SciELO - Scientific Electronic Library Online

 
vol.36 issue3TRANSLATION AS INICIATIONGuerini, Andréia; Torres, Marie-Hélène Catherine; Costa, Walter Carlos (Orgs). Vozes tradutórias: 20 anos de Cadernos de Tradução. Florianópolis: PGET/UFSC, 2016, 273 p. author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand

Journal

Article

Indicators

Related links

Share


Cadernos de Tradução

On-line version ISSN 2175-7968

Cad. Trad. vol.36 no.3 Florianópolis Sept./Dec. 2016

http://dx.doi.org/10.5007/2175-7968.2016v36n3p214 

TRANSLATED ARTICLES

EXPLORING TRANSLATION THEORIES

Anthony Pym

Translation:

Eduardo César Godarth1  *

Yéo N’gana1  **

Bernardo Sant’Anna1  ***

1Universidade Federal de Santa Catarina

Additional chapter: Descriptions – the intellectual background

This material explains the historical background of the concepts presented in chapter 5 of the printed book.

If we set out to describe a translation or an act of translating, the simple description might seem to require no grand theory. In fact, it could be considered too simple to be taken seriously by scholars. Some of the most significant concepts in European translation theory have nevertheless come from what we shall call a broad “descriptive paradigm,” and this chapter describes the ways that paradigm developed in the twentieth century. This background should help connect translation theory to some of the main anti-humanist currents of the day. It is also intended to correct some common misunderstandings, particularly with respect to the many ways the various schools and centers were interconnected. We place some emphasis on the Russian Formalists, even though they did not produce any major works on translation. This is because the key ideas of the Formalists can be traced through various paths throughout the century, reaching several points at which major translation theories did develop. The first connection is with the work done in Prague, Bratislava and, more loosely connected, Leipzig. The second link is with the “Tel Aviv school” (Even-Zohar, Toury and the development of Descriptive Translation Studies). And the third link is through Holland and Flanders. When literary scholars from those three areas met and discussed their projects at a series of conferences, Translation Studies started to take shape as an academic discipline. That is why the history is important—this particular paradigm does not come from the same roots as the others mentioned in this book. The second half of the chapter describes the main concepts used within descriptive studies: translation shifts, systems and polysystems, “assumed translations,” and a focus on the target side. In the next chapter we look more closely at some of the findings that have come from the general descriptive approach.

Special thanks to Itamar Even-Zohar, Gideon Toury, Zuzana Jettmarová, Jana Králová and Christina Schäffner for their help and advice with this chapter.

The main points covered in this chapter are:

  • Descriptive Translation Studies developed from a tradition in which objective scientific methods were applied to cultural products.

  • Those methods were often applied to translation by literary scholars working in smaller cultures.

  • Rather than prescribe what a good translation should be like, descriptive approaches try to say what translations are like or could be like.

  • Translation shifts are regular differences between translations and their source texts. They can be analyzed top-down or bottom-up.

  • Translations play a role in the development of cultural systems.

  • The innovative or conservative position of translations within a cultural system depends on the system’s relation with other systems, and may correlate with the type of translation strategy used.

  • When selecting texts to study, translations can be considered facts of target culture only, as opposed to the source-culture context that is predominant in the equivalence paradigm.

  • Translators’ performances are regulated by collective “norms,” based on informal consensus about what is to be expected from a translator.

  • The descriptive approach was instrumental in organizing Translation Studies as an academic discipline with an empirical basis.

5.0 What happened to equivalence

Equivalence went out of fashion. German-language Skopostheorie made it even more unfashionable by arguing that since “functional consistency” (the closest thing they had to equivalence) was no more than one of many possible requirements, translation usually requires transformations of a rather more radical kind. For those theorists, equivalence became quite a small thing, a special case. At almost the same time, however, other theorists were dismantling equivalence in precisely the opposite way. For this second very broad group, for what Gideon Toury would eventually construct as “Descriptive Translation Studies,” equivalence was a feature of all translations, simply because they were thought to be translations, no matter what their linguistic or aesthetic quality (cf. Toury 1980: 63- 70). That changed everything. If equivalence was suddenly everywhere in translations, or almost, it could no longer be used to support any linguistics that would help people create it, nor could the concept directly serve the prescriptive training of translators. Translation Studies was thus moved into a realm that was relatively unprotected by any parent discipline; it had to become its own discipline. The descriptive approach emphasized the need to carry out research on translation, mostly research of the kind done in structuralist literary studies, rather than expound principles and opinions. The theories associated with the research were thus positioned problematically out of touch with the growing number of training institutions; they were in an institutional context quite different from that of Skopostheorie. Here we will follow the adventures of that historical move. In the following chapter we will consider what the descriptions might actually have discovered.

5.2 Origins of the descriptive paradigm

The name “Descriptive Translation Studies” (with the capitals) was never fully consecrated as such until Gideon Toury’s book Descriptive Translation Studies and beyond (1995; Spanish translation 2004). It has since become a flag of convenience for a loose flotilla of innovative scholars. Around that particular name there is now a rather large body of thought and research. On the surface, this would seem to be a general paradigm in which scholars have set out to describe what translations actually are, rather than just prescribe how they should be. Those terms, though, are simplifications. If the aim were merely to describe, there would be little need for any grand theory. And yet what we find in this paradigm is a host of theoretical concepts: systems, shifts, norms, universals and laws, to name the most prominent, plus a long ongoing debate about how to define the term “translation” itself. Despite the emphasis on description, this remains very much a paradigm for theoretical activity.

In the historical context, the shift from prescription to description involved a clear challenge to the institutionalization of the equivalence paradigm. Rather than just tell people how to translate well (which is what and most equivalence-based linguistic analyses set out to do, along with Skopostheorie and hopefully most training institutions), descriptivist theories aim to identify how people actually do translate, no matter what the supposed quality. The equivalence paradigm mostly came from scholars who worked in linguistics or professional training; the descriptive paradigm was mostly peopled largely by researchers with a background in literary studies. This division appeared in the 1970s and early 1980s, roughly in parallel with the development of Skopostheorie. The intellectual genealogies of the descriptive paradigm might nevertheless be traced back to at least the early twentieth century.

5.2.1 Russian Formalism and its legacy

In the first decades of the twentieth century, the movement known as Russian Formalism set out to produce scientific descriptions of cultural products and systems, particularly in the field of literature. The basic idea was that science could and should be applied to the cultural sphere. As simple as that might appear, it was something that had never been done before in any consistent way. Nineteenth-century applications of empirical science to literature were mostly limited to prescribing the way novels should describe society (such was the ideology of Naturalism), along with some attempts to analyze artistic language within what became known as the Symbolist movement. Indeed, it may well be from that broad Symbolism that the seeds of Russian Formalism were sown (cf. Genette 1976: 312). In 1915 a group of young university students who met at the courses of Professor Vengerov founded the “Moscow Linguistic Circle.” This brought together Roman Jakobson, Petr Bogatyrev and Grigori Vinokur, who sought to study the specificity of literature in with the help of concepts borrowed from the emerging pre-structural linguistics (especially the notion of “distinctive features” in language). In 1916 the Society for the Study of Poetic Language (known by the acronym Opojaz) was founded in Saint Petersburg, bringing together Viktor Shklovsky, Boris Eikhenbaum, Boris Tomashevsky and later Yuri Tynianov. These were mostly literary historians in search of the underlying laws and principles of literature. One project was within linguistics, the other was concerned with poetic language; but at that stage the two sides could develop substantial common ground. Both projects were based on a very simple idea: as we have said, the methods and goals of science were to be applied to culture. Both sought to develop explicit models, defining terms carefully and using observations to verify or falsify hypothesized principles or laws of artistic language, independently of the psychology of authors, the emotions of readers, or any supposed representation of societies. According to a powerful Formalist principle, the object of study was not the literary work in itself, nor its contents, but the underlying features that made it literary (“literariness,” or literaturnost’, as Roman Jakobson put it). This literary language had its own artistic techniques (priyómy in Shklovsky’s terminology, sometimes rendered as devices in English, or procédés in French); it presumably had its own underlying systemic patterning, and, especially in the work of Tynyanov, specific dynamic relations with other cultural systems, both synchronically and diachronically. In describing process of change within literary systems, Tynyanov recognizes that a new “constitutive principle” may start from a series of chance occurrences or encounters, but in order to become substantial the principle may need the transfer of models and materials from beyond itself (1924: 19-20). That observation was not actually accompanied by any consideration of the role of translations, although elsewhere Tynyanov did write a critical account of Tyutchev’s renditions of Heine (study dated 1921, included in Arxaisty i novatory in 1929 and in the French translation Formalisme et histoire littéraire of 1991 but not in the partial German translation of 1967). A framework for the study of literary translation was certainly there, but the study itself would seem not to have been part of the main agenda of Russian Formalism. Any potential insights about translation would remain without immediate impact within Russian theory, although some students of Tynjanov’s, like Andrei Fedorov, became major theorists of translation in the Soviet era, and Jakobson would go on to write several seminal papers on translation, as we have noted in previous chapters.The legacy of the Formalist moment would have been passed on, in various forms, to the sociolinguist Valentin Vološinov, perhaps in part to the cultural theorist Mikhail Bahktin, and more obviously to the semioticians Yuri Lotman and Boris Uspenski, whose names might be more familiar. None of those cultural theorists, however, produced systematic theoretical work on translation; nor did the later Formalists themselves. When Andrei Fedorov wrote his ground-breaking “Introduction to the Theory of Translation” in 1953, he had studied at the State Institute for the History of the Arts, where the Formalists had created a program (our thanks to Itamar Even-Zohar for this information), so something of the basic approach certainly lived on. The traces of that legacy might be divined from Fedorov’s highly systematic approach to basic principles (after paying due homage to Marx and Lenin) and his detailed investigation of the way different genres and stylistic features should be translated. The same can be said of Efim Etkind, whose work on Russian poet-translators (1973) drew attention to the role of translation in the development of cultures.

From Fedorov and others we do reach a certain Russian school of translation theory, which includes important work by Retsker and Shveitser . Their general principles, however, are not linked to the literary school; they are linguistic, prescriptive, and basically compatible with the equivalence paradigm. If we are seeking the way scientific descriptions of systems led to a new paradigm of translation theory, then we have to look elsewhere.

What concerns us more here is how the Formalist ideas moved out of Russian and reached other translation scholars. We can pick out three interrelated threads: through Prague and Bratislava, through Tel Aviv, and through Holland and Flanders.

5.2.2 Structuralism in Prague, Bratislava, and Leipzig

One strand led to the scholars who met from 1926 under the name of the Cercle linguistique de Prague (the Prague Linguistic Circle). The most obvious connection was the linguist Roman Jakobson, who had taken a position in Brno (and whose escape from the German occupied Prague took him to Copenhagen, Stockholm, New York and Harvard, stimulating intellectual curiosity as he went, eventually cultivating some fundamental insights into translation). Another Russian member of the Cercle was Nikolai Trubetzkoi, who actually held a chair in Vienna, and a further member of the group was Henrik Becker, who attended the first meeting but lived in Leipzig (see Dušková 1999). We note these details to indicate that the Prague circle clearly extended beyond the city of Prague. In 1928 Jakobson, Trubetzkoi and other members of the group attended the First International Conference of Linguists in The Hague, the Netherlands, where they signed a resolution calling for synchronic linguistic analysis. They actually signed alongside Charles Bally and Albert Sechehaye, who had compiled and edited Saussure’s Cours de linguistique générale (1916), the prime reference for the science of synchronic analysis. The strands of intellectual history crossed; they are not easily spun into national traditions. But was there any translation theory in the web?

The scientific approach of Russian Formalism provided an impulse for basic advances of the Prague Cercle in structuralist linguistics, working in areas from phonology to the study of poetic language, all potentially part of the general analysis of cultural signs. Although the development of phonemics was undoubtedly the great lasting success of the group (and indeed of structuralism in general, we shall argue), their interests extended to many aspects of culture, especially literature, and occasionally translation.

In the work of Jan Mukařovský of the Prague Circle we find clear awareness of the historical role of translation. In his 1936 article “Francouzská poezie Karla Čapka” (The French poetry of Karel Čapek), Mukařovský argues that translation is one of the ways in which national literatures can be transformed, since they seek and develop equivalents for foreign texts (see Králová 2006). This insight might be gleaned from the work of Tynyanov within the frame of Russian Formalism as such (or indeed from work by Zhirmunskij on Pushkin, or Vinogradov on Gogol), but in Mukařovský it is now clearly stated as such.

In terms of literary studies, the transformational role of translation became part and parcel of an approach that saw cultural systems (such as national literatures) as sets of structural relations developing not just in terms of their internal logic, as had mostly been the case mostly in Russian Formalism, nor exclusively from external influences, as might have been the case of traditional historical studies, but from the complex social context formed by dynamics on both sides at once. The interest of translation was that it necessarily cut across those two deceptively separate frames; it forced the literary historian to see the internal and the external in the one vision. We might argue that this was more likely to happen when dealing with a “minor system” like Czech literature than with a “major” and apparently more independent system like Russian literature. The Prague interest in translation was perhaps not entirely an accident.

Prague structuralism was properly a phenomenon of the 1920s and 1930s. There was nevertheless a tradition, apparently discontinuous, that saw its influence filter down through the decades, especially in the study of literature. In the 1960s and 1970s we find the Czech scholar Jiří Levý and the Slovak scholars František Miko and Anton Popovič setting out to describe the structural principles underlying literary translations (see Jettmarová 2005; Králová 1998, 2006). Importantly, these scholars explicitly limited their prejudices about what equivalence was, or about what a “good translation” might be; their ideas of science made them describe rather than prescribe. Levý was publishing in Czech in the 1960s and became more widely known in German (Levý 1969). His work shows a gift for applying models from the exact sciences, drawing not only on linguistics but also on game theory (as we shall see in our chapter on indeterminism). Miko (1970) proposed to focus on what happens to the formal features of a text in translation. Popovič (1970) recognized that since translations transform texts, the study of translation should focus on what is changed as much as what remains the same. He thus set out to describe the “translation shifts” that affected the level of expression. We will return to this key concept below.

Note should be made here of the loose “Leipzig School” of translation theorists, who were working in similar ways from 1964 (for historical details see Wotjak 2002; on the conceptual range, Jung 2000). Although we would hesitate to draw any direct line with Russian Formalism and its legacy, there can be no doubt that scholars of the order of Otto Kade (in social communications theory), Gert Jäger (in structuralist linguistics) and Albrecht Neubert (in pragmatics and text linguistics) sought a scientific approach to translation, requiring clear concepts. This led them to reshuffle and define many of the common German terms. For example, “linguistic mediation” (Sprachmittlung) became the wider object of study (see Kade 1980), rising above a narrow conception of translation, and Kade coined the neologism Translation, in German, to cover both written translation and oral interpreting. The work in Leipzig was also important for the re-definition of “translation shifts,” since the research by Kade and Neubert increasingly focused on text-level relations. One should also admit that the school’s relation with official Marxist ideology sometimes went beyond mere lip-service. When Kade approached linguistic mediation as a social phenomenon, he sought the causes of translation problems not in the mysteries of language but in the “non- corresponding” development of two historical societies. The systemic thought is clear, wide-ranging and important, as indeed it is in Marx. The main work of Leipzig, however, was on non-literary translation at text level, without major investigation of social systems. As such, it did not become an integral part of the way the descriptive paradigm developed (the early paradigm tended to be literary and systemic). It instead fed into the development of the equivalence paradigm, which is where we have noted Kade’s work on types of equivalence; it had a terminological influence on general purpose-based approaches, which adopted the German term Translation, as well as the general penchant for re-naming things; some of its terms and basic text-functional insights helped fuel the development of Skopos theory; and Kade had his word to say in the development of Interpreting Studies (see Pöchhacker 2004: 34-35). That said, the Leipzig School’s impetus and identity did not live far beyond the fall of the Berlin Wall in 1989, at least not within Germany.

Albrecht Neubert has helped to foster text-linguistic approaches in the United States, largely thanks to an exchange program between Leipzig and Kent State, and Christina Schäffner, in the next generation, went to the United Kingdom, where she specializes in functionalist text-linguistic approaches, especially with respect to the translation of political texts. The theorists remaining in Germany tend to argue the toss between equivalence and Skopos, without great interest in description as a separate paradigm.

There is little evidence of any profound influence leading from Prague or Bratislava to Leipzig, despite geographical and political proximity. We should remember, however, that the various Communist regimes of the period attached great importance to translation, both as a way of maintaining national languages and as a means of fostering the international dimensions of their cause. This concerned not just the role of Russian as a pivot language, but also translation policies for literary works from across the like-minded world, from Latin America and Africa, for example, as well as translations of ideological texts for the future liberation of oppressed peoples. Those policies required translators; the translators had to be trained; the training created institutional space for thought on translation. Whatever we might nowadays think of the official ideologies, the development of systematic translation theory owes a great deal to the Communist period in the Soviet Union and Central Europe. We cannot reproduce the myth of an enlightened pre-Revolution Russian Formalism that somehow struggled through the dark days of benighted regimes. Stalin certainly persecuted the formalist movement, which he regarded as anti-Marxist, but the history of the Communist period should not be reduced to that.

5.2.3 Polysystems in Tel Aviv

A second strand of contact led from Russian Formalism to Tel Aviv. The agent of transfer here was the Israeli scholar Itamar Even-Zohar, who became aware of the Russian texts as a student of Benjamin Harshav at the Hebrew University of Jerusalem and started to read them while studying in Copenhagen (see Even -Zohar 2008b; the networks of European scholarship are complex) and working on his doctoral dissertation. Even-Zohar has generally been concerned with the systemic descriptions of the ways cultures develop, and he explicitly follows the insights of Tynyanov, Jakobson and Eikhenbaum in studying cultural phenomena as systems, with their own principles and laws that await discovery. Rather than focus on literary systems in isolation, Even-Zohar has sought to see cultures as “polysystems,” roughly as large, heterogeneous and complex systems (such as “Israeli culture,” “French culture”) within which there are smaller systems like literature, language, law, architecture, family life, and so on (hence the “poly,” meaning “many”) . These smaller systems may also be complex and dynamic, warranting the “poly” prefix as well.

In his detached view of cultural systems, Even-Zohar has remained faithful to the tradition of scientific modeling and to the multiple strands of the European tradition, as was shown in his doctoral thesis on translation. Like the structuralists in Prague and Bratislava, he has worked from within a “minor” culture (Hebrew), and his interest in the pre-Israeli phases of Hebrew literature has led him to develop a view of it as a multiple-components system (a polysystem). Even-Zohar founded a section for Translation Studies at Tel Aviv University and was at the origin of what would become the “Tel Aviv School” of Translation Studies, which includes Gideon Toury (whose PhD was supervised by Even-Zohar), Rakefet Sela-Sheffy and Gisèle Sapiro. This strand thus leads more or less directly to the coining of the term “Descriptive Translation Studies.”

5.2.4 Descriptivism in Holland and Flanders

The third strand concerns a group of scholars working in Holland and Flanders (the Dutch- speaking part of Belgium), sometimes referred to as the “Low Countries” school. Names here would include American James S Holmes1 in Amsterdam, and the Belgians José Lambert, Raymond Van den Broeck, André Lefevere and Theo Hermans. Their connection with Russian Formalism is far less important or direct than in the other cases. Major texts of the Russian school had been translated into French (1965, edited by Todorov) and English (also 1965, edited by Lemon and Reis; then 1971, edited by Matejka and Pomorska), but the intellectual climate was by this stage imbued with the prestige of structuralism anyway. The tenets of east- European thinking certainly also reached the Low Countries scholars through a series of personal contacts, particularly with Anton Popovič, who also met Even-Zohar and Toury. This led to a rather broader meeting of minds.

5.2.5 A European descriptivism

These three strands came together from the late 1960s, especially following a conference in Bratislava, Slovakia, in 1968 (see Holmes ed. 1970). Collective work was then carried out in the 1970s (see Holmes et al. eds. 1978), with some of the main scholars meeting as the Translation Committee of the International Comparative Literature Association. Toury (1978) built the bridge with Even-Zohar’s work on the way cultures develop. A series of influential papers by most of the scholars was then brought together in the volume The Manipulation of Literature (ed. Hermans 1985), and for some time the group was half-jokingly dubbed the “manipulation” school, although the term says very little about what they were doing.

As the diverse backgrounds would suggest, this was far from a group of scholars sharing the same theories (see Hermans 1999 for a detailed survey). They would all nevertheless agree that a scientific approach should be used to find out about the world, rather than to evaluate or criticize what is found. They would thus more or less agree that the previous work on translation, including many of the theories elaborated within the equivalence paradigm, was “pre-scientific” (a harsh term, but it was used freely enough) . And they all agreed, obviously, that translation was worth studying seriously, and that this opposed them in part to literary studies that had mostly seen translations as marginal products, inherently inferior to originals, and thus of little interest. As for the rest, each theorist’s precepts and interests tended to work in very different ways and on various different levels.

A shortlist of ideas in the development of the descriptive paradigm

In the following we indicate the names of scholars who were instrumental in the development and use of the propositions, although many other names could also be listed and most names should be associated with far more than one idea:

  1. The relations between source and target texts can be described in terms of “translation shifts” (Levý, Miko, Popovič).

  2. The innovative or conservative position of translations within a cultural system depends on the system’s relation with other systems, and correlates with the type of translation strategy used (Even-Zohar, Holmes, Toury).

  3. Translation Studies should be an empirical descriptive discipline with a hierarchical organization and a structured research program (Holmes, Toury).

  4. When selecting texts to study, translations should be considered facts of target culture (Toury).

  5. To understand not just translations but all kinds of “rewriting,” we have to consider the social contexts, especially patronage (Lefevere).

  6. Translation scholars need to look at more than just literature (Lambert).

As we have seen, the descriptive paradigm cannot really be tied to a coherent geographical location like the “Low Countries.” To be sure, all the scholars mentioned above are or were from relatively small cultures, and this could explain some of their insights. The smaller the culture, the more important translation tends to be in that culture (a descriptive hypothesis formulated in Pym 2004 but clearly deducible from Even-Zohar 1978). One should thus not be surprised that precisely those scholars decided to study translation seriously. However, there were also some significant changes of location. Theo Hermans moved to London and has had a decisive influence on the development of Translation Studies in the United Kingdom; André Lefevere moved to the United States, where the legacy has been less influential (no doubt because of his unfortunately early death in 1996, but perhaps also because the United States has a very large complex culture). The growth of research-based Translation Studies has since spread the basic descriptive paradigm virtually all over the world. For that same reason, the paradigm cannot really be restricted to a historical moment like the 1960s or 1970s (as suggested in Venuti 2000). Many of the fundamental research questions formulated by the original group are still being answered today, no doubt because the paradigm remains eminently suited to empirical research. Numerous PhD dissertations use these ideas and are effectively contributing to our knowledge about translations.

We will now attempt to outline the main research models developed within the descriptive paradigm. In the next chapter we will consider some of the findings actually produced by this research.

5.3 The attraction of structuralism

We have met the term “structuralism” several times in the above pages, especially with reference to the equivalence paradigm. We now take some time to explain what the term means, and why it was so important in the twentieth century.

At its simplest level, structuralism means that instead of studying things in themselves, as one might do under the influence of positivism, we study the relations (“structures”) between things. The trick is that, while the things are visible to everyone, the relations are hidden. Structuralism thus invites us to discover the secret logics that lie somewhere beneath the surface of cultural products. In retrospect, it offers the same appeal as did Marx when discovering the relations of production at the base of the way societies work, or Freud revealing the principles of the unconscious mind. For most structuralists of the first half of the twentieth century, the structures are really there, within our languages and cultures; structures do not come from the subjectivity of the individual researcher. Structuralism invites us to reveal objective verifiable truths, reachable through patient discovery procedures. It offers a scientific approach to culture (as in Russian Formalism). That was and remains a very appealing and powerful invitation, extended to anyone in search of knowledge.

Examples of these underlying structures can be found in many of the approaches that see languages as “world views.” One instance would be Saussure’s example of the way English sheep and French mouton enter into different structures within their language systems. We have seen how this idea initially created problems for the equivalence paradigm, which had to argue that translation was nevertheless somehow possible. For the descriptive paradigm, however, structuralism was something to learn from, not to oppose. Rather than ask if sheep could really translate mouton, the initial task in this paradigm was to describe the way historical translators effectively resolved the problem.

Structuralism enters the descriptive story in much the same suitcases as Russian Formalism, and more particularly through the Prague Cercle (which was indeed in touch with the legacy of Saussure) . As we have mentioned above, the Prague development of phonology would be the one great success story of structuralism. For example, in English we hear the sounds /b/ and /v/ as significantly different, since their difference helps us to distinguish “bat” from “vat.” In spoken Iberian Spanish, however, there is generally no significant difference, to the extent that the spoken language will not help people spell words like “vota” or “bota.” This is because English has two separate phonemes for these sounds (quite apart from the many different ways of pronouncing those sounds), whereas Spanish only has one. People speaking the languages can certainly pronounce the different sounds, but the underlying structure of each language divides up the sounds in different ways. That underlying structure is acquired when we learn a language, even though we are not aware of it. For Prague synchronic linguistics, and for structuralism in general, the object of study should thus be the underlying structure (the phonemes), not the surface-level phenomena (the details of phonetics). This was basically the same insight as Saussure’s analysis of sheep and mouton, except that in phonemics the structures formed complete and relatively stable systems. If you change one term (especially a vowel in English), the other terms actually do tend to change. In that sense, phonemics moved the focus of ideal research from structures to strong systems.

Once you understand this view of what a structure is, it is relatively easy to see a system as a network of structures where, ideally, a change in one term implies some kind of change in all others. There are actually very few cultural systems where this is the case. Most have parts where changes are connected, as in a particular region or genre, and others that remain relatively undisturbed. The introduction of a new lexical item does not alter the entire language (lexical fields are segmented, and the repertoires are mostly open-ended), although a change in tense usage would normally affect all the tenses in the language (verb tenses form systems with very few terms). The reigning idea, however, was that structures could indeed connect everything to everything, giving rise to numerous plans to explain the whole world.

The basic idea of structuralism went traveling around the West in several different guises. Many parts of the humanities were applying elements of the approach. In anthropology, the tradition leading from Mauss to Lévi-Strauss had tapped into structuralist linguistics; structuralism was in the scientific epistemologies of Bachelard and Merleau-Ponty; it would be behind the linguistics of Benveniste and Chomsky (who sought its Cartesian connections). Virtually across the humanities, researchers set out to study relations between things, in search of hidden principles. Translation Studies was no exception. But what would the basic structures of translation look like? Were translations in any way necessary for cultural systems? And could there be anything like a system of translations?

5.4 Theoretical concepts within the descriptive paradigm

Within this intellectual context, there is nothing obvious or banal in wanting to describe translations, rather than “prescribe” what translations should ideally be. Descriptive approaches were setting out to discover relations and laws that were in some way hidden; they were working on the problems of structuralism. Further, given the terms and concepts that had been accumulated in many related disciplines since the Russian Formalists, the researchers working on translation were able to draw on a broad range of established categories (no one can simply describe in language what they see with the eye) and indeed of hypotheses about what they would find. The underlying approach was certainly empirical, in the sense that these researchers were going out to test their ideas on actual data. But it would be quite misleading to reduce the approach to empiricism alone (as is done, for example, in Merino and Rabadán 2004), as if there were no properly theoretical concepts at work.

In the following sections we will briefly describe the main concepts at work within the descriptive paradigm. In the next chapter we will look at the larger concepts of norms and laws.

5.4.1 Translation shifts and their analysis

The most obvious way to apply structuralism to translation is to see the source and target texts as sets of structures. We can compare the texts and see where the structures are different, we then have specific structures (the differences) that somehow belong to the field of translation. That idea is as simple to understand as it is difficult to apply.

The structural differences between translations and their sources can be described as “translation shifts,” a term found in many different theories. For Catford, shifts are “departures from formal correspondence” (1965: 73), which sounds clear enough. If formal correspondence is what we find between “Friday the 13th” and “viernes y 13,” then any other rendition will be a “shift” of some kind. The range of possible shifts might thus include all the things that Vinay and Darbelnet (1958) found translations doing, or indeed anything detected by anyone within the equivalence paradigm. A shift might come from the translator’s decision to render function rather than form, or to translate a semantic value on a different linguistic level, or to create the correspondence at a different place in the text (using a strategy of compensation), or indeed to select different genre conventions. Much research can be carried out in this way: compare the texts, collect the differences, then try to organize the various kinds of shifts.

There are at least two ways of approaching this task: bottom-up analysis starts from the smaller units (usually terms, phrases or sentences) and works up to the larger ones (text, context, genre, culture); top -down analysis goes the other way, starting with the larger systemic factors (especially constructs such as the position of translations within the sociocultural system) and working down to the smaller ones (especially categories like translation strategies) . In principle, it should make no difference which end you start at: all roads lead to Rome, and there are always dialectics of loops and jumps between levels. Yet, perhaps surprisingly, the difference between bottom-up and top-down has a lot to do with the role of theory in description.

5.4.1.1 Bottom-up shift analysis

The range and complexity of bottom-up analysis is most completely seen in the comparative model developed by Kitty van Leuven-Zwart (1989, 1990), where shifts are categorized on many levels from the micro (below sentence level) to the macro (in her case, text-scale narrative structures). A useful summary is in the first edition of Munday’s Introducing Translation Studies (2001: 63-65) (and Hermans 1999: 58-63), however the model is omitted from the second edition of Munday (2008) since it is rarely used any more. Here we are interested in the underlying reasons why it is no longer used.

In Leuven-Zwart, the basic textual units entering into comparison are called “transemes” (cf. the “translemas” in Rabadán 1991). For example, the two corresponding units might be English “she sat up suddenly” and the Spanish “se enderezó,” which basically means that she sat up. What these two transemes have in common would be the “architranseme.” Once you have identified that, you can start to look for shifts, which can then be categorized in much the same way as Vinay and Darbelnet had proposed from within the descriptive paradigm. For example, you might note that the two phrases occupy corresponding positions in the two texts but the English has a value (suddenness) that seems to be absent in the Spanish. So we write down “absence of aspect of action,” and we call this absence a shift. Eventually we will have compiled a notebook full of such shifts, which we hope will form patterns (manifesting structures of some kind) that can tell us something about the translation. What could be wrong with that? Since this “sit up” example is presented as being relatively uncomplicated in both Hermans and Munday, it is worth spending some time on the difficulties it might actually involve:

  • For a start, how can we be sure that the value of “suddenly” is not in the Spanish? The verb “enderezó” is in the preterit tense (actually the pretérito indefinido), which in Spanish has a value in opposition to the past imperfect (the pretérito imperfecto, giving the form “enderezaba”), a tense that does not exist as such in English. That is, both languages can say “He was in the process of sitting up,” but English does not have a simple past tense for such drawn-out actions; Spanish does. One could thus argue, in pure structuralist mode, that the selection of the Spanish preterit in itself represents the value “suddenness.” The shift would then be from the English adverbial to the Spanish tense, and it would be regulated by the differences between the two tense systems.

  • Alternatively (although possibly for similar reasons), we might check large corpora of general English and Spanish and note that the English verb “sit” is associated with adverbials and phrasal particles far more than is the case for the Spanish verb “enderezarse” (none the least because “sit up” and “sit down” have no formal equivalents in Romance languages). In that case, the translator might have omitted the value “suddenly” (which could be expressed as “de repente,” for example) simply because it did not sound right in Spanish; it would have been an unusual collocation (for comparisons of verbs of movement in Spanish and English, see Mora Gutiérrez 2001, Slobin 1996, 2003). We might thus find an alternative non-structural justification for the translator’s decision, albeit without denying the underlying logic of structures.

  • More worryingly, if we try to apply this type of analysis to our “Friday the 13th” example, how can we be sure that the non-shift involves the form or the function? In a context framed by superstition, surely “martes y 13” (Tuesday the 13th) would be the expected translation, the normal one, the non-shift? What right do we have to pick one rendition and call it the “proper” or “expected” translation, and thereby relegate all the other possible renditions to the category of “shifts”?

  • Finally, there are many cases where formal correspondence itself implies some kind of shift. For example, the American English term democracy certainly corresponded formally to the East German term Demokratie (as in the Deutsche Demokratische Republik), but with a remarkable shift of ideological content (the example is used by Arrojo in Chesterman and Arrojo 2000). So why should the formal correspondence itself not represent a shift?

In all these ways, we find that bottom-up shift analysis presupposes far too quickly that the meanings of language are clear and stable (i.e. not subject to interpretation), and that there is thus one stable common core (the “architranseme”) in relation to which all the rest would represent “shifts.” On that score, the approach has far more to do with the equivalence paradigm than with the precepts of scientific description. Even without questioning the ultimately arbitrary way in which transemes are identified, there must remain some doubt about the identification of the shift and of its causation. The bottom-up accumulation of shifts tends to be methodologically murky, and the long lists of differences only rarely congeal into firm findings at the higher level of analysis. This approach can produce much doubt and even more data. At the end of the day, it requires orientation from a few reductive theories. That is one of the reasons why the descriptive paradigm is actually full of theories.

5.4.1.2 Top-down shift analysis

The descriptive work in central Europe tended to be much more theoretical than the bottom-up description of shifts outlined by Catford and substantiated by van Leuven-Zwart. In Leipzig, Kade (1968) explicitly argued that a bottom-up approach (“induction”) had to be accompanied by top-down analysis (a “hypothetico-deductive” approach) if theoretical results were to be achieved (that is, if the “necessity” and “regularity” of translation were to be understood). In Bratislava and Nitra the analysis of “shifts of expression” was also happening in roughly the same years as Catford (cf. Popovič 1968, 1970; Miko 1970) but the focus was not at all the same. For many of the Europeans, especially those coming from literary studies, shifts could be made quite independently of any simple desire to maintain equivalence. They could thus be approached in a top-down way, starting from major hypotheses about why they might exist and how they could form tendencies.

Popovič, for instance, claimed that there are “two stylistic norms in the translator’s work: the norm of the original and the norm of the translation” (1968/70: 82) . This seems so simple as to be obvious. Yet consider the consequence: as soon as the two “stylistic norms” are announced, the multiplicity of shifts is already theorized in terms of coherent patterns (“norms” is a term we will meet further below) . This kind of approach could connect quite easily with the study of literary stylistics, where one might see the two interacting “norms” as the voices of author and translator. On another level, shifts could be patterned differently because of historical factors (the nature of the receiving system, patronage, new text purpose, different ideas about what translation is, etc.). Or again, some shifts might come about simply as a result of the translation process as such (these would later be dubbed potential “universals”). On all those levels, the top-down approach to shifts seeks causal factors (the reasons for the shifts) that are quite different from those of the equivalence paradigm. These descriptive approaches could obviously join forces with the bottom-up analyses carried out by linguists, but their theoretical frame was fundamentally different. In effect, despite the misnomer “descriptive,” these were theories about the possible causes (personal, institutional, historical) explaining why people translate differently.

As an example of the top-down analysis of historically bound translation shifts, consider the basic problem of what to do with a source text that is in verse. This is analyzed in a seminal paper by James S Holmes (1970), first presented at a conference on “Translation as an Art” held in Bratislava, Slovakia, in May 1968 and published in a volume co-edited by Holmes himself (an American resident in Amsterdam), Frans de Haas (Amsterdam) and the Slovak Anton Popovič (making the book of the key publications where various strands come together).

We know that in some target cultures (notably in French, at least until the late nineteenth century), foreign verse forms can consistently be rendered in prose. So the problem is solved: translators know what to do (translate into prose), and readers know what to expect (verse is for only texts originally written in French). That would be one huge kind of shift, and it has remarkably little to do with equivalence of the linguistic kind. In other cultural situations, however, alternative shifts may be deemed appropriate. Holmes (1970) formalizes these further shifts in terms of four available options (in addition to the blanket rendering of verse as prose): the translator can use a form that looks like the source-text form (“mimetic form”); they can select a form that fulfils a similar function (“analogical form”); they can develop a new form on the basis of the text’s content (“organic form”); or they could come up with their own individual solution (“extraneous form”).

A model of options for the translation of verse (from Holmes 1970)

  1. Verse as prose: All foreign verse is rendered as prose, as has been the norm in translations into French.

  2. Mimetic form: The translator chooses a form in the target language that is as close as possible as the one used in the source language. For example, an English sonnet can be rendered as a Spanish sonnet well enough, even though the metrics of feet in English will not correspond to the syllabic metrics of Spanish. Often this involves introducing a new form into the target culture, as was done when English terza rima was modeled on the Italian verse form.

  3. Analogical form: The translator identifies the function of the form in the source-language tradition, then finds the corresponding form in the target-language tradition: “Since the Iliad and Gerusalemme liberata are epics, the argument of this school goes, an English translation should be in a verse form appropriate to the epic in English: blank verse or the heroic couplet” (Holmes 1970: 95). This option might be an application of the equivalence paradigm at a high textual level. It is to be distinguished from the blanket “verse to prose” option to the degree that it requires identification of the way the specific source-text form functions in the source culture.

  4. Organic or content-derivative form: The translator refuses to look solely at the form of the source text (as is done in the above options) and instead focuses on the content, “allowing it to take on its own unique poetic shape as the translation develops” (Holmes 1970: 96).

  5. Extraneous form: In some situations, the translator may adopt a form that is totally unconnected with the form or content of the source text, and that is not dictated by any blanket form for translations in the target culture. In other words, anything can happen.

Holmes sees these options as being appropriate to different historical situations. Mimetic form tends to come to the fore “in a period when genre concepts are weak, literary norms are being called into question, and the target culture as a whole stands open to outside impulses” (Holmes 1970: 98). This might be the case of German in the first half of the nineteenth century. On the other hand, “the analogical form is the choice to be expected in a period that is in-turned and exclusive” (Holmes 1970: 97), such as the neoclassical eighteenth century in France. As for the use of “organic” form, Holmes sees it as being “fundamentally pessimistic regarding the possibilities of cross-cultural transference” (1970: 98) and thus associates it with twentieth-century Modernism. “Extraneous” form is then regarded, not very convincingly, as having “a tenacious life as a kind of underground, minority form […] resorted to particularly by metapoets who lean in the direction of the imitation” (1970: 99).

Holmes’s analysis here suggests that translators’ decisions are always culture-bound, give or take a few unruly “metapoets.” When asked how any decision should be made, the descriptivist will thus always be able to say, “it depends on the (target) culture.” But then, how many different things can a decision actually depend on? Is there any way to model the huge range of variables covered by phrases such as “the translator’s sociocultural situation”? Descriptivists have made use of at least three concepts that are of some help here: systems, norms, and (for want of a better term) target-sidedness.

5.4.2 Systems of translations?

What Holmes does in his brief study is in a sense systematic: he identifies and classifies the available options, and he gives them a certain logical symmetry, largely thanks to some blunt distinctions between form, function and content. This is theory with a very top-down function: the theorist conceptualizes the alternatives, then goes looking for historical examples. One must be careful, though, about the status of this systematization. What Holmes does here is systematic (ordered, thorough, complete), but not necessarily systemic (in the sense that might be related to a system where all terms in some way depend on all other terms).

If we were talking about a language system (as in the work of the systemic functionalist Halliday, for example), we would see the speaker producing a string of words such that at each point there is a restricted set of what words can follow. The language system limits the choices that can be made. The same is true of the translator as a language producer, since the target language imposes limited sets of choices, which vary as we go about doing the translation. However, does the same kind of decision-making concern how to render a foreign verse form? The translator may certainly select one of Holmes’s five options, and that choice might have meaning in terms of the overall history of European verse forms, yet is it a decision like those where we are obliged to select a certain kind of verb or adverbial? Is it properly systemic? To a certain extent, yes: all receiving cultures have literary genres, and they mostly maintain structural relations between themselves. Then again, no: those sets of genres need bear no resemblance at all to the five translational alternatives outlined by Holmes. The receiving culture is one thing; the sets of theoretical alternatives are something quite different. In this case, the kind of choice process outlined by Holmes surely cannot be considered a psychological reality. If the translator was working into German at the beginning of the nineteenth century, there were all kinds of social and cultural factors that not only made the use of mimetic form appropriate, but also made Holmes’s alternatives relatively invisible. Germanic culture, without a state, was prepared to draw on other cultures in order to develop. Translations of Homer brought hexameters into German, and translations of Shakespeare brought in blank verse. Indeed, speaking in 1813, Schleiermacher saw this capacity to draw from other cultures as the key to foreignizing translations, regarded as being a particularly Germanic strategy. A literary translator trained in that cultural environment would then see “mimetic form” or “foreignizing” as the normal way to go about translation. The translator might even see it as the true or correct way in which all translations should be done, in all sociocultural environments. Prescriptive theorizing may result (“All translations should use mimetic form!”); some structural oppositions might be proclaimed in theory (“German mimetic form is better than French translations into prose!”); but the choices are not made within an abstract system comprising purely translational options.

As Toury would later clarify (1995a: 15-16), the system here belongs to the level of the theorist (the options theoretically available), which is to be distinguished from the alternatives actually available to the translator at the time of translating, which are in turn quite different from what the translator actually does. Toury thus distinguishes between three levels of analysis: “all that translation […] CAN involve,” “what it DOES involve, under various sets of circumstances,” and “what is it LIKELY to involve, under one or another array of specified conditions” (1995a: 15)

Three levels of analysis in Descriptive Translation Studies

Delabastita (2008: 234) elaborates on Toury’s three levels of analysis as follows, relating them to the notion of norms:

Level of system: theoretical possibilities (“can be”) For each translation problem or source text, it is possible to envisage a whole range of possible or theoretical solutions or target texts [as does Holmes].
Level of norms: culture-bound constraints (“should be”) On the intermediate level of norms, some of these possible relationships will be recommended or even required as being the only ones that can generate “genuine” translations, whereas others will be dismissed or even simply ignored.
Level of performance: empirical discursive practice (“is”) We can then observe which relationships have actually materialized in a given cultural setting. By definition, these empirical relationships constitute a subset of the possible relationships; their degree of frequency in a given cultural situation is a crucial indication that certain norms have been at work.

The top-down thinking is fairly clear here (even though, once again, one could presumably work upwards at the same time). Note, however, that the term “system” is used here only in the sense of “theoretical possibilities.” This is quite different from the kind of social or cultural system presented as the context in which translations function. The relative importance of this second, more general sense of “system” varies from theorist to theorist. Can the levels of “should be” and “is” be properly systemic in any strong sense?

When Holmes tries to explain why a particular translation option is associated with a particular period, he cites a range of quite profound phenomena: “genre concepts,” “literary norms,” “cultural openness/closure,” “pessimism/optimism about cross-cultural transfer,” and so on. This are all things placed in the target culture; they do not belong to any “system of translations” as such. Holmes mentions them in a fairly off-hand way; they seem to be quite separate, isolated phenomena. However, it is possible to see such things as being bound together to some extent, as different aspects of the one culture. This second vision requires us to see cultures as being systemic in themselves. In Holmes, those systems appear to hang together rather loosely; there is no necessary homogeneity or determinist fatality. In other theorists, particularly those more closely in touch with the legacy of Russian Formalism, cultural systems can impose quite strong logics. Lotman and Uspenski (1971: 82), for example, talk about entire cultures being “expression-oriented” or “content-oriented” (along with various more complex classifications), never doubting that such orientations characterize the entire cultural system. The stronger the logic by which the system is presumed to operate (i.e. the more systemic it is seen to be), the more that system can be seen as determining the nature of translations.

Here we return to the way Even-Zohar has worked with the idea of “polysystems.” The “poly-” part of the term may be seen as an indication that, unlike the approach of Lotman and Uspenski, there is a lot of flexibility involved. The internal logics of a culture are not going to determine everything that can be done within that culture. For Even-Zohar, translated literature can be seen as a kind of sub-system occupying a position within the literary polysystem that hosts it. The relations are nevertheless strong enough for certain general tendencies to be observed. The translations can become a key element in the literature (and thus “innovative” and “central” in position), or they may be secondary or unimportant (“conservative” and “peripheral”). In these terms, translation is seen as one of the ways in which one polysystem “interferes” with another, where the verb “to interfere” does not carry any pejorative sense (see Even-Zohar 1978 and subsequent papers on his website). Even-Zohar proposes, among much else, that translations play an innovative, central role when

(a) a polysystem has not yet been crystallized, that is to say, when a literature is “young,” in the process of being established; (b) when a literature is either “peripheral” within a large group of correlated literatures) or “weak,” or both; and (c) when there are turning points, crises, or literary vacuums in a literature. (1978: 47)

These three types of conditions are described as “basically manifestations of the same law” (1978: 47), the nature of which we will return in the next chapter.

Even-Zohar’s mode of thought, although expressed in a very lapidary way, goes well beyond Holmes’s concern with explaining why translations are the way they are. His conceptualization of systems as dynamic and pluralist allows Even-Zohar to ask what translations can actually do within their target cultures, and how they evolve from relations between cultures (particularly in terms of inferiority and prestige). He thus adds many elements to early insights such as Mukařovský’s awareness that literatures develop through translation. Even-Zohar’s general finding is in fact rather negative, since he concludes that “the ‘normal’ position assumed by translated literature tends to be the peripheral one” (1978: 50), that is, that translations tend to have a conservative, reinforcing effect rather than a revolutionary, innovative one. That kind of finding is unlikely to be popular within a discipline disposed to see translations as a hidden and maligned cause of change. Even-Zohar nevertheless stresses that translation is an essential element to the understanding of any cultural system, since no culture is an entirely independent entity.

The term “system” thus varies in meaning and importance from theorist to theorist. In each case, it pays to read the descriptions closely, paying particular attention to the verbs and the agents of the verbs (who is supposed to be doing what). In strong systems theory, you will find that the systems themselves do things, as if they were people. In other approaches, people are portrayed as doing things within systems of constraints. That is a big difference, bearing on fundamental issues such as human liberty, the determinist logics of history, and sometimes even the role and nature of translations.

While on the terminological difficulties, we should note a related problem with the term “function.” For descriptive studies, the “function” of a translation is generally correlated with its position within its corresponding system, in accordance with an extended spatial metaphor. When we say that, within a given cultural system, a translation is relatively “central” or “peripheral” (or things in between), we effectively mean that its function is either to change or to reinforce (or things in between) the receiving language, culture or literature. The function here is what the text does in the system. For the purpose paradigm, on the other hand, the “function” of a translation is generally conflated into the Skopos, the action that the translation is supposed to enable in a specific situation, just as the function of a source text is assumed to be the action in which the text is used (to teach, to express, to sell, etc.). Although both paradigms would claim to be “functionalist,” the term “function” means one thing in relation to systems theory (a position and role within a large-scale set of relations) and something else in relation to action theory (an action within a situation comprising various agents). There obviously must be common ground between the two usages, yet few theorists have actually sought it. Here is one way we might think about this relationship: On the surface, it would seem that the purpose of the translation, the Skopos, varies with each translation situation. All the situations are different, yet they always occur within wider social and cultural constraints that limit and orient them. One should thus be able to connect some wider systemic function to the smaller situational function.

Skopostheorie has remained relatively indifferent to top-down descriptivism, just as structuralist descriptive studies traditionally distanced themselves from the close-up dynamics of situations (Lefevere’s 1992 analysis of patronage being a significant exception). Both paradigms are strongly relativist; both refuse to see the source text as the only factor determining a translation. Yet they have long been looking in separate directions.

If there is a significant historical bridge between the two notions of function, it might lie behind the notion of norms.

5.4.3 Norms

In his three-level schema (the one we have reproduced above), after the level of what “can be” Toury opens a space for what “should be,” which he describes in terms of “norms.” Norms are thus positioned somewhere between abstract possibilities (such as Holmes’s alternatives) and what translators actually do (the kinds of pragmatics that Skopos theory deals with). For Toury, norms are

the translation of general values or ideas shared by a community […] into performance instructions appropriate for and applicable to particular situations, specifying what is prescribed and forbidden as well as what is tolerated and permitted in a certain behavioural dimension. (1995a: 55)

The term “performance instructions” here might suggest that a norm is the same thing as a client’s brief or a Skopos. It could also misleadingly be associated with a set of rules or official regulations (which would indeed be called normas in Spanish). In the descriptive paradigm, however, the term norm usually operates at a wider, more social level. For example, we could say that in the nineteenth century the norm for translating foreign verse into French was to render it into prose. There was no official rule stating that this had to be done, but there was an informal collective agreement. When translators approached the foreign text, they would accept as a matter of course that their work was not to imitate what the text looked or sounded like. When publishers hired translators, that is what they expected them to do. And when readers approached a literary translation, they would similarly accept that foreign poetry simply had to be in prose. Of course, the norm was not respected by all translators; norms are not laws that everyone has to follow. Norms are more like the common standard practice in terms of which all other types of practice are marked. That much is relatively unproblematic.

Why did the norm of “verse into prose” exist? On several different levels, it no doubt embodied the general idea that French culture was superior to other cultures. In Toury’s terms, it conveyed at least that much of the society’s “general values and ideas.” Given this assumed superiority, there was no reason to accept any foreign influence on the existing system of neo-classical literary genres. In Even-Zohar’s terms, we would say the perceived prestige of the target system allocated translation a peripheral role and hence a very conservative range of acceptable forms. Further, if we follow Toury, there would be some kind of social (though not juridical) penalization involved whenever a translator did not adhere to the norm. For instance, a text that differed radically from the established genres might be considered peculiar, ugly, or simply not worth buying. In every culture, the nature of a good translation is determined by such norms, since “bad translations” are penalized in some way, even if only by hurling adjectives like “bad.” Of course, in milieux governed by an avant-garde logic, the breaking of norms might mark a superior translation, rather than an inferior one. Norm-breaking might thus mark not only translations that are bad, but also those that are exceptionally good.

The concept of norms thus covers quite a few related but different things. Toury (1995a: 58) makes a basic distinction between “preliminary norms,” which concern the selection of the kind of text and the mode of translation (direct/indirect, etc.), and “operational norms,” which would cover all the decisions made in the act of translating. However, as our “verse into prose” example shows, norms also have different social and epistemological dimensions. They concern what translators think they are supposed to do, what clients think translators ought to do, what text-users think a translation should be like, and what kind of translations are considered reprehensible or especially laudable within the system. Chesterman (1993) organizes these various aspects by distinguishing between “professional norms,” which would cover everything related to the translation process, from “expectancy norms,” which are what people expect of the translation product. If translators in a given society usually add numerous explanatory footnotes, that might be a professional norm. If readers are frustrated when such notes do not appear, or if the notes are in an unusual place (perhaps at the beginning of the text rather than at the bottom of each page), then that frustration will be in relation to expectancy norms. Ideally, the different types of norms reinforce one another, so that translators tend to do what clients and readers expect of them. In times of cultural change, the various types of norms might nevertheless be thrown out of kilter, and considerable tension can result. Indeed, in systems of self-induced change, an extreme logic of the avant-garde may mean that all text producers, including translators, set about breaking norms, and text users thus expect norms to be broken. That is, norm-breaking can become the norm, as in extreme Modernism.

The idea of norms and norm-breaking has been important for the way descriptive research relates to the other paradigms of translation theory. If we apply the concept of norms seriously, we should probably give up the idea of defining once and for all what a good translation is supposed to be (although it is perhaps still possible to say what a good or bad social effect might look like, and thus evaluate the way norms work, cf. Pym 1998b). In fact, the very notion of what a translation is must become very relative. As we have said, this relativism would be a major point of compatibility with the Skopos paradigm (and indeed with the paradigm of uncertainty that we will meet in a later chapter). However, the same relativism runs counter to much of the linguistic work done in the equivalence paradigm. When a linguist analyzes a source text to see how it can or should be translated, the basic assumption is that the answers will come from the nature of that source text, and the nature of translation is thus a very clear thing; there is not much relativism involved. In the Skopos paradigm, the answers will come from the situation in which the translation is carried out, to the extent that it matters little whether a text is a translation or a liberal re-write. In the descriptive paradigm, however, any questions about the borders between translations and non-translations can be answered in terms of norms, which in turn express values from the wider system within which the translator is working. In this sense, the theory of norms positions translation somewhere between the relative certainty of equivalence and the relative indifference of Skopos theory.

Such comparisons of paradigms could be exploited in the 1980s, when the various approaches were starting to congeal into a tentative discipline called Translation Studies. Scholars working in the descriptive paradigm, usually with a background in literary studies, could legitimately criticize the narrow “prescriptive” work being done in the equivalence paradigm. How could a theory set out to tell someone how to translate, when the very notion of translation varied so much from epoch to epoch and from culture to culture? The call for descriptions was thus initially a more or less direct negation of the kind of prescription associated with the equivalence paradigm. Similarly, whereas the equivalence paradigm invited analysis to start from the source text and its role in the source situation, the descriptive paradigm tended to favor the target text and its position in the target system. Toury (1995a) explicitly recommends starting analysis from the translation rather than from the source text; he thus creates space for research that takes no account of the source text at all. For example, you can simply compare different translations, or compare translations with non-translations within the target system. That kind of full-frontal opposition helped to make Toury the enfant terrible of his day.

The notion of norms, however, allowed a kind of prescriptivism to be introduced into descriptive studies, almost through the back door. Even if the role of theory was not to tell translators how to translate, a descriptive approach could identify the norms by which a translation could be considered good by people in a certain place and time. This has allowed for a certain application of descriptive studies in the training of translators and interpreters. Toury (1992) has suggested, for example, that trainees be asked to render the same text according to different norms (e.g. translate as one might have done in twelfth-century Toledo, or under conditions of censorship). The trainee will thus be made aware that there are many different ways to translate, each with certain advantages and disadvantages. Of course, the same kind of exercise can be recommended within the purpose- based paradigm: translate the one text in different ways in order to achieve different purposes. The different paradigms can lead to the same kind of training activity.

Seeking an alternative mode of compatibility, Chesterman (1999) proposes that the study of norms will enable the teacher and learner to predict the relative market success of one strategy or another. No teacher can tell any student there is only one way to translate (since many norms are available), but empirical research can make it possible to predict success or failure when dominant norms are met or violated. Chesterman (1999) formulates his compatibilist position as follows:

Statements like “In principle, in authoritative and expressive texts [original metaphors] should be translated literally” (Newmark 1988: 112), or “translations should aim to have the same effect on their target readers as the source texts had on the source readers,” or “translators should translate transgressively, not fluently”) can be paraphrased approximately like this: “I predict that if translators do not translate in the way I prescribe, the effect will be that readers will not like their translations / that the publisher will reject the text / that intercultural relations will deteriorate” or the like.

In all these ways, the concept of norms has helped bridge some of the gaps between descriptivism and prescriptivism.

A more methodological problem concerns the way norms can be discovered. A bottom-up approach might gather together many translations, look for the shifts, and regard any high-frequency patterning of the shifts as a “norm.” That is a lot of work; it cannot say very much about why the norms are there; but it might be a valuable contribution. Alternatively, Toury (1995a) pays special attention to “pseudotranslations,” understood as texts that are presented as translations but are in fact original creations. In Hungarian, for example, science fantasy novels are usually presented as translations from American English, even though they are written straight in Hungarian, with invented authors, invented biographies, and all the trappings of a foreign product (Sohár 1999). Such pseudotranslations are found in a wide range of cultures, with numerous different functions (Santoyo 1984). Their particular interest for Toury, however, is that they can indicate what a target culture expects translations to be like, and often how that culture relates to other cultures in terms of prestige. This may provide a short-cut to the identification and possible explanation of norms.

A more top-down approach to the discovery of norms would start from peri-textual data such as reviews and critiques, which would tell us about the expectancy norms involved in the reception of a translation. More highly focused research can economize resources by concentrating on particular public debates about norms and norm-breaking, thus identifying and analyzing moments when norms are undergoing change (cf. Pym 1997). Such an approach helps connect descriptive theory with more dynamic (and perhaps less systemic) views of cultural history.

The concept of norms has thus helped bring several approaches closer together, at the same time as the empirical discovery of norms has undoubtedly increased our historical understanding of the way translations operate. The fundamental concept, however, is not as clear-cut as it may seem. Consider, for example, the way the German sociologist Niklas Luhmann (1985) describes legal norms as “counterfactual expectations,” in the sense that they do not take account of the way people actually behave. When these expectations are defeated (we find that there are criminals), the legal norms do not adapt accordingly (criminals must still be punished, no matter how many criminals there are). Many expectancy norms concerning translations could be of this counterfactual kind. For example, no matter how often we find that translations are domesticating (or foreignizing, or explanatory, or full of shifts, etc.), users of translations might still insist that they should not be. If some norms are working like this, then the bottom-up counting of facts and frequencies will never connect with the social pronouncements of what is acceptable or unacceptable. This is one very basic reason why a descriptive approach to norms requires theoretical concepts. And that is another reason why, in turn, the descriptive paradigm is full of theories.

Whenever theorists tell us about norms, we should ask exactly how they have discovered those norms. If bottom-up, the empirical patterns may not all have equal status as psychological or social facts. And if top-down, then we should ask where the theorist found the categories of analysis, and why.

5.4.4 “Assumed” translations

Here is another theoretical problem that cuts to the heart of empirical methodologies. If we set out to discover the historical and cultural diversity of translation norms, can we pretend to be sure from the outset what is meant by the term “translation”? If so, exactly what criteria should we use for collecting a set of things called “translations”? And if not, how can we possibly avoid imposing our own translation norms on other cultures and periods?

This is one of the classical theoretical aporias that tend to worry researchers in dominant Western cultures.

Toury’s solution to the problem has been to leave the defining to the people we study. For him, “a ‘translation’ will be taken to be any target-language utterance which is presented or regarded as such [i.e. as a ‘translation’], on whatever grounds” (Toury 1995a: 20). In other words, we wait to see what each culture and each period has to say about what is or is not a translation. The solution is the operative concept of “assumed translations,” which simply means that a translation is indeed a translation only for as long as someone assumes it to be one. A pseudotranslation, for example, might then be held to be a translation only for as long as the trick works, and it becomes a non-translation for those aware of the false pretence.

That solution remains fraught with logical difficulties. For example, if each language has different words for “translation,” how do we know those words are translations of each other? In order to select the words, we would surely need our own concept of translation, if not some clear ideas about what good and bad translations are. The debate over that issue has been one of the most fundamental but recondite activities in Translation Studies (cf. among others Gutt 1991; Toury 1995b; Hermans 1997, 1999; Halverson 2004, 2007; Pym 1998a, 2007a). For some, the problem is basically without solution, since if we use our normal terms to describe another culture’s term “we naturally translate that other term according to our concept of translation, and into our concept of translation; and in domesticating it, we inevitably reduce it” (Hermans 1997: 19). At the other extreme, we might argue that the empirical data are so diverse and unruly that we have to make some initial imposition and selection, simply in order to get research moving (cf. Pym 2007a; Poupaud et al. forthcoming). The best we can do is to be honest and self-critical about our initial principles and criteria, and open to the discovery of new concepts in the course of the research process. As different as these two options may appear, they both accept that concepts of translation are culturally and historically relative and can be described in explicit terms. They are thus both within the descriptive paradigm.

Interestingly, both approaches differ from Toury’s concept of assumed translations. The fundamental difference between the two sides has more to do with the role attributed to indeterminism, which we will discuss in a future chapter.

5.4.6 Target-side priority

As we have noted, Toury upset linguistics-based studies of translation not only by opposing prescriptivism, but more profoundly by insisting that translations should be studied in terms of their target contexts rather than in relation to their sources (see Toury 1995b: 136). This led to an extreme position: in Toury’s words, “translations should be regarded as facts of target cultures” (1995b: 139; cf. 1995a: 29). This proposition should be understood as part of a specific research methodology; it does not mean that translations somehow never have source texts (which would absurdly imply that all translations are actually pseudotranslations). Toury’s argument is that the factors needed to describe the specificity of how translations work can all be found within the target system. This is based on the assumption that translators “operate first and foremost in the interest of the culture into which they are translating” (1995a: 12), either in order to reinforce the norms of the target culture or to fill in perceived “gaps.”

Those general methodological precepts have born fruits. When one studies, for example, a corpus of English theater translated into Spanish (Merino 1994) or censored translations in Franco’s Spain (Merino and Rabadán 2002), even when the material is organized in terms of English-Spanish pairs, the shifts make sense in terms of the norms of the Spanish host system, especially when it comes to the Franco regime’s systemic censorship and its various historical avatars (for notes on the wider project on translation and censorship, see Merino Álvaraz 2005; for research projects associated with the earlier development of the descriptive paradigm, see Lambert 1988, 1995). In these and similar case studies, translations are indeed approached as facts of target cultures, and much quantitative data has been produced in those terms.

The principle of target-side priority has nevertheless been contested. The researchers working on literary translation at Göttingen in the 1990s generally preferred a “transfer” model, which explicitly traced movements between particular source and target cultures. Others have objected to the separation of the two cultures, arguing that translators tend to work in an “intercultural” space in the overlap of cultures (cf. Pym 1998a). More generally, as with the problem of defining translations, the binary opposition of source and target has been increasingly criticized from within the indeterminist paradigm, as we shall see later.

5.5 Translation Studies as an academic discipline

The descriptivist call to science is in many respects a structuralist aspiration, crafted in the belief that methodological research will reveal hidden relations. There is supposed to be a wider logic beneath observable facts. That call to science is sometimes taken further; “Sciences qua sciences,” says Toury, “are characterized by an incessant quest for laws” (1995a: 259, finding support in Even-Zohar 1986). The aim of Translation Studies is thus assumed to be to discover laws, and in the next chapter we will consider a few of the laws proposed so far. What interests us here is more the way this orientation has been able to shape a movement. On Toury’s view, Descriptive Translation Studies not only has a starting point (the methodological identification and analysis of facts) but also a general collective goal (the formulation of abstract laws based on numerous observed facts). This is a paradigm able to lead somewhere.

In its historical setting, the general belief in science and its goals allowed strangely little space for self-critical analysis of the scholarly community, or indeed of the social effects of the research itself. At the time the descriptivist paradigm was developing, such questions were of little concern. There was such confidence in the project, and presumably self-confidence in the researchers, that this became the first paradigm able to position itself in relation to other paradigms. Indeed, it was from this positioning that the discipline of Translation Studies itself was envisaged as a coordinated collective undertaking. This can be seen in Figure 5, which shows Holmes’s original proposal for Translation Studies (although the diagram was actually drawn by Toury):

Figure 5 Holmes’s conception of Translation Studies (from Toury 1991: 181; 1995: 10) 

We reproduce the diagram here in order to note three things. First, the initial division of “Pure” vs. “Applied” means that the place of the equivalence and purpose-oriented paradigms would lie quite far from descriptive work: they are presumably somewhere near the “applied” side of business, while the “Descriptive” branch is “pure” enough to form a pair with “Theory” all by itself. If the quest for laws is seen as the prime purpose of the discipline (rather than the improvement of translations or of translators, for example), then the discipline becomes purer as its categories become more abstract. In fact, the diagram justifies the very reasons why translators and trainers tend not to like translation theory. Second, we are very hard-pushed to find published work for many of the slots allowed for here. Even within the Descriptive branch, for example, we have remarkably little that could be called “function oriented,” presumably dealing with what translations actually do within cultures and societies, or with how translations are actually received. As for the series of “Partial Theoretical” compartments, are there any studies that slide in neatly? And third, there is no real place for people in the schema, neither for translators nor for researchers or theorists. The descriptive paradigm thus seems fundamentally ill-equipped to reflect on its own epistemological shortcomings.

Not surprisingly, the descriptive paradigm has not been able to impose its disciplinary map on all other paradigms. As an academic discipline, Translation Studies was indeed born from within this paradigm, but the space thus created was soon described as an “interdiscipline” (after Toury and Lambert 1989: 1), as a place where many other models and methodologies can be drawn on. The proponents of description were not entirely closed to the rest of the world.

5.8 Frequently had arguments

We will now try to gather together a few general observations concerning the historical development of the descriptive paradigm.

The following are points that would generally be considered positive aspects of the paradigm:

  1. The historical variety and vitality of translation has been revealed.

  2. The paradigm has played a central role in the legitimization of Translation Studies as an academic discipline.

  3. It has created knowledge that is potentially useful for all aspects of Translation Studies, including the prescriptive approaches it originally opposed.

  4. It breaks with many of the prescriptive opinions of the equivalence paradigm, albeit at the expense of creating its own illusions of objectivity.

The counterweight to these positive points must be a series of arguments about the apparent failings of the paradigm:

  1. The descriptivist enterprise is ultimately positivist, without awareness of its own historical position and role. It suffers the same drawbacks as the rest of structuralism.

  2. The definition of “assumed translation” is circular, and must at some stage rely on the theorist’s own criteria.

  3. Descriptions do not help us in the training situation, where we ultimately need prescriptions.

  4. The models all concerns texts and systems, but not people (see the Holmes map, where there is no room for studies of translators).

  5. The target-side focus cannot explain all relations (particularly the case of translation in postcolonial frames, or wherever power asymmetries are so great that the source side simply cannot be hidden from view).

  6. The focus on norms promotes conservative positions, allowing “ought” to be derived from “is.” This blocks off work on critical ethics.

Various scholars have responded to these points. Toury (1992), for example, points out the usefulness of descriptions in the training situation, since one can always present alternatives in order to illustrate that “everything has its price.” We have noted above how Chesterman (1999) also argues that empirical research should reinforce training, since it can be used to predict the success or failure of certain strategies. As for the apparent promotion of conservatism, Toury proposes that we train students how to break norms, as indeed he himself has done within Translation Studies.

With respect to the supposed lack of a human dimension, Toury’s abstract concept of norms is offset by serious interest in how translators become translators (1995: 241-258), and recent moves within the descriptivist project have been toward the incorporation of sociological models, particularly Bourdieu’s concept of “habitus” (variously after Hermans 1999 and Simeoni 1998). This would meet up with moves to write the history of translation as a history of translators (cf. Delisle and Woodsworth 1995, Pym 1998a). It also connects with the many translation scholars who have been engaged in writing literary history, often in a humanist mode where translators play leading roles.

Those arguments notwithstanding, there has been considerable resistance to descriptivism within training institutions, which have generally obtained better mileage from the Germanic theories of purpose. At the same time, the basic thrust of target-side studies threatens to relieve traditional departments of modern languages of what they do best (teaching source languages and literatures), and is thus unlikely to curry favor there. Descriptivist theory has thus tended to operate on the fringes of the more established training communities, guiding PhD theses useful for employment purposes.

So where will the descriptive paradigm go from here? Recent calls have been for a “sociological turn,” for some kind of alliance with a discipline better equipped to handle contextual variables. Theo Hermans (1999), for example, closes his account of the paradigm by pointing the way to the sociologies of Bourdieu and Luhmann. And so one turns that corner; but what do we find? Usually a plethora of data, on numerous levels, with very few categories able to organize the data in terms of cross-cultural communication. The great Modernist sociologies are based on the same structuralism that informed the history of the descriptive paradigm itself, albeit now with more scope for self-reflexivity (the sociologist can do the sociology of sociologists). More problematic, these sociologies are overwhelmingly of single societies only, of systems in the “one side or the other” sense that has reigned within the paradigm. They fit in so well with the target-side orientation of descriptive approaches that they risk bringing in little that is new. Indeed, the descriptive literary studies of the 1970s and 1980s were already doing systematic sociology of a kind. A new “sociological turn” could risk bringing us back full-circle.

Summary

This chapter has sketched out the historical and intellectual background of the descriptive paradigm of translation theory. The paradigm was mainly developed by literary scholars working in relation to smaller cultures. Although based on empirical research, it has a set of properly theoretical concepts, many of which can be traced from the Russian Formalists to work done in central Europe, to the Tel Aviv School, and to scholars in Holland and Flanders. The academic discipline of Translation Studies began to take shape from the exchanges between those groups. In general, the descriptive theories oppose the equivalence paradigm in that they aim to be non-prescriptive, their prime focus is on “shifts” rather than types of equivalence, and they do not undertake extensive analysis of the source culture. They tend to be like purposebased Skopos approaches in that they emphasize the target-culture context and the function of translations within that context. They nevertheless differ from purpose-based approaches in that they see functions in terms of the positions occupied by translations within the target systems, rather than with respect to a client or a brief. Descriptive theories also tend to concern what translations are usually like in a particular context, rather than he ways in which particular translations might differ. They are thus able to talk about the consensus-based “norms” that inform the way translations are produced and received. The paradigm is thus relativistic in that it is very aware that what is considered a good translation in one historical context may not be rated so highly in a different context. The research based on those concepts has done much to reveal the vast diversity of translation practices in different historical periods, different cultures, and different types of communication.

Sources and further reading

A good historical account of systems-based theories can be found in the first chapters of Theo Hermans’ Translation in Systems (1999). The proceedings of the various conferences in Bratislava in 1968 (ed. Holmes, de Haan, Popovič, 1970), Leuven in 1976 (ed. Holmes, Lambert, van den Broeck, 1978), Tel Aviv in 1978 (ed. Even-Zohar and Toury 1981) are full of ad hoc insight into the disjointed development of the paradigm, although the books are hard to find. The same could be said of the seminal collection The Manipulation of Literature (ed. Hermans 1985), which is rather more profound than its misleading title. Anyone undertaking empirical research on translations should have tackled Gideon Toury’s Descriptive Translation Studies and beyond (1995), if only to use it as a point of reference. Numerous papers on various aspects of methodology are available online at the sites of Itamar Even-Zohar (http://www.tau.ac.il/~itamarez/) and Gideon Toury (http://www.tau.ac.il/~toury/). A more entertaining descriptive approach to literary translation is André Lefevere’s Translation, Rewriting, and the Manipulation of Literary Fame (1992). For insights on the various sociocultural aspects of descriptive studies, see the selection of José Lambert’s articles in Functional Approaches to Culture and Translation (ed. Delabastita, D’hulst & Meylaerts, 2006). For a critical account of systems and norms, see Pym (1998a). A broad update on recent work in the descriptive paradigm can be gleaned from the volume Beyond Descriptive Translation Studies (ed. Pym, Shlesinger, Simeoni, 2008).

Suggested projects and activities

  • 1. Consider the all the language situations you participate in on a typical day, not only in newspapers, television and web sites but also in shops, banks and public services. How much of this linguistic material must have been translated in one way or another? (Consider news events that have happened outside of your language.) How much of that material is actually marked as translational?

  • 2. Where do translators and interpreters work in your town or city? What laws or policies orient their work?

  • 3. Look up translations in your language of John 1, similar to these (taken from Nord 2001):

a) In the beginning was the Word, and the Word was with God, and the Word was God.

b) Au commencement était le Logos; et le Logos était près de Dieu, et le Logos était dieu.

c) En el principio existía el Verbo, y el Verbo estaba con Dios, y el Verbo era Dios.

d) Al principio era el Verbo, y el Verbo estaba en Dios, y el Verbo era Dios.

e) No principio era o Verbo, e o Verbo estaba com Deus, e o Verbo era Deus.

f) In principio era il Verbo, e il Verbo era presso Dio e il Verbo era Dio.

g) Im Anfang war das Wort, und das Wort war bei Gott, und Gott war das Wort.

h) Zuerst war das Wort da, Gott nahe und von Gottes Art.

Which translations make sense, and which do not? Could these differences be described in terms of norms?

The last-listed German translation (h) is from Berger and Nord (1999). It could be translated into English as something like “First the Word was there, near God and in the manner of God.” This radically changes the widely memorized phrases of the Lutheran version (g), which might be rendered as “In the beginning was the Word, and the Word was with God, and God was the Word.” What might be the reasons for such a change? Could those reasons be described in terms of norms?

  • 4. Use the Internet to find out about the Mexican interpreter La Malinche (also called Malineli Tenepatl or Doña Marina). What systems would she have been operating within? What was her relation with the systems? What norms would have regulated her work? Are these systems and norms different depending on whether her story is told by feminists, or by Mexican nationalists? (The same exercise can be done for any number of high-profile translators, preferably working in situations of conflict.)

  • 5. Find a code of ethics for translators. Could any of the principles be described as norms? If so, what kind of norms are they? How would they relate to an empirical study of what translators actually do? (For a critical analysis of codes of ethics, see Pym 1992a, Chesterman 1997.)

  • 6. Find an authoritative history of your favorite national literature (e.g. French literature, Russian literature). Are translations part of the history? Are they mentioned in a separate chapter? In the index? Should they be? Would the inclusion of translations make any sense in the case of minor literatures in major languages (e.g. Australian literature)? Can periods of great change, such as the Italian Renaissance, really be written without reference to translations?

  • 7. Locate one page of a literary text and a professional translation of it. Try to divide it into paired segments (one ST unit corresponds to one TT unit) and identify the translation shifts. Are the shifts easily categorized? Can they all be described in terms of equivalence? For how many of the shifts could we say there are social or political factors involved? Should we talk about “shifts” or “variations,” or perhaps “deviations,” or even “errors”?

  • 8. Find out about The Works of Ossian (1765). Could this text be described as a translation? If not, what is it? Should it be analyzed within the field of Translation Studies?

  • 9. Check the definition of pseudotranslations. Can you find any pseudotranslations in the literatures of your languages? What would their cultural function be? Why have they been presented as translations?

1A strange piece of Translation Studies folklore maintains that the middle S in the name “James S Holmes” stands for his mother’s name; it should apparently be written with no stop after it. The writings of Gideon Toury consistently omit the stop.

Creative Commons License Este é um artigo publicado em acesso aberto (Open Access) sob a licença Creative Commons Attribution, que permite uso, distribuição e reprodução em qualquer meio, sem restrições desde que o trabalho original seja corretamente citado.