The Translator ISSN: 1355-6509 (Print) 1757-0409 (Online) Journal homepage: https://www.tandfonline.com/loi/rtrn20 Translation Quality Assessment Where Can Theory and Practice Meet? Susanne Lauscher To cite this article: Susanne Lauscher (2000) Translation Quality Assessment, The Translator, 6:2, 149-168, DOI: 10.1080/13556509.2000.10799063 To link to this article: https://doi.org/10.1080/13556509.2000.10799063 Published online: 21 Feb 2014. Submit your article to this journal Article views: 1686 View related articles Citing articles: 5 View citing articles Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=rtrn20 The Translator. Volume 6, Number 2 (2000), 149-168 ISBN 1-900650-31-2 Translation Quality Assessment Where Can Theory and Practice Meet? SUSANNE LAUSCHER Institut für Translationswissenschaft, University of Innsbruck, Austria Abstract. Despite increased interest within translation studies to provide orientation for translation quality assessment (TQA), academic efforts in this area are still largely ignored, if not explicitly rejected by the profession. The purpose of this paper is to investigate why scientific models for evaluating translations are difficult to apply and to outline a number of ways in which the gap between theoretical approaches and practical needs may be negotiated. Following a critical analysis of some TQA models, the paper suggests that a reductionist view of translations as products and a neglect of the conditions under which translations are produced ultimately result in evaluation criteria which cannot account for the individuality of target texts. An examination of an actual instance of English-German translation demonstrates that the translation process is guided by case-specific values. These values, as well as the strategies employed to realize them, are set and agreed by the interested parties during the translation process. In order to judge the quality of a translation, the values should be made accessible to the evaluator and operationalized as evaluation parameters. Because the application of evaluation parameters depends on situational and individual factors, translation quality is ultimately a matter of agreement and consensus. The growing importance of translating and translations in a globalized world has turned translation quality and translation quality assessment into topics of public interest. More people use translations than ever, and more people than ever make (or try to make) a living from translation. Slowly, it seems, translation is developing from a “profession of last resort” (Inglis 1998:43) into a “healthy, booming, and extremely profitable sector of economy” (Bonthrone 1998:14). This may explain both the growing number of initiatives relating to quality and quality assessment on the international level (e.g. ISO certification of translation businesses) as well as the national level (e.g. DIN 2345 on organizing translation projects, issued by the German Standardization Institute in 1998). ISSN 1355-6509 © St Jerome Publishing, Manchester 150 Translation Quality Assessment Practical efforts to grasp translation quality and make it operational for practical purposes often focus on developing a catalogue of criteria for a ‘good’ translation. The ultimate goal seems to be the establishment of a conclusive list applicable to all translations, and the criteria mentioned typically refer to correct language use in terms of the target language system (Schmitt 1998). Once that definitive list has been created, it is assumed, it should be possible to make prescriptive judgements about individual translations being ‘good’ or ‘bad’. However, practical definitions of translation quality, such as “delivering the right translation, at the right time, at the right place, and at the right price” (Bonthrone 1998:13), suggest that such lists are not sufficient for determining translation quality in a professional setting, where quality seems to depend on a variety of very diverse factors. Translation quality and translation quality assessment are also topics which are addressed in translation research. In his outline of translation studies as an independent discipline, James Holmes (1988:78) defined translation quality assessment as part of translation criticism, a branch of applied translation studies.1 Like many translation scholars before and after him (e.g. Reiß 1971:7, Wilss 1977:281, Ammann 1993:433-34, Gerzymisch-Arbogast 1994:573, Toury 1995:18-19), Holmes saw the main task of translation criticism as improving what he perceived to be a generally arbitrary and subjective evaluation practice. However, despite the increased interest in evaluation among translation scholars, especially since the mid 1990s, academic approaches to translation quality assessment are still widely criticized by both translation theorists and practitioners for avoiding the subject of quality (Schmitt 1998), for their inability to determine the nature of quality (Bonthrone 1998), and for neglecting the requirements of various evaluation scenarios (Hönig 1998). Scholarly models of translation quality assessment could become more applicable in practice if the translation process were better integrated into the evaluation procedure and if the relative nature of the evaluation procedure itself and, hence, of any prescriptive judgement were investigated more closely. After reviewing some of the existing models of translation quality assessment in terms of their general applicability to practical evaluation (section 1), I will illustrate the impact of the translation process on the quality of the translation product by using a practical example (section 2). In section 3, I will propose a tentative description of the evaluation process and briefly discuss the implications for the possibility of prescriptive judgement. I will use translation evaluation and quality assessment as synonyms, both referring to a procedure that allows for prescriptive judgement. Following Ripfel (1989:84-85), I define prescriptive judgement as a judgement that can claim validity in a specific evaluation situation. 1. Models of translation quality assessment Translation scholars have tried to improve practical translation quality as- Susanne Lauscher 151 sessment by developing models which allow for reproducible, intersubjective judgement (e.g. Reiß 1971:12-13, Wilss 1977:251, Ammann 1993:433-34, Gerzymisch-Arbogast 1997). They hoped to achieve this goal by building their models on scientific theories of translation, which can provide a yardstick, and by introducing a systematic procedure for evaluation. In the following paragraphs, I will discuss the practical applicability of several translation quality assessment models in terms of the translation theories they rely on, the evaluation parameters and the procedures they propose. 1.1 Equivalence-based approaches Equivalence has been a central concept in translation studies, especially in early translation theory. It has served as a descriptive and prescriptive category for defining the relationship between source and target texts, and for distinguishing translations and translating from other types of texts and textproducing activities. Equivalence-based translation theories view translation as the attempt to reproduce the source text as closely as possible. Because the target text can never be equivalent to the source text on all levels, researchers have distinguished different types of equivalence, for example formal and dynamic or functional equivalence (Nida 1964, de Waard & Nida 1986), and denotative, connotative, pragmatic, textual and formal-aesthetic equivalence (Koller 1997). In 1971, the German translation scholar Katharina Reiß introduced one of the first systematic approaches to translation quality assessment. Reiß defines translating as “a bipolar procedure aiming at producing a target text while constantly referring back to the source text” (1971:11; my translation).2 Translating means finding equivalents for source text items in the target language at the level of text and individual text units. A translation is deemed good if it achieves optimum equivalence, that is if “considering the linguistic and situational context, the linguistic and stylistic level and the intention of the author, target text and target text units have the same ‘value’ as the text unit in the source language” (ibid.:11-12; my translation).3 In order to assess optimum equivalence between source text and target text, the critic analyzes both texts by using the categories of (1) text type, which captures the main function of language in a text, (2) linguistic instructions, that is, the semantic, grammatical and stylistic properties of the source text, and (3) extralinguistic determinants, which refer to the impact of contextual factors on verbalization strategies, including different levels of knowledge of subject matter among source and target readers, their different perception and evaluation of particular phenomena, etc. According to Reiß, evaluating a translation means reversing the translation process and reconstructing the translation strategy. The analytical concepts of text type, linguistic instructions and extralinguistic determinants are integrated into a set of hierarchical rules which govern both the evaluation and translation processes. 152 Translation Quality Assessment By merging translation and evaluation process, the evaluator assumes the translator’s role and accounts for the complex linguistic and situational factors that condition translation solutions. Reiß proposes an evaluation procedure that involves two steps: (1) an analysis of the target text in order to evaluate the appropriateness of target language use, and (2) comparison of source and target texts based on the analysis of both texts, so as to establish the degree of equivalence between them. Reiß’ model has been very influential, especially in Germany and Austria, because it highlights the active role played by the translator in the translation process, and because it provides a comprehensive, systematic model of text analysis for both translation and translation evaluation (Nord 1996). The three process-governing rules (the determination of text function and the analyses of the linguistic and extralinguistic determinants) ensure that the procedure of quality assessment refers to both the macro- and the microlevel of a text. Thus, translation evaluation can no longer be a matter of word-by-word comparison of target and source texts. Through the category of extralinguistic determinants, Reiß also links source and target texts to their social, non-linguistic context, and draws attention to the impact of context on the linguistic make-up of texts. As a result, the wording of the source text can no longer be considered the only point of reference for translation and quality assessment. However, in terms of practical application, Reiß’ approach suffers from the vagueness of “optimum equivalence” as the yardstick for good translations. In order to make her model operational, we need a precise definition of “optimum equivalence”. Reiß (1971:12-13) offers a somewhat circular definition of equivalence in terms of texts and text elements being of equal value. It seems that equal value mostly refers to ‘correct’ language use in target culture situations; but can we clearly define which linguistic means are appropriate for which situation? Some of Reiß’s practical examples (e.g. Reiß 1971:54-55, 78) suggest that equivalence is established at least to some extent by bilingual dictionaries. Optimum equivalence will be achieved if the translator is able to choose the appropriate word or phrase from the dictionary by applying the process-governing rules, i.e. the determination of text type and the analyses of the linguistic and extra-linguistic determinants. If this is the way to proceed, it will still be difficult to assess the appropriateness of neologisms, which are a common problem, especially in technical translation. Likewise, it would be hard to assess the translation of creative language use, for example in terms of rendering metaphors. In addition, Reiß’s model cannot be applied to all texts which, in practice, are considered translations. Reiß herself points out that her approach is not applicable to translations that serve other functions than providing an equivalent reproduction of the source text, or to translations that address a different audience in the target culture. According to her, these target texts no longer Susanne Lauscher 153 constitute translations, but are “transfers” (Reiß 1971:91). Following her definition, Bible translations would not be translations, nor could they be evaluated using her approach. Juliane House published her approach to translation quality assessment in 1977, and a revised version in 1997. For House (1977/1981:29-30; 1997:31), translating is a linguistic procedure aiming at “the replacement of a text in the source language by a semantically and pragmatically equivalent text in the target language”, i.e. at functional equivalence on the text level. Following Lyons, she defines function as “the application or use which the text has in the particular context of situation”, which is reflected in linguistic properties of the text (1977/1981:37; 1997:36). For the purpose of translation quality assessment, function is further divided into a (a) primary level, and (b) secondary level function. In order to fulfill the primary level function, the target text must reproduce the function of the source text, whereas reaching the secondary level function means to allow the target reader access to the function of the original. Both primary and secondary level functions are related to two translation strategies, i.e. covert and overt translation. In covert translation, the translator reproduces the source text function by using an empirically established cultural filter to adapt the target text to the communicative preferences of the target audience. In overt translation, however, the translator tries to reproduce the function of the source text by staying close to the source text. Functional equivalence serves as the yardstick for a “good”, or “appropriate translation” (House 1997:31-32). It is made operational by two parameters, genre and register, which are designed to capture the linguisticsituational characteristics of the source text (ibid.:105-110). While House (ibid.:107) considers genre a pre-scientific, socially established category, she subdivides register further (following Halliday) into field, tenor and mode, and correlates it with lexical, syntactical and textual elements (ibid.:42). The evaluation procedure consists of: (1) establishing a “source text profile” (ibid.: 42) along the operational-enabling parameters against which the target text is measured; (2) establishing the function of the source text; (3) comparing source text profile with target text; and (4) providing a statement of quality that lists, in addition to errors, the matches and mismatches along the parameters of genre and register, and comments on the translation strategy. Several of House’s concepts, such as the distinction between overt and covert translation as well as the notions of cultural filter and source text profile, have become standard terminology in translation studies and have proved useful in didactic approaches (e.g. Nord 1991, Kußmaul 1995). In terms of translation quality assessment, House – like Reiß – tries to link context, text and word level through the concept of text function. Unlike Reiß, House specifies text function as a semantic-pragmatic category which can be identified by such linguistic properties of texts as theme-rheme structure, linguistic 154 Translation Quality Assessment means of expressing coherence, etc. (House 1997:43-45) In terms of practical application, House’s basic assumptions about translations and translating seem restrictive. She allows only for two target text functions, i.e. a target text function identical to the function of the source text or a target text function identical to the function ascribed to the source text by the contemporary source text audience. Again, Bible translations could neither be considered translations nor evaluated on House’s terms because they account for the needs of the target audience beyond communicative preferences (de Waard/Nida 1986). For assessing the quality of overt translations, we would have to determine who constitutes the actual source text audience – e.g. the general public, the literary establishment, men, women etc. – and what function that audience ascribes to the source text. In the case of Bible translation it would be hard to define the contemporary audience of the source text and the impact of the source text on this audience. Also, it seems unlikely that a target audience would be able to understand the function of the source text for the source text audience without access to at least some additional information. House’s yardstick of functional equivalence seems problematic because she assumes that text function is expressed by linguistic means. Her assumption raises two questions: (a) is text function inherent in linguistic expression?, and (b) do different languages use the same linguistic means to express text function? With respect to (a), van Dijk and Kintsch (1983; quoted in Risku 1998:184) argue that text comprehension is a strategic and interpretative procedure influenced by the knowledge and expectations of readers, and also by the situation in which a text is read, the purpose for which it is read, and so on. Their findings imply, first, that a text does not exist outside the interpretations of readers, and, second, that interpretation is subject to influences which are cultural, social and individual, and hence they lie outside the text. If we accept the view of van Dijk and Kintsch, we have to concede that the function of a text does not lie in the text itself or any aspect of it, but is attributed to the text by readers in specific contexts.4 Although there are linguistic means which are typically used to fulfill certain functions in certain types of text, for example metaphoric language in literary texts, the same means can be put to different uses, for example in an advertisement to sell cars. With regards to (b), whether different languages use the same linguistic means to express text function, I would like to quote a few examples from House’s practical application of her model. In her evaluation of the English to German translation of a children’s book, House (1997:128) argues that the omission of the children’s first names in the German text is an instance of mismatch, implying lack of politeness or disrespect. However, linguistic means for expressing politeness vary significantly from culture to culture. In the case of a parent telling a child not to do something in a German-speaking context, the child’s first name is often used to put stronger emphasis on the Susanne Lauscher 155 parent’s stance. It cannot be interpreted as lack of respect because it would be perceived as such in an English-speaking context. Similarly, House (ibid.:133) interprets the translation of poor bastards by arme Kerle as loss of humour. Surely, it is only by looking at the target culture and the text that we can decide if the rendering is humorous or not. A literal translation into German is certainly unacceptable, but House does not indicate alternatives. She also considers the non-use of anacolutha in a German translation as an instance of syntactical mismatch without taking into account rules of the German written code (ibid.:136). It seems to me that by focusing on the linguistic means the source language uses to express text function, the wording of the source text becomes the yardstick for judging the appropriateness of the linguistic means employed in the target text. Although House (ibid.:115) agrees that there are cultural differences in verbalization strategies, she considers them to be applicable only if they have been verified by empirical research. However, empirical research into culturally determined “communicative preferences” (ibid.:7999) seems to be at an early stage with respect to both method and issues covered, and House does not provide a sufficient basis for distinguishing communicative preferences in different cultures. Because of the association between text function and the linguistic means of the source culture, and the vagueness of the cultural filter, translations appear as a series of mismatches in House’s model. Following that model, a good translation would be one that respects scientifically established cultural differences in verbalization strategies, but otherwise tries to reproduce the linguistic properties of the original as closely as possible (which raises the question of what ‘close’ means). In practice, many target texts would be considered inappropriate translations because they respect target culture and target language conventions. It seems, though, that House herself does not consider her approach to be evaluative in the sense of providing prescriptive judgement. Rather, she characterizes it as “prepar[ing] the ground for the analysis of a large number of evaluation cases”, and “a basis for judgement” (ibid.:118). According to her, judgement depends on social factors, but these factors are not and should not be accounted for in a model of translation quality assessment (ibid.:165). Van den Broeck’s model of quality assessment focuses on literary translations (Van den Broeck 1985) and aims to establish the degree of ‘factual equivalence’ between source and target texts, that is the degree to which the two texts can be related to each other along “functionally relevant features” (Van den Broeck 1985:57). Following Toury, Van den Broeck proposes the concept of “adequate translation”, i.e. the “hypothetical reconstruction of the textual relations and functions of the source text” (ibid.) as the yardstick and parameters for the comparison of source and target texts. Textual relations and functions can be correlated with any aspect of the source text (ibid.:58). 156 Translation Quality Assessment To evaluate a translation, the critic first establishes the adequate translation and compares it with the target text by describing the mandatory and optional shifts observed in the target text on the basis of the results of contrastive linguistics and stylistics. The critic then compares his or her translation of the adequate translation (the critic’s norm) with the target text (the translator’s norm). When evaluating the translator’s norm, the critic tries to account for the reasons of optional shifts in the target text. Van den Broeck’s model implies that there is no such thing as the ‘one, best translation’. Rather, assessing the quality of a translation means comparing different translations, that is, the translator’s and the critic’s norm. Like Reiß, Van den Broeck acknowledges the active role of the translator in the translation process. He draws attention to the fact that target text deviations from the ‘adequate translation’ can be related to a variety of linguistic and extralinguistic factors, and cannot be easily attributed to the translator’s incompetence or their intention to manipulate the source text. However, Van den Broeck’s yardstick of ‘factual equivalence’ remains unclear because he leaves the concept of function undefined and does not explain how to determine the functional elements of the source text. Although the critic is expected to take into consideration the complexity of translating by considering the reasons for alterations, Van den Broeck does not specify how these reasons might be discovered. 1.2 Functional approaches Functional approaches to translation evaluation are based on the assumption that translating is not so much determined by the source text as by factors relating to the target culture (Reiß & Vermeer 1984; Holz-Mänttäri 1984; Vermeer 1989, 1996; Nord 1993). Translation is seen as a process of text production on the basis of a source text, and the target text is considered as a text in its own right. ‘Function’ is a prospective concept, which is determined for each translation by the translation commissioner and the translator with regards to the use of the target text in the target culture situation. The concept of function includes the situation of the prospective target readers. Function is the point of reference for the translation strategy developed and applied by the translator. The functions of target texts are basically arbitrary (Vermeer 1989:102-103, 104, PrunĨ 1997), although they can be restricted by socially determined concepts of translating. Ammann (1990) limits her model to the evaluation of literary translations within a research context. Her approach is based on skopos theory (Reiß & Vermeer 1984; Vermeer 1989, 1996), which assigns the key positions in the translation process to the target text and the target audience. For Ammann, a good translation is one that allows the target reader to reach a coherent interpretation of the target text. She makes the yardstick of coherent text Susanne Lauscher 157 interpretation operational by using scenes-and-frames semantics (cf. Vannerem and Snell-Hornby 1986, Vermeer and Witte 1990). The concept of scene refers to the non-linguistic perception of a phenomenon by an individual, while frame refers to the linguistic expression used to communicate the scene. The individual’s perception is co-determined by the society and culture to which they belong. Different cultures use different frames to express the same scenes. Texts consist of scenes made of sub-scenes. Ammann proposes an evaluation procedure in which the critic first determines the function and coherence of the target text, as well as the function and coherence of the source text, before proceeding to compare coherence between source text and target text. By using the concepts of scenes and frames, Ammann tries to dissociate the reproduction of life phenomena from the linguistic means of realizing these phenomena, thus making it possible to integrate target culture and target language conventions into translation quality assessment. The scenes-andframes approach also integrates into quality assessment the target audience, its world knowledge and its expectations. However, the problems of Ammann’s approach with respect to practical application lie with her concept of function and the vagueness of the scenes-and-frames approach. Ammann equates function with coherent reader interpretation. Yet, especially in the case of translated literature, incoherent interpretation may very well be what is intended. Although Ammann (ibid.:230, 237) tries to specify function in terms of contextual properties of the target text – such as the target text publisher, the position of the source text author in the target culture, etc., we cannot automatically assume that the target text function actually consists of producing a text which allows for coherent interpretation. With respect to the scenes-and-frames approach, Ammann does not really explain how we can identify scenes and how we can assess whether a scene is rendered by an ‘appropriate’, i.e. a coherent frame. Her practical examples suggest that coherence is related to ‘correct’ situational language use on the level of parole (ibid.:230-46). But language use is constantly subject to change, and in literary texts it is creative and often aims at expressing scenes in unusual frames. D’Hulst’s model (1996) is restricted to evaluating texts for special purposes, which tend to fulfill the same function in source and target cultures. For D’Hulst (ibid.:103), a translation is “an independent text functioning, by definition, in the target culture”. A target text is considered “good” if it fulfills the function it has been intended for. D’Hulst (ibid.:102-103) equates function with the concept of text act, a category similar to speech act. Text act (or function) is further specified by two types of text structure, a topic-centred and a hierarchical text structure. Text structure is further explained in terms of textual connectivity, which relates macro-structural text characteristics to the micro-structural level. Textual connectivity is achieved through coherence relations and referential connectivity. Coherence-relations on the macro-level 158 Translation Quality Assessment are realized by either a paratactic or a hypotactic organization of utterances, while referential connectivity is realized by cohesive links and theme-rheme structure. D’Hulst assumes that certain text structures can be correlated to certain text acts, i.e. the text act of description can be correlated with a topiccentred structure, and the directive text act with a hierarchical text structure. In D’Hulst’s model, the critic first evaluates the structure of the target text and the source texts separately by examining the appropriateness of the text structure to the text function. Then the critic compares the source text and the target text and comments on the differences. The model differs from other models discussed previously in that it allows for the possibility of defective source texts. Also, by relating text function to structural characteristics of the source and target texts, D’Hulst tries to overcome the problem of linking text functions to specific linguistic means, which proved to be one of the weaknesses in House’s approach. D’Hulst’s approach is applicable under two conditions: (1) a text’s function can be clearly and unambiguously related to text structures; and (2) there is a typology of text acts and related text structures for each language. With regards to (1), I believe that in reality many texts, e.g. advertising texts or product documentation, consist of both topic-centred and hierarchical text structures, which would make the identification of text acts on the basis of text structures difficult. With regards to (2), we have to admit that such a typology does not exist (yet), and therefore we cannot make it operational for practical purposes. It seems to me that scholarly approaches to translation evaluation have not yet been able to provide help for practical quality assessment because they do not account for the reality of translating and translations. Equivalencebased approaches are built on definitions of translating and translations that are too restrictive to account for all texts considered translations in practice. Functional models overcome this problem by defining translation as any target text that fulfills a function in the target culture; but the yardsticks these models propose are based on general assumptions which are not examined in the light of the actual functions assigned to texts nor the means chosen to realize them. In other words, the models neglect the link between the context in which a translation takes place and the impact of this context on the target text. 2. Quality assessment and the translation process In order to examine the impact of the context of translation on the target text, let us look at a practical example. In August 1996, the German publisher Siedler Verlag published the translation of Daniel Jonah Goldhagen’s book Hitler’s Willing Executioners, a historical investigation into the reasons for the Holocaust. Siedler Verlag is a small, prestigious publishing house founded in 1983 by Wolf Jobst Siedler, an experienced and well-known personality in the German publishing indus- Susanne Lauscher 159 try.5 Although part of the Bertelsmann Group, one of the leaders in the international commercial media industry, the profile of the publishing house is still largely shaped by Siedler himself. Siedler Verlag specializes in the publication of non-fiction for an educated lay audience, focusing mostly on historical and political subjects. It is also known for the publication of memoirs of prominent statesmen, such as the former German Chancellor Helmut Schmidt and the former Russian President Boris Yeltsin. Translations make up only a small part of Siedler Verlag’s list. Foreign texts are selected in terms of the publisher’s programme and their potential interest for the prospective audience. Most of the translations are from English. The reception of both the original Hitler’s Willing Executioners and the German translation was accompanied by heated debate and conflict. In the United States, the book’s main hypothesis – powerful, violent German anti-Semitism resulted in the eager participation of thousands of ordinary Germans in the most cruel persecution and murder of Jews – led to polarized reactions ranging from praise to severe criticism of the author and his work (e.g. the reviews printed in Schoeps 1996). In Germany, the debate started as soon as the original was published (about four months before the publication of the translation). The debate was highly publicized and involved academics as much as journalist and the general public. The demand for copies of the American original was apparently high (Frankfurter Allgemeine Zeitung No. 187, 13 August 1996).6 When the German translation was published in August 1996, the first edition of 40,000 copies was sold within five days (Frankfurter Allgemeine Zeitung, ibid.). The media attention was moderate compared to the reactions to the original. However, the German weekly Der Spiegel, an influential and widely-read political magazine, published an anonymous review which criticized the German translation severely for adapting the hypotheses of the study to German sensibilities (Der Spiegel 33, 1996:42). Based on the analysis of the translation of the title and a number of passages from the book, the reviewer accused the publisher of having flattened the message and language of the original author. In a subsequent letter to the publisher, the original author rejected the Spiegel’s criticism (Der Spiegel 34, 1996:14). The review nevertheless stirred some interest in the translation among both target audience and translation scholars. The translation was briefly discussed by a German historian (Pohl 1997:15) in a well-known historical quarterly; and Juliane House (1997:147-57) used some passages from the book as a test case for her translation quality assessment model. Both agreed with Der Spiegel’s verdict. All three reviewers expected the translation to reproduce the content of the book in the author’s original style. According to the editor in charge of the German translation,7 the purpose of the German translation was to “reproduce the message of the original as faithfully as possible”.8 The purpose was set in the licensing agreement, which, 160 Translation Quality Assessment upon the request of the author, held the publisher responsible for keeping editorial changes to a minimum. The author was guaranteed the right to revise the translation before publication. Both translator and editor experienced strong pressure during their work due to the contractual restrictions and intense public attention. Because of the tight deadline, the translation and editing process of the 600-page book had to run in parallel. In the course of the translating and editing process, a number of textspecific problems had to be resolved. One such problem concerned the frequent repetition of certain terms and phrases. Keeping the repetitions in the German text was perceived as producing a negative effect. The problem was resolved by using a number of synonyms, in agreement with the original author. Another problem was the translation of terms which had no equivalent in German, such as genocidal, and technical terms such as cultural discourse. In the first case, translator and editor agreed on a number of paraphrases which would be used throughout the target text. For the translation of technical terms the editor consulted experts in the subject matter . Despite the contractual restrictions, some structural changes were introduced in the translation. The editor suggested partially integrating footnote 5 of the introduction to the American original (Goldhagen 1996a:475-76), which explains the generalized use of the Germans throughout the book, into the introductory chapter of the German text (Goldhagen 1996b:19). The German translation also contains a preface to the German reader, in which the author explains the purpose of his project and the particularities of his approach; it also includes a bibliography. Despite the time pressure, special attention was given to the revision process. A group of German sociologists, psychologists and historians proofread the German text. Their critical observations were integrated into the translation. Finally, the text was submitted to the original author for approval. In the acknowledgement section of the German translation, the author described the translation as a rendering of the original “faithful to the word”9 (Goldhagen 1996b:552). Of course, the translation of the Goldhagen book constitutes an isolated case, described from the perspective of only one of the parties involved. The case would need further investigation, and should not be used to make generalizations.10 For the purpose of this paper, I am mostly interested in what the case tells us about the translation context and its impact on translation quality, as well as the consequences for quality assessment. Most importantly, I think, the example of Hitler’s Willing Executioners suggests that every target text is unique, a “concrete affair” (Salevsky 1996:304), the result of a strategy designed by parties who have different interests in the translation. It seems that, before the actual translation process starts, the parties involved design and agree on some sort of ‘ideal’ target text for the target situation. The ideal translation corresponds to what the parties deem ideal in the con- Susanne Lauscher 161 text of the translation situation (cf. Vermeer 1996:12-14), including the communicative intention of the source text author with respect to the target audience, the status of the subject matter of the source text in the target culture, as well as the organization of the specific translation process in terms of its stages, and the time frame allotted for the production of the final version. The translator tries to realize the ideal target text – or, more precisely, his or her interpretation of the ideal target text – by applying a preliminary translation strategy. In the course of the translating and editing process, the strategy is defined more precisely on both the macro- and micro-structural levels by the translator and his or her co-operating partners, e.g. editor, proofreader, source text author. In fact, the case of Hitler’s Willing Executioners suggests that the translation strategy can be fully formulated only at the very end of the translation process. What are the consequences for scholarly models of translation quality assessment? First, they need to be based on a yardstick which is flexible enough to integrate translation strategies designed in actual translation processes. At the same time, the yardstick needs to be precise enough to integrate the various levels of translation strategies. The concept of function can serve as such a yardstick if it is defined more precisely (cf. Risku 1998 for a recent attempt ). Given the complex origin of translation strategies it seems to me that scholarly approaches to translation quality assessment are limited to establishing a comprehensive conceptual tool. House (1997:119) seems to reject this idea by stating that “the reconstruction of the translator’s choices and her/his decisions is ... a highly complex, and ... probabilistic undertaking”. However, translation criticism, and translation studies in general, cannot claim to improve our knowledge about translation and phenomena related to it if they refuse to study actual translation practice as a basis for developing theoretical concepts (cf. Salevsky 1998). Complexity should be a challenge, rather than a hindrance, for further scientific investigation. In order to improve our knowledge of actual translation processes, their impact on translations and translation quality, the profession should make them more accessible, e.g. in the form of translation comments, forewords, product specifications, etc.; this is something that translation scholars have repeatedly called for (e.g. Holz-Mänttäri 1984:114-15, Vermeer 1989:64, 104, Nord 1991:188, Vermeer 1996:14-15). In other words, the profession, too, has a responsibility for making translation quality assessment more reliable. As long as our knowledge of actual translation processes remains limited, proponents of scholarly models of translation quality assessment must acknowledge the speculative side of those models. Comparing source and target texts can tell us about differences between them, but it does not allow us to identify such differences as mistranslations or errors. It does allow prescriptive judgement in our own terms, as long as these terms are explained and 162 Translation Quality Assessment justified (cf. Nord 1991:192, House 1997:119), but it does not allow prescriptive judgement in the terms on which the target text was designed. Neither does the comparison of source text and target text allow for attributing the responsibility for differences. At least in the case of published translations, target texts are not the result of one person’s decisions but of complex cooperation. Hence, we cannot use them to criticize the translator’s qualifications nor his or her intentions or ethics (cf. Lefevere 1996). 3. Translation quality is a matter of agreement When dealing with practical translation quality assessment translation criticism faces the problem of prescriptive judgement. In the context of translation practice, e.g. in quality control (cf. Schmitt 1998) or translator training, prescriptive judgements are necessary. On the other hand, House (1997:119) recommends that prescriptive judgement be avoided because it requires knowledge of the ideals and beliefs of the parties involved concerning translations and translating, and because the decisions which shape a target text are too complex to be reconstructed. According to her, avoiding prescriptive judgement is a sign of respect for translation and translators (ibid.). However, it seems to me that it is precisely by inquiring into the possibility of prescriptive judgement that translators and translations could gain more recognition and respect, because prescriptive judgement makes the complexity of the translation process visible. According to the German sociologist Max Weber (1914/1988), any science dealing with social phenomena should make a clear distinction between facts and what he called “value judgements”. Weber believed that the social sciences should study values and value judgements as social phenomena, but they should refrain from pronouncing value judgements in a scientific context because value judgements could not be derived from scientific evidence.11 Studying the evaluation process could be one way to approach the question of prescriptive judgement. If we abstract from the translation quality assessment models discussed in section 1, we can roughly describe translation evaluation as a procedure in which an evaluating person compares an actual target text to a more or less explicit, ‘ideal’ version of the target text, in terms of which the actual target text is rated and judged. The evaluation procedure consists of three elements which influence judgement: the evaluating subject, the object and the model target text. With respect to the evaluating subject, Pöchhacker (1994:123) states that translations can be evaluated from various perspectives, for example by clients, the target addressees, translators, and outside observers like translation researchers, and this has an impact on how the target text will be judged. He also points out that translation evaluation depends on the evaluating person’s experience and knowledge of translation. Susanne Lauscher 163 Reiß (1971:106-109) argues that the object of evaluation does not exist as such, but only in the interpretation of the evaluating person. However, text interpretation is necessarily individual. Even if we were to mandate a specific interpretation of a text, the individual component cannot be entirely eliminated because, by definition, this interpretation has to be reinterpreted by the evaluating person. The evaluating person’s interpretation of the target text constitutes the object of evaluation and therefore decisively influences his or her judgement. The judgement of a target text also depends on the model target text to which the actual target text is compared. The practical example discussed in section 2 suggests that the model target text is built on a general concept of translating (cf. Ammann 1993:434, House 1997:1). However, there is no concept of translating on which everyone agrees, nor one on which translation scholars agree. Research into the history of translation (e.g. Vermeer 1992, Lefevere 1992) shows that concepts of translating have changed over time. They are culture-specific (cf. Venuti 1995), and are not uniform even within one culture at one particular moment of time (Lauscher, in progress). Judgement also depends on how the theory of translating is made operational. The case of Hitler’s Willing Executioners suggests that even when the critics adopt similar translation theories they may deem different means appropriate for realizing them. Finally, several researchers have suggested that the judgement of a target text depends on the context in which evaluation takes place. The judgement itself fulfills a purpose (Ammann 1990:211). It may serve to examine a translator’s qualification for a particular translation job, to assess whether he or she has satisfied the requirements for a specific translation task, to inform a translation student about his or her progress, to inform the reader about the quality of the translation of a new work of fiction, etc. (Hönig 1998). A judgement is also oriented towards a prospective addressee. It will look different depending on whether it targets professional translators, the audience of the target text, clients or translation students (Gerzymisch-Arbogast 1997:576-78). Given the inherent subjectivity of the evaluation procedure and its dependence on the context in which it takes place, it seems to me that translation criticism has the following task with respect to prescriptive judgement. First, it must investigate the evaluation process further in order to determine the levels on which the influence of the subjective occurs, and it must emphasize that agreement on these levels is a prerequisite for prescriptive judgement. Second, translation criticism could consider and build into its models general knowledge about prescriptive judgement. For example, Ripfel (1989:85) argues that prescriptive judgement is possible only if the parties concerned accept judgements as binding. Hence, the prescriptive character of judgement depends either on consent among the parties concerned or on their power to impose their interpretation of an ‘ideal target text’. 164 Translation Quality Assessment 4. Conclusion In this paper I have explored various ways to bridge the gap between scholarly approaches to translation quality assessment and practical quality assessment. Translation criticism could move closer to practical needs by developing a comprehensive conceptual tool, which would make it possible to grasp the complex reality of translation processes, and by inquiring further into evaluation phenomena and into the possibility of prescriptive judgement. At the same time, those involved in translation practice, i.e. people who need and are involved in the production of translations, could become more aware of their respective roles and responsibilities in the translation and evaluation processes. Translation quality assessment and the judgement of translations are a matter of communication, co-operation and consent. SUSANNE LAUSCHER 19, Chemin Beaume Mele, 06650 Le Rouret, France. 102104.3236@ compuserve.com Notes 1. 2. 3. 4. 5. 6. The evaluation of translations is also the subject of other academic disciplines, such as literary studies. Literary studies focus on literary translations and place strong emphasis on the reception of translated texts in the target culture (cf. Frank 1987). The critique of translations in a literary context does not aim at influencing translation practice and generally abstains from prescriptive judgements. Translation studies, as defined by Holmes, refers to translating in all social contexts and covers all types of translated texts. Translation criticism in the context of translation studies aims at improving translation and evaluation practice by developing systematic methods of evaluation. “Jeder Übersetzungsprozeß ist ein bipolarer Vorgang, der sich in der Gestaltung eines zielsprachlichen Texts unter ständiger Rückbindung an einen ausgangssprachlichen Text erfüllt.” “... die zielsprachliche ‘Entsprechung ist als optimal äquivalent zu betrachten, wenn sie nach Maßgabe des sprachlichen und situationellen Kontextes, der Sprach- und Stilebene, der Intention des Autors etc. gewählt ist und im Sprachsystem der Zielsprache die gleiche ‘Wertigkeit’ besitzt wie die ausgangssprachliche Einheit in der Ausgangssprache.” Reception theory and reader-response-theory hold similar positions. The information on Siedler Verlag is taken from an interview with Wolf Jobst Siedler, published on Siedler Verlag’s website (www.mountmedia.de/ mm/verlage/siedler/index.shtml) in 1997. For more details about the reception of the book in the United States, Israel, France, Italy and Germany see Heil and Erb (1998). Susanne Lauscher 7. 8. 9. 10. 11. 165 I would like to thank Ms. Andrea Böltken for allowing me to interview her over the telephone. These are the terms the editor used to describe the translation project. “Sie [the editor] hat mit ihrer beispielhaften, unermüdlichen Arbeit, ihrem sicheren Urteil und ihrem feinen Sprachgefühl dafür gesorgt, daß die deutsche Ausgabe den englischen Text wortgetreu wiedergibt.” House (1997:155) mentions an interview with the German translator which supports the hypothesis of the German text flattening the original; she does not indicate a source. At the beginning of the 20th century, the British philosopher G. E. Moore argued that values cannot be derived from facts by comparing the logical structure of descriptive and prescriptive utterances. His findings, known as the ‘naturalistic fallacy’, lead to different views of prescriptive judgement as either totally subjective, i.e. devoid of factual, descriptive content, or containing descriptive information because of a specifically human ‘value intuition’, or because of the social nature of values. Today, it seems to be generally agreed that descriptions themselves are prescriptive because they refer to a selection of characteristics the subject deems necessary to describe an object. The selection of characteristics to describe an object can be conventional and internalized to such an extent that it appears ‘objectively true’. The ‘naturalistic fallacy’ contradicts House’s (1997:118) assumption that “social evaluative judgements ... are dependent on, or rather follow from, the analytic, comparative process in translation evaluation”. References Ammann, Margret (1990) ‘Anmerkungen zu einer Theorie der Übersetzungskritik und ihrer praktischen Anwendung’, TextconText 5: 209-250. ------ (1993) ‘Kriterien für eine allgemeine Kritik der Praxis des translatorischen Handelns’, in Justa Holz-Mänttäri and Christiane Nord (eds) Traducere Navem. Festschrift für Katharina Reiß zum 70. Geburtstag, Tampere: University of Tampere, 433-46. Baker, Mona (ed) (1998) Routledge Encyclopedia of Translation Studies, London & New York: Routledge. Bonthrone, Robin (1998) ‘The Quality Wars. Barbarians at the Gate: The Abuse of ‘Translation Quality’ in Europe’, Language International 10(2): 12-15. De Waard, J. and E. Nida (1986) From one Language to Another. Functional Equivalence in Bible Translating, Nashville, Camden, New York: Thomas Nelson Publishers. D’Hulst, Jacqueline (1997) ‘Focus on the Target Text. Towards a Functional Model for Translation Quality Assessment’, in Kinga Klaudy and János Kohn (eds) Transferre necesse est, Budapest: Scholastica, 102-107. Fleischmann, Eberhard (ed) (1997) Translationsdidaktik: Grundfragen der Übersetzungswissenschaft, Tübingen: Narr. 166 Translation Quality Assessment Frank, Armin Paul (1987) ‘Einleitung’, in Brigitte Schultze (ed) Die literarische Übersetzung: Fallstudien zu ihrer Kulturgeschichte, Berlin: Erich Schmidt, IX-XVII. Gerzymisch-Arbogast, Heidrun (1997) ‘Wissenschaftliche Grundlagen für die Evaluierung von Übersetzungsleistungen’, in Eberhard Fleischmann (ed) Translationsdidaktik: Grundfragen der Übersetzungswissenschaft, Tübingen: Narr, 573-79. Goldhagen, Daniel Jonah (1996a) Hitler’s Willing Executioners. Ordinary Germans and the Holocaust, New York: Alfred A. Knopf. ------ (1996b) Hitlers willige Vollstrecker. Ganz gewöhnliche Deutsche und der Holocaust, Berlin: Siedler Verlag. Heil, Johannes and Rainer Erb (eds) (1998) Geschichtswissenschaft und Öffentlichkeit. Der Streit um Daniel J. Goldhagen, Frankfurt a.M.: Fischer Taschenbuch Verlag. Hermans, Theo (ed) (1985) The Manipulation of Literature. Studies in Literary Translation, London & Sydney: Croom Helm. Holmes, James S. (1988) Translated! Papers on Literary Translation and Translation Studies, Amsterdam: Rodopi. Holz-Mänttäri, Justa (1984) Translatorisches Handeln. Theorie und Methode, Helsinki: Suomalainen Tiedeakatemia. ------ and Christiane Nord (eds) (1993) Traducere Navem. Festschrift für Katharina Reiß zum 70. Geburtstag, Tampere: University of Tampere. Holzer, Peter and Cornelia Feyrer (eds) (1998) Text, Sprache, Kultur: Festschrift zum 50jährigen Jubiläum des Institus für Übersetzer- und Dolmetscherausbildung der Universität Innsbruck, Frankfurt a.M.: Lang. Hönig, Hans (1998) ‘Positions, Power and Practice: Functionalist Approaches and Translation Quality Assessment’, in Christina Schäffner (ed) Translation and Quality (Current Issues in Language and Society), Clevedon: Multilingual Matters, 6-34. House, Juliane (1977/1981) A Model for Translation Quality Assessment, Tübingen: TBL Verlag Gunter Narr. ------ (1997) Translation Quality Assessment. A Model Revisited, Tübingen: Gunter Narr. Inglis, Neil (1998) ‘Hanging in the Balance. The Battle over US Translation Standards Is Far from Over’, Language International 10(5): 41-43. Koller, Werner (1997) Einführung in die Übersetzungswissenschaft, Heidelberg: Quelle und Meyer. Klaudy, Kinga and János Kohn (eds) (1997) Transferre necesse est, Budapest: Scholastica. Kußmaul, Paul (1995) Training the Translator, Amsterdam& Philadelphia: Benjamins. Lauscher, Susanne (in progress) Die Bewertung von Übersetzungen in Presserezensionen. Versuch eines praxisorientierten Modells auf empirischer Basis, Innsbruck: Institut für Translationswissenschaft der Universität Innsbruck. Levefere, André (1996) ‘Übersetzen, Übersetzung und Macht’, in Heidemarie Susanne Lauscher 167 Salevsky (ed) Dolmetscher- und Übersetzerausbildung gestern, heute und morgen, Akten des Internationalen Wissenschaftlichen Kolloquiums anläßlich des 100jährigen Jubiläums der Dolmetscher- und Übersetzerausbildung Russisch an der Berliner Universität, Frankfurt u.a.: Lang, 171-80. ------ 1992) Translation, Rewriting and the Manipulation of Literary Fame, London & New York: Routledge. Nord, Christiane (1991) Textanalyse und Übersetzen: theoretische Grundlagen, Methode und didaktische Anwendung einer übersetzungsrelevanten Textanalyse, Heidelberg: Groos. ------ (1993) Einführung in das funktionale Übersetzen, Tübingen: Francke (UTB 1734). ------ (1996) ‘Text Type and Translation Method. An Objective Approach to Translation Criticism’, The Translator 2(1): 81-88. ------ (1997) Translating as a Purposeful Activity: Functionalist Theories Explained, Manchester: St. Jerome Publishing. Pöchhacker, Franz (1994) Simultandolmetschen als komplexes Handeln, Tübingen: Narr. Pohl, Dieter (1997) ‘Die Holocaust-Forschung und Goldhagens Thesen’, Vierteljahreshefte für Zeitgeschichte 1: 1-48. PrunĨ, Erich (1997) ‘Translationskultur (Versuch einer konstruktiven Kritik des translatorischen Handelns’, TextconText 11: 99-127. Reiß, Katharina (1971) Möglichkeiten und Grenzen der Übersetzungskritik. Kategorien und Kriterien für eine sachgerechte Beurteilung von Übersetzungen, München: Max Hueber Verlag. ------ and H. J. Vermeer (1984) Grundlegung einer Allgemeinen Translationstheorie, Tübingen: Niemeyer. Ripfel, Martha (1989) Wörterbuchkritik. Eine empirische Analyse von Wörterbuchrezensionen, Tübingen: Max Niemayer Verlag. Risku, Hanna (1998) Translatorische Kompetenz: kognitive Grundlagen des Übersetzens als Expertentätigkeit, Tübingen: Stauffenburg Verlag. Rübberdt, Irene and Heidemarie Salevsky (1997) ‘New Ideas from Historical Concepts: Schleiermacher and Modern Translation Theory’, in Mary SnellHornby, Zuzana Jettmarová and Klaus Kaindl (eds) Translation as Intercultural Communication, Amsterdam & Philadelphia: John Benjamins, 301312. Salevsky, Heidemarie (1998) ‘Translationsmodelle – Basis für die Bewertung von Übersetzungs- und Dolmetschleistungen?’, in Peter Holzer and Cornelia Feyrer (eds) Text, Sprache, Kultur: Festschrift zum 50jährigen Jubiläum des Institus für Übersetzer- und Dolmetscherausbildung der Universität Innsbruck, Frankfurt a.M.: Lang, 55-68. ------ (ed) (1996) Dolmetscher- und Übersetzerausbildung gestern, heute und morgen, Akten des Internationalen Wissenschaftlichen Kolloquiums anläßlich des 100jährigen Jubiläums der Dolmetscher- und Übersetzerausbildung Russisch an der Berliner Universität, Frankfurt u.a.: Lang. Schäffner, Christina (ed) (1998) Translation and Quality, Clevedon: Multilingual Matters. 168 Translation Quality Assessment Schmitt, P. A. (1998) ‘Qualitätsmanagement’, in Mary Snell-Hornby, Hans G. Hönig, Paul Kußmaul and Peter A. Schmitt (eds) Handbuch Translation, Tübingen: Stauffenburg-Verlag, 394-99. Schoeps, Julius (ed) (1996) Ein Volk von Mördern? Die Dokumentation zur Goldhagen-Kontroverse um die Rolle der Deutschen im Holocaust, Hamburg: Hoffmann und Campe. Schultze, Brigitte (ed) (1987) Die literarische Übersetzung: Fallstudien zu ihrer Kulturgeschichte, Berlin: Erich Schmidt. Snell-Hornby, Mary (ed) (1986) Übersetzungswissenschaft - Eine Neuorientierung. Zur Integrierung von Theorie und Praxis, Tübingen: Francke. ------, Zuzana Jettmarová and Klaus Kaindl (eds) (1997) Translation as Intercultural Communication, Amsterdam & Phialdelphia: John Benjamins. ------, Hans G. Hönig, Paul Kußmaul and Peter A. Schmitt (eds) (1998) Handbuch Translation, Tübingen: Stauffenburg-Verlag. Toury, Gideon (1995) Descriptive Translation Studies – and beyond, Amsterdam & Philadelphia: John Benjamins. Van den Broeck, Raymond (1985) ‘Second Thoughts on Translation Criticism. A Model of its Analytic Function’, in Theo Hermans (ed) The Manipulation of Literature. Studies in Literary Translation, London & Sydney: Croom Helm, 54-62. Vannerem, Mia and Mary Snell-Hornby (1986) ‘Die Szene hinter dem Text: ‘scenes-and-frames semantics’ in der Übersetzung’, in Mary Snell-Hornby (ed) Übersetzungswissenschaft – Eine Neuorientierung. Zur Integrierung von Theorie und Praxis, Tübingen: Francke, 184-205. Venuti, Lawrence (1995) The Translator’s Invisibility. A History of Translation, London & New York: Routledge. Vermeer, Hans J. (1989) Skopos und Translationsauftrag – Aufsätze, Heidelberg: thw. ------ (1992) Skizzen zu einer Geschichte der Translation, Frankfurt a.M.: thw. ------ (1996) A Skopos Theory of Translation (Some arguments for and against), Heidelberg: TEXTconTEXT-Verlag. ------ and Heidrun Witte (1990) Mögen Sie Zistrosen? Scenes & Frames & Channels im Translatorischen Handeln, Heidelberg: Groos. Weber, Max (1988) Gesammelte Aufsätze zur Wissenschaftslehre. 7.Aufl, Tübingen Mohr (UTB für Wissenschaft:Uni-Taschenbücher 1492). Wilss, Wolfram (1977) Übersetzungswissenschaft. Probleme und Methoden, Stuttgart: Klett.