A. Greimas The Social Sciences: A Semiotic View For a long lime, and still today, the Saussurian dichotomy of tongue and parole seemed to provide the explanatory framework that allowed us to account for the permanence of a structure underlying the totality of message-events that are both contingent and justified. This concept of system, which was immanent to a large number of linguistic activities, was rounded off by the recognition of the temporal linearity of discourse. Structure, which was timeless, could produce sequences of significations that were both evenemential and temporal. It generated historical events. With respect to the problem of how temporality is produced from structures, difficulties almost always arise when the primary elements have to be exploited and integrated as explanatory elements in partial analyses. Hence the temporal nature of discourse becomes indistinct when describing the syntax of a natural language. We know that this type of description functions only with units of discourse that do not exceed the limits of the sentence, for it is a fact that syntactic structures do not organize the totality of discourse but only very limited segments of it. Discourse is therefore not the articulation of successive structures but the redundancy of the utterance, a single hierarchical structure. From this limited point of view, an author perceives signification not as being spread out over time but rather as the iteration of a certain number of permanencies. Even if we set aside the grammatical manifestation of linguistic reality to analyze the transphrastic level of signification whose elements seem distributed along the continuum of time and constitute discourse as the temporal manifestation of meaning, we encounter the same conditions that transform temporality, considered as the means of transmission, into simultaneity, the extralinguistic conditions of the reception of messages strung together into discourse. Hence, every apprehension of signification transforms histories into permanencies; whether we happen to question the meaning of life (or of history), the interrogation (i.e., the fact that one assumes the role of receiver of messages when faced with a linguistic manifestation) has the following consequence: historical algorithms appear as states, in other words, as static structures. We can reserve judgment as to the restricting value of the Brondalian notion, according to which the synchronization of information, a necessary condition for its setting into structure and consequently, for its capacity to signify, cannot exceed the simultaneous apprehension of more than six terms. It does seem impossible. however, not to take into account the fact that at the syntactic level, the utterance always appears as a minute spectacle with a very limited number of actors (subject, object, sender, receiver), and the fact that the fundamental signification of a story (narrative, myth, tale, etc.) can be reduced to a simple homologation. In the case of written discourse, temporality or spatiality at the level of expression is simply the means by which signification manifests itself, which is nonetheless neither temporal nor spatial. Therefore, the problem must be formulated in a different way. We cannot infer their historicity from the obvious temporality of linguistic phenomena. Since linguistic description is directed only at structures, we must attempt to understand whether and how they are anchored in history. Duration and Hierarchies To ask oneself about the “meaning” of a film one has just seen consists, within the framework of one’s own proper internal language, in organizing a limited number of elements making up the narrative, with a totalizing aperception in mind. Every subsequent step could only consist in choosing one of these elements and decomposing it. A new level of signification would be situated at a hierarchically inferior level and would correspond to the analysis of one of the terms already posited. All theories of language agree on this point: language is a hierarchy. It matters little, following one’s acquired habits or the exercise of such and such a discipline, whether by process of visual symbolization one designates the elementary level of this permanence as analogical or substructural or, on the contrary, as situated at the level of the metalinguistic or superstructure! pyramid. The elementary signification of any story, taken within the limits of its entire duration (which, in the historical sciences, would correspond to Braudel’s “long durations”), can be posited as an invariant, “medium durations”: being considered as variables, and “short 1

durations", as stylistic and conjectural variations. Such a correlation of durations and structural levels can appear tempting. A unique hierarchical model would therefore enable us to account for all sorts of diachronic transformations, interpreted as paradigmatic substitutions of variables situated at a specific structural level. Historical duration would not be completely abolished, however, but transcoded into a new descriptive language, and history itself would be integrated into a wider semantic universe, finally, periodization, a procedure inherited from the nineteenth century, could be made more flexible and reinterpreted as an intermingling of manifestations belonging to different historical structures. Unfortunately, such a conception does not bear up under examination. …. Synchrony and Diachrony The difficulties encountered in integrating the temporal dimension in our reflection on the mode of existence of the structures of signification, for us, simply underscore the nonpertinence of the Saussurian dichotomy between synchrony and diachrony. Whether we take them etymologically or in the historical sense of their formulation still permeated with nineteenth-century historicism, both of these antonymic concepts are essentially conceived of as two complementary aspects of temporality, the “chronic" axis being logically prior to the opposition they are supposed to establish. This is not the case with postSaussurian theories of language. For them the structure of any language contains no temporal reference and the term “synchrony" is simply maintained by tradition. The description of a structure basically corresponds to the construction of a metalinguistic model dependent on its internal coherence and within the manifestation, able to account for the functioning of the language it proposes to describe. For this type of model, the historical dimension is simply a backdrop on which linguistic behaviors are inscribed, the study of which does not appear to be pertinent at first glance. Considering the distance that linguistics has taken in relation to diachrony, a real misunderstanding takes place when historians decide to incorporate synchrony into the main body of concepts they are used to manipulating. For historians synchrony signifies the assembly of a series of events taking place at the same time, and the description of a linguistic synchrony would ultimately signify the recording of all the words spoken at the same time by thousands of speaking subjects. Even if they are forced to acknowledge a certain staggering of messages in duration, nothing allows them to fix its limits. Do sentences, paragraphs, chapters constitute synchronic units? Will they attribute one or two years to the duration of a synchrony? Although this is what is commonly done, the exercise is rather pointless.

2