Henry Schogt Semantic Theory and Translation Theory - Классики о переводе - Каталог статей

Henry Schogt Semantic Theory and Translation Theory Although there is an undeniable and very important link between semantics and translation, the one dealing with meaning, the other with transfer of meaning, semantic theory and translation theory are not closely connected, and often translation specialists are even bothered rather than helped by the tenets of semantic theory. It is interesting to examine the two domains of semantics and translation in order to find out whether explanations can be given for this state of affairs where there is little positive interaction between the two. Saying that semantic theory is concerned with meaning does not tell us more than a concise dictionary would do. So it is necessary to list some of the major questions semantics investigates. Without any claim to being exhaustive or to establishing a hierarchical order of importance, we can mention the following points: (1) the philosophical, epistemological problems of the relation between language, thought and the outside world; (2) the relation between a meaningful element of a language and the other elements of the same level of analysis one finds in that language; (3) the communication between individuals who speak the same language, the communication being either oral or written. 1. The first problem, although of a general character and not tied to any specific language, has important implications for the translator. If there is a link between language, thought and reality (Language, Thought and Reality is the title of a volume of selected essays by Benjamin Lee Whorf),1 different realities engender dif ferent languages, but also different languages shape different realities. In its most extreme form the so-called Sapir-Whorf hypothesis leads to the view that communication between two people who do not share the same native language is impossible, even if one of them has learned the language of the other. Even those who think they have learned a foreign language remain prisoners of their mother tongue's value system, and are therefore incapable of truly communicating with those whose language they think they have mastered. This is a typical example of a clash between theory and practice: nobody accepts the Sapir-Whorf hypothesis completely and yet it is difficult to deny that the language one speaks focuses on elements of the outside world and creates abstract notions that other languages may leave either unnoticed or, in the case of abstract notions, unconceptualized. Verbal systems may differ in that one puts great emphasis on temporality and the anterior/posterior scale, whereas another has an overt marking of aspect, a category that does not have the deictic character of temporality.2 Even if the first language is able to express aspect and the second language temporality, these categories do not have a prominent position, so that there is no real equivalence between the two languages as to temporality and aspect. 2. This leads to the second question of semantics concerning the immanent analysis by which each element is defined with respect to the other elements of the same level and thanks to them. This principle of structuring linguistic elements and describing them from a functional point of view, rather than referring to any physical features they may have in isolation is best known from phonology/phonemics as presented by Prague structuralists, in the first place by Trubetzkoj.3 Attempts to structure meaningful units in the same way as phonemes have been only partially successful until now. When discussing the verbs craindre and redouter and the verbal expression avotrpeur Saussure already pointed out that the value of each of the three units was determined by the existence of the two others.4 Unfortunately Saussure does not specify in what respect they differ and whether the differences are part of the intellectual content or not. This opens up the question of connotation and denotation and the division into noncognitive and cognitive elements of meaning that plays such an important role in translation discussions. Semantic or notional field studies illustrate more clearly than the example given by Saussure the possibilities and the limitations of the structural method as applied to the lexicon. The field idea is partly to be seen as an answer to the problem of structuring units belonging to an open, unlimited inventory. Contrary to phonemes, that are for each language a restricted, closed category with a number of units rarely exceeding 75—and mostly much lower than that—lexical units are so numerous that the principle according to which each unit influences all the other units at the same level is invalid for the lexicon as a whole, or if one wants to maintain it, without any verifiable significance. In the much smaller and more homogeneous lexical field, however, one can observe a situation similar to that prevailing in a phonological system. There are, however, some differences in spite of the similarity: a. Phonemes are discrete units, whereas the signified of lexemes is often much fuzzier. It is difficult to give general rules: while in some instances the meaning is discrete (e.g., kinship terminology), in most cases there is no clear delimitation of the different meaning-areas, but rather fuzzy edges and overlap. (Saussure's craindre, avoir peur and redouter fall into this category). b. Within a speech community there are only slight variations in the phonological system. Those variations are a group phenomenon, age, sex, social class, and regional origin being determining factors. Lexical field structure knows much wider deviations, and whereas the number of units in the phonological systems of a speech community is for all variations very close to the average (in other words, the difference between maximum, average and minimum is small), the number of lexical units in a given field may vary greatly from one individual to the next. This has important implications for the precise, lexical meaning of each item: the fewer terms there are, the less specific their meaning becomes. It should be noticed that grammatical categories such as person, number, case, tense, mood, and aspect come much closer to the phonological model. c. Outside reality, social structure, way-of-life vary from one speech community to the other. Semantic fields reflect these differ ences (sometimes no longer pertinent in contemporary society) A unit may belong to different fields in different languages. For Dutchman potatoes belong to the same field as noodles and rice (starch component of the meal), whereas in France they are listed with vegetables. If the fields provide generic terms, those terms will be different in cases such as the potato one. d. Immanent analysis has to be complemented by knowledge of reference. As reference brings in the outside world as well as the mental constructs that are without material denotation, the experience of each individual speaker colors his or her frame of reference in a special way, even within one and the same speech community. If, as is very often the case, the semantic or lexical field study deals with one language, individual variations are not taken into account and the field that is constructed is supposed to be representative for common usage. Comparisons between fields in two or more languages are based on the same assumption of general validity of each field within its own speech community. By adhering to the principle of abstraction and generalization of the linguistic sign, the semantic field studies stay in the realm of langue, or competence. However, translators work with texts, and operate at the level of parole, or performance. 3. Whether they are called speaker and hearer, sender and receiver, or encoder and decoder, the communication model postulates someone who forms a message and someone who receives that message and interprets it. If both persons follow the same grammatical rules, and have the same lexical inventory, the message gets across without change, and the communication is successful. This rosy picture does not take into account all the complexities that have been investigated by speech act descriptions and pragmatics. A simple example will suffice to illustrate the problems one may encounter. When I needed some crucial information for a paper I was writing, I phoned a friend for help. His three-year-old 5. Russian kinship terminology offers a good example. The intricate system, still functional in rural society, lost its pertinence elsewhere in the nineteenth century. Tolstoy's unexpected shift from brother-in-law (brother of the wife): Surin to brother-in-law (husband of the sister): zjat' in section 3 ofThe Death of Ivan Ilič may be a result of terminological confusion. In Besy (The Possessed, pt. 2, chap. 1, sec. 5) Dostoyevsky makes one of his characters confuse the terms for mother-in-law (mother of the husband): svekrov' and daughter-in-law: snoxa. son answered the phone, as he loved to do. To my question, "Is your father home?" he answered "Yes," and left the phone dangling, without calling his father. The semantic interpretation of my question was correct, but the communication was not really successful. Semantic theory focuses on cognitive meaning and leaves the complexities of intention and innuendo to other disciplines. It should be mentioned that John Lyons in his book on semantics deals extensively with the problems of illocutionary acts, paralin-guistic phenomena, and multiple interpretation levels.6 He seems to use "semantics" as a generic term, and then use the same signifier as a hyponym for the branch dealing with cognitive meaning, thus falling in line with traditional usage. The Geneva school, with Charles Bally and Albert Sechehaye and the Dutch author of a French syntax, Cornelis de Boer,7 puts emphasis on the distinction between old and new information in the theme and rhème or thème and propos theory. This theory foreshadows the notion of "foregrounding" that was introduced as a refinement of transformational description. Although the constituent elements are the same, different word order, different stress, or different construction result in a shift of focus, thus signaling different elements to the special attention of the listener/reader. When source language and target language do not have the same devices to create these special effects, the translator may find himself at a loss. Studies such as the one by Claude Hagege on the typology of languages show the extent of this problem.8 Before discussing any further the connections between semantics and translation, a few words should be said about the models presented by Klaus Heger9 and Luis Prieto.10 Both make an attempt to go beyond the langue / parole dichotomy. That dichotomy has proven to be a major stumbling block when the paradigmatic linguistic sign has to be integrated into a syntagmatic message. For Heger the Saussurean sign is what he calls a signème, stress- ing by the choice of this term the -emic character of the unit. Its signified contains all the elements (sèmes and noèmes, the difference between which is explained by Heger, but is not relevant for his purpose) that may become pertinent in any instance of actual use of the term in question. Only part of the inventory is pertinent in each separate instance of actual use, depending on context and situation. Heger calls sémeme this actualized part of the signème. One could also call it, according to his theory, the signified of the sig-neme monosémisé. Heger does not indicate any procedure for establishing the inventory of all the sèmes (and noèmes) of a given sig-nème. He may have Bernard Pottier's method11 in mind or some other kind of componential analysis, it does not matter which one, the problem with all methods being that no formal means of verification can be given. (One should remember in this respect the criticisms made of Katz and Fodor and their bachelor example.)12 Luis Prieto stays, as does Heger, within the Saussurean structuralist tradition, but his approach is different. His sign is not the isolated lexical unit of which Saussure gave "arbor" as an illustration, but rather what Prieto calls a complete message. Linguistically the message may not be a complete sentence if circumstances allow one to leave out elements. Prieto starts from the signified, assuming somewhat optimistically that people want to say something before they start to speak. How they say what they want to convey depends on a number of factors, the first and most obvious one being the language that is used. Furthermore context and cx-tralinguistic situation are taken into account. Prieto's originality lies in the way he interprets the notion of extralinguistic circumstances. For him not only the physical environment, but also the personality of the speaker, his assessment of knowledge and personality of his interlocutor and cultural traditions are part of the extralinguistic setting of the linguistic act. So, depending on context and extralinguistic circumstances (or, in North American terminology, linguistic and nonlinguistic context) the same message may take different forms, be expressed by way of a different signi-fier. As for the interpretation Prieto follows the same method. Interpretation depends on the same factors and will vary according to context and extralinguistic circumstances. The same signifier, then, may correspond to different signifieds. Prieto thus builds a framework in which he can accommodate sentence and utterance. He operates as a true Saussurean within the limits of one language and, although the flexible circumstances allow for the incorporation of generation differences, the description is basically syn- chronic. Prieto's starting point, the desire to say something, is reminiscent of the transformational semanticists who put the semantic component in the deep structure, contrary to what the interpretative school is used to doing. Prieto does not go into detail about the relationship between syntax and semantics. The most exhaustive study on that topic is undoubtedly by the Dutch linguist and slavicist Carl Ebeling.13 Ebeling adheres to the principle of intentionality, including word order and stress, but leaving out sociolinguistic and regional involuntary indices. The intricate system of interrelationships and the abstract form in which the more than two hundred rules of the discovery procedure are presented form, however, a major stumbling block for those who want to apply Ebeling's important theory to solve practical problems.14 The picture one gets from this rapid and incomplete survey— the Saussurean principle of arbitrariness of the sign and the objections raised by others against that principle were not even mentioned—is not very encouraging for the translator for the following reasons: 1. immanent analysis yields for each language a unique set of components and there is never a one-to-one relationship between components belonging to different languages; 2. no method can guarantee the obtention of an exhaustive inventory (smallest meaningful elements not necessarily formally expressed in the signifier); 3.even within one speech community there are considerable differences of interpretation as well as of formulation (Prieto); 4.expressive, emotive, and social elements, though very important in the communication process, are often not included in semantic description because they are of doubtful intentionality; 5.being of a synchronic character, analyses are only valid for a certain point in time; being linked to a—sometimes idealized—form of speech of a given speech community, they are also geographically determined; 6.and then, there is the Sapir-Whorf hypothesis to deal a final blow to the translator. It is interesting to see how those who are not only translators, but also theoreticians of translation, cope with this mass of negative views on the possibility of interlanguage communication and translation, and how they incorporate some elements of semantic theory in their work, while discarding others. To begin with the point just mentioned: as we said before, no translator accepts the Sapir-Whorf hypothesis in its most extreme form. The mere fact that interlanguage communication and translation have been going on for thousands of years is considered sufficient proof that Sapir and Whorf were wrong. Sometimes univer-sals of language are used as a counterargument. However, apart from some basic principles (André Martinet's double articulation, the conventional character of language, the fact that all languages express actors and actions and can qualify those elements) most universals are of a relative type. If a language expresses A, then it is (almost) certain that it also expresses B. For the translator this sort of universal is not so helpful, because he may have to deal with two languages, one of which has A and B, while the other has only B or even neither B nor A15 So even when the Sapir-Whorf view is rejected, the translator has to deal with differences in structure and of means of expression. Most translation manuals stress the fact that, in spite of differences in value in the immanent analyses, there may be equivalence in actual signification at the level of utterance. In other words, even if the total inventories of semes of the -emic signs are different, the pertinent semes (Heger's sememe du signe monosémisé) may be identical. If some semes are lost, they may be added in other parts of the sentence by way of compensation. A special case arises when a distinction that is optional in the source language and compulsory in the target language is not expressed in the source text. Roman Jakobson, who pointed out this problem where the translator has to add information not provided by the source text, believes that the passage from compulsory to not compulsory can always be made, and that lacunae can always be filled by circumscriptions or neologisms.16 Whether or not, in the case of literature, the aesthetic value of the source text is destroyed by such solutions is a moot point. The translator has to decide whether to maintain the cumbersome element, to replace it, or to leave it out. His choice will often depend on the public he is translating for. His role as encoder in the target language after decoding the source language puts him in the position that Prieto describes, where the precise form of the message is adapted to presumed knowledge and background of the receiver and, one might add, in some instances to his expectations. Do all these comments form a sound basis, or at least a starting point for a theory of translation? If existing work by linguists is any indication, the answer to this question turns out to be either a qualified 'Yes" or a qualified "no," depending on the kind of translations one has in mind. For texts where cognitive meaning prevails and formal expression of that meaning has no other function than expressing that meaning, the semantic theories are helpful, and are adequate for describing the double process of encoding and decoding. For literary texts, it is precisely the literariness that falls outside the domain of semantics. That fact reduces the importance of semantics considerably, although it keeps its significance for the denotative element of the text. So it is not surprising that a linguistic theory of translation of general applicability does not exist. In most instances the authors of theoretical linguistic studies on translation do not give a general theory, but make a series of theoretical observations often in connection with specific types of translation. Sometimes this approach is reflected in the title: Problemes théoriques de la traduction;17 Tra-duire: Théorèmes pour la traduction.™ In other cases such as A Linguistic Theory of Translation,19 the title does not give away the fact that the book contains more or less independent chapters dealing with various problems without tying them together. Nida20 and Nida and Taber21 give a much more coherent and comprehensive analysis of translation problems, but the authors are able to do so by limiting themselves to one specific text, the Bible, seen not as a work of art, but as a text written to educate and instruct, to convince and convert. Therefore some of the choices that the translator has to make do not cause any problems. They do not maintain what Antoine Berman calls l'étrangeté du texte,22 but adapt their translations to the system of the target language, and replace idiomatic expressions and metaphors by more or less equivalent ones or circumscriptions. The clarity of the message prevails: level of speech and accessibility should be adapted to the audience the message is intended for, regardless of whatever the level of the original passage may be. So with Nida and Taber we are already entering the area of specific texts and specific audiences (hearers or readers), and have left behind the general level on which linguistics operates. And that is precisely the reason why semantic theory and translation of literary texts have so little common ground. Whereas semantic description looks for what is generally valid and systematic, literary translation requires the analysis of the idiolect of the source text, not only from a semantic point of view but also with respect to all the intentional and nonintentional indices that are deemed important in that text. It is obvious that Brian Fitch's megatype, where more than one language is involved, falls outside what can be described in linguistic terms.23 There is no general linguistic theory cither for François Péraldi's r érectil,24 for Michael Riffaterre's purée septembrale,25 or, as Solange Vouvé has so clearly demonstrated, for translating Finnegans Wake.26 There are, of course, other translation theories, but they deal most of the time with specific types of literary texts, and put much emphasis on the aesthetic, the psychological, the subconscious. Very often they take almost for granted the semantic denotative component of translation. One has to admit that linguists dealing with semantic theory do not have the same preoccupations as literary analysts and translators of literary texts. This does not mean, however, that it is not useful and even important that linguists and literary specialists listen to each other and try to understand each other's point of view.
1 2 3 4 5 Категория: Классики о переводе \| Добавил: Voats (20.09.2011)
Просмотров: 2523 \| Рейтинг: 2.0/1