Introduction
Linguistics has had something of a chequered history in translation studies. In the mid-twentieth century, linguistic analysis occupied what Lawrence Venuti (2000:69) has called the “optimistic extreme” in translation theory. In the face of philosophical scepticism, linguistically-oriented commentators cheerfully proclaimed that translation was possible, and outlined whole batteries of procedures used by translators to overcome target language lexical gaps, or problems caused by structural differences between languages. Several books were published on the subject, including Vinay and Darbelnet’s (1958) Stylistique comparée du français et de l’anglais, Mounin’s (1963) Les problèmes théoriques de la traduction, Nida’s (1964) Towards a Science of Translating, Catford’s (1965) A Linguistic Theory of Translation, and Wilss’(1977) Übersetzungswissenschaft. Probleme und Methoden. Translation theorists like Nida and Wilss were quick to adopt the “truly scientific perspective” that Chomsky had purportedly gifted to linguistics (Newmeyer 1980:20), and both included whole chapters on machine translation, a burgeoning field where the usefulness of formal descriptions of contrastive linguistic structures seemed incontrovertible. The confidence was palpable. Linguistically-oriented translation theory had set itself up on a scientific basis and was convinced that its results could be applied in the practices of human and machine translation. It was ready for a fall.
By the late 1980s, proponents of linguistic approaches to translation were coming under fire from several quarters (see, especially, Godard 1990; Hermans 1985a; Lefevere and Bassnett 1990; Snell-Hornby 1988a, 1990). They were criticized for their scientistic posturing and derided for restricting their analyses to lower linguistic ranks. They were described as doggedly clinging on to a naïvely representational view of language, one that could sustain the ‘illusion of equivalence’ between languages, but that necessarily effaced the role of the translator in creating meaning in the target language. They were caricatured as myopic drones, conducting painstaking analyses of source and target texts, but indifferent to any wider cultural import that translation might have. Linguistics was further considered inadequate to the task of dealing with “the manifold complexities of literary texts” (Hermans ibid:10). And, as if things weren’t bad enough, the “triumph of machine translation” (Lefevere and Bassnett ibid:4) had turned out not to be just around the corner after all.
The onslaught against linguistically-oriented approaches to translation continued well into the 1990s, gathering momentum with the continuing rise of cultural studies in translation. Rosemary Arrojo (1998), Edwin Gentzler (1993) and Lawrence Venuti (1996, 1997) in particular used the philosophical apparatus of postmodernism to critique linguistically-oriented translation studies, questioning linguists’ pretensions to objective neutrality, their promotion of scientific models, and their view of language itself.
But if linguistics was coming under fire in translation studies in the 1980s and 1990s, these decades are also remarkable for the volume of work that continued to be published by linguistically-oriented translation scholars. Baker (in press) lists Delisle (1980, 1993), House (1981, 1997), Blum-Kulka (1981, 1986), Wilss (1982), Hatim and Mason (1990, 1997), Mason (1994), Bell (1991), Baker (1992, 1993, 1995, 1996b), Neubert and Shreve (1992) and Harvey (1998) as examples of such work. It could also be added here that some earlier work never lost its appeal – Vinay and Darbelnet (1958) and its 1995 translation into English is one such volume – and Baker’s own linguistically-oriented coursebook on translation (1992) became an academic best-seller.
The parallel development of two seemingly incompatible approaches to translation became the subject of much debate in the late 1990s. Mona Baker, who was to become something of a spokesperson for linguistic approaches (see Baker 1996a, 2000, in press), warned that the rift between the cultural studies and linguistic paradigms threatened “to reduce the discourse on translation studies into a series of fault finding exercises and divisive oppositions” (1996a:9) and argued for a place for both perspectives in translation studies. And Lawrence Venuti, whose sympathies clearly lay with cultural studies, nevertheless argued not that linguistic approaches be abandoned, but that they be supplemented in specific ways (1996: 104, 109). The tone of his landmark Translation Studies Reader (2000) is even more conciliatory: cultural, linguistic, and other approaches to translation sit side-by-side in the book, and even in some of the same articles.
So how can linguistic approaches to translation have been so right, and then so wrong, and then alright again? These questions have been addressed to varying extents by other writers, in particular Baker (1996a, 2000, in press) and Fawcett (1997). One picture that emerges from these critiques is that of a fledgling discipline – the one that became known as translation studies – attempting to secure its place in the academy by aligning itself with the ‘science’ of linguistics and initially focusing on the idealized, decontextualized and simplistic interlingual operations this seemed to entail. As time went on, linguistic approaches to translation began to recognise their own shortcomings and shift their focus to real translation behaviour, translation as communication, whole texts, and their ideological import. By chronicling the evolution of linguistically-oriented translation studies, scholars like Baker and Fawcett could counter criticisms of linguistics emanating from within cultural studies and even point to a convergence of interests between the two supposedly competing paradigms. But they did more than that. By highlighting the diversity not only within what could be considered linguistic approaches to translation, but also within linguistics itself, they began to undermine a basic position implicit in much of the postmodern critique of linguistics in translation: the assumption that there is an undifferentiated monolith called ‘linguistics’.
That there is diversity within linguistics, and that linguists can occupy diametrically opposed positions when it comes to such basic issues as what their object and mode of inquiry should be, are points that have been well made by linguists themselves (see, for example, Beaugrande 1994, 1997; Sampson 1980). Michael Stubbs’ (1993, 1996) treatment of the deep and systematic differences between Chomskyan and neo-Firthian linguistics is of particular interest here, as it juxtaposes the work of arguably the most influential twentieth-century linguist, Noam Chomsky, with that of the linguist whose legacy is felt in the work presented in this book, J.R. Firth. Below I follow Stubbs’ lead in setting Firth off against Chomsky, but my analysis will have an added, translation studies dimension, in that it will attempt to show how the ideas of each were taken up in linguistically-oriented theorizing on translation. If more space is given to Firth’s ideas, this is because they are more applicable to translation, and in particular corpus-based translation studies, than are Chomsky’s. Ironically though, I will argue how ultimately it was mainly Chomsky’s influence that was acknowledged, implicitly or explicitly, by critics of linguistics in translation studies.
Chomskyan linguistics
Chomsky’s oeuvre is extensive, spanning nearly fifty years of publication and taking in linguistics, philosophy and dissident politics. I will not attempt to do his intellectual achievements justice here. Instead, the reader is referred to Smith’s (1999) sympathetic survey of Chomsky’s work to date. The technical details of his linguistic theory have also changed constantly over the last half century, so much so that Smith likens Chomsky to Picasso for his ability to overthrow his own established systems with “startling” frequency (ibid:1). But certain fundamental positions have remained constant in Chomskyan linguistics, and problematic for his detractors. These positions include Chomsky’s stance on the goals of linguistic theory, the nature of linguistic evidence, and of language itself.
Chomskyan linguistics has long been concerned with native speakers’ knowledge of their language, their linguistic ‘competence’ (Chomsky 1965) or ‘I-language’ (Chomsky 1986). Chomsky originally contrasted competence with ‘performance’, “the actual use of language in concrete situations” (1965:4), but noted that only competence could be the object of inquiry in a properly constituted linguistic discipline. From the outset, Chomsky has been far more concerned with the mental structures underlying human language than with the actual use of language:
in the technical sense, linguistic theory is mentalistic, since it is concerned with discovering a mental reality underlying actual behavior. Observed use of language or hypothesized dispositions to respond, habits, and so on, may provide evidence as to the nature of this mental reality, but surely cannot constitute the actual subject matter of linguistics, if this is to be a serious discipline. (1965:4)
Coupled with Chomsky’s mentalism is his staunch defence of innateness. Chomsky argues that much of what we know about language is hard-wired in our brains, part of our biological endowment (see, especially Chomsky 1988). Part of the evidence for this claim comes from Chomsky’s conviction that there are structural properties that are common to all languages, linguistic universals that can only be explained with reference to a language acquisition device that is shared by all humans.
Chomskyan linguistics is further based on a number of idealizations that allow linguists to ignore speakers’ idiosyncrasies, the nebulousness of speech communities, and irrelevant features of performance. This basic position is summed up in one of Chomsky’s (1965:3) most famous passages:
Linguistic theory is concerned primarily with an ideal speaker-listener, in a completely homogenous speech-community, who knows its language perfectly and is unaffected by such grammatically irrelevant conditions as memory limitations, distractions, shifts of attention and interest, and errors (random or characteristic) in applying his knowledge of that language in actual performance.
Despite objections from other quarters, which will be dealt with below, these idealizations continue to be underwritten by the Chomskyan commitment to linguistics as a natural science. In Smith’s words, “all science involves idealization”, and in order to see the principles that underlie linguistic behaviour, “it is necessary that some things be fruitfully ignored” (1999:12-13).
Like Saussure (1916) before him, Chomsky thus bases his linguistics on a dichotomy between an abstract system that underlies language use (Chomsky’s competence; Saussure’s langue), and real-life language as used by real people (performance; parole). There are differences however: Chomsky rejects the Saussurian notion of langue “as merely a systematic inventory of items” and advocates a return to “the Humboldtian conception of underlying competence as a system of generative processes” (1965:4); and whereas Saussure’s langue exists perfectly only within a collectivity, Chomsky’s competence resides very much in the minds of individuals. Nevertheless, both privilege the study of an abstract linguistic system, which is not amenable to direct observation, over real spoken and written language, which is. For Saussure this has the effect of insulating langue from the massive variability of parole, of maintaining langue as a well-defined object in the heterogeneous mass of speech facts. It also allows him to defend the autonomy of linguistics from the other disciplines that might be interested in the “heterogeneous mass of speech facts”, including anthropology, sociology and psychology. And by fencing off synchronic from diachronic linguistics, he manages to fix langue as a stable, ahistorical entity.1 Chomsky likewise, by constructing his object of inquiry so narrowly, carves out a niche for an autonomous discipline of linguistics. But nowhere in language does the idea of autonomy work quite as well as it does with syntax.
Beaugrande (1997, 1998) has suggested that by disconnecting linguistic theory from real data and the life-histories of speakers, formalist linguists – as exemplified by Chomsky – privilege syntax. While phonology and morphology typically retain contact with the reality of language, “by matching their theoretical units (phonemes and morphemes) against the practical units (sounds, inflections, etc.) discovered in fieldwork data” (1998:774), syntax can readily disengage from reality by assuming that word-order “can be most rationally accounted for by an independent system of deterministic ‘rules’ situated entirely on the side of ‘language’ [Saussure’s langue]… and unaffected by the contexts of real speech, the motivations of real speakers, and perhaps even the meaning of utterances” (ibid:773). Beaugrande’s final point here is an allusion to Chomsky’s (1957:17) insistence that “grammar is autonomous and independent of meaning”, a position that informed Chomsky’s earlier work. Chomsky now holds that much conceptual information associated with words is contained in the entries of our mental lexicons, and constitutes a large part of our knowledge of language. He calls this conceptual information ‘syntax’ however, “because it has to do with mental representations and the structure of mental representations” (in Olson and Faigley 1991:93). What Chomsky is adamant about, however, is that linguistics cannot have anything to say about reference, the relationship between words and things in the real world. In Chomskyan linguistics, there is no bridging the internal structures of language with the outside world. Ad hoc extralinguistic contexts may be set up by hearers to enable them to interpret linguistic utterances, but this has little to do with linguistics, and everything to do with what Chomskyans regard as the separate discipline of pragmatics. (Verbal) communication in general may be parasitic upon language, but it does not determine language, and Chomsky has little to say about the issue (see Smith 1999:151-154). Likewise, speakers and writers may use language to act upon the world in certain ways, for example to impose authority, but this has little to do with the structures of language, and remains uninteresting to Chomsky, who claims the question has “no intellectual depth to it at all, like most things in the social sciences” (in Olson and Faigley 1991:88).
A final word will be said here about Chomsky’s methodology: Chomsky’s focus on competence, or I-language, has implications for how linguistic research is carried out, and for what constitutes evidence in a linguistic theory. Competence is studied through introspection, and if the object language is the linguist’s own native tongue, then the linguist’s intuitive judgements about the well-formedness of particular sentences constitute valid data. The fact that the sentences invented by linguists to test their judgements are often outlandishly convoluted does not pose a particular problem for Chomskyans, who hold that there is “no requirement in science that the phenomena used as evidence should occur spontaneously under natural conditions” (Smith 1999:33).
