The Routledge Encyclopeida of Translation Studies

Language function, text function, communicative function There have been many attempts to classify the functions of language. Among the most influential are those of Bhler (1934), Jakobson (1960) and Halliday (1973). Bhlers Darstellungsfunktion, Ausdruckfunktion and Appelfunktion refer, respectively, to the representation of objects and phenomena, the attitude of the text producer towards such phenomena and the appeal to the text receiver. These three functions correspond broadly to Jakobsons referential, expressive and conative functions, although the latter additionally distinguishes phatic (the use of language to create and maintain social contact), metalingual and poetic functions. Halliday distinguishes three macrofunctions: the ideational (representation of experience), the interpersonal (the speakers expression of attitude) and the textual (the internal organization of language, or the way links are established within the text and between the text and its context of situation). There is, then, a degree of consensus among these alternative formulations. Rei (1971, 1976, 2000) developed a translationoriented text typology with the aim of deriving strictly objective criteria for assessing the QUALITY of translations. Based on Bhlers three functions of language, Rei identified three corresponding text types (informative, expressive, appellative) which she linked to translation methods. In the translation of informative texts (examples of which would be reports and textbooks), the aim is invariance of content and the translation is deemed successful if the information has been transmitted in full. In the case of expressive texts (e.g. literary texts), the aim is the communication of

artistically organized content and the translation method involves identifying the artistic and creative intention of the ST author and conveying it in an analogously artistic organization. The translation of appellative or operative text types (e.g. ADVERTISING) aims to provoke in the target readers identical behavioural reactions to those of the reader of the source text, and the translation method called for is ADAPTATION. Reis approach is sourcetext based, i.e. she judges translation quality with reference to the source text (type). Her translationoriented text typology is thus, strictly speaking, not a functionalist theory of translation in the more specific sense in which this label is now used in translation studies. There can be no doubt that language functions and communicative functions impinge significantly on the translators task. However, no actual text will exhibit only one language function, and many texts cannot be assigned to one specific text type only. Hatim and Mason (1990a), who add pragmatic and semiotic dimensions to their characterization of the communicative domain of context, argue that all texts are multifunctional, even if one overall rhetorical purpose will generally tend to predominate and function as the ultimate determinant of text structure. Semiotics and translation Several authors have called for drawing on semiotics to enrich translation theory. Major contributions in the structuralist tradition include Toury (1986), who made a compelling case for the semiotic nature of translating and attempted a definition of it as an act (or a process) which is performed (or occurs) over and across systemic borders (ibid.: 1112). The issue of the boundaries

between semiotic systems is also taken up in the work of Torop (e.g. 2000). With reference to Lotmans concept of the semiosphere (2005), Torop described the boundary not as a limiting factor but as a mechanism that translates external messages into the internal language of the semiosphere, discriminates ones own from the alien, [and] turns external nonmessages into messages (ibid.: 605). Chesterman (2002a), in contrast, used insights from another semiotic tradition Greimass generative semiotics to investigate translation causality. In the interpretive semiotics camp, a notable early voice was Roman Jakobson, who identified the nexus between Peirces theory of signs and the theory of translation in what is probably the single most quoted essay in the field (Jakobson 1959). Jakobson, who regarded Peirce as the deepest inquirer into the essence of signs (ibid.: 233), based his call for a semiotic understanding of translation on Peirces insight that the meaning of any linguistic sign is its translation into some further, alternative sign (ibid .: 232). In other words, translation was identified as a crucial element of all meaningmaking and of ordinary language use. This allowed Jakobson to extend the scope of translation beyond interlinguistic translation or translation proper (cheese Russian syr/tvorog), to include intralinguistic translation (bachelor unmarried man) and intersemiotic translation (sunrise the picture of a rotating planet). Two authors began to engage more extensively with interpretive semiotics towards the end of the 1980s: DeledalleRhodes (19889, 1996) and Gorle (1989, 1994). More recent contributions include Nergaard and Franci (1999), Eco (2001), Cosculluela (2003), Petrilli (2003) and Stecconi (2004a). These

authors have taken on the task of redefining the traditional categories of translation studies in the belief that interpretive semiotics represents the future of translation theory. For instance, Gorle proposed to change the core image of translation in the West from transfer to growth: Indeed, the image of translation that emerges from a Peircean semiotics is one of change and growth, of expansion through transformation (Gorle 1994:231). Finally, the term semiotics is also used as shorthand for research that goes beyond verbal language (Poyatos 1997), for instance, in studies that explore the translation of advertising (Adab and Valds 2004) and multimedia/multimodal material (Gottlieb 2005), without necessarily engaging with either structural or interpretive semiotics as described above. Translatability The universalist view considers the differences between languages to be surface phenomena only. They can cause practical problems for translation, but in principle translatability is guaranteed by biological factors and cultural considerations. All human brains are wired in the same way, hence there is a common human rationality. Moreover, we all inhabit the same physical world, hence there is a common core of human experience. Different languages may package meaning differently, but ultimately all languages are able to convey all possible meanings. In Roman Jakobsons words, All cognitive experience and its classification is conveyable in any existing language and [l]anguages differ essentially in what they must convey and not in what they may convey (1959:234, 236 emphasis in original). In the univerPage 301 salist perspective, language is typically seen as comprising two layers, a surface

and a deep structure. Ideas and meaning are generated at the deeper layer and can be represented by a variety of surface linguistic structures. This view was held in the medieval period by Roger Bacon and dominated Early Modern and Enlightenment thinking it is echoed in Noam Chomskys transformational grammar of the 1960s. The idea of language as twolayered promotes a dissociation between form and meaning, or, in Saussurean terms, signifier and signified (see SEMIOTICS). Form is material and perceptible, and varies from language to language, while meaning is invisible and can be extrapolated from the form that carries it. This is what Reddy (1979/1993) described as the conduit metaphor of language. It holds that meaning is transmitted, and can be preserved intact, as it travels along its conduit. The conduit metaphor also guarantees translatability, as translation transfers meanings by substituting one carrier for another. Translation thus constantly practises the separation of signifier and signified, as Derrida has pointed out (see DECONSTRUCTION). It makes us assume that different signifiers somehow convey a signified that remains identical to itself. This points up an aporia in Saussures concept of language, in which signifier and signified are like two sides of a piece of paper and hence inseparable. Translation, however, is predicated on the separability of signifier and signified (Derrida 1972a/1981). In the West, translatability was taken for granted from Roman antiquity onwards, the Wests first copying culture, as Kelly (1979) called it. After the Roman empire, translatio studii accompanied the westward translatio imperii, again providing little ground to doubt translatability, whether linguistic or more broadly intellectual (Stahuljak 2004). The Christian Bible was overwhelmingly read and subsequently exported in translation, giving rise to the

idea that its truth could be expressed in any language and therefore existed independently of language.

Poetry

The central question that all studies of the translation of poetry have asked, implicitly or explicitly, is whether poetry can be translated (see TRANSLATABILITY). It may seem obvious that it can, for poetry has always been widely translated, and some poets, such as Catullus or Rilke, have been translated many times. In fact, translated poetry plays such a large part in the literature of most cultures that it is taken very much for granted (Honig 1985:1). English readers of Virgil or Omar Khayym or Alvarezs (1992) anthology Modern European Poetry, for example, might see the poems as foreign without necessarily reading them as translations. This could be taken as evidence that they have been successfully translated, if translation is viewed as a type of writing which avoids drawing attention to itself. The opposite view that poetry translation is difficult or even impossible arises from the coincidence of two assumptions: (i) translated poetry should be poetry in its own right (see, for example, Coleridge 1990:200) (ii) poetry is difficult, cryptic, ambiguous and exhibits a special relationship between form and meaning (Furniss and Bath 1996:13). These two assumptions together have led many writers such as Weissbort (1989: x) and Raffel (1988: vii) to suggest that the translation of poetry, more than that of any other genre, demands both special critical abilities and special writing abilities. One way of negotiating this difficulty is to translate poetry into prose, an approach sometimes favoured (see, for example, Arnold 1954:316 Selver 1966:13ff. Weissbort 1989: xii) for

writers such as SHAKESPEARE. This might be because prose is seen as easier to write, although Scott (2000:163) argues that prose translations of poetry have their own resourcefulness and their own freedoms. Prose translations are, however, the exception. Another way of dealing with the supposed difficulty of poetic translation is to move away from the original, producing what Lowell called Imitations (1958) or what Paterson calls versions (2006:73ff.). Hamburger (1989:51) sees such deviation from the original as an admission of defeat yet many translators of poetry feel it is the only way to produce translated texts which aim to be poems in their own right (Paterson 2006:73). The skopos of poetic translation One way of expressing the fact that translated poetry aims, in general, to be itself poetry, is to say that the aim or skopos (Nord 1997:27) of its translation is to carry over the source text function into the target text it is thus an instrumental translation (see FUNCTIONALIST APPROACHES). However, if it is to avoid being what Hamburger saw as merely a springboard for ones own work, then it must aspire also to be documentary, to give some idea of what the original is actually like (Honig 1985:177, 179), and especially to allow its readers to see those very difficulties which make it poetic. The common tendency to publish translated poetry bilingually, especially in recent years, points to this documentary aspect. Especially for the bilingual reader, the relationship of the translated poems with the source text is highlighted by a similar layout in both languages. Thus recent books such as the Welsh anthology by Minhinnick (2003), Crucefixs version of Rilkes Duino Elegies (2006) or if it is to avoid being what Hamburger saw as merely a springboard for ones own work, then it must aspire also to be documentary, to give some idea of

what the original is actually like (Honig 1985:177, 179), and especially to allow its readers to see those very difficulties which make it poetic. The common tendency to publish translated poetry bilingually, especially in recent years, points to this documentary aspect. Especially for the bilingual reader, the relationship of the translated poems with the source text is highlighted by a similar layout in both languages. Thus recent books such as the Welsh anthology by Minhinnick (2003), Crucefixs version of Rilkes Duino Elegies (2006) or Gardners translations of Dutch poet Remco Campert (2007) suggest that successful translation of poetry does not depend upon the readers belief that the translated poem is an original. Yet translators like Minhinnick point out that they Page 195 attempt to restyle (2003: x) the poems where necessary. The notion that translation means in essence documentary writing, and therefore we need a new term (version or imitation) to describe translation of poetry which is also instrumental, was behind Jakobsons suggestion that what poetry required was not translation but creative transposition (1959/2000:118). Other writers do not see the need for instrumentality in translated poetry as running counter to the idea of translation. Gutt, for example, argues that poetic texts demand direct translation (1991/2000:167): they must preserve the stylistic qualities of the original. The focus on poetic style as a way of combining documentation of the poetics of the source text with the necessary instrumentality of the target text (even if not put in the same terms) is shared by a number of theorists of poetic translation (e.g. Tabakowska 1993 de Beaugrande 1978 BoaseBeier 2006a) who argue that the translation of poetry

must take into account the special nature and language of poetry and the type of reading it demands. Translation and the nature of poetry The idea that there is something peculiar to poetry which, if captured in translation, will allow the poetic effects (Gutt 1991/2000:164) of the original to be recreated is implicit in descriptions of poetic translation as writing which captures what Pope called the spirit (Lefevere 1992b: 64f.) or Rowan Williams the energy (2002:8) of the original poem. One way of making this abstract notion more concrete is to equate it with style, because style can be seen as the result of the poets choices (Verdonk 2002:9), and therefore the embodiment of poetic voice (Stockwell 2002b: 78) or mind (BoaseBeier 2003a), as well as that which engages the reader (BoaseBeier 2006a: 31ff.). This focus on style as central to poetic translation is found especially in the writings of: (i) translators who are themselves poets and can be assumed to have an inherent (perhaps unconscious) knowledge of how poetry works (e.g. Pope, Paterson or Williams), and (ii) critics who take the view that a theoretical understanding of poetry is essential not only to the reading of translated poetry but also to the act of translation (e.g. Tabakowska 1993 BoaseBeier 2006a). There have been many debates about the characteristics of poetic style and whether they distinguish poetry from prose or indeed literary from nonliterary texts (e.g. Fowler 1981:162ff. see LITERARY TRANSLATION). Some of the elements that have been put forward as distinctive of poetic style are: its physical shape (Furniss and Bath 1996:13), including use of lines and spaces on a page its use of inventive language (Eagleton 2007:46) and, in particular, patterns of sound and structure (Jakobson 1960:358) its openness to different interpretations (Furniss and Bath 1996:225)

 its demand to be read nonpragmatically (Eagleton 2007:38) The layout in lines can be seen as a signal to read the text in a particular way: as a text in which style is the main repository of meaning (BoaseBeier 2006a: 112). Typically, writers will speak of recreating particular aspects of style such its openness to different interpretations (Furniss and Bath 1996:225) its demand to be read nonpragmatically (Eagleton 2007:38) The layout in lines can be seen as a signal to read the text in a particular way: as a text in which style is the main repository of meaning (BoaseBeier 2006a: 112). Typically, writers will speak of recreating particular aspects of style such as metaphors (Newmark 1988/1995:10413), repetition (BoaseBeier 2003b) and ambiguity (BoaseBeier 2004) all these are stylistic resources which, though present in nonpoetic language, are used in greater concentration in poems and add up to Eagletons sense of inventiveness. Ambiguity, in particular, is a stylistic device which allows for different interpretations and thus its preservation in translation enables the poem to retain its ability to fit different contexts (Verdonk 2002:6f.). Discussions on the nature of poetry suggest that there might be poetic characteristics that are universal yet poetic traditions vary from one culture to another and, as Connolly (1998:174) points out, this is also an important consideration in translating poetry.

Untranslatability The daytoday

practice of translators appears to show overwhelmingly that translation is possible. If it happens all the time, surely it can be done? The argument against translatability does not usually posit absolute untranslatability but rather questions whether fully adequate translation can be achieved. The monadist case may be summed up as follows. In their different grammatical and lexical structures, individual languages embody and therefore impose different conceptualizations of the world. The structural asymmetries between languages prevent conceptual mapping from one language to another due to the lack of analogues and the absence of a languageindependent mapping tool. The way different languages divide up the colour spectrum or organize kinship terms are among the classic examples of such asymmetries, but they affect all aspects of language. The French linguist Emile Benveniste (1958) even argued that the supposedly universal logical categories of the ancient Greeks were based on features of their language. The consequence, subsequently explored by ethnographers and philosophers, is that different languages may give rise to incommensurable logics (Winch 1964). Languages, that is, are embedded in the cultural environment of which they are a constitutive part. This reciprocity between language and CULTURE and the asymmetries between different lifeworlds, which are also languageworlds, make translation impossible (see language. The French linguist Emile Benveniste (1958) even argued that the supposedly universal logical categories of the ancient Greeks were based on features of their language. The consequence, subsequently explored by ethnographers and philosophers, is that different languages may give rise to

incommensurable logics (Winch 1964). Languages, that is, are embedded in the cultural environment of which they are a constitutive part. This reciprocity between language and CULTURE and the asymmetries between different lifeworlds, which are also languageworlds, make translation impossible (see also CULTURAL TRANSLATION). The monadist view was articulated by the German Romantics, notably Johann Gottfried Herder, Wilhelm von Humboldt and Friedrich Schleiermacher, and taken up in the twentieth century by Edward Sapir and Benjamin Lee Whorf (hence the SapirWhorf hypothesis). For Herder, all crosscultural comparison was deeply problematic because each culture, and its language, had to be assessed on its own terms. Von Humboldt paradoxically asserted the impossibility of translation in the preface to a translation (the Agamemnon of Aeschylus, 1816) and presented untranslatability as a challenge to be taken up. In his famous 1813 lecture, Schleiermacher too asked whether translation was not a foolish undertaking, and went on to outline it as a task as unending as Page 302 that of hermeneutic understanding (Lefevere 1977). In doubting the possibility of translation, the monadist view clearly invokes the quantitative measure mentioned above. It does not hold that we cannot learn another language or explicate in one language concepts proper to another. It claims that, due to the asymmetries between languages and cultures and the organic link between language and culture, translation understood as a linear discourse replicating another discourse with regard to both length and meaning is not possible. Approximate renditions can be achieved, or explanatory paraphrase texts may also be translatable up to a point or in certain limited

respects. Untranslatability, then, mostly appears in relative form, as a matter of aspect, kind or degree. There always remains an untranslatable rest, for instance in the shape of connotation, nuance or poetic quality. Among the least translatable texts would be those that consciously exploit the idiomatic resources of a given tongue, or those that are encoded in multiple ways. In POETRY, for example, words may be woven into semantic, metrical, rhyming, intertextual and other patterns. This led Jakobson (1959) to claim that poetry is untranslatable and only creative transposition is possible leaving wide open the question of exactly how creative transposition differs from translation. Since the case for untranslatability bears both on linguistic structure and on the relation between language and culture, it is often subdivided into two kinds, linguistic and cultural. For J. C. Catford (1965), linguistic untranslatability occurs in cases where ambiguity or polysemy is functionally relevant in a text, cultural untranslatability when situational features that are referred to in an original (for example, sauna, igloo) are absent in the culture of the translating language. Catford wondered whether cultural untranslatability should not be treated as simply a type of linguistic untranslatability: for any item unknown in the receptor language a loanword could be imported or an explanatory phrase devised, even if that would result in an unusual, linguistically marked collocation. If a prohibition against translating certain texts or kinds of text is regarded as instancing untranslatability, then perhaps institutional untranslatability is a species of cultural untranslatability that is not reducible to linguistics. In this sense, the QURN is institutionally untranslatable to the extent that the Islamic world will

not recognize a version in another language as having religious authority. Adolf Hitlers Mein Kampf has become untranslatable because the current copyright holder, the German State of Bavaria, routinely refuses permission (Hermans 2007). In this case, the prohibition can only affect publication under the terms of international copyright law. Conversely, a language may be deemed incapable of accommodating certain concepts or be regarded as an inappropriate host for them due to its ties with an alien culture. Christian missionaries in the colonial Philippines or Spanish America reckoned some of their key doctrinal terms could or should not be rendered into native languages but were to be used by the natives in Latin or Spanish to avoid contamination by pagan beliefs (Rafael 1993). The colonial context also provides examples of how cultural incommensurability, however radical in theory, may be overridden in practice to enable translation nonetheless. Colonial settlement meant that European ideas of property were imposed on native populations so as to allow the colonizers to claim territorial could or should not be rendered into native languages but were to be used by the natives in Latin or Spanish to avoid contamination by pagan beliefs (Rafael 1993). The colonial context also provides examples of how cultural incommensurability, however radical in theory, may be overridden in practice to enable translation nonetheless. Colonial settlement meant that European ideas of property were imposed on native populations so as to allow the colonizers to claim territorial sovereignty and transfer ownership on their terms, regardless of the way the natives related to the land on which they lived (Cheyfitz 1991 Patton 2000).

Using the somewhat unfortunate example of an anglophone linguist encountering a native speaking a jungle language, Quine (1959) explored the possibility of translation in situations of radical linguistic and cultural difference from a philosophical angle, suggesting that the construction of meaning across such divides remained hostage to an ineradicable indeterminacy.