Translation and Language

Peter Fawcett

• Publisher: Routledge
• Year: 2014
• eBook ISBN: 9781317642305
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

Translation Studies and linguistics have been going through a love­-hate relationship since the 1950s. This book assesses both sides of the relationship, tracing the very real contributions that linguists have made to translation studies and at the same time recognizing the limitations of many of their approaches. With good humour and even­handedness, Fawcett describes detailed taxonomies of translation strategies and deals with traditional problems such as equivalence. Yet he also explains and assesses the more recent contributions of text linguistics, sociolinguistics, pragmatics and psycholinguistics.

This work is exceptional in that it presents theories originally produced in Russian, German, French and Spanish as well as English. Its broad coverage and accessible treatment provide essential background reading for students of translation at all levels.

1. Introduction

DOI: 10.4324/9781315760483-1

A troubled relationship

Modern linguistics began in the early twentieth century with the work of the Swiss linguist Ferdinand de Saussure. He focused on the notion of language as a system at a given moment in time (a synchronic approach) at a highly abstract level that uncovered powerful principles about the way in which language in general is structured. The structuralist model he produced was to prove immensely influential when, much later, it was taken up by anthropologists, literary critics and philosophers as the one model that would apparently explain what we had always wanted to know about life, the universe and everything.

Since linguistics is the study of language and has produced such powerful and productive theories about how language works, and since translation is a language activity, it would seem only common sense to think that the first had something to say about the second. Indeed in 1965 the British scholar John Catford opened his book A Linguistic Theory of Translation with the words: “Clearly, then, any theory of translation must draw upon a theory of language — a general linguistic theory”. In exactly the same year, however, the famous American theoretical linguist Noam Chomsky was rather more sceptical about the implications of his own theory for translation, saying that his theory “does not, for example, imply that there must be some reasonable procedure for translating between languages” (1965:30). Although no expert in translation, Chomsky nonetheless divined that there was something about the activity that put it beyond reason. Perhaps he had read what the academic Ivor Richards (1953:250) said about translation: “We have here indeed what may very probably be the most complex type of event yet produced in the evolution of the cosmos”.

This uncertain relationship between linguistics and translation theory continued to be reflected in the literature. Eight years after Catford’s and Chomsky’s pronouncements, the German theorist Jörn Albrecht (1973:1) expressed regret and astonishment that linguists had not studied translation; yet the Soviet linguist Aleksandr Shveitser, writing in the same year (although quoted here from the later German translation), made the opposite claim: many linguists had long since decided translation could indeed be an object of linguistic study (1987:13). He rejected the idea that linguistics can explain only the lowest levels of translation activity, saying this was based on too narrow a view of linguistics. He did, however, refer briefly to the furore caused by the first major attempt by a Russian scholar to produce a linguistic description of translation (Fedorov 1953), which provoked lively polemic and liberal accusations of ‘deviation’ (see Cary 1957:187).

The intervening years have not resolved the tension. Almost thirty years after the Catford-Chomsky declarations, the English academic Roger Bell (1989:xv) claimed that translation theorists and linguists were still going their own separate ways. The French scholar Maurice Pergnier has pointed out that even though linguistics has developed in ways that make it much more relevant to the concerns of translation, there are still those who would like to liberate translation completely from its sway (1993:9). Indeed, his compatriot Marianne Lederer is just one among many who dismisses linguistics from translation studies: “I hope in this way to bring out the reasons why translation must be dealt with on a level other than the linguistic” (1994:87).

Such a position is provocatively extreme. Linguistics quite clearly does have something to offer the study of translation, and in these pages we shall be exploring what that is. At the same time, however, we shall be pointing out the limitations of the discipline, especially if people want to see translation as an entirely linguistic activity or want to use linguistics as a recipe giving ready-made solutions to specific translation problems rather than as a resource for extrapolating general problem-solving techniques from specific concrete problems.

The relationship of linguistics to translation can be twofold: one can apply the findings of linguistics to the practice of translation, and one can have a linguistic theory of translation, as opposed, say, to a literary, economic or psychological theory of translation.

In the first approach, a subdivision of linguistics such as sociolinguistics might have something to say about the way in which language varies in relation to social status, age, gender and so on. It will enable us to recognize these variations and describe them. And when we have to deal with sociolinguistic variation in a text to be translated, linguistics can provide one input in deciding how to cope with the situation.

In the second approach, rather than applying linguistic theory to elements within the text to be translated, one can apply it to the entire concept of translation itself. Thus the theory of dynamic equivalence put forward by the American scholar Eugene Nida, which we consider below, can actually be seen as nothing less than a sociolinguistics of translation, describing the way translators can adapt texts to the needs of a different audience in the same way we all adjust our language to suit the people we are talking to.

Both of these approaches are found in writings on linguistics and translation, and we shall try to signal them as we go along. For the remainder of this chapter we shall follow the first approach, giving an overview of the basic concepts and main divisions in structural linguistics to see how important they are in translation, then coming back to them in greater detail in later chapters.

Langue/parole

Saussure made it possible to see language as a set of structured systems rather than a ragbag of bits and pieces. Some parts of language, such as grammar, have always been thought of as systems, of course. But the structuralist linguistics that emerged from Saussure’s work attempted to uncover the systematic and structured nature of other parts of language: the sound system (phonetics and phonemics), the grammar system (syntax, which is word order, and morphology, which is word shape) and the meaning system (semantics).

For linguistics to make progress in describing these systems, Saussure thought it necessary to distinguish between what he called ‘langue’ and ‘parole’ (the terms are often used in their French form in other languages, because, ironically, it can be difficult to find translation equivalents). This is the difference between the abstract language system (langue or ‘a language’), which Saussure saw as the object of linguistics, and actual uses of language (parole or ‘speaking’), which were thought to be too variable for systematic, ‘scientific’ study because the factors involved were too numerous and too random.

An example might illustrate this: After a certain amount of alcoholic intake you might say I’ve got a shore head when you mean sore head. Now, although linguistics can describe the difference between s and sh in phonemic theory, in this particular case the difference has no linguistic meaning; it is a matter of parole; it is a one-off event that has no function in the language system. By contrast, the difference between sore and shore in the non-alcoholic I got a bit sore sitting on the shore does have a function in the language system: the sound opposition in this case serves to mark out a change in meaning, and it does so on a systematic basis (single/shingle, sin/shin etc.) These differences are a matter of langue.

This distinction between langue and parole, and the insistence that linguistics should study only langue, led to tremendous progress in the discipline. Yet the early linguistic approaches to translation that tried to follow the same line led to considerable dissatisfaction. To many translators and translation theorists the findings seemed sterile, leaving out many things of interest to translation. The German scholar Dieter Stein (1980), for example, went so far as to declare that the linguistics of langue had little or nothing to offer translation studies (which is to forget that language structure can be a serious problem in translation).

The langue-oriented approach can certainly produce useful comparative descriptions of language systems, and, as the Canadian translation theorist Jean Delisle says, such things must be a part of every translator’s knowledge (1988:78). I can scarcely envisage being a translator if I don’t have that basic command of my languages. But these things by no means exhaust the problems of translation. They belong to what the German theorist Werner Koller (1979:185) calls ‘foreign language competence’, knowledge that is basic to, but not the whole of, ‘translator competence’, because simply knowing two languages is not all that is needed to be a translator, as these pages will make abundantly clear.

Stein advocated what he called the ‘Sit/Text’ approach, which involved data of a textual and situational nature. This would require a linguistics of parole rather than of langue and would allow us to account for such things as the drunkard’s shore head, which is vital for translation. The French theorist Jean-René Ladmiral also claimed that “translation is a communication operation guaranteeing identity of parole through differences of langues” (1979:223), while Albrecht reminded us that “what is being translated are not ‘codes’ or languages but ‘messages’ or texts” (1973:26), in other words parole not langue. For Koller (1979:183), translation theory is “a science of parole”.

The problem was that parole-oriented linguistics was scarcely developed. There was thus a fear that abandoning the langue-oriented approach would mean giving up any attempt to turn translation theory into a scientific theory that would rescue it from the earlier dilettante approaches. Even though the linguistics of parole is now better developed, it makes use of what the Croatian scholar Vladimir Ivir (1996:153) calls ‘ad hoc categories’ that do not have “theoretical coherence and scientific rigour” because they are “not amenable to … theoretical treatment”. The irony is that by the 1990s the whole idea of a scientific approach to translation had come under fire anyway.

The view that translation must be studied as parole (a communicative event) rather than langue (an abstract system) is now widely accepted, to the extent that an author like Pergnier (1993:223) can refer to it as a ‘fact’, and an important fact, since, as he says, it is because translation is a fact of parole that there is no such thing as the one ‘right’ translation of a message.

The langue/parole distinction is a very high-level distinction, concerning as it does the entire language. Saussure’s other major distinction concerned one of the lowest levels of language, but was, if anything, even more revolutionary in its consequences.

Signifier/signified

If language is a structure, it must have component parts. The most important of these is the sign, a technical concept intended to get away from the notion of ‘word’, which is notoriously difficult to define. The sign itself is a structure that has two parts: the signifier and the signified, both of which are mental states. The signifier is a mental image of the physical sound that you make when you say, for example, cat or koshka (or mimi if you speak Tahitian), while the signified is a mental concept or representation of physical cats in the real world.

One of Saussure’s key claims is that the link between the signifier and the signified is not given by some Supreme Being or by Nature, as many nonlinguists believe, but by society. The relation between the two is an arbitrary social construct. A tubular object consisting of meat and other ingredients wrapped in a casing is assigned the signifier sausage in English and Wurst in German. Neither of these is more ‘right’ than the other. There is nothing magic in the object itself that makes sausage a ‘better’ word than Wurst. It might seem that way, especially to people who speak only one language, but in reality the link is purely arbitrary and no particular language has the ‘right way’ of saying things.

Perhaps one of the easiest ways to understand this is to consider the phenomenon of political correctness. In a short space of time the supporters of this ideology managed to create a whole new set of words to talk about things for which they previously used the same words as everybody else. They then tried, with varying success, to make their new words the social norm. Altering an entire vocabulary in this way can only be possible because of the arbitrariness of the signifier-signified link.

Paradoxically, the same PC phenomenon demonstrates the deep-rooted belief among many people that there is a special link between signifier and signified. The attempt by some feminists to write the word man out of existence, even as a component of words where it has no connection with ‘male adult’ (so that emancipate becomes ewomancipate), suggests a very strong belief in word magic, in the power of the signifier to shape the way we think about the signified. There is a joke about two farmers watching pigs wallowing in the mud. After a time one says to the other, ‘No wonder they be called pigs’. This view is put with admirable succinctness by the comic novelist Terry Pratchett (1989:132): “All things are defined by names. Change the name, and you change the thing”.

This kind of belief is not entirely irrational. If the signifier-signified link is arbitrary, then translation would be very simple: you would identify the signified, strip away the source-language signifier, and replace it with the target-language signifier. According to this primitive theory of translation we would read the sign sausage, identify the language-independent signified denoted by the signifier, find its German signifier, and make a simple substitution: sausage would become Wurst. Translation would be a job for computers (a vast topic in itself which would be too technical for the present book to cover).

However, things are not that simple. Signs do not just signify (point to things in the real world). They also have value derived from language-internal structuring which is not the same from one language to the next. The words cat and koshka don’t have the same range of meanings, so their value is different. English has two words wood and forest for the one Russian word lies, so again the values are different.

But words also carry a superstructure that is often referred to by the term ‘connotation’. We think of some words as ‘good’ (grandmother, baby, chocolate) and other words as ‘bad’ (spider, snot, slug). But these connotational meanings are highly variable even within a language (some people don’t like babies, others may have a fondness for spiders, while grandma may be the proverbial ‘grandmother from hell’) and they are often different between cultures. French people sitting down to eat remain calm in front of a plate of snails. Many English people would react differently, and so for them the menu may offer escargots, promising the exotic not the slimy. Connotation has proved difficult for linguistics to formalize, but we shall look at one useful attempt below.

Paradigmatic and syntagmatic: word sets and collocations

In addition to having its own internal structure, the sign can be structured in two other ways. Signs can be joined up in a string, and they can be grouped in a bundle. This is often called the ‘chain and choice’ model, and we shall see examples of how a translation problem that cannot be solved at one point in the chain may be solved by an appropriate choice at some other point. In the first case (making the chain) we produce word sequences: in the restaurant we can string words together to say ‘I’d like sausage and chips, please’. The order in which we put the words is not normally random. It is governed by ‘syntax’, the rules of our language whi...

Table of contents

• Cover Page
• Half Title Page
• Title Page
• Copyright Page
• Table of Contents
• Foreword
• 1. Introduction
• 2. Sub-Word Components
• 3. Semantics
• 4. Translation Techniques
• 5. Equivalence
• 6. Beyond the Word
• 7. Beyond the Sentence: Context and Register
• 8. Text Structure
• 9. Text Functions
• 10. Sociolinguistics
• 11. Pragmatics
• 12. Psycholinguistics
• Conclusion and Perspectives
• Glossary
• Bibliography