Anthony Pym


_Video material

Limits and Frustrations of Discourse Analysis in Translation Theory

Anthony Pym 1992

      First published in Fremdsprachen 2-3 (1991), 29-35. Revised version in Revista de Filología de la Universidad de La Laguna 11 (1992), 227-239

        Abstract: The general attention to discourse analysis developed in the 1970s has found applications in translation theory in the 1980s and into the 1990s. However, a survey of the linguistic approaches concerned shows that many kinds of analysis are inappropriate to the study of translation quite simply because they cannot say if a source text and a target text can or should belong to the same discourse. That is, most theories cannot describe the limits of any particular discourse within or across different tongues. A more pertinent approach is to define discourse as a set of constraints on semiosis at a non-lingual level, and then use this definition to recognise translation as a possible index of intercultural discursive constraints. 


        The fact that translating operates according to texts and usage rather than rules and analysis has been widely recognised in many recent approaches to its theorisation. However, this insight has not come from translation theory itself. It has instead been little more than a series of attempted applications of developments within linguistics, particularly after the limitations of structuralist semantics led, in the 1970s, to serious empirical and theoretical interest in pragmatics, sociolinguistics, text linguistics and numerous uses and abuses of the term discourse. Historically, these latter terms have entered translation studies from strictly non-translational concerns. Why then should they be immediately important or pertinent to a coherent theory of translation? 

        I believe that most existing forms of discourse analysis are fundamentally inadequate to the problems of translation and therefore potentially misleading for the development of translation theory. My arguments will take the form of three general points: 1) Contemporary use of the term discourse is in a mess and probably deserves to be abandoned; 2) The only kind of discourse analysis strictly pertinent to translation is that which sees translating as discursive work; 3) Far from passively receiving externally derived analyses, translation itself should become a discovery procedure for the location and delimitation of discourses. That is, the limits and frustrations of most forms of discourse analysis might profitably be overcome through a judicious application of translation analysis.

        Two basic questions to be answered

        Let me begin with two minor frustrations. First, there are peculiar approaches like Jean Delisle's L'Analyse du discours comme méthode de traduction (1984) which never actually get around to saying what a discourse is or might be, and thus do little more than misuse a stupidly modish title. Second, there are equally misleading texts like Hatim and Mason's Discourse and the Translator (1990) which present a glossary of functionalist linguistics relating terms like discourse, genre, speech act, text act, text type, register and the rest, but never actually find room to define what translation is or might be (their glossary omits the term), thus leading to confusion or suspicious collusion between descriptive and normative theory: "Discourses are modes of thinking and talking," says Hatim quite reasonably on the level of description, but then comes the normative rider: "...which have to be preserved in translation" (1990, 85). This normative application is problematic to the extent that it discounts translation itself as a significantly variable mode of discursive work (don't translators think and talk?) and uses the notion of discourse merely to add to the authoritative commands mystically enshrined in and around the source text, independently of whatever particular communicative situation might correspond to the translation itself. 

        I suspect that what Delisle has to say about translation could be said without any reference to discourse at all. Hatim and Mason, however, have so much to say about discourse and associated terms that the "translator" of their title is merely an occasion for a theoretical performance that could equally have been provoked by any literary or political text and in fact has very little to do with any specificity that translation might be able to claim. 

        These minor frustrations lead to two fundamental questions. First, from the descriptive point of view, how might the term discourse be defined in a way useful and pertinent to translation theory and practice? Second, with respect to the normative application of such definitions, should a source text and its corresponding target text form or conform to one or two discourses? In other words, should the notion of discourse be limited to only the source side of the translator's task, or should the before and after of translational labour be seen in terms of two distinct discourses? Exactly where are the limits of a discourse as it affects translation?

        Ideally, the answer to the first question (how might a definition of discourse be useful?) should implicitly or explicitly answer the second question (one or two discourses?), just as, inversely, the clarification provided by application of a definition should then justify its initial selection. The result could even become a pertinent application of discourse analysis to translation. But where might such illumination be found?

        Use and abuse of the term discourse

        Does the relation between a source text and a target text involve one or two discourses? This is a fairly simple question of definition. But if a definition of discourse cannot provide a fairly simple answer to it, this alone may be sufficient basis for declaring the defining theory impertinent to the problems of translation. The question could thus become a fairly radical device for clearing the field of inadequate approaches. 

        I should stress that the criterion of translational pertinence does not mean that impertinent definitions are automatically wrong or misguided. Nida and Taber, for example, transgress no divine law when they gloss "discourse" as:

        A specimen of linguistic material displaying structural and semantic coherence, unity, and completeness, and conveying a message; also called text. [The entry "Text" says "see Discourse"] (1969: 200, 208) 
        I can of course say that I am more interested in the level of "coherence, unity and completeness" than in the "linguistic material" itself; I can suggest that the above definition forecloses possible categorisation by allowing for as many discourses as there are pieces of coherent language in the world; I can point out that the assumption of a kind of completeness that is at once semantic and material is an ideologically self-serving negation of progressive interpretation ("the Bible says it all" contradicts the historicity of discourse as a temporal approximation to identity or truth), but none of these objections can make the definition wrong. These factors do however make the definition fundamentally inadequate to translation, since they cannot determine whether two non-identical texts-source and target, or indeed any pair that manifest the same coherence and unity-belong to the same discourse or to two different discourses. This argument would of course apply to all simple equations of the terms "discourse" and "text".1 

        The need to distinguish between discourses and texts is of theoretical as well as pragmatic importance, since the necessary singularity of the latter items raises doubts about the generality of the conclusions that can be drawn from their study. To see why this should be so, it is perhaps worthwhile going back to the reasons why discourse was originally excluded from structuralist linguistics.

        It is well known that Saussure considered parole to be non-collective, heterogeneous and thus unavailable for scientific study (1916, 19 ff.). More precisely, he considered manifest language to be no more than a sum of speech acts (actes de parole), represented as:

        (x + x1 + x2 + x3)

        Such series, like series of material texts, were considered unable to attain the mode of existence of tongue, in which Saussure saw all items as identical replica of a common collective item:

        i + i + i + i = I (collective pattern)

        The fundamental problem posed by discourse analysis would then be whether it is possible to describe concatenations of manifest language in terms that do not imply the base identity (I) of the tongue but do exhibit a rational organisation on a level somewhere between the above formulae, somewhere between the absolutely individual item and the normatively collective tongue. Saussure seems to have believed that there was no such level: "As soon as we give the tongue first place amongst the facts of speech, we introduce a natural order that lends itself to no other classification" (1916, 9). But surely the very notion of discourse relies on the existence of other levels of classification?

        The problem with theories of discourse is not that they have failed to locate any other level, but that they have located far too many non-Saussurean orders, both natural and artificial, on every level from the individual to the collective. Since the beginnings of discourse analysis towards the end of the last century2, there has been an accumulation of different approaches on different levels, none of which appears to exclude any other: to my knowledge, no partial theory (such as "discourse = text") has ever been explicitly rejected. The resulting terminological and pedagogical chaos has provided ample fodder for intellectual fashions, becoming in itself a process of unarrested addition (x + x1 + x2 + x3) which has remained remarkably resistant to identity equations. A brief log-book of French, English and German approaches might demonstrate the point:

        - In French or French-inspired tradition, "discourse" has come to mean rather more than mere dissent from the primacy of the tongue. Theoretical usages of the term may be grouped according to at least five problematics: 1) the relation between the persons represented in linguistic utterances (Benveniste 1966; Joly 1988); 2) the relation between utterances and texts (Derrida 1967, 149ff.; Metz 1970; Barthes 1970); 3) the definition of literary voice or speech modes as opposed to "story" (from Benveniste; but also Todorov 1971, 1978; Chatman 1978; and Todorov's translations of the Russian Formalists, especially Shklovski 1917, 6); 4) the correspondence between language and social structures (Pêcheux 1975); and 5) the nature of semiotic processes in general, increasingly associated with degrees of use-related competence (as noted in Greimas and Courtès 1979). But from none of this does there emerge any clear consensus as to whether translation involves one or several discourses: as much as one might suppose that traditional categories such as "discours direct / indirect / indirect libre" would allow both source text and target text to share the same mode, increasing awareness of the interrelatedness of discourses has led to the theoretical construction of things like "discursive formations" as sociocultural units (Pêcheux 1975; Cros 1983, resting on Foucault 1969), which would seem to imply that any translation that goes to another sociocultural unit must enter another discursive formation, and thus, possibly, become another discourse. But no one seems very sure about the point. 

        - In English-language research, it has become very difficult to separate adaptations of French usages from similarly generalised notions of "code" (Bernstein 1970, reinforced by Eco 1976) "system" (often from Lotman) and "text grammar" (especially van Dijk 1985). The latter tendency has considerably strengthened the identification of discourse with text, although, in view of our comments above, there is clearly some justification for maintaining the general distinction proposed by Widdowson, for whom text analysis investigates "the formal properties of a piece of language", whereas the object of discourse analysis is "the way sentences are put into communicative use in the performing of social actions" (1979, 52). This distinction benefits from Austin's analysis of performatives and their subsequent drafting into theories of speech acts (Searle 1969). However, later associated with a mystique of deconstruction, these apparently benign terms were then grouped around "discourse" to present an epistemological challenge to the notion of grammar itself: as Hopper puts it in his summary of discourse analysis in the 1980s, "...if meaning is bound to context, and contexts themselves are unbounded, in what shall linguistics (or any other endeavour, of course) be anchored?" (1988, 22). Saussure already had one answer prepared, and a good second answer to this question would be the sociological readings of unequal "discursive formations" as found in Bernstein or Pêcheux. But none of this can unambiguously answer our questions about translation, nor am I able to cite any convincingly exhaustive classification of codes, systems, texts, speech acts or discourses.

        - German theory appears not to indulge in excessive notions of speech or discourse. Instead, it seems particularly focused on the material presence of text. For instance, Lewandowski's dictionary (1985) correlates the French discours with Sprechen, which correlates with the English speech, which is also used for Rede, of which the classifications or Redearten are glossed as Textsorten. One cannot escape the notion of text! Something similar happens in Reiss's approach to translation criticism, where Bühler's three language functions are quickly reduced to "text types" as if the transition were scarcely problematic (1971, 33). Reiss's approach thus shares the blunt morality we have noted in Hatim: if the translator has worked correctly, the source and target texts should ideally belong to the same text type. Which is to say that translation itself is not a text type or series of types, nor, of course, a discourse. 

        Remarkably enough, throughout the above profusion of terms, none of the projects willing to propose classification systems are immediately of discourses: almost all of them concern "language functions". Yet one need only refer to the wide variety of attempted classifications-mainly derived from Ogden and Richards (1928), Bühler (1934), Austin (1955), Jakobson (1956) or Halliday (1973)-to suggest that the categories concerned are as yet not demonstrably derived from the object of study itself. Why should Jakobson's six or Austin's five categories be preferred to Bühler's or Halliday's three? Why should there only be five (as available as the hand), or three (as mysterious as the Trinity), or a lucky seven (according to Hymes 1968), or as many as sixteen tentatively universal speech-act components conveniently analysable in terms of the eight letters of the English word SPEAKING (Hymes 1974)? 

        More importantly, how sure can we be that Saussure was entirely wrong in rejecting the possibility of a systematic discourse analysis when, in the cited approaches, all manifest language is generally recognised as fulfilling several if not all language functions, the actual classification then being based on the relative dominance or focus of attention (Jakobson says Einstellung) on one function or another? This effectively begs the question of who or what is focusing the attention (the researcher or the "native"?) and overtly admits that linguistic material can always furnish substantial complex examples able to straddle and thus annul the theoretical distinctions proposed. It might be concluded that these categories by no means contradict Saussure's belief that language itself resists such classification. Indeed, a similar although more practical conclusion might be drawn from Bakhtin's late essay on "speech genres" (1952-53), where the existence of such categories is forcefully asserted, but not one is actually named or delimited.

        What is the basic problem with these approaches to discourse and associated terms? From the perspective of our fairly brutal questions concerning translation, it is that there is no general agreement about the limits of a discourse, nor about the intersubjective level on which its identity should be located. This is why there is no general consensus on whether translation involves one or two discourses. 

        A notion of discourse that seems pertinent to translating

        As much as the term discourse might mean all things to all theorists, our initial questions require the formulation of at least a working definition. Moreover, such a definition should say something more than Hatim's description of discourses as "modes of thinking and talking", which says nothing about limits and remains unclear as to the discursive status of the translator's thinking and talking. 

        A more fruitful approach would seem to be of the kind that sees discourses as sets of constraints on the process of semiosis. This means, first, that discourses are not to be confused with signifiers or utterances, and second, that they can be related to a level of meaning where something happens: "semiosis" is Peirce's term for the dynamic displacement of meaning from symbol to symbol (we would perhaps say signifier to signifier) through the capacities of the interpretant (usually taken to to be a further symbol itself). A simple application of this theory is Jakobson's statement that "the meaning of any linguistic sign is its translation into some further, alternative sign" (1959). A more practical application of semiosis might be what we have done in our commentary on Lewandowski, looking up the dictionary definition of a term, then the definitions of the defining terms, and so on until, according to certain theories, the exercise will exhaust the entire dictionary and take so long that the language itself will have changed, the dictionary will have to be rewritten, and the process should begin again. Or at least, such is the endless process that might be projected by a naïve reading of Saussurean tongue as a synchronic system in which "tout se tient", everything holds together. To say that something called discourse is able to constrain semiosis is thus to posit that, in practice, nobody in their right mind would exhaust an entire dictionary to interpret an isolated term, and that the "holding together" must thus occur at levels of lesser dimension than the tongue. In the case of Lewandowski's dictionary, our short burst of semiosis was effectively terminated by the return to the base term "text". Discourses are perhaps not as open-ended as some theories would like, but they may certainly direct the potentially unlimited process of semiosis in one way or another, and do so with a certain fidelity to the term's etymological value as dis-currere (3) By incorporating at least a nominal notion of translation into the very definition of discourse (for Jakobson, semiosis is translation), this view moreover posits that discourses can bridge the frontiers between different tongues, extending themselves or finding their limits through processes of translation. 

        The reason why such a general notion of discourse could be crucial in the context of translation studies may be appreciated through approaches like Werner Koller's:

        The study of translation investigates conditions of equivalence and describes the coordination of utterances and texts which are in two languages and to which the criterion of translational equivalence applies; it is the study of parole. Contrastive linguistics, on the other hand, investigates the conditions and preconditions of correspondence (formal similarity) and describes corresponding structures and sentences; it is the study of langue. (1979, 183-4) 

        This distinction is of strategic importance in that it detaches translation studies from the search for semantic universals and thus from the obverse problematics of linguistic relativism. But the borderline thus drawn suffers from being equated with langue and parole. The risk with these terms in this context is not only that Saussure declared parole to be unavailable for systematic analysis, but that they might be aligned with something like the more dangerously precise concepts of Chomskyan "performance" and "competence", ultimately reviving the universalist/relativist problematic originally to be avoided. Koller's terminology specifically tends to defeat his strategy when the above categories are used to distinguish between bilingual competence and translational competence (1979, 40ff, 185), implicitly taking something from langue and placing it in the field of parole. Yet parole cannot be equated with the "coordination (Zuordnung) of utterances and texts", nor with the "conditions (Bedingungen) of equivalence". There are thus good reasons for preferring the term discourse.

        Koller's contribution is useful to the extent that it opens a space where discourses, as sets of constraints on semiosis, may involve coordinations and impose conditions of equivalence. It is clear that these coordinations and conditions are not co-extensive with tongues, but it remains to be discovered how they might become pertinent to translation. 

        If, let us say, all discourses existed in all tongues, such that the translator merely had to fill in the linguistic material complying with the coordinations and conditions, then discourses themselves would present no problem for the translator and would thus be impertinent to translation theory. So much for Hatim's normative application. If, on the other hand, discourses are considered to be of lesser dimensions than tongues and entirely determined within particular tongues, their existence must deny interdiscursive translation and thus-since cultures are presumably not equal discursive formations-the very possibility of interlingual translation. Once again, the categories are not useful and tend to block translation theory in the same way as visions of the tongue as a closed conceptual system once did.

        This in effect means that neither of the simple answers to our "one or two discourses" question is useful. If a source text and a target text are equivalent because within the same discourse, translation is a banal phenomenon. And if they are different discourses because in different tongues, then translation would appear to be an unthinkable phenomenon. 

        The only way to cut across this dilemma is to regard translation as the active movement by which discourse may be extended from one cultural setting to another. What translation theory would then want to know about discourses is the relative degree of difficulty and success involved in their extension and the degree to which they may undergo transformation through translation. It is here that translation could become a discovery procedure of some importance to intercultural discourse analysis. 

        Translation as an index of discursive constraints

        There are obvious cases in which the same discourse can be extended across several tongues. To take the example of formal letters, "Dear Sir" is not translated by "Querido señor", but by "Muy señor mío", or "Estimado amigo", or perhaps "Distinguido amigo / colega" etc. The variants are not due to lingual constraints but to discursive conventions. In fact, one could say that even when literalism gives a "correct" salutation (as for example the Castilian "Estimado amigo" rendered by the Catalan "Estimat amic"), the certainty of the translation is due more to discursive constraints than to dictionary equivalents, the choice of the language itself being by no means neutral for the discursive force of the utterance. More precisely, one could say that translational uncertainty in such cases can only be solved in terms of specific contexts-that is, the interrelationship projected by the entire letter-and not by the analysis of structures internal to any one tongue. Although it is thus strictly impossible to say whether exactly the same discourse is always operative in both an original and a translated letter (rules like "use Spanish expression X for English expression Y" would be inadequate), it is certainly by asking if the same discourse is possible that conjectural solutions are reached and discourses are extended one way or another.

        The relatively simple example of formal letters would tend to suggest that translation works on the assumption that the same discourse can indeed be manifested in different cultures. It moreover suggests a fairly simple two-part strategy for isolating pertinent data determined by discursive conventions:

        - First, if translation according to lingual components fails to give corresponding source and target terms, the unit concerned is determined by discursive and not lingual constraints. The fact that "Dear Sir" can be rendered as "Estimado amigo" but not "Querido señor" is thus indicative of potentially equivalent discourses, and possibly of the same discourse.

        -Second, if back-translation fails to attest the equivalence of the pairs thus isolated, then the unit concerned is determined by non-equivalent discourses. For example, the possible circuit "Muy señor mío" into "Dear Sir" into "Estimado amigo" reveals that the Spanish and English discourses, although potentially equivalent for the purposes of a particular translation, are in fact non-equivalent when no translation is involved (although they could possibly become equivalent as a result of repeated translation). 

        If discourse analysis were applied to translation only at the level of potentially equivalent discourses-the first of the above steps-, its findings could not be expected to go beyond those of any reasonable bilingual dictionary, where various target terms are listed according to the fields of knowledge and general situations in which a given source term might be used. Moreover, such attempts to categorise global semantic space willfully overlook phenomena of cultural alterity and historical change. That is, they cannot explain why translation should be necessary, nor accept that translation actually does anything as a discursive act. It is for this reason that the second step given above actively questions the idea that all discourses are possible in all cultures, thereby according translation a specific role in the discovery and challenging of discursive limits. 

        A further, perhaps less banal example might illustrate this critical potential and enable us to close on a suggestive rather than definitive note. 

        The Australian Aboriginal chant

        Nabira-mira, Dumuan-dipa, Namuka-madja, Aï-aïjura 

        has been translated as

        At the time of creation, the Nabira-miras, a father and son, tried unsuccessfully to fish with spears. Their spears were transformed into cliffs known as the Dumuan-dipa, and they themselves became the cliffs of Namuka-madja, which are near the island of Aï-aïjura. (Mountford 1956, 62-63) 

        On the linguistic level, the chant is simply a list of proper names: a pair of mythological characters and three place-names, all of which are reproduced in the translation. But these proper names evoke an entire narrative that relates a mythological event to some very specific geological features. 

        Blithe application of the first of the above steps reveals that the non-correspondence is due to discursive and not lingual constraints. But what cannot immediately be expressed in English is the specificity of these features and the familiarity that allows the associated narrative to remain unexpressed in the chant. The need to explain thus manifests a previously only potential discourse, that of the target text. 

        For obvious reasons, back-translation in this case fails to attest equivalence on the level of discourse. But should one then consider these texts as representing two different discourses, perhaps elaborating the same narrative base in terms of two mutually exclusive cultures?

        If familiarity is untranslatable, this does not mean that its modalities and conventions are necessarily incomprehensible for the translating culture: Australian army battalions habitually hoist banners bearing series of toponyms like "Gallipoli, Ypres, Tobruk", familiar names able to evoke narratives as functionally complex and institutionally pervasive as the Aboriginal chant. Similarly, explanatory discourse exists in Australian Aboriginal culture, albeit often with social limitations for initiatory purposes. As difficult as the translation might seem, it cannot be said that the discursive formations concerned are mutually exclusive. The point is rather that the translation in this case functions as a bridge between initially non-equivalent discourses, at once explaining the chant and thereby, potentially, allowing the same chant to be read as an English text in itself--since it is nothing more than a list of proper names--, the familiarity of which may then increase with repetition. Moreover, since such a translation retrospectively alters the discursive status of the source text, it should clearly count as legitimate discursive work on the same level as the source text, extending the original discourse from tongue to tongue despite quite massive initial non-equivalence. 

        If difficulty in translation may thus be used as a general index of non-equivalent discourses, it does not necessarily follow that translations are simple mirrors of discrete cultures. To take seriously Jakobson's statement that "the meaning of any linguistic sign is its translation into some further, alternative sign"--to read the Aboriginal chant as an English-language text--is to accept that translation may play an active role in our understanding of discourses themselves. 


        1. Ducrot and Todorov (1972, 376) also refer "analyse du discours" to the entry "texte"; discourse analysis as found in van Dijk et al. is similarly freely described as text linguistics, and van Dijk himself indeed suggests that the term "discourse analysis" translates the German "Textwissenschaft" (1985, I, xi). My comments here do not exclude the possibility that these usages be made pertinent to translation through a reinterpretation of the material side, as found, for example, in Robel's equation of "texte" with "l'ensemble de toutes ses traductions significativement différentes" (1973, 8). 

        2. Joly (1988) claims that discourse analysis can be dated from Tobler's study of free indirect discourse in Zola (1887) and the debate that later developed, after 1912, concerning the forms this discourse may take in French and German. The contributions of Vendryès (1923) and Bakhtin/Voloshinov (1930) are also mentioned to offset Todorov's citing of Malinowski (1923) as the first to posit the necessity of discourse by describing language as action (1970, 4). 

        3. Greimas and Courtès (1979) cleverly incorporate this notion into their final definition of "discours" as "une sélection continue des possibles, se frayant la voie à travers des réseaux de contraintes". If I prefer to describe discourse as the constraints rather than the process of semiosis itself, it is because I do not believe that human work should be attributed to abstract concepts. 

        WORKS CITED 

        AUSTIN, J. L. (1955). How to Do Things with Words (The William James Lectures delivered in Harvard University in 1955), ed. J.O. Urmson and M. Sbisà, Second Edition, Oxford: Clarendon Press, 1975. 

        BAKHTIN, Mikhael M. (1952-53). "The Problem of Speech Genres", trans. V. McGee in Bakhtin, M. M., Speech Genres and Other Late Essays, ed. Emerson, C. and Holquist, M., Austin: University of Texas Press, 1986. 

        _____ / VOLOSHINOV, Vladimir (1930). Marxism and the Philosophy of Language, trans. L. Matejka and I. R. Titunik, New York: Seminar Press, 1973 

        BARTHES, Roland (1970). S/Z. An Essay, trans. R. Miller, New York: Hill & Wang, 1974. 

        BENVENISTE, Emile (1966). Problèmes de linguistique générale, I, Paris: Gallimard. 

        BERNSTEIN, Basil (1970). Class, Codes and Control, vol I: Theoretical Studies Towards a Sociology of Language, London: RKP. 

        BÜHLER, Karl (1934). Sprachtheorie. Die Darstellungsfunktion der Sprache, Stuttgart / New York: Gustav Fischer Verlag, 1982. 

        CHATMAN, S. (1978). Story and Discourse: Narrative Structure in Fiction and Film, Ithica and London: Cornell University Press. 

        CROS, Edmond (1983). Théorie et pratique sociocritiques, Montpellier, CERS, Université Paul Valéry. 

        DELISLE, Jean (1984). L'Analyse du discours comme méthode de traduction: théorie et pratique, Ottawa: Editions de l'Université d'Ottawa. 

        DERRIDA, Jacques (1967). De la grammatologie, Paris: Minuit. 

        DUCROT, Oswald, and TODOROV, Tzvetan (1972). Dictionnaire encyclopédique des sciences du langage, Paris: Seuil. 

        ECO, Umberto (1976). A Theory of Semiotics, London and Basingstoke: Macmillan, 1977. 

        GREIMAS, A. J., and COURTES, J. (1979). Sémiotique: dictionnaire raisonné de la théorie du langage, Paris: Hachette. 

        HALLIDAY, M. A. K. (1973). Explorations in the Functions of Language, London: Arnold. 

        HATIM, Basil (1990). "Intertextuality and Idiolect as Intended Meaning: a concern for both translator and literary critic alike: with special reference to Arabic", Parallèles. Cahiers de l'Ecole de Traduction et d'Interprétation, (Université de Genève) 12, Hiver 1990-1991. 

        _____ and MASON, Ian (1990). Discourse and the Translator, London and New York: Longman. 

        HOPPER, Paul J. (1988). "Discourse Analysis: Grammar and Critical Theory in the 1980s", in Profession 88, New York: Modern Language Association of America. 

        HYMES, Dell H. (1968). "The Ethnography of Speaking", in Readings in the Sociology of Language, ed. J.A. Fishman, The Hague, Mouton. 

        _____ (1974). Foundations in Sociolinguistics: An Ethnographic Approach, Philadelphia: University of Pennsylvania. 

        JAKOBSON, Roman (1956). "Closing Statement: Linguistics and Poetics", in Style in Language, ed. T.A. Sebeok, Cambridge MA, New York and London: MIT Press-John Wiley & Sons, 1960. 

        _____ (1959). "On Linguistic Aspects of Translation", in On Translation, ed. R. Brower, New York: Oxford University Press, 1966. 

        JOLY, André (1988). Essais de systématique énonciative, Lille: Presses Universitaires de Lille. 

        KOLLER, Werner (1979). Einführung in die Übersetzungswissenschaft, Heidelberg: Quelle & Meyer. 

        LEWANDOWSKI, Theodor (1985). Linguistisches Wörterbuch (3 vols), Neue bearbeitete Auflage, Heidelberg-Wiesbaden: Quelle & Meyer. 

        METZ, Christian (1970). Langage et cinéma, Paris: Larousse. 

        MOUNTFORD, Charles P. (1956). Arnhem Land: Art, Myth and Symbolism, Melbourne: Heineman. 

        NIDA, Eugene A., and TABER, Charles R. (1969). The Theory and Practice of Translation, Leiden: E.J. Brill. 

        OGDEN, C. K., and RICHARDS, I. A. (1928). The Meaning of Meaning, 8th edition, London: RKP, 1946. 

        PÊCHEUX, Michel (1975). Les Vérités de la palice, Paris: François Maspero. 

        REISS, Katharina (1971). Möglichkeiten und Grenzen der Übersetzungskritik, Munich: Max Heuber Verlag, Heuber Hochschulreihe 12. 

        ROBEL, Léon (1973). "Translatives", Change 14 (Transformer, traduire), Paris: Seghers Laffont. 

        SAUSSURE, Ferdinand de (1916). Cours de linguistique générale, ed. C. Bally and A. Sechehaye; trans. W. Baskin: Course in General Linguistics, Glasgow: Fontana/ Collins, 1974. 

        SEARLE, John R. (1969). Speech Acts. An Essay in the Philosophy of Language, Cambridge: Cambridge University Press. 

        SHKLOVSKI, Victor (1917). "L'Art comme procédé", trans. T. Todorov, Théorie de la littérature: Textes des Formalistes Russes, ed. T. Todorov, Paris: Seuil, 1965. 

        TODOROV, Tzvetan (1971). Poétique de la prose, Paris: Seuil. 

        _____ (1978). Les Genres du discours, Paris: Seuil. 

        VAN DIJK, Teun A. (ed.) (1985). Handbook of Discourse Analysis, 4 vols, San Diego: Academic Press. 

        VENDRYES, Joseph (1923). Le Langage. Introduction linguistique à l'histoire, Paris: Albin Michel, 1968. 

        WIDDOWSON, H. G. (1974). "The deep structure of discourse and the use of translation", in The Communicative Approach to Language Teaching, ed. C.J. Brumfit and K. Johnson, Oxford University Press, 1979 

        _____ (1979). "Directions in the teaching of discourse", in The Communicative Approach to Language Teaching, ed. C.J. Brumfit and K. Johnson, Oxford University Press. 

© Anthony Pym 2014
URV. Av. Catalunya, 35
45002 Tarragona, Spain
Fax: + 34 977 299 488