Anthony Pym


_Video material

{Koller's Äquivalenz (review article)

Einführung in die Übersetzungswissenschaft (Uni-Taschenbücher 819). Werner Koller. Heidelberg: Quelle & Meyer, 1979. 290 pp. ISBN 3-494-02089-2; 4., völlig neu bearbeitete Auflage, Heidelberg, Wiesbaden: Quelle & Meyer, 1992. 343 pp. ISBN 3-8252-0819-2. 36.80 DM.  

This review was first published in The Translator (Manchester: St Jerome) 3/1 (1997), 71-79.  

Werner Koller’s Einführung in die Übersetzungswissenschaft (Introduction to Translation Studies/Science) has been a point of reference for German-speaking students and teachers ever since it was first published in 1979. As a comprehensive survey of basically linguistic theories of translational  equivalence it has also long been used as a convenient enemy by the newer German approaches that choose to privilege communicative purposes of various kinds. Hence a good reason for revisiting this classic: it provides an occasion for flying past almost two decades of German-language debate, locating a major landmark that might help explain dialectics that non-German readers could otherwise consider merely Germanic. Yet Koller’s is also a text that has revisited itself. The fourth edition, published in 1992 (a slightly revised fifth is to appear in 1997), is more like a new book, including extensive rewriting, adding some fifty pages, replying, often implicitly, to the criticisms of equivalence; and an English version of the essential positions has been published a 1995 article in Target, thus doing much of the revisiting for me. Two further reasons for returning to Koller: first, to pay homage to what is still a fine textbook; second, to support what might appear to be his rearguard defence of an outmoded way of approaching translation. So let’s look at the 1979 edition, at a few criticisms of it, then the 1992 rewrite with its 1995 polish.  
 As chance would have it, I read Koller’s first edition in its day. As a student stranded at the end of the world I found the book a tremendously useful introduction to the approaches of the 1960s and 1970s: theorists of the order of Mounin, Nida, Catford, Wilss, Jumpelt, Reiss, Levy, Kade, Jäger and Neubert were all encapsulated in lengthy re-citable citations, although names like Toury, Ladmiral and Seleskovitch were only marginally noted and a George Steiner was nowhere to be seen. This enabled wandering readers to pursue paths they wanted to follow, to ignore others that appeared less appealing, and to gain a glimpse of theorists that were interesting but mostly unavailable (notably Kade). Koller’s rationale for presenting his state-of-the-art survey was broadly as follows: Attempts had been made to establish an independent academic discipline called Übersetzungswissenschaft (understandable as the study/science of written translation) but they had run into a ‘legitimization crisis’ because too little had been done to mark the discipline’s relations with neighbouring areas like “contrastive/comparative linguistics/stylistics, comparative literature, text analysis, communication studies and computational linguistics” (1979:10). Koller’s underlying plan was thus to go through the available theories, to extract what seemed to be the most properly translational part of them, and to make that part the rock on which order could be built and legitimization attained. A more social aim was also suggested: Koller’s implicit promise (well, his last chapter) was to develop a scientific mode of criticism that could eventually improve the quality of translations (cf. Reiss 1971). As a student reader I found these goals as unexciting as they were acceptable; the usefulness of Koller lay more in the path he took to get there. I similarly accepted the name he gave to his apparently legitimizing centrepiece: ‘equivalence’, or more exactly the ‘concept of equivalence’ (Äquivalenzbegriff) as something that exists on the level of language use (parole) rather between language systems (langues), leaving the latter to nontranslational linguistics (183-85). Fair enough, I thought. Of course, what surprises me now is just how much of this apparently straightforward procedure was actually open to question. I’m also surprised to see that the proposed hard-core solution, this Äquivalenzbegriff, was directly discussed in only six rather dull and example-bereft pages (186-191), in fact the very pages translated into English in Andrew Chesterman’s extremely useful anthology of classical readings (1989:99-104). Let me now summarize the German original of that solution:  
 Bringing together definitions of translation by Oettinger, Catford, Winter, Wilss, Jäger, Nida and Taber, Koller concludes that since equivalence is their common ground, equivalence must be what is most specific to translation (a non sequeter, but wait...). At the same time, recognizes Koller, to say that translations must be equivalent to some original is to posit a relation devoid of content (1979:186). What is required is a definition of various kinds of equivalence, in fact descriptions of what a target text should have in order to satisfy various equivalence conditions. There are at least five frames for these equivalence relations: denotative (based on extra-linguistic factors), connotative (based on way the source text is verbalized), text-normative (textual and linguistic norms), pragmatic (with respect to the receiver of the target text) and formal (the formal-aesthetic qualities of the source text). Given this plurality, the translator must establish a hierarchy of different types of equivalence for each text, using a “translation-oriented text analysis” in order to determine an appropriate weighting of the five sets of criteria (191). And that’s about it: a six-page position. Note the following:  
- The heading for the pages I just summarized is ‘Äquivalenz in der Übersetzungswissenschaft’ (Equivalence in Translation Studies/Science). If read carefully, Koller is not talking directly about what translation is or should be (there are no examples) but about equivalence as a concept in theories of translation. He seeks to name what a certain group of theorists have selected as their object of study, since he believes that a science (well, Wissenschaft) should be able to name its object and set its boundaries in this way. As such, Koller can’t really be wrong: the theories he names were indeed there. The only trouble is that in 1979 there was no substantial alternative ‘translation science’ in sight, thus allowing considerable slippage, especially in the eyes of student readers like myself, between the object of the science and the eternal true nature of translation: when Koller talks about ‘translation in its proper sense’ I still tend to believe he is talking about what is specific to translations rather than what belongs to their ‘proper’ study. The slippage is also institutional: since the theories are used to train translators and criticize translations, they can’t help effectively producing their object and legitimizing it over and above possible alternative objects. No longer a student reader and now relatively unmoved by new sciences and slippery distinctions, my response is quite different: And what if the theorists were all wrong? Surely we should be looking at the thing itself rather than at our institutional coteries?  
- Koller’s five ‘equivalence frameworks’, which have remained strangely intact over the years (see Koller 1992:228-258; 1995:197), have no firm theoretical reason for cutting the cake into just five portions, neither in 1979 nor now. They do, however, suggest a healthy plurality that some readers have performed acrobatics to overlook: five is important because it is more than one (not just source-text) and more than two (not just source-against-target). True, source-text criteria get the biggest helpings and the most attention, but in 1979 the mentions of different text norms and target readerships were important and progressive within the general discussion of equivalence. As Chesterman notes (1989:99) the bit about norms incorporated the thrust of Reiss, and the pragmatics of readerships had “close points of contact” with Vermeer. Koller’s pluralism was truly promising; this could have been a story of Germanic theorists playing happy families. Yet things seem not to have turned out that way. Meanwhile, somewhere on the sidelines, what worried me at the time was that Koller didn’t look closely at the equivalence relation itself, which simply became a synonym for ‘translation’: merely cutting up the cake tells us very little about what we are actually eating.  

- Having established a pluralist notion of equivalence, Koller actually says very little about how any translator is to obtain the correct hierarchy of frameworks for each situation. His call for a ‘translation-oriented text analysis’ appears to involve looking at all the factors that could be looked at. What surprises me now is that this kind of text analysis, first sketched out by Reiss, is precisely the practical recommendation we find most elaborated in Nord’s Textanalyse und Übersetzen (1988; English version as Text Analysis in Translation). Also surprising is the ease with which the pragmatic notion of Skopos, the translation purpose, might indeed organize hierarchies of Koller’s frames, complementing rather than destroying his approach, and leaving the translator just as responsible for devising a set of strategies for each new situation. With hindsight, I suggest, great and even accumulative continuity can be found in the historical sequence of translation theories, with Koller providing a bridge between  structuralist linguistics and pragmatics. Once again, happy families could have been the order of the day. 

- By the bye, the pages I just summarized repeatedly refer to translating as the Herstellen (‘production’ or perhaps ‘fabrication’) of one kind of equivalence or another. Chesterman (1991:101-103) renders this Herstellen as ‘attainment’, which could have the translator trying to reach some pre-established goal, ultimately making Koller a theorist of what Mossop (1983) justly criticized as ‘equivalence seeking’, the kind of activity that condemns the translator to look ever backward. Yet a theory that places equivalence on the level of parole, a theory that includes target norms and readers among its parameters, is more fairly read as a tale of how translators effectively produce this specificity called equivalence. True, Koller’s creative potential in this area is often squashed by authoritarian modes of deciding what is or is not a translation, of deciding where equivalence has been found or missed. But the term Herstellen is very much there; it could have become a spark for rather more exciting things.  
 What went wrong? I don’t really care about the small-town politics, but here are a few details gleaned from a distance: Although Koller, a German-speaking Swiss, wrote most of his book while at Heidelberg, from 1978 he has been Professor of German Linguistics at Bergen, Norway. As the Einführung became—and remains—a standard textbook for linguistic approaches to translator training, the Germanic translator-training institutes, notably in Heidelberg, saw throughout the 1980s a strong move to break with that fundamentally linguistic framework, thus effectively redefining the institutional location and academic power relations of translation institutes. This involved seeking alternative inspiration in action theory or a diffuse functionalism, privileging the purpose of the translation and calling it, with many technical terms and much theorizing, Skopos. Easy to say, now, that those moves broadly followed the trends of general linguistics, which itself became increasingly pragmatic and functionalist. At the time, though, the political call was for some kind of general rupture, for the one true beginning of ‘modern’ translation theory (everyone else was already postmodern). The discussion of Koller’s concept of equivalence appears to have been unfortunately embroiled in that process.  

 Some of the critiques were substantial: Holz-Mäntttäri (1984:15-16) correctly saw the shortcomings of Koller’s definition-plus-example mode of argument, but I’m not sure she fared much better on that score. The more common treatments of Koller involved simply ignoring him or attaching his name to a list of ‘linguists’ and pretending to dump the entire discipline: “Linguistics alone won’t help us...,” declared Vermeer (1987:29), “ let’s look somewhere else” (cit. Nord 1997). Others cited the Einführung as a flagship of ‘equivalence’ and then proceed to sink it simply because that term had been used, by others, in too many different ways (cf. Snell-Hornby 1986:15) or because the term, in the hands of uncited authors, “presents an illusion of symmetry between languages” (Snell-Hornby 1988:22). That is, the concept of equivalence was either too vague (yet Koller’s was reasonably precise for its age) or too bound to language systems (Koller’s certainly wasn’t). Either way, it couldn’t win. Something similar happened in Nord’s Textanalyse (1988), where Koller was criticized for having too pluralist a ‘text-typology’ (Nord’s strange name for the five equivalence frames recited above) and for thus allowing too many possible rules (23), and in the next breath Nord somehow saw Koller’s approach as leading to a type of equivalence that was too narrow, so narrow in fact that Nord reduced it to something called ‘functional equivalence’, described as just one of the possible goals of a translator (25-26). More recently Nord (1997), now allowing for rather greater compatibility, has come up with the nit-picking claim that Koller’s pluralist notion of equivalence can’t legitimize very word-for-word ‘philological’ translations, even though no theoretical earthquake is required to give such translations an explicit place on the list. In short, Koller’s approach has been fed into a topsy-turvy little world that makes one wonder if the theorists actually read what the Einführung said, or if they did so with any degree of charity. The level of debate has hardly been inspiring. And no one in this camp, as far as I can tell, has really proposed any alternative way of delimiting the study of translation. I will return to this in a moment.  

 The 1992 version of the Einführung might be read as a response to the discussions of the 1980s. Of the numerous changes with respect to the 1979 version, the ones that most interest me are on the level of the reorganization of the book itself: gone is the final chapter on translation criticism (gone with it is the implicit promise to help improve the quality of translations), and the original 15 or so pages discussing equivalence have now become the entire second half of the book, some 140 pages. As for the first half of the book, its historical element has been considerably beefed up, notably through attention to (German) ‘translators and their theories’. And everywhere, in contradistinction to the 1979 version, there are numerous examples, with analyses, discussions, more long citations and what Newmark (1995:77) has called a “refreshingly and unfashionably didactic” attitude (I think that’s a complement coming from Newmark). I note the following as implicit or even hidden strategies:  

• Far from retracting the concept of equivalence, as one might have expected, Koller has embarked on a classical fuite en avant, elaborating the concept, making it stronger, more visible, more teachable.  

• At the same time, Koller’s explicit historicization of theory logically makes him disown any essentialist view of translation. The theorist thus becomes aware of his own place in history, eschewing neutrality and openly elaborating a concept of equivalence suited to his own particular set of centuries (described as ‘modern’, to be sure). Equivalence-based translation apparently only became possible from the mid seventeenth century, when the vernaculars (well, German) could be considered on the same footing as Latin (1992:66). The dating is a little vague, especially since Koller elsewhere locates the rupture from the beginning of the seventeenth century (1995:202), and the vision remains wilfully Germanocentric. In Spanish translation history, for example, the corresponding concept can fairly safely be dated from the mid sixteenth century. More important, however, is the way this positioning historicizes and specifies Koller’s own approach, forcing the theory to be named not just ‘modern’ but also ‘linguistic-textual’, thus recognizing alternatives and, with them, other possible objects of study. Of course, this is so much the worse for ages and cultures where the concept of equivalence does not reign, and especially for other theories that want to touch the object Koller claims for himself.  

• Within this perspective, the many examples in the 1992 rewrite surely go beyond the commercial need to be reader-friendly and to feed the massive growth in translator training (the world had some 109 translator-training institutions in 1979, at least 234 in 1992). I suspect the examples are also there because the 1979 mode of argument fell into a hole: How could Koller still extract the limits of translation from summaries of other theories when there were now so many theories that did not talk about equivalence? Some alternative epistemology was needed, a bit of which seems to have been borrowed from descriptive translation studies (hence, I suspect, the relative absence of ‘translation criticism’). The use of numerous examples thus becomes part of a claim to empiricism, as if the categories of equivalence had (partly) been derived from the study of translations themselves, specifically of the kind of translations providing Koller’s direct historical embedding. Such is the claim we read in the 1995 article (“these equivalence frameworks...are based on theoretical and empirical studies...”, 198). In 1979 those same frameworks were derived from other theories, with scarcely a smudge of empiricism.  

• A kind of empiricism is similarly involved in Koller’s response to the really curly question of separating translation (i.e. text reproduction, governed by equivalence) from nontranslation (i.e. text production, not subject to equivalence). Koller admits a wide intermediary space where the two textual products are mixed, such that a basically translational text may contain the translator’s notes, insertions, elaborations or improvements, all of the latter being islets of nontranslation. Apparently we should just set about measuring how much equivalence and non-equivalence are used in particular text genres (1992:196), as if empiricism really provided an answer to the theoretical problem of knowing what we are looking for in the first place. Koller’s examples here ultimately reveal little more procedural finesse than the use of commutation and back-translation tests, with the theorist declaring that since a given element is not to be found in the source (no matter how implicit the element), it is not a translation. Of course, it all depends on what the theorist is disposed to find; this is still Koller’s Achilles’ heel. Much, much more theoretical work is needed on this question, specifically with respect to what actual readers are disposed to believe about a text they accept as a translation. More attention might be paid to Gutt’s application of relevance theory to equivalence (although the latter is never named as such), and a general subjectivization of equivalence might fruitfully be connected with the more relaxed formal conditions of similarity (cf. Chesterman 1996). But such directions require distinctions between translation and nontranslation, and the making of such distinctions has long been the theoretical work of Koller’s equivalence.  

 If Koller’s theory is so weak at this point, why should I praise him for sticking to his guns? Precisely because he has the courage to insist, despite the ‘modern’ trends of the 1980s, that there is a difference between translation and nontranslation, and that this difference is historically functional rather than eternally essential. Whether or not that difference need be described in terms of equivalence is of less importance to me than the search for the difference itself. The problem is that theories seeking to by-pass equivalence do so by using nonrestrictive definitions of translation, thus doing little to stop the term ‘translation’ from being applied to every text that has been produced from an anterior text, which could logically be extended to all moments of intertextuality and thus to every text except perhaps God’s Word. The problem is not that we lack descriptions of translation; what we need is a good working definition of nontranslation. True, this problem is recognized in the Skopos distinction between ‘translational action’ (which includes everything translators can be called upon to do) and a ‘translation’ as just one kind of textual result of such action. True, Koller (1995:194) unfairly overlooks this terminological distinction when criticizing Ammann (1989). But still, the fact that a theorist carries around two or more different terms does not provide any procedural means for distinguishing between them. Doing translation theory is rather more demanding than multiplying names-for-things. None of Koller’s critics, I repeat, has really come to terms with the problem of defining nontranslation. Koller deserves praise for insisting that they should do so.  

 Why? The reason is perhaps not obvious. If what is at stake were merely the solidity of an academic discipline or the size, scope and relative independence of translator-training institutions, I would willingly opt for the wider frame without further ado: Yes, let us embrace illegitimacy and talk freely about all forms of crosscultural communication, and let us train translators and interpreters to do much more than translate and interpret, since what we really need are people with the competence to tell us when not to translate. In fact, were that all, I would immediately follow Vermeer (1998) in suggesting that the term ‘translation’ be dropped altogether, along with ‘equivalence’, if and when it should cause “a psychological barrier” to the wider need for crosscultural communication; I would willingly start talking about “technical rewriting”, and do more to train people to do such rewriting. The problem is that this term ‘translation’ is in fact one hell of a psychological barrier, in terms not only of social expectations but also of language policies, especially in Europe. For as long as business and politics sees translation, specifically equivalence-based translation, as a major solution to their communication problems, for as long as governments are prepared to pay translators and interpreters to maintain the illusion that our languages, the big ones, have equal status, then that idealized translation must be named, objectified, criticized and localized as a luxury solution with a hopefully luxurious price-tag. Rather than waffle in the vagaries of ‘translation as many different things’, rather than dissipate ourselves into the study of any text whatsoever, we should confront current preconceptions as honestly as possible by objectifying translation as one thing only, a very illusory, deceptive and labour-intensive thing, surrounded by many alternative modes of crosscultural communication that allow more highly visible mediators and more consistently cost-effective transactions. In short, I support Koller’s defence of equivalence because I don’t really like the current social and political ideals of translation, because I want to name and defend the varieties and virtues of nontranslation, and because the concept of equivalence is needed to prise those two fields apart, be it only for a moment of illumination, so that we know what we’re talking about. It is a question of strategy, not of essence: only through a restrictive definition of translation can we properly propose alternatives to translation. This is one of the reasons why the journal you’re reading names its subject as ‘The Translator’ and then promises not ‘translation studies’ but ‘studies in intercultural communication’, at the suggestion, if I remember correctly, of yours truly,  


Ammann, Margret (1989) Review of Jörn Albrecht et al., Translation und interkulturelle Kommunikation, TEXTconTEXT 4:1-2, 106-129.  
Chesterman, Andrew (1989) Readings in Translation Theory, Helinski: Oy Finn Lectura Ab.  
Chesterman, Andrew (1996) ‘On Similarity’, Target 8(1):159-164.  
Gutt, Ernst-August (1991) Translation and Relevance. Cognition and Context, Oxford: Blackwell.  
Holz-Mänttäri, Justa (1984) Translatorisches Handeln. Theorie und Methode, Helsinki: Academia Scientiarum Fennica.  
Koller, Werner (1995) ‘The Concept of Equivalence and the Object of Translation Studies’, Target 7(2):191-222.  
Mossop, Brian (1983) ‘The Translator as Rapporteur: A Concept for Training and Self-Improvement’, Meta 28(3):244-278. 
Newark, Peter (1995) ‘Paragraphs on Translation - 36’, The Linguist 34(2):76-77.  
Nord, Christiane (1988) Textanalyse und Übersetzen, Heidelberg: Julius Groos Verlag. English version: Text Analysis in Translation, Amsterdam/Atlanta: Rodopi, 1991.  
Nord, Christiane (1997) Translating as Purposeful Activity. Functionalist Theories Explained. Manchester: St Jerome.  
Reiss, Katharina (1971) Möglichkeiten und Grenzen der Übersetzungskritik, Munich: Max Heuber. 
Snell-Hornby, Mary (1986) ‘Übersetzen, Sprache, Kultur’, in Mary Snell-Hornby (ed) Übersetzungswissenscahft. Eine Neuorientierung. Tübingen: Francke, 9-29.  
Snell-Hornby, Mary (1988) Translation Studies. An Integrated Approach. Amsterdam and Philadelphia: John Benjamins. 
Vermeer, Hans J. (1987) ‘What does it mean to translate?’, Indian Journal of Applied Linguistics 13(2):25-33.  
Vermeer, Hans J. (1998) ‘Didactics of Translation’, Mona Baker (ed) Encyclopaedia of Translation Studies, London and New York: Routledge.  

Last update 11 March 1999  

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