Andrew Chesterman (2003c)


Does translation theory exist?

(Summary of a plenary lecture in Turku, April 4, 2003)


Kääntäjä 6, 4.



The question in the title is one of definition and classification.  In other words:

  What do we mean by “theory”? And what do we mean by “translation”?

  What do we mean by “translation theory” (or theories)?

  Is “translation theory” in a class of its own, distinct from other theories? I.e. does translation have its “own” theory or theories?

How important are such questions? How good does a definition or classification have to be? What is the point of them? Definitions and classifications are tools, they are means; they are not ends in themselves. We need definitions in order to form concepts, categories and classifications; and we need classifications in order to be able to make descriptive generalizations, which in turn may lead to explanations, predictions, perhaps even applications. Definitions and classifications only need to be as precise as their purpose requires.


What do we mean by theory?


A theory is, etymologically, a way of looking at something, contemplating it. The notion of a “theory” is fuzzy, but we can recognize different kinds of theory. All theories aim to increase understanding.

Theory as myth: a myth can be seen as a narrative that aims to explain a mystery (death, the seasons, the multiplicity of languages in the world (Babel)...)

Theory as metaphor: a metaphor (or simile) enables us to understand something better, by comparing it to something else. (Translating is “copying”, “like cannibalism”,...)

Theory as model: on this view, a theory is a model showing the main elements and relations of a given phenomenon; a simplified representation. (E.g. the common communication model of translation, with sender, message, receiver, new sender, etc.)

Theory as hypothesis: this is e.g. Popper’s view, according to which theories are tentative hypotheses about solutions to problems; we then need to test them.

Theory as structure: Lakatos argues that a theory is a structured set of explanatory principles (or a research programme), with a central core of basic assumptions surrounded by a “protective belt” of hypotheses that are open to testing.



What do we mean by translation?


The notion of “translation” is also fuzzy. Let’s assume it includes both product and process, and also interpreting. We only need a working definition, so let’s say that a translation (product) is a text in one language (target language) that counts as another text in another language (source language) for some purpose.


What is a translation theory?


There are many theories in Translation Studies, but no single theory or research programme that covers “everything”. People use the term in different ways. We talk of skopos theory, polysystem theory, and interpretive theory, for instance, but of the manipulation school. But there are dozens of approaches and models that could be called theories: equivalence-based, sociological, cognitive, textlinguistic, systemic, cultural... Loosely speaking, a translation theory is simply a way of looking at translation, a way that helps us to understand it.


Do these theories share anything?


There is some shared ground – i.e. areas of research that all scholars seem to recognize as interesting and relevant, even though they might not be working in a given area themselves. (See the Forum debate in Target, from 2000 to 2002.) Here are some central theoretical questions. (I do not include questions about applications, such as translator training.)


1. What do we mean by “translation”, “käännös” etc.? What kind of concept is it (classical / prototype / cluster / ...)? How do we distinguish it from related concepts? Is there a “universal” concept? [Assumption: “translation” and “käännös” (etc.) are, to some extent at least, translations of each other.]

2. What kinds of translation are there? How do translations vary under different conditions? Can we form a universal typology? What are the most useful principles of classification? [Theories of text-types, translation functions...]

3. How can we best describe the relation between translations and source texts? [Theories of equivalence, shifts, strategies...]

4. How can we best describe the relation between translations and their target languages? E.g. how to measure the degree of naturalness or “fit”? [Theories of rhetoric, style, textlinguistics, discourse analysis.]

5. Are there general charactistics of translations across different languages, text-types etc.? Do we find repeated patterns? Might there even be universal characteristics? If there are, how might they help to explain why translations are the way they are? What do translations tell us about the nature of language? [Corpus linguistics...]

6. How can we best describe the cognitive process of translating / interpreting? How might information about this process help us to explain certain features of translations? What do translations tell us about the mind? [Think-aloud protocols, relevance theory, psycholinguistics...]

7. How can we best describe the sociological process of translating, how do translators work in practice? How might information about this process help us to explain certain features of translations? How do translations help us to understand the nature of communication in general? How do amateurs become professionals? (Translator’s role, status, client relations, networking, teams...) [Skopos theory, functional theories, communication theories...]

8. What are people’s attitudes / reactions to translations of different kinds, to different aspects of translations? What ideas do people have about translation quality, under different conditions? How can translation quality best be assessed? [Theories of quality assessment, prescriptive theories, discourse of translation]

9. How can we best describe the cultural context and significance of translation? How are translations affected by, and affect, cultural norms, languages, repertoires and perceptions? How do translations help us to understand more about culture in general? [Polysystem theory, norm theory, reception theory...]

10. What is the ethical dimension of translation? [Moral philosophy, professional ethics, language policies...]

11. What is the history of translation? What was (not) translated when, by whom, how, why, with what effect? What do translations tell us about history?

12. Can machines translate? To what extent? How can we make better translation aids? How do computers affect the way translators work?


Do we have a shared paradigm?


What we have so far, in translation theories, is lots of conceptual analysis,

lots of ways of describing translations, and lots of ways of explaining. Does this add up to a shared paradigm, containing several complementary theories?

         No. We have a shared research object: translations and translators. To some extent we have shared goals, but only at a very general level: goals of definition, description, explanation, application. We do not have shared kinds of data: our data may be translations, source texts, non-translated texts, archives, think-aloud protocols, questionnaires, interviews, keystrokes, reviews, paratexts, contracts, revisions... We do not share our methods, either, although the field does have at least two core methods: contrastive text analysis and conceptual analysis. We do not really share a conceptual framework.


Several paradigms?


These days we seem to have at least four paradigms: linguistic, sociological, cultural and cognitive. Machine translation research could be seen as a fifth paradigm. Our research topics often overlap between these paradigms.




Many (all?) translation theories are applications from other fields; many of our actual concepts also come from elsewhere (equivalence, strategy, norm, universal...). Does this matter? It has both advantages and disadvantages.

It may help to give the discipline itself a respectable status as a new field bridging more traditional disciplines. It allows us to propose richer and more powerful explanations for translational phenomena. Interdisciplinarity pushes us towards the goal of consilience, the unity of all knowledge, because it helps us to make new connections across disciplines.

         On the other hand, interdisciplinarity easily leads to fragmentation, superficiality, and misunderstanding, as we borrow concepts and methods from other fields that we may not adequately understand.

         Many research problems in translation studies are themselves interdisciplinary, as are some of the social problems that translation theory might be expected to help with. What would be the best language policy for the EU, for instance?


So is the question in the title actually a good question? Yes and no...


Recommended reading


Chalmers, A.F.  1978/1982. What is this thing called Science? Milton Keynes: Open University Press. (A good introductory discussion of the notions of theory and method, the ideas of Popper, Lakatos etc.)


[Back to publications list]