative discourse seems to give us a more correct representation of the process of communication than a artifact constructed on the basis of the extra-linguistic "structure of communication," and furthermore why we propose to interpret an "exchange of messages," at least at the semantic level, as a discourse with two, or several, voices.
7. This "humanization" of communication, a preoccupation of most of the recent theories in this field, does, however, raise new problems for which as yet there are no definitive solutions. Let us first note the problem of participative communication. Contrary to what happens during ordinary communication, where the attribution of an object of value is concomitant with renunciation, discourses that are in nature ethno-literary, philosophical, or juridical (cf. constitutional law), display communication structures where the transcendent Sender (absolute, sovereign, original, ultimate, etc.) offers values that are modal (being able, for example) as well as descriptive (material goods), without really renouncing them and without any loss of its being. This case of the transcendent Sender must obviously be distinguished from that of the giver of knowledge which, during communication, transmits a cognitive object while its own knowldge is not in any way diminished. This peculiarity is explained by the fact that the subject of the enunciation is a syncretic actor, subsuming the two actants, enunciator and enunciatee; in other words, it is its own enunciatee and thus regains what it had given out as enunciator [self-communiaction].
8. Another question, still unanswered, is that raised by the distinction - fairly easy to recognize but difficult to explain - between received communication and appropriated communication. Psychoanalytical discourse has brought to light the distance between mechanisms which insure that the signification will be grasped, and the little-known procedures governing its appropriation, its integration into the already-existing axiology. It is as though the receiver subject cannot gain full possession of the meaning unless it is previously equipped with a form of wanting and of being-able-to-accept; unless, in other words, the receiver subject can be defined by a certain type of receptive competence that in turn constitutes the first and last aim of the enunciator's discourse. If to appropriate another person's word is believed in it to some extent, then causing to appropriate is speaking in order to be believed. Considered in this way, communication is far less causing-to-know - as it is too hastily conceived - than causing-to-believe and causing-to-do.
9. Another of the many possible problems is that of the frequent concomitance, and resulting confusion, of productive doing (which can be formulated as narrative doing) and communicative doing. A ritual is programmed doing, whose goal is its own signification: the installation of an observer (the public, for example), perverts it not only because it transforms it into a spectacle, but also because the behavior of the observed become equivocal and splits into two. The conversation of two people ceases to be what it is if the participants know they are being listened to. We are not dealing here merely with the issues of the semiotics of the theatre but more widely with the "spectacular" dimension of our cultures and our signs, the study of which has been poorly approached and about which little is as yet known.

1. The study of narrative discourses has led us to distingush, at a superficial level, the cognitive dimension and the pragmatic dimension, the latter serving in a way as an internal referent for the former. The pragmatic dimension, recognized in narratives, corresponds roughly to the descriptions which are made there of signifying somatic behaviors, organized into programs and taken by the enunciatee as "events" independently of their possible utilization at the level of knowing. Pragmatic objects are recognizable as descriptive values (such as objects which can be stored up or consumed) by contrast to modal values. In this sense, pragmatics would be homologated with the third function of G. Dumézil. It is under this meaning that one will distinguish correlatively pragmatic doing and cognitive doing, pragmatic subject and cognitive subject, pragmatic performances and competences and cognitive ones.
2. One sees the gap which separates our conception - which takes into account the set of human activities such as they are described in discourses, by interrelating them according to the dichotomy pragmatic/cognitive - from that which has developed in America, beginning in particular from the works of Ch. W. Morris. Pragmatics, in the American sense, essentially seeks to set out the conditions of (linguistic) communication such as, for example, the way in which two interlocuteurs have an effect on each other. For us, this "pragmatics" of language which has reference to the characteristics of its utilization constitutes one of the aspects of the cognitive dimension; for it concerns in fact the cognitive competence of the communicating subjects, such as it can be recognized (and its simulacrum reconstructed) within the utterance-discourse. Thus, persuasive doing and interpretive doing do not constitute "extra-linguistic" parameters, as a certain mechanical conception of communication would have it understood, but they enter totally into the process of communication - as it is envisaged by semiotics - where the addressor and the addressee, for example, are not empty domains (such as sender and receiver) but are competent subjects. It is self-evident that, in the very perspective of American "pragmatics," a semiotics of "genuine" communication (as a describable object) can be drawn up by extrapolating in particular the models of cognitive semiotics, which have sprung from the analysis of narrative discourses.

1. The terms sender and receiver (usually written with a small first letter) - taken over from R. Jakobson's schema of linguistic communication - designate, in their most general sense, the two actants of the communication (also called in information theory source and receptor, but in a mechanical and non-dynamic perspective). When they are viewed as logically presupposed implicit actants of every utterance, they are called enunciator and enunciatee. When, by contrast, they are explicitly mentioned and are thereby recognizable in their utterance-discourse (for isntance, "I"/"you") they are called narrator and narratee. Finally, when the discourse reproduces the structure of communication by simulating it (cf. dialog), they are called interlocutor and interlocutee. In these last three cases it is clear that we are dealing with an act of delegation originating from the sender and the receiver.
2. Considered as narrative actants, Sender and Receiver (in this case usually written with a capital letter) are actantial domains characterized by a relation of unilateral presupposition (between Sender as presupposed term and Receiver as presupposing term). Consequently the communication between them is asymmetric. Paradigmatically, the Sender is in a hypernymic relation with the Receiver, while the latter is in a hyponymic position. This asymmetry is amplified in the syntagmatization of these two actants, when they appear as two subjects concerned with a single object. This is what happens, for instance, in the case of participative communication. Sender and Receiver are stable and permanent narrative actans, whatever might be the roles of communication actants that they can assume. Thus the subject-Receiver communicates, and sender, knowledge about its own performances.
3. Often positioned as belonging to the transcendent universe, the Sender is the one which communicates to the subject-Receiver (belonging to the immanent universe) not only the elements of modal competence, but also the set of values at stake. The Sender receives the communication concerning the results of the subject-Receiver's performance; it falls to the Sender to sanction this performance. From this point of view and in the framework of the narrative schema, the manipulatory Sender (the initial Sender) can be opposed to the judicatory Sender (the final Sender).
4. Given the polemical structure of the narrative, the presence of a subject and an anti-subject presupposes the existence of a Sender (S1) and of an anti-Sender (S2). This axis of contraries can then - according to the semiotic square - unfold and produce as contradictories two new actantial positions: the position of non-Sender (S1) and that of non-anti-Sender (S2). It happens, for instance, that S1 assumes, on the pragmatic dimension, the role of active and performing Sender (as communicating the elements of modal competence) in the framework of the positive deixis, while S2 is, on the cognitive dimension, the passive Sender (receiving the knowledge about the subject-Receiver's doing and sanctioning it) and thus belongs to the negative deixis. In such a case the active Sender is conclusive, gathering in the fruits of the action (in the framework of the sanction). Yet it is not sure that this organization of these actants on the semiotic square is actually part of the narrative structure properly speaking.
5. During the analysis of narratives it is sometimes necessary to make a distinction between individual Sender - as manifested in the case of vengeance - and social Sender, called on to render justice. Each of these two actants may manifest a having-to-do either compatible or incompatible with that of the other.

1. The term code was first used in information theory, where it designates an inventory of arbitrarily chosen symbols, accmpanied by a group of rules for the composition of coded "words," and often compared to a dictionary or lexicon of the natural language (cf. the Morse code). We are dealing here, in its single form, with a derived artificial language. In this sense, the alphabet, together with spelling rules, can be considered as a code.
2. In the automatic treatment of information, the code splits into a set of symbols containing instructions and capable of being understood by a machine (cf. machine language), and into the automatic code proper, which is by nature binary (current on/current off) and allows the data to be recorded in the memory bank and to be processed, and enables information to be supplied on demand.
3. The naive application of the concept of code to the problems of communication (in the well-known statement of Wiener, Chinese is only American coded into Chinese) and the fleeting success of research into the field of automatic translation have generalized the use of this term in linguistics.
4. The theory of linguistic communication has attempted to exploit the opposition code/message (R. Jakobson), which is no more than a new way of formulating the Saussurian dichotomy (natural) language/speech. Code is then taken to mean not only a limited set of signs or units (belonging to a morphology), but also the procedures by which they are arranged (their syntactic organization). The articulation of these two components enables messages to be produced.
5. If language is considered as the combinatory arrangement of minimal pertinent features (semes and/or phemes) it can be seen that the inventory of semic categories, for example, together with rules of sememic construction and of the projection of discoursive isotopies, forms a semantic code which will be manifested at the level of linguistic signs in a lexematic dictionary. In certain cases it is even possible to speak of a partial code to designate a particular semic system, a sort of sub-code the constituent elements of which enter into the composition of different semes.
6. As a borderline case, some semioticians use the term code to cover an undefined set of units which have only a slender tie with each other, based on association, with no appeal made to an underlying logico-taxonomic organization (cf. R. Barthes in S/Z).

1. Seme commonly designates the "minimal unit" of signification (comparable to the pertinent - or simply distinctive - features of the Prague School). Located on the content plane, it corresponds to the pheme, the unit of the expression plane. By maintaining the parallelism between the two planes of language, we can say that semes are elements constituting sememes, just as phemes are elements constituting phonemes. It can also be said that a semantic system may be postulated, hypothetically, so as to account for the content plane of a semiotic system, comparable to the phonological system, the articulation of which constitute the expression plane.
2. The seme is not an autonomous, atomistic element; it exists only because of the differential gap that opposes it to other semes. In other words, the nature of semes is purely relational and never substantial, and the seme cannot be defined as the end-term of the relation that one sets up and/or grasps with at least one other term of the same relational network. Thus, we acknowledge that semic categories (=semantic categories that constitute the content plane) are logically anterior to the semes that make up these categories and that semes can be apprehended only within the elementary structure of signification. It is by giving a precise logical status to constituent relations of such a structure (contradiction, contrariety, implication) that the concept of seme can be determined and made operational.
3. Because semes are only terms, i.e., points of intersection and of meeting of signifying relations (only rarely corresponding to lexical realizations in natural language), they must be named arbitrarily during the procedure of analysis: verticality/horizontality, for example, ade denominations of a metalinguistic type, to which we should give a coherent organization; we are not dealing here with mere paraphrases in natural language. This is a theoretical position that opposes semioticians (such as ourselves) to generative semanticists (and even to B. Pottier): semioc analysis is for us a metalinguistic construction.
4. The approximate definition of the seme as the "minimal unit" of content must be challenged, not only in its status as a unit, but also as being "minimal": (a) Theoretically, it is easy to imagine that a combinatory system of some twenty semic categories (a number comparable with the number of phemic categories used by a given natural language) may produce as many semes as necessary for the needs of a natural language or any semiotic system. These semic categories, once an inventory has been made of them, would probably contain the universals of language. This is what we mean when we speak of semes as minimal units of meaning. However, without a complete inventory of "primitive" semes, no semic analysis can operate; (b) Therefore, the "minimal" character of the seme must be understood in a very relative sense, as minimal with respect to a chosen field of inquiry. Thus, faced with a given terminology of kinship or with a syntagmatic class of determinatives in an enclosed paradigm, semic analysis will call upon only the minimal number of differential features (or semic categories) necessary to exhaust all oppositions among the morphemes being studied. The same situation obtains for the analysis of the semantic components of one or several discourses. The minimal nature of the seme (we must not forget that the seme is a construct) is, therefore, relative and is based on the criterion of the pertinence of the description.
5. An examination of different semic categories allows us to distinguish several classes: (a) figurative semes (or exteroceptive semes) are entities on the content plane of natural languages, corresponding to elements of the expression plane of the semiotics of the natural world, i.e., corresponding to the articulations of the sensory classes, to the perceptible qualities of the world; (b) abstract semes (or interoceptive semes) are content entities that refer to no exteriority, but which, on the contrary, are used to categorize the world and to give it meaning: for example, the categories relation/term, object/process; (c) thymic semes (or proprioceptive semes) connote semic micro-systems according to the category euphoria/dysphoria, thus setting them up as axiological systems.
6. Two types of organization of semic sets may be distinguished: (a) taxonomic (or systematic) structures representing the organization of homogeneous semic categories as hierarchies (based on hyponymic relationships); (b) morphematic structures resulting from the integrating articulations originating in different semic categories and micro-systems and appearing as figures (possessing diverse elements in hypotactic relationships).
The distinction set up by B. Pottier between generic semes and specific semes stems from a taxonomic conception and a semic organization; our own conception of semic figures (constituting sememic kernels) is based on morphematic organization.
7. The establishment of the semic combinatory system produces a large number of semes. These are now, however, mere collections of semes, but hypotactic constructions, obeying a set of formation rules. Within a sememe, we may distinguish contextual semes (which the sememe possesses in common with other elements of the semantic utterance) and kernal semes, that characterize the sememe (and, possibly, the lexeme it belongs to) in its specificity.

1. In B. Pottier's terminology, the sememe is defined as the set of semes identifiable within the minimal sign (or morpheme). This unit of signification thus delimited is composed of three semic sub-sets: the classeme (generic semes), the semanteme (specific semes), and the virtueme (connotative semes).
2. With respect to this definition, our own conception of the sememe is different in severla fundamental aspects: (a) Whereas Pottier attributes to the sememe the totality of investments of the signified of a morpheme, the sememe - for us - corresponds to what everyday language calls an "acceptation" or "particular meaning" of a word. Pottier's sememe, then, corresponds to our lexeme, the latter being made up of a set of sememes (a set that may even be monosememic) that are held tofether by a common semic kernel. For example, the lexeme "table" possesses, in addition to the sememe given by dictionaries as "a flat surface held up by one or several legs," other identifiable semes in expression such as "table talk," "water table," "table of contents," "multiplication table," etc. The lexeme, as a set of sememes, is the result of the historical development of a natural language, whereas a sememe is a structural phenomenon, a unit of the content plane; (b) The semanticism common to several sememes included in the same formant, but distinct from the semic investments of contiguous sememes of a same string, constitutes the kernel of the sememe (cf. Pottier's specific semes, or semantemes). This kernel - or semic figure - is peculiar to the sememe, everything else coming from the context (usually from the minimal contextual unit, made up of at least two sememes) and constituting its classematic base. In other words, the sememe is not a unit of menaing delimited by the dimensions of the minimal sign; "in language" it is only a semic figure. it is only at the moment of its manifestation in a discourse that this figure is joined with its classematic base (made up of contextual semes) and thus selects a sememic trajectory which actualizes it as a sememe, excluding other possible trajectories which remain virtual, capable of producing other sememes of the same lexeme in other discourse contexts. Where Pottier decomposes the sememe into
sememe = semanteme + classeme
we would prefer to substitute another break-down:
sememe = semic figure + classematic base
These two formulations are based on different theoretical foundations. (The problem of the virtueme remains to be solved.); (c) The distinction made between lexeme (bound to its formant) and sememe (the unit resulting from the articulation of the content plane alone) frees the semantic analysis from the constraints of the sign and allows us to uncover similar or comparable sememic contents under different lexematic covers. By defining in advance the level of analysis considered pertinent, and by suspending (see suspension) those semic oppositions judged non-pertinent, it is possible to go from the parasynonymy of sememes to the recognition of their synonymy, and thereby, to constitute classes of sememes (or of constructed sememes), joining together many occurrences of sememes dispersed throughout a discourse and belonging to different lexemes; (d) Finally, the sememe cannot be considered as a collection of semes, the product of a mere combinatory system. It is a syntactic organization of semes; semic figures often contain, implicitly, actantial structures (for example, "to give" implies a presence of at least two actantial positions) and/or more or less complex thematic configurations ("to moan" = "to make a low, prolonged sound, while in pain or in sorrow").

1. Semic analysis and componential analysis are most often placed together, despite their different origins (one is European, the other American), and despite their autonomous development and their divergent goals (the first seeking to account for the semantic organization of a lexical field, the second to describe as economically as possible the temrinology of kinship). They have in common the fact that they are taxonomic procedures that attempt to expose the paradigmatic organization of linguistic phenomena on the semantic plane, by establishing distinctions with the help of pertinent features (oppositions of semes in one case, the "components" or constructive elements in the other).
2. semic analysis may be correctly considered as the extension of distributional analysis, but with the contribution of the tools of semantics. For example, the class of noun determinatives, once it has been established thanks to their distributions, will be treated as a closed paradigm, made up of sub-classes such as articles, demonstratives, possessives, etc., and which can only be defined by semic oppositions; a later analysis of these sub-classes, considered one by one, permits their articulation into grammatical categories, etc.
3. The complexity grows if we want to treat open classes (nominal or verbal radicals) in the same way. The criteria chosen to determine the limits of a sub-class made up of lexemes are weak and often intuitive (for this reason, B. Pottier, who inaugurated this type of analysis with a taxonomy of types of "seats," makes reference to the vague concept of "field of experience," the fragility of which he recognizes); in addition, the nature of the semes establishing the necessary distinctions (for "seats": "to sit in/on," "with or without arms," "with or without a back," etc.) is problematic. Such an approach involves the risk of passing, without realizing it, from the analysis of a semantic field to that of a field of experience (psychological), ending up finally with the description of a field of "reality" (physical). The example of Pottier's classification of means of transportation, for example, followed upon his "seat" taxonomy, proves this point very well.
4. componential analysis, at the outset, takes as its object of study a microsystem within natural languages that is constituted by kinship terminology. The strange, unique character of this microsystem - the functioning of which can be compared only to that of person - presents as many advantages as disadvantages for analysis. The main advantages, which guarantee for componential analysis its homogeneity and rigor, are the purely paradigmatic nature of this code and its purely arbitrary and semantic character (the ego, which serves as a referene point for the whole system, cannot be identified with any referential human being). By using only a small number of semic categories - consanguity/marriage, lateral relationship/distant relationship (for the calculation of degrees of kinship), etc. - componential analysis succeeds in building an almost perfect taxonomic model. But its major disadvantage lies in the restricted nature of its field of applicability: attempt at extrapolation outside of this immanent microsystem - for example, for the study of botanical or zoological ethnotaxonomies in ethnolinguistics - run into difficulties comparable to those present in semic analysis.
5. Semic and componential analysis, inasmuch as it is defined as a clarification of paradigmatic relationships and an establishment of taxonomies considered as results of the combinatory principle alone, appears to be autonomous discipline with its own specificity and, as an indirect consequence, with a limited domain of application. The broadening of this field of research depends largely on the progress of semantics itself, which is slow in developing. Indeed, semantics, built upon the phonologicla model, is finding it difficult to introduce into its analyses the principles of syntagmatic and syntactic organization of the semantic universe.

1. When any knowable object is semiotic in character, it can be perceived either as system or as process - its two fundamental aspects. In such a case, the term syntagmatic is used to designate the process. In opposition to the paradigmatic axis (which is defined by relations of the type "either ... or") that identifiable entities have among themselves, the syntagmatic axis is characterized, in a first approach, by a network of relations of the type "both ... and."
2. We should emphasize the purely relational nature of syntagmatics in order to remove all ambiguity from this concept that suffes from unfortunate confusion. When it is identified with Saussurian speech, syntagmatics is considered as the realization of language, i.e., as provided with a different mode of existence, more "real" than paradigmatics; but such is certainly not the case. In addition, syntagmatics is often defined by lineraity. But this is only one mode of manifestation, either temporal or spatial, of the logico-relational structure - and is, therefore, atemporal and spatial - that syntagmatics is. The "both ... and" relationship is futher confused with the notion of "material" contiguity, whereas it should be interpreted only as the co-presence of entities within an utterance (sentence or discourse). As for contiguity (word order), it corresponds to one of the regulated constraints concerning the expression plane (i.e., the plane of suprasegmental phonology). We already know the important role played by the syncretism of the notions of linearity and contiguity in distributional analysis.
3. It is against this background, established by the relational network of co-presence (or of combinations), that more precise and more restrictive syntagmatic relations are worked out. Thus, Hjelmslev distinguishes three types of possible relations by recognizing, in addition to mere combinations, relations of selection (by which the presence of one term presupposes that of the other, but not inversely) and of solidarity (by which two terms mutually presuppose each other). Such an elementary typology then leads to the identification and formulation of syntagmatic units (or syntagms), definable by the relations that the constituent elements maintain among themselves and with the unit that subsumes them. Consequently, syntagmatics appear as a hierarchy of relations, a hierarchy arranged into successive levels of derivation.
4. Since every process presupposes the existence of a semiotic system, it is not possible to speak of different semiotic systems that are purely syntagmatic, for the simple reason that all types of discourse, although they possess a syntagmatic organization, are inscribed in intertextuality and are therefore in correlation with other discourses. At present, the difficulty in establishing a typology of discourse comes from the limitations of our knowledge; but from this gap, we cannot infer the absence of paradigmatic networks.

1. The terms of the dichotomy system/process, universal in character, are designed by Hjelmslev, when they are applied to semiotics, paradigmatic and syntagmatic. This dichotomy is essentially and solely based on the type of relation which characterizes each of its axes. The functions betweenthe entities situated on the paradigmatic axis are "correlations" (logical disjunctions of the type "either ... or"), while those which are located on the syntagmatic axis are "relations" (logical conjunctions of the type "both ... and"). Paradigmatics is thus defined as the semiotic system consituted by a set of paradigms joined together by disjunctive relations. This gives to it, in a first approximation, the form of a hierarchy, taxonomic in character.
2. Paradigmatics can be considered as the reformulation of the Saussurian concept of natural language, with this exception, however, that the Hjelmslevian system is not constituted by simple correlations between categories (defined at the same time by their mode of syntagmatic behavior). While for Saussure, "making sentences by putting words together" belongs to speech, the simultaneously paradigmatic and syntagmatic definition of category brings Hjelmslevian paradigmatics close to Chomskian competence (which contains the rules of sentence formation).
3. Literary semiotics places great emphasis on the projection of the paradigmatic axis upon the syntagmatic axis, a procedure which, as R. Jakobson claims, characterizes the mode of existence of a large number of poetic discourses. The fact is that the terms in paradigmatic disjunction can appear in conjunction (copresence) on the syntagmatic axis (for example, the antiphrasis can be manifested under the form antithesis). The generalization and more rigorous formulatio of this principle intuited by Jakobson has brought tolight the role of paradigmatic projections in the organization of narratice discourses, particularly in the narrative schema.

1. The sign is a unit of the manifestation plane constituted by the semiotic function, i.e., by the relation of reciprocal presupposition (or solidarity) that is established between entities on the expression plane (or signifier) and on the content plane (or signified) during the language act.
2. For F. de Saussure, who raised the issue of the linguistic sign, the latter results from the union of signifier and signified (which, in his first analysis, he identifies with acoustical image and concept). Although, in developing his theory, he was consequently led to refine these two notions by considering the signifier and the signified only inasmuch as they serve to constitute linguistic form (like the front and back of a sheet of paper), the term sign was commonly identified for a long time - and still is today - with the minimal sign, i.e., with the "word" or, more rigorously, the morpheme (or moneme for A. Martinet). it is in this sense that the all-purpose definition of language as "system of signs" is used.
3. The contribution of L. Hjelmslev to the theory of the sign is twofold: (a) by presenting the sign as the result of semiosis taking place at the time of the language act, he demonstrates that the dimension of the units of manifestation is not pertinent for the definition of the sign; in other words, in addition to minimal signs ("words"), we can also speak of utterance-signs or discourse-signs; (b) by postulating for each of the two planes of language (expression and content) the distinction between form and substance, he is led to specify the nature of the sign as the union of the form of expression and the form of content (thus, on the expression plane, it is the phonological, and not the phonetic, structure that enters into the constitution of signs).
4. The use of language produces, then, semiotic manifestation in the form of strings of signs. The analysis of signs, produced by the articulation of the form of expression and that of content, is possible only if the two planes of language are first disassociated the one from the other in order to be studied and described, each one separately. In other words, if the analysis of manifestation (aming at the recognition and establishment of minimal signs) constitutes the necessary first step, semiotic exploration actually begins with the study of units smaller than the minimal sign, and must be pursued separetely for each language plane, where the constituent units are no longer signs, but figures.
5. The extra- or para-semiotic meaning of the term sign does nonetheless exist and is sometimes introduced into semiotic or linguistic literature. In this case, sign commonly designates "something that is there in order to represent something else." As used in semiotics, it then denotes any form of expression assigned to translate an "idea" or a "thing" - corresponding to the concept of formant. Such a use of the term presupposes a particular conception of language, i.e., as a stock of "labels" to be attached to preexistent objects, as a pure and simple nomenclature (Hjelmslev).
6. Anglo-American linguistics, influenced by behaviorism, has either been scarcely interested in the issue of the sign, or else it has, under the influence of positivism, sought to introduce the notion of referent into the definition of sign, by constructing a triangular model of its interpretation (Ogden and Richards, following Charles S. Peirce). The three terms of the model consist of: (a) the symbol (= the signifier, or representamen for Peirce), (b) the reference (= the signified, or interpretant for Peirce), and (c) the referent (the denoted "reality," or object according to Peirce). Saussurian linguistics, as we know, considers the exclusion of the referent as the necessary condition of the existence of linguistics.
7. The problem of the referent broadens even more the gap that continues to separate two conceptions of linguistics, and especially of semiotics. Whereas the analysis of signs is for European semiotics but one step toward a description of the articulation network of forms, American semiotics (T. Sebeok) tends to stop at the level of signs and to proceed to a classification of these signs, based for a large part on the type of relation existing between the sign and the referent (for example, the icon is defined by a relation of resemblance, the idnex by a relation of "natural" contiguity, the signal by an artificial relation, etc.).
8. Another distribution of signs, of an instrinsic nature, seems possible; it would specify signs according to whether they belong to a given type of semiotic system (monoplanar, biplanar, pluriplanat).

1. By transformation can be meant, in a very general way, the correlation (or the estalishment of a correlaton) between two or among several semiotic objects: sentences, textual fragments, discourses, semiotic systems, etc.
Because of its origin, the term tranformation, in the European tradition, refers to linguistic comparativism, whereas, in the American context, it refers to procedures developed in mathematics; whence, especially in semiotics, frequent confusion and misunderstandings occur.
2. From the point of view of their field of application, we can distinguish, independently of their intrinsic nature, intertextual transformations (established between two or among several autonomous semiotic objects - be they paradigmatic or syntagmatic) and intratextual transformations. The latter are of two kinds: (2) transformations located at the level of deep semiotic structures, and (b) those that are established or identified between the deep levels and the surface levels of a semiotic object. For simplicity's sake, and following T. Pavel's example, we shall designate intertextual transformations as L-transformations (formulated and used by Lévi-Strauss and his disciples), horizontal intratextual transformations as G-transformations (which it is our task to define in paragraph 5 below), and vertical intratextual transformations as C-transformations (Chomskian and post-Chomskian).
3. Among intertextual transformations, we should first set aside the Proppian transformations. After having described the "morphology" of the Russian folktale, V. Propp tried to place his narrative model back into the historical dimension, attempting to identify the tranformations that the model can undergo in the course of its evolution. These transformations are described by Propp with the help of parameters of historical evolution. But Propp's parameters have a very dubious character, since, according to them, the marginal comes before the rational, the heroic comes before the humorous, the coherent before the incoherent. Propp's transformations are thus oriented transformations. In addition, the transformations described by Propp are local (they affect only one class of equivalences according to a sub-segment of his concept of "function"), isolated (the transformation produced in one place in the text does not affect the other syntagmatic positions), and superficial (they are located at the level of surface variants). One example will suffice to make us realize the imprecision and ineffectiveness of such "transformations": the house of the donor, representing at the surface level as a thatched cabin in the forest, having chicken legs on which it turns around, is "transformed" into a thatched cabin which, all other things remaining equal, does not turn around. Even from an atomistic point of view, such "transformations" cannot be compared with the historical changes described, in the 19th century, int ehf rom of "phonetic laws."
4. The concept of transformation, such as it was progressively developed and applied by C. Lévi-Strauss, possesses on the contrary a definite heuristic value. Since it is applied to very complex and diverse linguistic phenomena, it cannot receive a precise and homogenous formulation, as its author admits. Therefore, we can give only its main characteristics. The Lévi-Straussian transformation falls within the framework of a linguistic comparativism which would have been carried to its final conclusion:
(a) Thus, myth, for example, is defined by Lévi-Strauss neither as an ideal form, nor as a prototype that is historically or logically prior to all its variants, but as a structure of transformations (or of formal correlations) that undegirds all the variants of the myth, whether the variants be known or unknown, realized or not. The Freudian interpretation of the Oedipus myth is thus but one of the variants of this myth and is in a transformation relation with the other variants.
(b) Defined in this way, myths maintain, at a higher level, transformation relations with other myths (myths about the origin of fire are "transformed" into myths about the origin of water; those dealing with fire for cooking are "transformed" into myths about the origin of edible meats, etc.) in order, finally, to constitute "mythical systems" that are closed and circular (a continuous reading of the mythical transformations brings the reader back to the point of departure).
(c) These transformations are neither local nor isolated (as with Propp), but concomitant. A transformation affecting one segment of text (stemming from a paradigmatic class of equivalences) brings about, under conditions that remain to be specified, a concomitant transformation of another textual segment (belonging to another class of equivalences). It is clear that the noted concomitance makes it possible to imagine the possibility of a formal definition of the narrative syntagm.
5. As far as we are concerned, the transformations that we acknowledge, in the framework of narrative semiotics, are intratextual and syntagmatic: they complement Lévi-Straussian transformations, without contradicting them, since the latter are intertextual and paradigmatic. When located at the level of deep semiotic structures, transformations are considered as logical operations. On the logico-semantic plane, they are defined as the shifting from one term of the semiotic square to another, carried out through the operations of negation and assertion. On the narrative plane, closer to the surface, transformations correspond to operations of conjunction and disjunction between subjects of being (see state) and objects of value: these are elementary transformations. If we conceive of narrative discourse - and perhaps discourse in general - as "something that happens," i.e., as a trajectory leading from an initial state to a final state, then a transformation algorithm should be able to account for this trajectory: discourse then appears as a string of transformations.
While keeping the term transformation for these horizontal logical operations, and in order to avoid any ambiguity, we use the term conversion (conversions are similar to Chomsky-type transformations, but cannot be identified with them) to refer to the vertical reformulations of structures, created by the shifting from one level of semiotic depth to another.
6. In the typological framework that is thus constituted, we can try to locate the transformations of generative grammar. Setting aside their more or less formal character and considering them only from the point of view of conceptual theory, we may say that they are intratextual, vertical, oriented (going from deep structures to surface structures), and paradigmatic class of the sentence). As a conversion of deep structures into surface structures (or the shifting from one syntagmatic indicator to another, derived, indicator), transformations are here presented in the form of rewriting rules, that intervene only after the syntagmatic rules are and carried out on strings produced by syntagmatic rules (obviously, inasmuch as they allow for transformations, as determined by their structural analysis). Traditionally, there is a differentiation made between optional and obligatory transformations on the one hand, and, on the other hand, single and binary ones (the latter are generalized, in the case of embedding and coordination), depending on whether they concern one or two strings generated by the base.
The status of Chomskian transformations is difficult to define, and for several reasons: (a) they are "supplementary" rules with respect to the syntagmatic rules; (b) they are often heterogeneous in nature (a rule that is syntagmatic in itself may become "transformational" simply by its position in the grammar); (c) the very order of the rules (or their being placed into an algorith) is sometimes a problem as J. Lyons has pointed out, and the deep structures have to be disrupted so that the transformational system may be safeguarded.

1. The term arbitrariness (of the sign) is somewhat imprecise in Saussurian theory. It designates the character of the relation which unites the signifier and the signified. This relation, which consistutes the linguistic sign, is non-founded and unmotivated (i.e., impossible to interpret in terms of causality). This conception has played an important role historically, allowing F. de Saussure, among other things, to found the autonomy of (natural) language considered as form.
Though there exists no causal or "natural" relation between the signified "table" and the signifier "table," it is nonetheless impossible, when considering how language (or any semiotic system) functions, not to recognize the existence of a necessary realtion (E. Benveniste) - or of a reciprocal presupposition (L. Hjemslev) - between the signifier and the signified. This relation is called the semiotic function (L. Hjemslev) and its establishment (or semiosis) defines first and foremost the language act. Thus, logically this relation is necessary. It is also necessary from the social point of view: although the signs of a natural language are conventional (another term proposed by Saussure), they are not arbitrary, since speaking subjects cannot by themselves substitute either the signifiers or the signifieds.
2. The arbitrary or somewhat motivated character of signs is not a result of the nature of sign, but of its interpretation, i.e., of the feeling or attitude that a linguistic community or an individual has with respect to the signs they use. In this case we are dealing with meta-semiotic, and not semiotic, facts.
3. Another type of confusion can be avoided by situating the question of the arbitrariness of the sign only within the framework of bi-planar semiotic systems, to the exclusion of mono-planar semiotic systems, the units of minimal manifestation of which are not signs but signals (L. Hjelmslev).
4. The question of the arbitrariness of the sign, which deals with relations internal to semiotic systems, is bound up with the very different question of the external relations between a given semiotic system and the "reality" of the outside world, or of the relations between two different semiotic systems (for example, the problem of the "nameable" in pictorial semiotics). In the first case, we are dealing with problems concerning the status of the referent, in the second with the peculiarity of natural languages.
5. In somewhat the same way, L. Hjelmslev introduced the dichotomy arbitrariness/adequacy. He uses the term arbitrariness to designate theory - to the extent that, as a pure, coherent construction, it does not depend on the data of experience. By contrast, when a theory, or some of its premises, is applicable ot the data of experience, it is known as adequate (or consistent with its goal).
6. Finally, the question of the arbitrariness of the sign comes up once more when we deal with the construction of a meta-language (or a meta-semiotic system): the units, recognized and defined during semiotic description, are pure relational networks, and the names which may be conferred on them are arbitrary. However, if such a metalanguage is applied to an object semiotic system, the names chosen should carry as much information as possible about the manifestation.

1. The term metalanguage was introduced by the logicians of the Vienna school (R. Carnap) and, particularly, by those of the Polish school, who have shown the need "for distinguishing clearly between the language that we speak about from the language that we speak" (Tarski). The concept so created was subsequently adapted to the needs of semiotics by L. Hjelmslev, and of linguitics by Z. S. Harris. The morpheme "meta-" serves therefore to distinguish between two linguistic levels. that of the object-language (semiotic system) and that of metalanguage.
2. It suffices to observe the function of natural languages, to discover that they have the peculiarity of being able to speak not only of "things" but of themselves as well; to see that, according to R. Jakobson, they have a metalinguistic function. The existence of a multitude of metalinguistic expressions within natural language raises at least two kinds of problems: (2) On the one hand, when this set of expressions is brought together, could it constitute a metalanguage? In other words, would it possess the fundamental characteristics that define a semiotic system? (b) On the other hand, would the exclusion of all metalinguistic sentences permit us to obtain a pure language of denotation?
It is difficult to give a positive reply to these questions. What we can affirm, with some certianty, is the extremely complex character of natural languages: they can embrace a number of micro-universes that produce diverse and quasi-autonomous discourses.
3. Once Z. S. Harris had recognized the richness and importance of the metalinguistic elements of natural languages, he postulated the possibility that a given language has of describing itself, and the possibility as well for the linguist to construct a grammat as a metalanguage with the aid of materials found within the object-language (semiotic system). Such an attitude has no doubt left its mark upon American linguistics and explains a certain indifference on the part of generative semantics, for example, to a rigorous conceptualization of the descriptive language that it uses.
4. Benveniste as well holds metalanguage to be "the language of grammar"; but the consequences that can be drawn from such an affirmation are quite different. If we wanted to take over completely the heritage of comparative grammar instead of constructing new linguistic theories ex nihilo, then reflection upon the conditions of comparability of these languages would force us to admit that grammatical concepts used to this end must necessarily transcend the natural languages being compared. The possibility of comparison raises, for its part, the problem of the existence of language universals. In this case metalanguage can only be external to the object-language; it must be conceived of as an artificial language that contains its own construction rules. It is in this sense that we must interpret the theoretical efforts of L. Hjelmslev, for whom metalangbuage is a semiotic system, that is, a hierarchy, not of words or sentences, but of definitions which can take the form of a semiotic system or semiotic process. The hierarchy construction leading to the inventory of the latter, non-definable, conepts (which we can consider to be hypothetical universals) enables us then to construct an axiomatic system on the basis of which deduction will to some extent produce linguistics as a formal language, that is as a "pure algebra."
5. Metalanguage as conceived appears then to be a descriptive language (in the broadest and most neutral sense of the term). As such it may be represented in the form of several superimposed metalinguistic levels. Each level is meant (in the Polishs chool tradition) both to question and to found its immediately inferior level. We have in the past proposed a three-level distinction: descriptive, methodological, and epistmeologicla. The epistemological level controls the elaboration of models; the methodological level in turn spervises the conceptual tools used in the description stricto sensu.
6. It is equally fitting to maintain a distinction between metalangauge and the language of representation which we use in order to manifest the metalanguage. We know that different modes of representation - such as the use of parentheses, tree diagrams, rewriting, etc. - are homologus; they merely offer differrent ways of representing the same phenomenon or "reality." It is as if these languages of representation were, with respect to the metalanguage, in a relation comparable to that of Latin, Greek, or Arabic alphabets with respect to the natural written languages that they translate.
7. The set of problems concerning metalanguage, as summarized above, is inscribed within a limited framework; it concerns only natural languages taken as object-languages, and the metalanguage we are concerned with here is more or less coextensive with grammar (or grammatical theory). Semiotics as a theory of the set of "signification systems" has no choice but to go beyond this framework. It is certainly banal, for example, to say that natural languages are able to speak not only about themselves but also of other semiotic systems (painting, music, etc.) as well. In this case it is obvious that certain regions within natural language must be considered metalinguistic, or rather metasemiotic, with respect to the semiotic systems of which they speak. The problem of non-scientific metalanguages then arises in semiotics concurrently with the elaboration of a metalanguage with a scientific aim, which semiotics needs. The set of relations between linguistics and general semiotics (or semiology) is thus once more called into question.

While the term function is frequent in linguistics and, more generally, in semiotics, it is often employed - sometimes even within the same theory - with at least three different meanings: (A) with a utilitarian or instrumental meaning; (B) in an organicist meaning; (C) finally, in a mathematico-logical sense.
(A) Instrumental interpretation
1. For A. Martinet, the principal function of language is the communication function, language being a "communication instrument." Such a conception, which claims to attenuate the formalism of structural linguistics, in fact restricts the scope of linguistic theory. Even if language is communiaction, it is also production of meaning, of signification. Linguistic theory with such a restricted scope can no longer be extrapolated and applied to other semiotic systems (with the possible exception of "true" communication systems, such as the signals used to regulate traffic). Functional linguistics, as Martinet conceives of it, is a "realist" linguistics.
2. it is with this same instrumental meaning that the expression functional definition is used when it contains information concerning the use or the finality of the described object or behavior (cf. "a chair ... is used to be sat in"). The semantic analysis of lexemes of this type makes explicit either instrumental values or instrumental programs that these imply.
(B) Organicist interpretation.
1. E. Benveniste uses the concept of function with a meaning inspired in biology. For him it is an element necessary for the definition of structure: "What gives the character of a structure to a form is the fact that the constituent parts fulfill a function." This effort at conciliating structure and function allows him to reinterpret the diachronic linguistics of the 19th century, but also to justify the conception of the sentence as a structure the constitutent parts of which are charged with syntactic functions.
2. By syntactic function is traditionally understood the role that such or such an element, defined beforehand as a morphological unit (adjective, verb, etc.) or syntagmatic unit (nominal, verbal syntagm), plays within all that a sentence is. Subject, object, predicate, for example, are names of particular functions. Even if the inventory of syntactic functions does not take into account the hierarchy of the elements (the subject and the epiteth are not situated at the same level of derivation), this dimension of syntactic organization is still pertinent and can give rise to new reformulations in the framework, for example, of our actantial syntax. Generative linguistics, which started from a division of the sentence into syntagmatic classes, has itself been obliged to reintroduce this level of analysis under the guise of syntagmatic markers, where the subject is defined, for example, as the nominal syntagm immediately dominated by ∑.
3. It is in the framework of an epistemologicla reflection that certain psychologists (K. Bühler) or linguists (R. Jakobson) have been led to separate the functions of language (sorts of spheres of action working together toward the same goal) the set of which would exhaustively define linguistic activity. Thus, Bühler recognizes three principal functions in language: expressive, conative (summons), and referential (representation). Setting out these functions along the axis of communication, R. Jakobson adds three more to them: phatic, metalinguistic, and poetic. Such a distinction has the advantage of giving a suggestive general view of the different sets of issues concerning language: it would be risky to see anything else there. This schema cannot be considered as an axiomatic system on the basis of which a whole theory of language could be elaborated, by way of deduction. Nor is it a taxonomy of utterances. At the very most one could see therein possibilities of connotations of the "denotative" messages, of the postulations of connotative signifieds whose markers would need to be recognized on the level of the discourse. Philosophy of language no longer seeks to determine the functions of language on the basis of a general reflection on its nature, but on the level of the speech act, inscribed in the framework of intersubjectivity. A pragmatic approach succeeds in constituting lists of "functions" (of the type "request," "wish," "order," "expectation," etc.) which, while renewing the way the problem of communication are treated, still appear, at the present moment, as so many unscientific paraphrases, and do not constitute a coherent set.
4. In his Morphology of the Folktale, V. Propp designates the name of functions syntagmatic units which remain constant despite the diversity of narratives, and whose ordered sequence (31 in number) constitutes the tale. Such a conception, which provides the possibility of postularing the existence of a principle of organization underlying whole classes of narratives, has served as starting point for the elaboration of different theories of narrativity. As for the notion of function, still vague in Propp, it can be made more precise and reformulated in terms of narrative utterances.
5. G. Dumézil uses the term function to account for the tripartide division of society itself into three classes (priests, warriors, and farmers-cattle-raisers). The tripartide articulation of the ideological functions permits the attribution of a particular semantic field (a sovereign sphere) to each of the functions, at the same time that it establishes a hierarchical relation between them.
(C) Logico-mathematical interpretation.
1. Conscious of the difficulty, if not the impossibility, of totally excluding from linguistics the organicist meaning of function (which reflects, imperfectly to be sure, the productive and dynamic aspects of the activity of language), L. Hjelmslev has tried to give to this term a mathematico-logical definition. For him, function can be considered as "the relation between two variables," and he adds that this relation is to be envisaged as "a dependance which fulfills the conditions of the analysis," for it participates in the network of reciprocal interrelations, constitutive of every semiotics. Such a relation, named function, is established between the terms, called functives. It can be seen that Hjelmslevian linguistics is indeed functional, but in a very different sense than that of Martinet.
2. A synthesis of the two conceptions of function - those of E. Benveniste and of L. Hjelmslev - seems possible for a definition of the elementary utterance. Keeping the name of function solely for the "syntactic function" called predicate, and designating as actants other syntactic functions which, as functives, represent the end terms of the relation constitutive of the utterance, its canonical formulation can be given: F(A1, A2, ...). Minimal semantic investments of function as thus defined can permit the subsequent establishment of a first typology of elementary utterances. Thus, at first, it has seemed economical to distinguish between, on the one hand, utterances constituted by a functon and, on the other, those the predicate of which would be a qualification (corresponding to existence propositions in logic). The application of this opposition to narrative analysis opened the way to two types of research. While the functional model accounted for the ordering of narrative utterances defined by their functions (= "functions" in the Proppian sense), the qualification model provided the possibility of describing the manner of being of semiotic objects, considered in their taxonomic aspect. However, the proposed distinction was in contradiction to the structural postulate according to which a relation, whatever kind it be, can be installed (or be recognized) only between at least two terms (in this case, between two actants). But qualitative utterances are precisely presented as single-actant utterances. It has thus proved necessary partially to rework the definition of the elementary utterance, assimilating qualificative utterances to utterances of state (characterized by junction between the subject and the object) and opposing them to utterances of doing (having transformation as function). In this pespective, function can then be defined as the constitutive relation for every utterance.
3. L. Hjelmslev calls semiotic function the relation which exists between the expression form and the content form. Defined as reciprocal presupposition (or solidarity), this relation is constitutive of signs and, by this fact, is creative of meaning (or, more precisely, of meaning effects). The speech act consists essentially in the establishment of the semiotic function.

1. In his classification of signs, C. S. Peirce opposes index both to icon (which involves a relation of resemblance) and to symbol (based on a social convention). For him, the index sets up a relation of "natural" contiguity, linked to a fact of experience not provoked by human beings.
2. For L. Prieto, who emphasizes the mechanism of indication (uner all its possible forms), the index is to be understood, in a much broader sense, as "an immediately perceptible fact which causes us to know something about something else which is not perceptible." From his point of view, the signal is only a particular form of the index.
3. If it be admitted, as in Saussurian linguistics, that the exclusion of the referent is an a priori which is necessary for the exercise of any semiotics, then it must be acknowledged that the index - in the two meanings indicated above - is in the category of non-signs.
4. In his concept of narrative, R. Barthes has proposed an opposition of index and informant. While the informant is a "realistic operator," which serves to authentify the referent's reality (for example, the exact age of a person), the index is constituted by a set of notations (for example, the notations relative to a personality type, to a sentiment) which, rather than being immediately signifying data (as in the case of the informant), have only "implicit signifieds." Thus, for example, the description of a landscape, of an object, is sometimes used to inform us indirectly concerning the psychology or the destiny of a character. As is clear, this meaning converges with the common use of the word index.

1. For L. Hjelmslev, the symbol is an entity of monoplanar semiotics, which can receive one or several interpretations. By contrast with bi-planar semiotic systems, the Danish linguist thus reserves the name of symbol systems to monoplanar semiotic systems. As a non-sign, the symbol is thus differentiated from the sign, which is an entity of bi- or pluri-planar semiotic systems.
2. One may retain the expression molar symbol (also termed by Hjelmslev - but improperly it would seem - isomorphic symbol) in order to designate, in the sense Saussure gives to symbol, an entity (possibly inscribed in a biplanar semiotic text, but having its own autonomous status) which, in a given socio-cultural context, can yield only one interpretation and which, contrarily to the sign, does not allow for subsequent analysis into figures (for example, the scale, symbol of justice). Such symbols may be catalogued, but they do not, strictly speaking, constitute symbol systems.
3. It is with a similar meaning that C. S. Peirce defines the symbol as founded on social convention, as opposed to the icon (characterized by a relationship of resemblance with the referent) and to the index (based on a relationship of "natural" contiguity). Ogden and Richards, on the other hand, attempted a clumsy synthesis between the Saussurian conception of the sign and the traditional definition of the symbol; in their triangular model, the symbol corresponds to the Saussurian signifier, the reference to Saussure's signified, whereas the referent denotes "reality."
4. In its non-linguistic and non-semiotic uses, the term of symbol accepts multiple and verying definitions, such as "that which represents something else by virtue of an analogical correspondence," or "absence made presence," etc. In all these cases, its sign nature is not challenged, since the complementary determinations which are added thereto refer at times to the pluriisotopic character of the discourse, and at other times they refer to the still ill-examined mehcanism of connotation, etc. The use of this syncretic term is ambiguous and should, provisionally, be avoided in semiotics.
5. In scientific metasemiotic systems, a symbol is a conventional type of drawing (using geometrical figures, letters, etc.) which is used to name univocally a class of entities, a type of relation and/or of operation. Symbolic notation is to be considered as a visual mechanism for representing constitutant units of a metalanguage. In the narrower sense, the term symbol applies, in the first place, to the representatives of the entity classes. Thus it is said that a finite set of symbols (from a to z, for example) constitutes the alphabet (which mroe or less corresponds to traditional "morphology"). Algebraic and logical notation had accustomed us to using letters as symbols of classes, whereas figures (equal signs, multiplication signs, etc.) are reserved for representing relations and operations: these figures are sometimes called operational symbols. In the case of tree representation, non-terminal symbols are used to label the nodes of all the levels, except for the final level, the symbols of which, called terminal, can be replaced by lexical items through the application of the rules of lexical insertion. The tree's bbranches are assimilated to operational symbols, charged with representing the operations of concatenation and of derivation. It is thus clear that Hjelmslev's definition of the symbol as a unit of a monoplanar semiotic system (par. 1 above) can be identified with the definition of the entities of a scientific metasemiotic system.

1. As a term belonging to rhetoric, metaphor designates one of the figures (called tropes) which "modify the meaning of words." Now this term is used in lexical or phrastic semantics to name the result of the substitution of one lexeme by another performed in a given context on the basis of semantic equivalence. Since the studies devoted to the area of metaphor could fill up a library by themselves, we cannot even pretend to sketch an overview of this literature here. We shall therefore limit ourselves to some remarks concerning its role and the way it functions in the framework of discourse semiotics.
2. When it is considered from the point of view of the reception structures, metaphor appears as a foreign body (as an "anomaly" in the generativist perspective). Its readability is always equivocal even when it is guaranteed by the discoursive trajectory in which the metaphor is inscribed (the contextual semes, by integrating it into the discoursive trajectory, make it into a sememe). Indeed, the metaphoric lexeme proposes itself as a virtuality of which the radings are multifold but suspended by the discoursive discipline. It is this virtuality, this number of suspended readings which produces a meaning effect of semantic "richness" or "thicness." For example, the rose which is substituted for "young woman" is, of course, read as "young woman," yet it also manifests for a moment the virtualities of color, form, scent, etc.
3. From the point of view of its origins, metaphor it obviously not a metaphor but a common lexeme. Separated from its context it should be viewed as a "nuclear" figure possibly bringing with it, when it is transferred, some but not necessarily all of the semes belonging to its original context. For example, in the above case of the transfer of "rose," this lexeme does not keep the contextual seme plant. This transferral of lexematic figures accounts for the fact that the discourse in which the metaphor is inserted tends to develop into a figurative discourse.
4. In the perspective of the generative trajectory of the discourse we are primarily interested in metaphorization (and not with metaphor) as a procedure of discoursive production. R. Jakobson righly emphasized the paradigmatic character of this procedure. Indeed, metaphorization, as the substitution of one semiotic entity for another, presupposes the existence of a paradigm of substitutions. In this sense it can be said that all the sememes of a natural language which shares at least one common (or identical) seme virtually constitute a paradigm of substituable terms. (This is what allowed F. Rastier to say that this iterative seme constituted an isotopy.) Yet - and it is at this point that Jakobso's thesis becomes dubious - the paradigmatic relations are meaningful precisely only insofar as they create meaning, in other words, insofar as they create differences by means of oppositions, in the framework of each paradigm, between what is kept by the discourse and what is excluded by it. The creation of differences is indeed the only way, since F. de Saussure, to conceive of the production and/or perception of signification. By contrast, Jakobson's "poetic function" consists of exploting, by the substitution procedure, paradigms of resemblances, and not paradigms of differences. In fact this amounts to abolishing meaning (Is it not to this totalizing of meaning that the Baudelairian "correspondences" tend?). It is possible that poetic discourse aims, by its redundances, at abolishing meaning. Yet it cannot do so, thanks to (or becayse of) the syntagmatic axis, which maintains signification unaltered, through the elaboration of figurative isotopies.
5. Metaphorization can thus be interpreted as a paradigmatic substitution of figures, which substitution is obtained, on a common semic basis, by the neutralization of the other semes of the same figure. This interpretation of metaphorization also permits an account to be made of the other "anomalies" of the utterance's semantic functioning. As is well known, a seme is not an atom of meaning, but is the term of a semic category. Consequently, the substitution procedure which, instead of taking up the same seme, aims at imposing the contrary (or contradictory) seme belonging to the same semic category, produces an antiphrasis. For example, one says, "Big deal!," to express contempt for the pettiness evidenced by someone else. Similarly , the semes participate in hypotactic (or hypertactic) seme belonging to the same sememe, then the result of the operation can be called metonymy (a kind of deviant metaphor). Of course, these are not "real" defintiions, but suggestions that semantics can offer concerning the way of formulating answers to the problems raised by the figures of rhetoric.
6. From the perspective of discourse semiotics, these procedures of semantic substituion are primarily of concern to us as isotopic connectors. While the metaphor usually functions in the framework of a sentence and can be apprehended and described in this context, it becomes a discoursive fact only when it is prolonged, that is, when it constitutes a transsentential figurative isotopy. In this case, the procedures of paradigmatic substitution that we have discussed above present themselves as initiators of isotopies, and then, at regular intervals, as maintainers or isotopic connectors, linking isotopies refer either to other figurative isotopies, or to more abstract thematic isotopies. Taking as starting point a semantic isotopy, one can then designate, according to the nature of the connection - metaphor, antiphrasis, metonymy, etc. - the other isotopies of the discourse as metaphoric, antiphrastic, metonymic, etc.

1. Traditionally, the rherotical figure known as metonymy (which includes the more specific instance of synecdoche) designates the linguistic phenomenon in which a given sentential unit is substituted for another unit to which it is "linked" (in a relation of contained to contianed, of cause to effect, of part to whole, etc.).
2. Interpreted within the context of discoursive semantics, metonymy is the result of a substitution procedure by which a given seme, for example, is replaced by another hypotactic (or hypertactic) seme; both semes in question belong to the same sememe. From this perspective we may consider metonymy a "deviant" metaphor. C. Lévi-Strauss has remarked that within mythic thought "every metaphor ends in metonymy", and that all metonymy is by nature metaphorical. His remark is easily interpreted if we take into account that in both of these rhetorical figures a substitution phenomenon is indeed produced on the basis of a semantic equivalence.

1. By contrast with anaphoras (or cataphoras) which, within the discourse, refer to given units or segments, deictics (or markers, for E. Benveniste) are linguistic elements which refer to the domain of the enunciation and to its spatiotemporal coordinates: I, here, now. Thus pronouns ("I," "you") can serve as deictics, as well as adverbs (or adverbial phrases), demonstrative adjectives, etc. It is clear that we are here dealing with uttered enunciation as it is manifested by the interplay of procedures of disengagement and engagement which simulate the established or the abolition of a distance between the discourse's utterance and the domain of its uttering.
2. It can be further noted that the use of deictics permits the referentialization of the discourse - that is, a simulation of the linguistic existence of an external referent which actually is a correlation between the specific semiotic system that natural language is and the semiotic system of the natural world, both having their specific organization.

1. The deixis is one of the fundamental dimensions of the semiotic square; by the relation of implication it links one of the terms of the axis of the contraries with the contradictory of the other contrary term. Two deixes cna thus be reorganized: the one (s1 - s2) is called positive; the other (s2 - s1) negative, although these qualifications do not involve an axiologic investment. This investment only appears following the projection of the thymic category, euphorica/dysphoria, upon the semiotic square.
2. In a given narrative, temporal positions (now/there) can be postulated as deixes of reference on the basis of which temporal, aspectual, and spatial categories can develop. Thus, what is sometimes termed the "time of the narrative" appears as a present (identifiable with a "then" deixis) by relation to which a past and a future can be installed, conforming to the logical system anteriority/concomitance/posteriority.

1. Taken as an utterance, text is opposed to discourse, as concerns the substance of their expression (graphic for the former, phonic for the latter) which is used in the manifestation of the linguistic process. According to certain linguists (for instance, R. Jakobson), oral expression (and, consequently, discourse) is the first given; writing would then be but a derivative, a translation of oral manifestation. For others (such as L. Hjelmslev), on the contrary, the genetic point of view is not pertinent, since a semiotic form can be manifested by different substances.
2. The term text is often used as a synonym of discourse, primarily because of terminological mixing with natural languages that do not posses the equivalent of the word discourse (as between French and English). In that case, textual semiotics is theoretically not distinguished from discoursive semiotics. The two terms - text and discourse - may be indifferently applied to designate the syntagmatic axis of non-linguistic semiotic systems; a ritual, a ballet may be considered as text or as discourse.
3. L. Hjelmslev used the term text to designate the totality of a linguistic string, unlimited because of the system's productiveness. It is the identification and choice of units having maximal dimensions and recurrent in the text that make it possible to undertake their analysis and to determine for example, the type of linguistics (or grammar) that can be constructed. If the recurrent unit adopted is the sentence, then the type of linguistics that is developed to account for it is called phrastic; the choice of discourse as the maximal recurrent unit of text gives rise to the construction of a discoursive linguistics.
4. The term text is sometimes used in a narrow sense, when the nature of the chosen object (the work of a writer, a set of known documents of collected accounts) imposes limits on the term. In this sense, text becomes synonymous with corpus.
5. In meanings (3) and (4), text designates an entity prior to its analysis. However, we already know that analysis always presupposes the choice of a level of pertinence and seeks to recognize only certain types of relation, exclusing those which could just as well be determined (substance or form, syntax or semantics, etc.). The outcome of this is a new definition, according to which a text is made up only of those semiotic elements fitting the theoretical goal of the description. it is in this sense that we can speak, for example, of the utterative text (obtained after the elimination of the marks of enunciation). It is also in this sense that it is possible to interpret "text as productivity" (J. Kristeva), a concept that subsumes the set of operations of production and transformations of the text, and which attempts at the same time to account for the semiotic properties of enunication and utterance.
6. When the generative trajectory is interrupted, it leads to textualization (linearizaion and junction with the expression plane). The text obtained through this procedure is the equivalent of the semantic representation of the discourse. From the point of view of generative grammat, the text, as semiotic representation, can then serve as the deep level for the linguistic structures which generate surface linguistic structures.

1. Given the richness of the semantic field covered by the concept of ideology and the numerous ambiguities which result from tis different possible interpretations and definitions, it can be hoped that the semiotic approach might make the conept a bit more precise.
2. Thus, it seems advantageous to distinguish two fundamental forms of organization of the universe of values: their paradigmatic and syntagmatic articulations. In the first case, the values are organized in systems and are presented as valorized taxonomies which can be designated by the name of axiologies. In the second case, their mode of articulation is syntactic and they are invested in models which appear as potentialities of semiotic processes. When these models are set in opposition to the axiologies, they can be considered as ideologies (in the restricted, semiotic, meaning of this word).
3. Values, participating in an axiology, are virtual and are the result of the semiotic articulation of the collective semantic universe; they thereby belong to the level of deep semiotic structures. By being invested in the ideological model, they are actualized and are taken up by an individual or collective subject which is modalized by wanting-to-be and, subsequently, by wanting-to-do. This is to say that an ideology belongs to the level of surface semiotic structures, and thus that it can be defined as an actantial structure which actualizes the values that it selects within axiological systems (of a virtual order).
4. An ideology is thus characterized by the actualized status of the values that it takes up. The realization of these values (that is, the conjunction of the subject with the object of value) abolishes, ipso facto, the ideology as ideology. In other words, ideology is a permanent quest for values, and the actantial structure which informs it must be considered as recurring in every ideological discourse.
5. Considered as a domain of the generative trajectory taken as a whole, the ideological organization presents values, that it takes up, under their abstract or thematic form. However, ideological discourse can, at any instant, be more or less figurativized and be thus converted into mythologicla discourses.

1. In contrast to operation (as an act of humans upon things), manipulation is characterized as an action of humans upon other humans with the goal of having them carry out a given program. In the first instance, what we have is a "causing-to-be," in the second a "causing-to-do." Both forms of activity, one of which is inscribed in the pragmatic domain and the other in the cognitive domain, thereby correspond to modal structures of a factitive sort. When projected upon the semiotic square, manipulation as a causing-to-do gives rise to four possibilities:
2. As a discoursive configuration, manipulation is undergirded at one and the same time by a contractual structure and a modal structure. We have, in effect, a communication (intended to cause-to-know) in which the sender-manipulator pushes the receiver-manipulatee toward a position wherein freedom is lacking (not-being-able-not-do-do), to the point that the latter is obliged to accept the proposed contract. What is at stake at first glance is the transformation of the receiver-subject's modal competence: If the latter connects not-being-able-not-do-do with having-to-do, we would have provocation or intimidation; if the subject adds a wanting-to-do to not-being-able-not-to-do it would be instead a case of seduction or of temptation.
3. Situated syntagmatically between the sender's wanting and the receiver-subject's actual realization of the narrative program (proposed by the manipulator), manipulation plays upon persuasion and thus articulates the persuasive doing of the sender and the interpretive doing of the receiver.
(a) The manipulator can exercise its persuasive doing by relying upon the modality of being-able: on the pragmatic side it will propose either positive objects (cultural values) or negative objects (threats) to the one manipulated. In other instances the manipulator will persuade the receiver with the help of knowing: on the cognitive side it will bring the receiver to know what the manipulator thinks of the former's modal competence, in the form of positive or negative judgments. Thus it appears that persuasion in terms of being-able characterizes temptation (where a positive object of value is proposed) and intimidation (presentation of a negative gift); persuasion in terms of knowing is peculiar to provocation (with a negative judgment: "You are incapable of...") and seduction (manigesting a positive judgment).
(b) The one manipulated is led correspondingly to perform a pragmatic doing and necessarily to choose, either between to images of its own competence (positive in the case of seduction, negative in the case of provocation), if it is a manipulation in terms of knowing; or between two objects of value (positive in the case of temptation, negative in the case of intimidation), if the manipulation plays upon being-able. (Of course, such an elementary typology of the forms of manipulation is only tentative; at least it sketches out an avenue of research.)
4. At the level of the receiver's modal competence, and taking into account only the modality of being-able-to-do we can anticipate four separate positions
On the basis of this approximate lexicalization (as indicated between the parentheses) of the modal structures, we can propose names (within our socio-cultural universe) for the ypoes of sub-codes of honor that the manipulation puts into play (that is, from the receiver-subject's point of view): the codes of "sovereignty" (freedom + independence + powerlessness). The action that the receiver-manipulatee will carry out following manipulation by the sender thus becomes a simple, practical narrative program for that receiver; its primary narrative program is conjunction with honor (in the case of a manipulation on the plane of knowing) or with a given object of value (if the manipulation is based upon being-able).
5. As a causing-to-do, it appeast that manipulation must be recorded as one of the essential components of the canonical narrative schema. The system of exchange or, more precisely, the contract that is registered there, is taken over, as it were, at a hierarchically superior level by the structure of manipulation. In this case, indeed, the relation between Sender and Receiver is not one of equals (as in the simple operation of exchange, which calls for two subjects of comparable competence), but of superior to inferior. Moreover, the manipulation effected by the sender will call for sanction by the Sender-judge; both operations are located on the cognitive dimension (in contrast to the performance of the receiver-subject carried out on the pragmatic plane).
6. Even if, as we have noted, manipulation is just beginning to be analyzed, we can nonetheless foresee, by transposing it from the plane of narratives to that of somatic practice, the development of a true semiotics of manipulation (corresponding to a semiotics of sanction and a semiotics of action); at the very least we know how important a place it holds in human relations. Such a semiotics must be establishable on the basis of the initial Sender's narrative trajectory and account not only for the manipulation of the subject (some possible forms of which we have just mentioned) but also that of the anti-subject (with the strategy of the trick which makes possible operations of "cooptation" and "infiltration," for example).

1. Sometimes opposed to the couple phonetics/phonology, sometimes to syntax (especially in logic), semantics is one of the components of the theory of language (or of grammar).
2. During the nineteenth century, linguistics dealth principally with the elaboration of phonetics and morphology; in the twentieth century, as if by a reversal of trends, it has especially undertaken the development of syntax and semantics. Indeed, it was not until the end of the last century that M. Bréal first formulated the principles of a diachronic semantics; its main task was to study changes of word meaning, by adapting to the social dimension of natural languages the tools of ancient rhetoric (especially tropology) and of nineteenth-century stylistics.
3. In abandoning the diachronic dimension of research in favor of a synchronic description of occurrences of signification, semantics, during the first half of the twentieth century, took on the task of recognizing and analyzing semantic (or notional, or conceptual) fields. Beginning with the work of J. Trier, who also made us of semasiological and onomasiological approaches, it received the name of lexicology (G. Matoré). Nevertheless, such a lexical semantics maintains the word as the basic unit of analysis, and is close to the hypothesis of Sapir and Whorf, concerning the categorization of the world with the help of the lexical mechanism of natural languages. This approach, with its taxonomical goal, only produced partial and limited results, for lack of criteria stemming from the immanent structure of language.
4. It was in the 1960's that the use of the phonological model (based on the more or less explicit postulate concerning the parallelism of the two planes of language) opened the way to what is commonly called structural semantics. Considering the expression plane to be made up of differential gaps of the signifier to correspond to gaps of the signified (interpreted as distinctive features of signification), this new approach thereby possesses a means to analyze explicit lexical units (morphemes or comparable units), by decomposing them into smaller underlying units (sometimes called minimal units), i.e., semantic features or semes. Whatever may be the theoretical presuppositions of those linguists engaged in this research (to give a few names, U. Weinrich, B. Pottier, A. J. Greimas, Aspresjan, Katz and Fodor), and without considering the diversely satisfying results obtained by these linguists individually, structural semantics undeniably constitutes a decisive step forward: its methodological experience has made possible new reflections on the theory of signification and has opened the way to semiotics.
5. As it exists today, semantics seems to have dismissed the apprehensions of a good many linguists, condensed in Bloomfield's famous formula: meaning does exist, but we cannot say anything meaningful about it. Indeed, although the "materiality" of the signifier serves as a sort of guarantee for its scientific description, the plane of the signified - which can only be presupposed - resisted a positive approach. In order for semantics to be admitted and recognized as a constructed language capable of describing an object-language, a revolution of thought had to take place. It is a matter of replacing the certitude of a description of "facts" of language by the idea that linguistics is but a theoretical construction, seeking to account for phenomena that are otherwise elusive (and that cannot be seized directly). Even then, one must realize that the status of semantics as a metalanguage creates a division among semanticists, of which they are more or less conscious. In addition to the demanding project of developing a scientific metalanguage (a project in which we are involved), semantic language is often considered as a mere paraphrase into natural language.
6. Among those problems that have been left unanswered and that semantics should resolve, let us first point out the problem of semic production. We can theoretically imagine that some twenty binary semic categories, considered as the taxonomic basis of a combinatory system, can produce several million sememic combinations; at first glance, this number is quite sufficient to encompass the semantic universe coextensive with a given natural language. Leaving aside the practical difficulty of setting up such a basic set of semantic universals, we run into another problem, just as difficult, when we attempt to define rules of semantic compatibility and incompatibility, which govern not only the construction of sememes, but also broader syntagmatic units (utterance, discourse). It is therefore clear that semic (or componential) analysis obtains satisfying results only when it produces limited taxonomic descriptions (which can be extended to the structuration of more open semantic fields). It is also clear that the idea that we have matrices at our disposal (for the purpose of semantic interpretation), that are comparable to those that phonology can furnish for its own interpretation, must be abandoned; finally, linguistic semantics (generative or logical, cf. O. Ducrot) is limited to clarifying only possible universals. Thus, the great illusion of the 1960s - i.e., the possibility of providing linguistics with the necessary means for an exhaustive analysis of the content plane of natural languages - had to be abandoned, since linguistics had gotten engaged, often without realizing it, in the extraordinary project of a complete description of all cultures, even embracing all of humanity.
7. In order to move beyond the development phase we have briefly outlined, semantics - as we have set out to develop it within the Groupe de Recherhes śemio-linguistiques - must satisfy, it would seem, at least three main requirements: (a) it must be generative, in the form of progressive investments of content, situated on successive levels, leading from the most abstract investments toward the most concrete and figurative, in such a way that each level may receive an explicit metalinguistic representation; (b) it must be syntagmatic, and not simply taxonomical, in order to explain, not particular lexical units, but the production and reception of discourse. Concerning this point, the importance given to contextual semes in the construction of sememes allows us to postulate the following hypothesis: the deepest semantic investments correspond to syntagmatic units having the broadest dimensions and serving as a base for establishing isotopies of discourse; in this way, additional layers of semantic investment will give rise to making contents more specific, breaking discourse down into smaller syntagmatic units in order, finally, to arrive at sememic combinations; (c) Semantics must be general: natural languages, just like natural worlds, are loci where many different semiotic systems appear and are produced; it is therefore necessary to postulate the unicity of meaning and to recognize that meaning may be manifested by different semiotic systems at the same time (in the case of theater, for example). That is why semantics stems from a general theory of signification.
8. Within the framework of semiotic grammar as we conceive of it, two complementary components - syntactic and semantic - are to be distinguished; these two components can be articulated at two different levels of depth. The generative process of discourse thus presents two semantic domains at the semiotic or narrative level: (a) fundamental semantics, equipped with an abstract logical representation; and (b) a narrative semantics, the investments of which are inscribed within the molds of the surface narrative syntax. The resulting semantico-syntactic representation is that of semiotic structures, which can be assumed by the domain of the enunciation for the purpose of producing discourse.

1. In a first perspective, the concept of discourse can be identified with that of semiotic process. In this way the totality of the semiotic facts (relations, units, operations, etc.) located on the syntagmatic axis of language are viwed as belonging to the theory of discourse. When one had in mind the existence of two macrosemiotic systems - the "verbal world" manfiested in the form of natural languages, and the "natural world" manifested as the source of non-linguistics semiotic systems - the semiotic process appears as a set of discoursive practices: linguistic practices (verbal behavior) and non-linguistic practices (signifying somatic behavior manifested by the sensory orders). When linguistic practices alone are taken into consideration, one can say that discourse is the object of knowledge considered by discoursive linguistics. In this sense discourse is synonymous with text. For indeed, certain European languages which do not have an equivalent for the French and English word "discours(e)" have been led to substitute for it the word "text" and to speak of textual linguistics. On the other hand - by extrapolation and as an hypothesis which seems to be fruitful - the terms discourse and text have also been used to designate certain non-linguistic semiotic processes (a ritual, a film, a comic strip are then viewed as discourses or texts). The use of these terms postulates the existence of a syntagmatic organization undergirding these kinds of manifestations.
2. In a somewhat different theoretical framework - but one which is not contradictory with the preceding one - discourse can be identified with utterance. The more or less implicit way in which the utterance (= what is uttered) is conceived determines two theorecial attitudes and two different types of analysis. For phrastic linguistics, the basic unit of the utterance is the sentence. The discourse is then viewed as the result (or the operation) of the concatenation of sentences. By contrast, discoursive linguistics - as we conceive it - takes as its basic unit the discourse viewed as a signifying whole. Consequently, sentences are only segments (or broken-up parts) of the discourse-utterance. Of course, this does not exclude the possibility that a discourse might have at times the dimensions of a sentence, as a consequence of the procedure of condensation.
3. When analysis of discourse is located in the wake of phrastic grammars, it seeks to recognize - and to construct models of - the discoursive sequences viewed as series of sentence-utterances. To this end, various procedures are devised or proposed, such as: (a) the establishment of an equivalence network among sentences and/or series of sentences (Z. Harris); (b) the formulation of rules - either logical or rhetorical rules - of sentence concatenation; (c) the determination of grammatical isotopies of the sequences (taking anaphorization into account); (d) the elaboration of deeper representations accounting for the series of surface sentences, etc. Although such procedures are pertinent, they remain nonetheless partial and do not seem to be based on any general theory of discourse. They remind us too much of those tasks of "paragraph construction" set down in programs for secondary education in France and could be followed, in the same turn, by the equally classic "discourse construction" with its three pints...
4. When, by contrast, it is postulated from the outset that the utterance-discourse forms a whole, then the procedures to be set up must be deductive - and no longer inductive. The procedures must consist of the analytic separation of the discoursive whole into its components. When, furthermore, a generative approach complements these procedures, semiotic theory is led to conceive of the discourse as a multilayered organization constituted by a number of depth levels, superposed on each other. Only the last one - the most superficial one - can receive a semantic representation which is approximately comparable to the "deep" linguistic structures (in the Chomskian perspective). Thus it appears that phrastic grammar is the natural sequel to discourse grammar.
5. For such a conception of discourse to be integrated into the general theory of language, it needs, on the one hand, to be homologated with the fundamental dichotomies language/speech, system/process, competence/performance (see these terms), and on the other hand, to be related to the domain of the enunciation. Keeping the term competence to designate all the necessary conditions for the exercise of the enunciation, a distinction between two autonomous configurations of this competence will be noted: semio-narrative competence and discoursive competence (in the strict sense). Semio-narrative competence is situated upstream, being anterior to the enunciation as such. In agreement with Hjelmslev and Chomsky, one can conceive of it as made up of articulations which are both taxonomic and syntactic - and not as a merely paradigmatic organization, as "language" is for Saussure. In agreement with Saussure, semio-narrative competence can be viewed as having a transcendental status: the semio-narrative forms are postulated as being universal - found in all linguistic and translinguistic communities - and persevere in the translation from one language to another; they are also recognizable in non-linguistic semiotic systems. Thus semio-narrative competence corresponds to what could be termed classificatory and programming forms of human intelligence: but this would be an irresponsible formulation. As competence, it can be described as a fundamental grammar of the discourse-utterance, preceding the enunciation and presupposed by it. By contrast, discoursive competence is located downstream from semio-narrative competence: it is constituted during the enunciation, and governs, by molding them, the uttered discoursive forms.
6. These brief remarks about the twofold nature of competence were necessary in order to establish a new meaning and a more restricted definition of discourse. For indeed, if enunciation is, according to Benveniste, the "putting into discourse" of language, then discourse is precisely what is put in place by the enunciation. When we substitute in Benveniste's definition the concept of semio-narrative competence for the concept of "language," we can then say that putting into discourse - or discoursivization - consists in taking over the semio-narrative structures and transforming them into discoursive structures. The discourse is the result of this manipulation of the deep forms, which brings about a surplus of signifying articulations. A discoursive analysis, distinct from the narrative analysis that it presupposes, can then be envisaged.
7. Such a conception of discourse nullifies the traditional opposition between between discourse (as transphrastic monologue) and communication (as dialogue and phrastic exchange). Thus communication ceases to be an extra-linguistic structure which serves as a basis for the exchange of messages. Communication appears rather as a domain, a stage of the generative trajectory of the discourse. At times, it brings about the appearance of a single subject-actor of the enunciation, which assumes and projects beyond itself various actantial roles. At other times it manifests a bipolar actorial structure which produces a two-voice discourse (= "communcation") but which is nevertheless located on a homogeneous semantic isotopy and the syntactic forms of which are comparable to those of the dialogue installed, after having been enunciated, in the discourse-utterance. Furthermore, the structure of communication no longer needs an exterior pragmatics (in the American sense) in order to be understood and described. The actants of the enunciation, because they assume a semio-narrative competence which transcends them and causes them to participate in the semiotic universe, are by definition competent: they "know how to communicate" without the help of any psycho-sociologicla factors.
8. The fact that the term discourse is progressively identified with semiotic process and that it is even used to designate, metonymically, one or the other entire semiotic system or semiotic process, raises the problem of the definition of the semiotic system (as both object of knowledge and object constructed by the description). Indeed, we must keep in mind that linguistics is at the origin of semiotic research. We must also remember that natural language is not merely defined as a semiotic system (as a language) but is also viewed, explicitly or implicitly, as a model according to which the other semiotic systems can and must be conceived. Yet, natural language, semantically coextensive with culture, is a huge domain. We consider it to be a macrosemiotic system which can only be compared with another macrosemiotic system having the same dimensions: that of the signifying natural world. Consequently the other semiotic systems appear as "minisemiotic systems" located or constructed within these universes. Soviet semioticians were perhaps the first to raise this possibility when proposing the ill-defined but quite suggestive conept of "secondary modeling systems" to designate these "minisemiotic systems" which, even though they belong to the "macrosemiotic systems," are presumed to have an autonomy of operation and/or of signification. The Soviet concept of "secondary system" (a metonym includes process) approximately corresponds to the concept of discourse (a concept which was developed in the French context and which must be interpreted as a process which presupposes a system).
9. In this new meaning, the term discourse nevertheless remains ambiguous. A semiotic domain can be named discourse (for example: literary discourse, philosophical discourse) because of its social connotation related to a given cultural context (J. Lotman says, for instance, that a medieval sacred text is viewed by us as literary) independently from and prior to its syntactic or semantic analysis. The typology of discourses which can be elaborated in this perspective is therefore connotative, that is, it is specific to a given cultural localization (circumscribed both geographically and historically) and it does not have any relation with the semiotic status of these discourses.
10. Even when one leaves aside the connotative definitions of discourse (according to which, for instance, literary discourse is defined by literarity), the problem of knowing what discourse is - in the semiotic sense - remains unresolved. When one examines the various semiotic systems from the point of view of their syntactic and semantic components, one notices that some of them - for instance, literary semiotic systems - are indifferent vis-á-vis their invested contents, while, by contrast, others are indifferent to the syntactic organizations they might have - for example, the "feminine narrative" formulated by C. Chabrol, viewed as a universal articulation of content, whatever its nature, can be taken over as "literary" content, literary discourse can only base such specificity it might have upon the syntactic forms which it manifests. Yet, the variety of forms is so broad that literary semiotics appears mainly as a vast repertory of discoursive forms and not as a syntactic structure which could be defined. While there are "many" literary discourses one cannot speak of "the" literary discourse. On the other hand when thinking about the "feminine narrative" - and also about the semantic fields called "political discourse," "religious discourse," etc. - one can say that there exist deep organizations of content which can be formulated as systems of values or as epistemes (i.e., as combinatory hierarchies). But, once more, these axiologies can be manifested in many kinds of discourses. This amounts to saying that the general semantic theory of discourses must be dealt with indepdendently of their syntactic typology which, when it will have been further developed, will certainly appear as very far removed from the present connotative typology of discoursive genres.
11. Coming back to the domain of the enunciation, which is the locus of the generation of discourse, one can then say that the form of the produced discourse is dependent upon the twofold selection taking place in it. When the semio-narrative structures are viewed as the repertory of the forms which can be altered, the enunciation has then as its task the selection of those forms it needs "to make a discourse." Thus, the choice between the pragmatic or cognitive dimensions of the envisaged discourse, the choice between the forms which are appropriate for a discourse characterized by the construction of a subject (cf. the Bildungsroman) and those which are necessary for a discourse characterized by the construction of an object (cf. for example, the recipe for vegetable soup), etc., determine in advance the type of discourse which will finally be manifested. On the other hand, the use of the mechanisms of disengagement and of engagement, which define the enunciation as an activity of production, can only be viewed as a selective operation which chooses, within the combinatory set of discoursive units which these mechanisms are able to produce, certain preferred units and/or a specific preferred arrangement of units. In both cases, and whether one is dealing with semio-narrative competence or with discoursive competence (properly speaking), the production of a discourse appears to be a continuous selection of possibilities, making its way through networks of constraints.

1. By semiotic square is meant the visual representation of the logical articulation of any semantic category. The elementary structure of signification, when define - in a first step - as a relation between two terms, rests only on a distinction of opposition which characterizes the paradigmatic axis of language. It is, consqeuntly, adequate for the establishment of a paradigm composed of n terms, but it does not thereby allow for the distinction, within this paradigm, of semantic categories founded on the isotopy (the "family relations") of distinctive features which can be recognized therein. A typology of relations is necessary, which will make it possible to distinguish intrinsic features, those which constitute the category, from those which are foreign to it.
2. The linguistic tradition which sprang up in the 1920s and '30s has imposed the binary conception of categor. In the wake of comparative research into morphological categories, it was rare for linguists (such as V. Brøndal, for example) to uphold the existence of multipolar structures containing as many as six inter-linked terms. However, even R. Jakobson, one of the defenders of binarism, was obliged to recognize the existence of two types of binary relations: the first, of the type A/A, characterized by the resultant opposition of the presence and absence of a definite trait, and the second, of the type A/non-A, which manifests to some extent the same trait, present twice in different forms. On the basis of this knowledge, the result of linguistic doing, it has been possible to establish a typology of intercategorial relations.
3. The first generation of categorial terms. It is sufficient to start with the opposition A/non-A, and, while considering that the logical nature of this relation remains undetermined, to call it the semantic axis, in order to realize that each of the terms of this axis may separately enter into a new relation of the type A/A. The representation of this group of relations is then given in the form of a square:
It remains for us to identify these various relations one by one: (a) The first - A/A - defined by the impossibility for two terms to be present together, is known as the relation of contradiction, which is its static definition. From the dynamic point of view, it can be said that this is the operation of negation, carried out on the term A (or non-A), which generates its contradictory A (or non-A). Thus, starting with the two primitive terms, it is possible to generate new contradictory terms (terms of the first generation); (b) The second operation is that of assertion: carried out on the contradictory terms (A, non-A), it can be presented as an implication and may cause the two primitive terms to appear as presupposed elements of the terms asserted (A ⊃ non-A; non-A ⊃ A). If, and only if, the effect of this double assertion is to produce these two parallel implications, we are right in saying that the two presupposed primitive terms are the terms of one and the same category and that the chosen semantic axis is constitutive of a semantic category. On the contrary, if A does not imply non-A and if non-A does not imply A, the primitive terms - A and non-A - with their contradictories, are dependent on different semantic categories. In the first case, the opertaion of implication established between the terms (A and non-A) and (non-A and A) is a relation of complementarity; (c) The two primitive terms are both presupposed terms; characterized moreover by the fact that they can be present concomitantly (or, in logical terms, that they can be true or false together - a criterion which is difficulty to apply in semiotics), they are said to enter into a relation of reciprocal presupposition or which comes to the same thing, a relation of contrariety.
It is now possible to five a definitive representation of what we call the semiotic square:
S1 - S2 : axis of contraries
S2 - S1 : axis of subcontraries
S1 - S1 : positive schema
S2 - S2 : negative schema
S1 - S2 : positive deixis
S2 - S1 : negative deixis
There remains, however, one point which requires classification; that of the existence of binary semantic categories stricto sensu (whose constitutive relation is not contrariety, but contradiction), such as, for example, assertion/negation. There is nothing to prevent such categories from being represented in a square:
It is clear from this that the negation of a negation is equivalent to assertion. Generalizing, it can be said that a semantic category may be called contradictory when the negation of its primitive terms produces tautological implications. Such a definition, taxonomic in nature, satisfies traditional logic, which can operate substitutions in both directions(non-oriented) by replacing assertion with negation, or inversely. In linguistics, things happen differently: discourse retains the traces of previously carried out syntactic operations
The term "certainly" is, of course, the equivalent of "yes," but at the same time it contains, in the form of an implicit presupposition, an operation of prior negation. Thus, it is preferable in semiotic descriptions to use - even for the contradictory categories - the canonical representation in square form.
4. The second generation of categorical terms. It has been shown how two parallel operations of negation, carried out on the primitive terms, enabled two contradictory terms to be generated, and how, following this, two implications established relations of complementarity by determining, at the same time, the relation of contrariety which thus has become recognizable between the two primitive terms. We will not take the time to go through, beginning with the network thus constituted, the same operations which, by the negation of the subcontraries, establish a reciprocal presupposition between them. It is important now to derive the first consequences of the relational model thus constructed: (a) Clearly, the four terms of the category are not defined in a substantial manner, but only as points of intersection, as the results of relations: this satisfies the structural principle laid down by F. de Saussure, according to which "in language, there are only differences"; (b) It will also be noted that, starting from the projection of the contradictories, four new relations have been recognized within the square: two relations of contrariety (the axis of the contraries and of the subcontraries), and two relations of complementarity (positive and negative deixes); (c) Given that every semiotic system is a hierarchy, it is established that the relations contracted between terms may serve, in their turn, as terms establishing between themselves hierarchically superior relations (functions playing the role of functives, in L. Hjelmslev's terminology). In this case it is said that two relations of contrariety enter into the relation of contradiction between themselves, and that two relations of complementarity establish the relation of contrariety between themselves. The following example illustrates this finding:
It can thus be seen that truth and falseness are contradictory metaterms, whereas secret and lie are contrary metaterms. The metaterms and the categories they constitute are considered as terms and categories of the second generation.
5. The third generation of categorial terms. The problem so far not tackled is that of the third generation of terms. In fact, the comparative research of V. Brøndal has shown the existence, within the network which articulates the grammar categories, of complex and neutral terms whicha re a result of the establishment of the relation "both ... and" between contrary terms. The complex term is seen as the joining of the terms of the axis of the contraries (S1 + S2), whereas the neutral term results from the combination of the terms on the axis of the subcontraries (S1 + S2). Some natural languages may even be able to produce positive complex terms and negative complex terms, depend on which of the two terms constituting them dominates.
Various solutions have been proposed to account for the formation of such terms. While not wishing to put forward yet another hypothesis, we consider that until more precise and more numerous descriptions are furnished, the problem area remains open. For all that, the importance of the problem is inescapable: we know that discourses of a sacred, mythic, poetic, etc., nature manifest a particular predilection for the use of complex and probably contradictory syntactical trajectories, which culminate in this kind of formation.
6. The semiotic square can be usefully compared to the hexagon of R. Blanché, and to the groups of Klien and Piaget. It is dependent, however, both on the epistemological question which deals with the conditions of existence and the production of signification, and on methodological doing applied to concrete linguistic objects. In this it is distinguished from logical or mathematical constructions, which are independent, as formulations of "pure syntax," from the semantic component. Under these conditions, any hasty identification of the semiotic models with the logico-mathematical models can only be hazardous.

(A) General meaning
1. Without entering into the philosophical and ideological controversies that the notion of structure continues to provoke, it is appropriate to specify the constituent elements of the definition of this concept, by placing it within the framework of structural linguistics, which has succeeded in giving it an operational character. By borrowing the broad lines of the formulation that L. Hjelmslev has given it, we will consider structure as an autonomous entity of internal relations, set up into hierarchies. In order to explain this definition, let us take a look at each one of its elements: (a) Such a conception implies that priority has been given to relations at the expense of elements: a structure is first of all a network of relations, the intersections of which constitute terms; (b) The network of relations that defines structure is a hierarchy, that is, an entity that can be divided into parts that, while being bound to one another, maintain relations with the totality that they constitute; (c) Structure is an autonomous entity, which means that, while maintaining relations of dependence and interdependence with the broader set of which it is a part, it is equipped with an internal organization of its own; (d) Structure is an entity, that is, its ontological status does not need to be questioned, but must, on the contrary, be put between brackets, so as to render the concept operational.
Thus, the question of whether structures are immanent to the object examined or whether they are constructions resulting from the cognitive activity of the knowing subject, is to be excluded from properly semiotic concerns, however fundamental this question may be from a philosophical point of view. Likewise, the philosophical presuppositions that underly the conception of the structure (and which are manifested especially in the way in which one considers the relations between structure and function and defines the latter), by giving it at times a slightly mechanistic (Bloomfield) or phenomenological (Hjelmslev) tint, and at other times a somewhat organicist (Benveniste) hue, rather enrich the epistemological-methodological tools, without harming their operational character.
2. Such a conception of structure constitutes a background for semiotic theory, a "scientific attitude" from which the approaches that the researcher makes can be sketched out. Considered in itself, structure is not the specific property of semiotics, nor even of the human sciences taken as a whole. With a few adjustments, one could say that it is implied in any project or approach with a scientific aim. What especially led linguistics, at a critical moment of its growth, to make explicit the principles upon which is grounded its own doing, was the difficulty experienced by the human sciences in moving from the stage of "opinions" to that of "disciplines." Let us add, moreover, that such a definition of structure is not directly operational: being too general, it applies to any set that one suspects of being organized or that one intends to organize. Defined as a network of relations, structure refers to the concept of relation and, in order to be semiotically efficient, presupposes a typology of relations. Considered as a network, it does not give us any information concerning either its amplitude or its complexity: the problem of minimal structural organizations, of elementary structures, is posed quite naturally, for only these structures can permit us to understand the modes of existence and of functioning of more complex sets.
(B) Elementary structure of signification
1. If we accept the definition of structure as a "network of relations," thinking about elementary structure must first consider only one relation, considered as a simple relation. By positing, within the same definition framework, that "objects of the world" are not knowable in themselves, but only by their determinations (or their properties) and that, in addition, the latter can be recognized only as values (that is, as they relate one to another), we are led to postulate that relation alone institutes "properties"); these properties, in turn, serve as determinations for objects and render them knowable. Such a relation, called elementary, is nonetheless present in a double aspect: it gives the basis for "difference" among values, but difference, in order to have meaning, can only rest on the "resemblance" that situates values with respect to one another. Interpreted in this way, the relation that sets up the elementary structure includes the two definitions of the syntagmatic axis (the "both ... and" relation) and of the paradigmatic axis (the "either ... or" relation) of language. Defined as a relation that establishes at least two value terms, elementary structure is to be considered, on the one hand, as a concept uniting the minimal conditions for the apprehension and/or the production of signification, and, on the other hand, as a model containing the minimal definition of any language (or, more generally, of any semiotic system or process) and of any semiotic unit. It is thus presented as a locus of convergence of gnoseological postulation for a subsequent axiomatic system.
2. The concept of elementary structure can become operational only if it is submitted to a logical interpretation and formulation. It is the typology of elementary relations (contradiction, contrariety, complementarity) that opens the way to new generations of interdefined terms, and which allows us to give a representation of the elementary structure in the form of the semiotic square.
3. Formulated in such a way, elementary structure may be considered as a constitutive model, and this for two reasons: as an organization model of signification (its morphological or taxonomical aspect) and as a production model (its syntactic aspects). As deep structure, it is the basis for the level of fundamental syntax.
4. Elementary structure must, in addition, be viewed as a locus of (semantic) investment and of formation (that is, putting into a form) of contents. These contents, syntactic or semantic (stricto sensu), projected upon the square, may be articulated in foreseeable positions and constituted in semantic categories. Thus, for example, any actant may be "split open" and give rise to an actantial category (actant, anactant, negactant, neganactant). Once a semantic category has been thus obtained, it may serve as a basis for a set of hypotactic sub-articulations, each one more narrowly defined; and for that very reason, it may incorporate a semantic microuniverse, (that is, a generator of discourse). Certain categories - abstract and very general - may hypothetically be considered as semantic universals, that is, as elementary axiological structures; it will be said that the category life/death articulates individual universes, and that the category nature/culture articulates collective universes. To these two elementary structures can be added, because of its large degree of generality, the figurative axiological structure which articulates, in the form of a square, the four "lements of nature" (earth, air, fire, water).
5. The elementary structure, as an articulatory model, has its main application at the level of deep and abstract structures. There it plays the role of description procedure (and, possibly, of discovery procedure), allowing semiotic phenomena to be represented prior to their manifestation (and, in the case of natural languages, prior to their lexicalization). Thus, the almost mechanical application of this model to surface phenomena usually constitutes but a caricature of semiotic procedures. However, that does not mean that elementary articulations do not appear at the surface, at the level of morpheme-signs, for example; but at that level the categories rarely lexicalize the set of their possible terms: in manifestation these categories present various forms that may be grasped as binary articulations (masculine/feminine, for example), ternary articulations (love/hate/indifference, for example), etc.
(C) Structural forms
1. In addition to the precise meaning that we have just acknowledged for the term structure, common usage has imposed a more general meaning which more or less corresponds to the one attributed to articulation, organization, device, mechanism, etc., and which emphasizes the relational character - supposed or established - of the semiotic sets of objects being considered. Therefore, in order to introduce more clarity into the presentation of the lexical items of this dictionary, we have found it preferable to bring together here a rather ill-sorted set of expressions currently used, providing each one with a few summary explanations and cross-references (which will allow a more in-depth examination).
2. Actantial and actorial structures. The distinction between actant and actor, derived from the intuitive notion of character (or from Vladimir Propp's dramatis persona), has had not a few reprecussions on the whole of semiotic theory. The actant, a syntagmatic unit of the surface narrative grammar, once it is placed on the narrative trajectory was divided into a set of actantial roles; the actor, a unit of discourse, has been redefined as the incarnation, the locus of semantic investment, within discourse, of a least one actantial role and, at the same time, of at least one thematic role. Consequently, the actantial mechanism - the set of actants taken on by the narrative grammar for the purpose of generating discourse - has proven to be not isomorphous with respect to the actorial organization such as it is constituted at the discourse level of the same text (the modality of being-able-to-do, for example, may be presented in the form of an independent actor, such as a magic object, or it may be integrated into the subject-hero, as an intrinsic property). Proceeding from such observations, we may speak of actorial structures, characteristic of such and such a type of discourse. The actorial structure os objectivized (and socialized) when the actorial mechanism is characterized by the establishment of a large number of independent actors; on the other hand, it is said to be subjectivized (or psychologized) if the number of actors present in the discourse is reduced and is, in certain cases, summarized within one actor subsuming a large number of actantial roles (yileding an interiorized, intense dramatization, well known in psychoanalysis).
3. Aspect structures and category structures. Located at the deep semiotic level, narrative grammar uses a categorial logic, based on the discrete characters of units and on the discontinuous character of states (an object of the world is either "black" or "not black," without transition). When narrative structures are thus formularized, they are temporalized during discoursivization and consequently receive complementary aspectual investments; therefore, to logical transformations at the deep level correspond, at the discourse level, diachronic "changes" which we can account for with the help of aspectual categories (articulating semes of punctuality, durativity, inchoativity, perfectivity, etc.). Such a conception of aspectual structures consequently allows us to reconciliate "story" and "structure" and to conceive of conversion mechanisms of category structures an aspectual (temporal) structures, and vice versa.
4. Modal structures. A somewhat deeper examination of modal categories (wanting to, having to, being-able to, knowing) has shown that their character as "governing term" would not permit them to be formulated independently of the "governed term"; in other words, it was not possible to speak of wanting or of being able, but only of wanting-to-do or wanting-to-be, of being-able-to-do or being-able-to-be, etc. Since the modality is an integral part of the utterance of state which it overdetermines, it is appropriate, in syntagmatics, to speak of modal structures, whereas in paradigmatics modalities may be considered as modal categories.
5. Narrative structures and discoursive structures. This distinction correspond to the two levels of depth which we consider as the fundamental domains of the overall generative trajectory, which ends in the production of discourse. The expression narrative structures or, more accurately, semio-narrative structures, is then to be understood in the sense of deep semiotic structures (which preside at the generating of meaning and bear general forms of discourse organization); they are distinct from discoursive structures (in the narrow sense) are located at a more superficial level and which, beginning with the domain of enunciation, organize the discoursivization (putting into discourse) of narrative structures. In addition, by narrative structures (in the narrow sense) is often meant simply the surface narrative syntax - this confusion stems from the fact that certain "grammars" or "logis" of narrative conceived of the deepest level of narrativity in a more or less comparable form.
6. Polemic structures and contract structures. Different textual analyses have come to the conclusion - which can seemingly be generalized - that all discourse at least implicitly carries with it a confrontation structure, bringing together at least two subjects. This confrontation often takes the form of a somatic or cognitive clash (we can then speak of polemic structures); or it may take the form of a transaction, in which case the structure organizing the discourse can be called contractual. These two forms, which, as can be seen, correspond to the level of socological theories, to the concepts of "class struggle" and "social contract," can be found together in the structures of manipulation. In addition, the polemico-contractual structure of discourse which has but one enunciator allows us to understand and interpret dialogue communication as a discourse having two voices.
7. Deep structures and surface structures. The distinction between deep structures and surface structures is totally relative, since semiotic theory can, depending on its needs, foresee along the overall generative trajectory as many levels of depth as it wants. Thus, for us discourse structures appear as surface structures with respect to semio-narrative structures, which are deeper. However, we use this dichotomy especially in order to establish a distinction, among semiotic structures (to which we give the form of a grammar), between two levels of depth: between the fundamental (deep) grammar and the narrative (surface) grammar in the strict sense (superficial), the former having a logico-semantic nature, the latter having an anthropomorphic nature.
8. Semio-narrative structures. The fact that semiotic theory has only been developing in a progressive and sometimes meandering way has been the source of certain terminological confusions. Such is the case, for example, for the concept of narrativity which, first applied solely to the class of figurative discourses (narratives), has proven to be an organizing principle of all discourse. Consequently, the content of the expression "narrative structures" has been transformed so as finally to designate, in opposition to discourse structures, the deep generative trunk which is in principle common to all types of semiotics and to all types of discourse, the locus of a general semiotic competence. A slow terminological substitution is therefore taking place; little by little the expression semio-narrative structures is replacing "narrative structures" in the broad sense.
9. System structures and morpheme structures. At first glance the semic organization of the semantic universe takes on two different forms: on the one hand, that of semic systems (that is, hyponymic subarticulations of a paradigmatic character, bearing only homogeneous semes (that is, stemming from several semic systems), and bound together by hypotactic relations of a syntagmatic nature. This distinction seems sufficiently important to tus to be mentioned here, since it probably allows us to account for the functioning of the figures of metaphor and metonymy, as of the relationship of contiguity.

All the previous entries are what we are supposed to learn for the exam. The following entries are what I consider possibly useful for my own academic work.

1. As opposed to both idiolect, which designates the semiotic activities of an individual actor, and to dialectic, which refers to the differentiation of these same activities considered from a social point of view (the differentiation is due to a geographic distribution of human groups), sociolect characterizes semiotic activity in its relation with social stratification. If the organizations of a given society are viewed as extra-semiotic phenomena, the semiotic configurations that correspond to them constitute the signifying face of these organizations, for they reveal how society, class, and social strata or groupings are distinguished from one another. Sociolects are thus sorts of sub-languages, recognizable by the semiotic variations that oppose them to one another (their expression plane) and by the social connotations that accompany them (their content plane). Sociolects are constituted as social taxonomies, underlying social discourse. The study of sociolects stems from a particular discipline, socio-semiotics.
2. Sociolectal variations can be pinpointed on the lexical surface level (cf. nomenclatures, terminologies, etc.) or on the level of discourse organization (writing can be assimilated to a sociolectical act, as opposed to style, which is of an idiolectical order). At the level of deep semantic structures, the sociolectal universe is characterized both by a particular use of the nature/culture category (by providing the collective semantic universe with specific hypotactic ivnestments) and by its articulation of the life/death category, which allows it to interpret the individual semantic universe in its own fashion. In the long run, it is a matter of accounting for the attitude that a socio-cultural community adopts with respect to fundamental issues that it faces.

1. A structure is said to be binary when it is defined as a relation between two terms.
2. A set of historical and pragmatic factors has given binary structures a privileged place in linguistic methodology. This may be due to the successful practice of the binary coupling of phonological oppositions established by the Prague School, or due to the importance gained by binary arithmetical systems (0/1) in automatic calculus, or to the operative simplicity of binary analysis in comparison with more complex structures, since more complex structures can be formally represented in the guise of hierarchy of binary structures, etc. Binarization, as a linguistic practice, should be distinguished from binarism, which is an epistemological postulate according to which the binary articulation of grasp of phenomena is one of the characteristics of the human mind. The name of R. Jakobson is linked, rightly or wrongly, to this claim: he gave a binary formulation of the phemic categories which he defined as phonological universals of natural languages.
3. Binary formulation remains valid so long as no attempt is made to define the type of relation which joins the terms: indeed Jakobson himself recognized the existence of two types of binary opposition, which we interpret as contradiction and contrariety. It is this sort of typology which has enabled us to postulate the existence, beyond the realm of binary, of a more complex elementary structure of signification.
4. Binarity characterizes only one type of structure: it is possible to consider as binary categories only those whose constitutive relation is contradiction (for example, assertion/negation; conjunction/disjunction).

1. The context is the entire text which precedes and/or accompanies the syntagmatic unit under consideration and upon which the signification depends. The context can be explicit; it is then called linguistic. It can also be implicit; in this case it is qualified as extra-linguistic or situational. The implicit context can be called upon in the semantic interpretation, because (a) the situational context can always be made explicit, since it is a living natural language producing an unlimited text (Hjelmslev); (b) the implicit elements of a linguistic text can eventually be reestablished through the homologation of this text with the non-linguistic text which belongs to the semiotic system of the natural world.
2. In his model of communication, R. Jakobson posits the context as one of the factors of the linguistic activity and identifies it with the referent (it is the referential function of language). Whether it is verbal or can be made so, the context is considered as necessary for making the message explicit.
3. The phrase contextual semes (or classemes) is used to designate the semes or bundles of semes which are recurrent in the unit under study and in its context. The contextual semes participate in the composition of a sememe (which can be related with the "word in its context").

1. The adjective somatic is generally used to qualify a figurative actor (character), located and acting in the pragmatic dimension of discourse. Somatic doing is either pragmatic (if it refers to a programmed bodily activity) or communicative (the human body being able to signify by gesture, attitudes, miming, etc.). It will therefore be useful, in the latter case, to distinguish between somatic communication and verbal communication.
2. Under certain conditions that remain to be determined (when a pragmatic narrative - for example, a miracle narrative in the Gospel - is inserted into another, broader, narrative), somatic doing is related (or accomplished) not only with respect to an assigned goal (a healing, for example), but also in regard to an actant observer (usually implicit) which is supposed to read and interpret this narrative (or this behavior) constructed as signification. Such a somatic doing, both pragmatic and communicative, brings about a meaning effect of "irreality" and is read on the cognitive dimension of discourse.

Entities belonging to non-linguistic semiotic systems are considered paralinguistic when they are produced concomitantly with the oral or graphic message of natural languages. Generally included under this label are, on the other hand, the phenomena of intonation, gestuality, bodily postures, etc., and, on the other, choice of fonts, page design, etc. The term paralinguistic (or even paralanguage) indicates a narrowly linguistic point of view which, while recognizing the existence of other semiotic practices, considers them as secondary or subordinate.

PHATIC (activity, function)
To Malinowski belongs the credit for being the first to seek to delineate the notion of phatic communion. In his eyes the communication of information, such as it takes place during verbal exchanges between human beings, is secondary to the desire to establish and maintian intersubjective solidarity and, more generally, social cohesion. This solidairty and cohesion are the foundation for phatic communion, thanks to which one can "speak of nothing and everything." Following him, R. Jakobson attempted to present this aspect of communicability by formulating it as a particular function, the phatic function of language. The phatic function is accpetable as long as it is considered as a general property of language; but it seems more debatable when it must be integrated into the structure of communication. Instead of speaking of the phatic function as one of the functions of communication, it would be better to say that it is the phatic intention which, on the contrary, lays the foundation for communication, and that the phatic act must be considered first of all as a somatic act (comparable to the glance or to gestures of greeting or of welcome), and, as such, capable of being integrated into proxemics (in the broader sense of the term).

1. Proxemics is a semiotic discipline - or rather a project for a discipline - which seeks to analyze the arrangements of subjects and of objects in space, and, more particularly, the use that the subjects make of space in order to produce signification. Thus defined, it appears as a problematic area of semiotic theory which intersects in part with semiotics of space but also with natural semiotics, theatrical semiotic,s discourse semiotics, etc.
2. The contours of this field still remain very uncertain. In a first approach, proxemics seems to be concerned with the spatial relations (proximity, distance, etc.) which the subjects maintain between themselves, and with the non-verbal significations which they draw from it. Sometimes when it is no longer a question of natural semiotic systems (that is to say, "real" behavior in the world) but of artificial or constructed semiotic systems (theater, liturgy, ritual, town-planning, etc.) and when one is led to foresee a domain of enunciation, the arrangements of the objects become bearers of meaning as much as do those of the subjects.
3. Proxemics should not be limited merely to the description of spatial arrangements, formulated in terms of utterances of state. It must also consider the movements of subjects and the "displacements" of objects, which are not less meaningful, for they are spatio-temporal representations of the transformations (between states). Such being the case, proxemics has overflowed the borders that it drews and sees itself obliged to integrate gestural languages as well as spatial languages into its field of analysis.
4. Independently of the limits that proxemics will set up on its own, the procedures of proxemization must be integrated here and now into that component of discourse semiotics known as spatialization.

1. As a specific research field, gestuality has been introduced into semiotic reflection only progressively and hesitantly, appearing at one time as a circumscibed and autonomous domain of significations, analyzable as a gestual language, at another as omnipresent, spilling out over the still imprecise borders of particular semiotic systems that are in the process of being constituted.
2. Gestuality has been - and still is - considered as a paralinguistic phenomenon which has an auxiliary function in intersubjective communication. This accompaniment gestuality has been judged, a bit too hastily, as "meager" - because it was incapable of producing disengaged utterances and of transmitting objective contents; certain scholars would, therefore, reduce its role to that of simple emphasis. Yet upon closer examination, it can be defined as frame gestuality, i.e., as gestuality framing the enunciation. The categories that take the form either of modal utterances (assertion, negation, doubt and certitude, etc.) or of utterances of quantification (totalization, division) and of qualification (euphoric and dysphoric states), or, above all, of phatic utterances (reception and rejection, openness to the world and closure of self, etc.) that transform communication into intersubjective communion.
3. Some scholars have tried to study gestuality as a language by applying the Saussurian formula of "sign system" to it. The signs would be recognizable through commutation tests; the system would be used for communication. Unfortunately, the inventories of communication gestures that have been established (cf. the gestures of the North American Indians) could not be structured into system: they correspond to no "phonological" structure and belong to no semantic organization (other than that of "centers of interest"). Therein are lumped together, pell-mell, accompaniment gestures, icons, and above all, gixed gestual syntagms, desemanticized and conventionalized, in short the whole Peircean classification in haphazard order. The existence of an autnomous "gestual language" seems, thus, far from being assured.
4. Another approach to gestuality consists in starting this time, not from gestures considered as signs, but from gestual texts (folk dances, ballet, acrobatic numbers, pantomime, etc.). The itnerest of this type of research is multifold. The approach is, first of all, analytic: the segmentation of the text raises, of course, the problem of gestual units of more or less large syntagmatic dimensions, but also that of the pertinence of the gestual features that characterize them. It is thus not surprising that this type of investigation should lead, on the one hand, in the area of the expression plane, to making evident the need for a language of description (the drawing-up of systems of symbolic notation of gestures has already made quite a lot of progress and raises new questions relating to their metalinguistic coherency). On the other hand, it has led to raising the problem of the singification of these gestual discourses, which appear both as programmed texts, thus undergirded by an implicit intentionality, and as theatrical utterances, produced in function of a reader-observer and, consequently, doubly significant; by themselves, and for others.
5. An examination of gestual texts leads not only to a distinction between signifying gestuality and meaningless gesticulation, but also makes obligatory a definition of "gestual substance" as that which is expressed by means of that particular matter that the human body is, as "volume in motion." Gestuality is not limited merely to the gestures of the hands and the arms or to facial expression, it is an integral part of people's somatic behavior and, in the long run, constitutes only one of the aspects of what can be called their somatic language. But whereas gestual texts, previously mentioned, appear to be processes of constructed (or artificial) somatic semiotic systems, there exists, alongside them, one suspects, one or more "natural" semiotic systms which account for human behavior programmed as signifying practices. An analysis of narrative discourse allows us, precisely, to distinguish a pragmatic dimension of the discourse, made up of descriptions of signifying somatic behavior and organized into programs and which are simultaneously designated as events for the reader. We have sought to define gestual texts by means of these two characteristics. Whence it can be seen that the narrative models constructed to account for pragmatic behavior "on paper," can be transposed for the formulation of a natural "pragmatic" semiotic system.
6. It will immediately be noted that "somaticity," as well as gestuality, are not concepts easily described: "speaking" or "singing" are behaviors quite as somatic as "walking" or "gesticulating." It can thus be said, finally, that the diverse semiotic systems are embedded into each other in their "natural" state (cf. different rituals and ceremonies, for example) as well as in their "constructed" state (theater, opera, circus, etc.), and that, most often, we have to deal with syncretic semiotic systems whose constitutive elements and their interconnections are to be disentangled.

1. Traditionally, the term referent is used for the objects of the "real" world which the works of the natural languages designate. The term "object" being clearly inadequate, referent came to be understood as including qualities, actions, real events. Furthermore, as the concept of "real" world is itself too narrow, the referent must also involve the "imaginary" world. The one-to-one correspondence between the linguistic universe, which is thus metaphysically presupposed, is nevertheless incomplete. On the one hand, certain grammatical categories - and especially logical relations - do not have valid referents. On the other hand, some deictics (for example, personal pronouns) do not have a fixed referent: each time they point to different objects. This isto say that as long as positivistic presuppositions, held as self-evident, are the starting point, it is possible to construct a satisfactory theory of the referent which could account for all the phenomena under consideration.
2. It is nevertheless in the above context that two attempts to integrate the referent in broader theories are to be located: (a) Ogden and Richards conceive the referent in terms of the Saussurian theory of the sign. They propose a triangular model aimed at accounting for the structure of the sign: the symbol (or signifier) is linked to the referent not directly but through the intermediary of the reference (or signified). In this interpretation, the reference, instead of being viewed as a relation, is reified and is transformed into a concept - a hybrid entity, neither linguistic nor referential - the expansion of which includes a class of referents; (b) R. Jacobson [sic] conceives the referent in terms of communication theory. For him the referent is specifically the aspect of context in the structure of communication. It is necessary for making the message explicit and can be apprehended by the receiver. It is neither verbal nor capable of being verbalized (i.e., it can be made explicit by means of language). R. jakobson thereby acknowledges the existence of a referential function of language (following K. Bühler's concept of representation): the discourse-utterance, after being disengaged (expressed in the third person), is used for describing the world, i.e., the referent.
3. Thus the linguistic context - either verbal or capable of being verbalized - becomes the locus of the text's reference and the specific elements of this context are then called referents. In this sense the term referent is then synonymous with what is anaphorized. At this point and in this way it appears that the question of the reference needs to be addressed so as to describe the network of references both within the utterance and the domain of the enunciation.
4. In order to establish a compromise between the autonomy of language (proclaimed by F. de Saussure) and the self-evidence of the "real" world, which is emphasized by the positivists, it is sometimes proposed that the referent be defined as constituted by "'things' in so far as they are 'named,' or 'signified,' by words" (J. Lyons), i.e., not things in and of themselves, but things named or capable of being named. Such a view is not without contradictions. If, indeed, the principle of the categorization of the world by language is accepted (cf. E. Benveniste and especially Sapir and Whorf), i.e., if the view that natural languages structure the world and constitute it into distinct "objects" is recognized, how thne can one call upon this world (which, in part, results from linguistic activity) in order to define the signs which make up these languages?
5. In our opinion, another solution seems possible. First, one needs to note that the extralinguistic world (the world according to "common sense") is given form by human beings and constituted by them as signification. Such a world, far from being the referent (i.e., the denotative signified of natural languages) is itself, on the contrary, a bi-planar language, a natural semiotics (or a semiotics of the physical world). The problem of the referent is then reduced to the question of the correlation between two semiotic systems (for example, natural language and natural semiotics; pictural semiotics and natural semiotics). This is a problem of inter-semioticity (cf. intertextuality). Conceived thus as a natural semiotic system, the referent loses its need to exist as a linguistic concept.
6. In such a perspective, the question of the referent of literary discourse can be circumscribed. Literary discourse is often defined as lacking a referent or as having a fictive or imaginary referent: the term fiction is even used in order to designate this textual genre. On the one hand, the impossibility of defining the "real" discourse (i.e., a discourse the signs of which would correspond to objects of the world), excludes the definition of the fictive discourse. These two types of discourse can only be characterized by veridiction, which is an intrinsic property of the saying and of the said. On the other hand, any discourse (literary discourse as well as legal or scientific discourse, among others) builds its own internal referent and establishes for itself a referential discoursive level which provides a basis for its other discoursive levels.
7. When one wants to deal with discourse from a generative perspective, the problem which needs to be addressed is not that of a referent given a priori, but rather the problem of the referentialization of the utterance. And this implies the study of the procedures through which is constituted the referential illusion - the meaning effect "reality" or "truth," proposed by R. Barthes. Among these procedures, which hae not yet been systematically studied, one can note, for example, spatio-temporal anchoring (the use of toponyms and/or chrononyms providing the illusion of "reality") or internal disengagement (which referentializes the discoursive segment on the basis of which the disengagement is performed: cf. the passage from dialogue to narrative or conversely).


Post a Comment