TEXT TYPE, GENRE, REGISTER AND STYLE IN RELATION TO TEXT ANALYSIS AND TRANSLATION
Introduction
Although the terms text type, genre, register and, to a lesser extent, style are widely used in both linguistics and Translation Studies, there seems to be little general consensus with regard to their meaning. The aim of the present paper is to discuss varying definitions of these terms and to suggest how they can be used to describe situational variation among texts. Our survey includes an overview of the disciplines of register studies and genre analysis, as well as efforts within Translation Studies to define text type. The discussion should be of relevance not only to those concerned with translation, but also to all those interested in the analysis of non-literary texts.
Style
Style has been defined in a broad range of ways over the years and has become a kind of umbrella term for a range of factors both textual and contextual. Generally, style is seen as concerned with choices made by individual writers, judged in relation to norms. The situational dimensions governing stylistic choices may be relatively permanent – for example, to do with dialect or time – or relatively temporary and/or localised. The latter relate to the professional or occupational activity involved, the status or relative social standing of the participants in the communication situation, the specific purpose of the communication, and the idiosyncratic preferences of the individual writer or speaker (cf. Crystal and Davy 1969). The other factors influencing stylistic choice are the fundamental features of language use – in particular, the medium involved and whether we are concerned with monologue or dialogue. Deviations from norms may be qualitative, i.e. a breach of a rule or convention, or quantitative, i.e. in terms of frequency of occurrence of particular features or items. However, norms are flexible (Leech and Short 1981) and may be creatively extended – which, of course, is the most difficult stylistic feature for a translator to deal with.
By contrast, register is concerned with group convention and in relation to many kinds of functional texts, individual writers and their choices do not have the same status: the more specialized the text, the more concerned we are with register rather than style and the more likely it is that writers will not even be named. Thus, to a large extent, style studies has been replaced by register studies, especially in relation to non-literary texts (cf. Hatim 1997: 13ff and Snell-Hornby 1988: 120ff).
Register
The most influential discussion of register is still that found in the study of cohesion by Halliday and Hasan (1976), who introduce the concept in order to deal with textual meaning. The register is “the set of meanings, the configuration of semantic patterns, that are typically drawn upon under the specified conditions, as well as the words and structures that are used in the realization of these meanings” (Halliday and Hasan op.cit. 23). Register has three elements: field, which covers subject matter, the purposive activity of the speaker/writer and the nature of the social action that is taking place (basically, 'what is happening?'); tenor, which covers the relevant role structure or social relations, both permanent and temporary ('who is involved?'); and mode,[1] or symbolic organization of the text and its function in the context, including the channel, which can be described as the axis spoken-written ('what part is language playing?').
This tripartite model fits neatly with the three sets of underlying options, strands of meaning potential, or macro-functions identified by Halliday (1970), to which the options in the grammar of a language are related and which can be combined in any utterance, as required. The ideational function is concerned with cognitive meaning (related to our experience of the world, both internal and external) as well as with basic logical relations; in serving this function, language "gives structure to experience, and helps to determine our way of looking at things" (Halliday op.cit. 143). The interpersonal function relates to uses of the language to express social and personal relations, or the varying roles that we adopt in communication situations. These roles are defined by language itself: every language offers options whereby the user can vary his or her own communication role, making assertions, questioning, giving orders, expressing doubts and so on; these basic speech functions are expressed grammatically by the system of mood and differ when the role adopted by the language producer differs. Finally, the textual function enables us to construct texts or "connected passages of discourse that is socially relevant" (ibid.); and for Halliday it is the text, rather than the word or the sentence, that is the basic unit of language. These discussions of different language functions which can be realised simultaneously provide an alternative to the assumption that language use, particularly in written text, is concerned only with the communication of information. The organization of a written text indicates how it is to be read, but there is much more involved than the distribution of factual or propositional information – language is multi-functional. Thus field corresponds with ideational meanings, which are realised in the choices made within linguistic systems such as transitivity; tenor with interpersonal meanings, which find expression in the mood and modality of text; and mode with textual meanings, reflected in factors such as Theme-Rheme progression or distribution of given-new information.
To hang together, texts need to display consistency of register: coherence of meaning is dependent not only on content, but on selection from the semantic resources of the language. A text is “coherent with respect to the context of situation, and therefore consistent in register; and it is coherent with respect to itself, and therefore cohesive” (Halliday and Hasan 1976: 23, my emphasis). However, cohesion is a necessary but not a sufficient condition for the creation of text and the relation of text to the context of situation is very variable. As it is hard to draw a line between the same and different situations, we cannot ask whether texts are in the same register, we can only ask in what respects they differ or are alike. In comparing texts, subject matter is no more or less important than other factors – indeed, the idea of a single register corresponding to any one situation is a myth (cf. Hatim 1997: 22ff). A text needs continuity of register – the pattern formed by the communicative event (field), the role-relationships of the participants (tenor) and the language acts within the event (mode). Moreover, register variables interact with each other so that, for example the overlap between tenor and mode gives rise to what Gregory and Carroll (1978: 53) call "functional tenor", which is "the category used to describe what language is being used for in the situation" – in other words, is the speaker/writer trying to persuade, exhort, inform and so on (or, to put it another way, what is the communicative purpose of the text?) Ultimately, the key variable in questions of register is the relationship between those communicating:
The language we use varies according to the level of formality, of technicality, and so on. What is the variable underlying this type of distinction? Essentially, it is the role relationships in the situation in question: who the participants in the communication group are, and in what relationship they stand to each other. (Halliday 1978: 222)
Although cohesive relations are general to all kinds of texts, the forms taken by the cohesive relations will differ according to the register – clearly, texture in conversation and in formal written language is very different. Thus it is register, associated with classes of contexts of situation, that defines what a text means. In other words, text as the basic unit of meaning in language has to be interpreted in context and from a functional point of view.
Register studies
The discipline of register studies is perhaps most strongly associated with the work of Biber (1988, 1992, 1994), who uses register as a general cover term for all language varieties associated with different situations and purposes (thus correlating with many definitions of genre, text type and style). This broad concept of register is similar to that held by others, such as Gregory and Carroll (1978: 4), who define it as "a contextual category correlating groups of linguistic features with recurrent situational features". The purpose of register studies thus becomes the description of situational and linguistic characteristics and the analysis of the functional or conventional associations among these. The methodological approach recommended by Biber (1994) is the analysis of co-occurrence patterns. The linguistic component of the analytical framework involves searching for register markers, i.e. the distinctive features of particular registers, as well as differing uses of core linguistic features. The lexical and grammatical features used for register analysis include: tense and aspect markers, pronouns and pro-verbs, questions, nominal forms (nouns, nominalisations, gerunds), passive forms, dependent clauses (complement clauses, relative, adverbial subordination), prepositional phrases, adjectives, adverbs, lexical classes (e.g. hedges, stance markers), modals, verb classes (e.g. speech act verbs, mental process verbs), reduced forms, coordination, negation, grammatical devices for structuring info (e.g. clefting), cohesion markers (including lexical chains), and the distribution of given and new information. Comprehensive linguistic analysis of a register involves consideration of a representative selection of these features. Such analyses are quantitative, because "register distinctions are based on the relative distribution of linguistic features, which in turn reflect differences in their communicative purposes and situations” (Biber op.cit. 35). This framework treats register as continuous rather than discrete; and while some registers (such as personal letters) have well-defined norms, others (such as academic prose) show considerable variation. Thus texts will vary considerably within some genres, but not in others, while the registers themselves vary with regard to how much they are specified in relation to different linguistic dimensions.
Accounting for the situational parameters of language variation is, of course, an extremely complex business, as the table in Biber (1994: 40-41) shows. This situational framework, which draws in particular upon the work of Hymes (1967, 1974), Halliday (1978) and Crystal and Davy (1969), has also been shaped by the thinking of corpus linguists. In order to specify the situational characteristics of registers so as to distinguish between any pair of registers, seven parameters are to be taken into account:
1. the communicative characteristics of the participants (e.g. single, plural, institutional)
2. relations between addressor and addressee (inc. social role, level of shared knowledge, interactiveness, personal relationship)
3. setting (inc. place and time of communication, domain)
4. channel (inc. written vs. spoken, recorded vs. transient, medium of transmission)
5. relation of participants to the text (inc. production and comprehension circumstances, personal evaluations, attitudinal stance)
6. purposes, intents, goals (e.g. persuasion, information transfer, entertainment)
7. topic/subject (inc. level of discussion, specific subject).
Without discussing it in great detail, a number of points need to made about this framework. One is that some of the parameters, such as topic/subject and aspects of the participants' relation to the text, are open-ended, rather than closed; moreover, many (such as purpose) are in reality clines or continuums, rather than discrete values, and can also shift within a single text. The ultimate goal of this framework within the context of register studies is the specification of quantitative values for each parameter, expressed either in terms of ordinal scales (more-or-less relations) or dichotomies (such as high/low, yes/no). Biber's (op.cit. 44ff) illustration of the situational characteristics of eleven 'registers', is striking because of the disparate range of what the latter term includes – from three types of letter (personal, professional, recommendation), through spoken modes (face-to-face conversation, lecture, sermon), novels, narration (which Biber himself admits is not well-defined situationally), and psychology article, to two types of prose (expository and academic). The quantitative measurement of the correlation between linguistic dimensions and situational parameters (or communicative functions) is further illustrated in relation to four of the registers just mentioned, plus business telephone conversation and newspaper article. This shows the correlations, expressed in numerical terms, between four situational parameters (mode, interactiveness, careful production and informative purpose) and five varied linguistic features (contractions, that-deletions, prepositional phrases, passives and varied vocabulary – expressed as type/token ratio) (cf. Biber op.cit. 48ff).
Such studies of statistically significant lexico-grammatical features of different situational varieties of language are a valuable analytical tool in that they provide empirical evidence to confirm or undermine intuitive or impressionistic statements about registers. However, a checklist analysis does not represent the ways writers and readers actually use texts: readers do not scan texts for particular elements, while producing an acceptable text of a particular type is not a simple matter of introducing the features required to mark it as a member of that type. The latter point naturally applies equally to the process of translation, which is perhaps one reason why such an approach to register has not had more influence within Translation Studies (further reasons are given when we discuss text type, below). Moreover, as Bhatia (1993) points out, such analyses fail to provide adequate explanations of why particular features are present or absent, or how information is structured within a particular variety; nor do they explain how social purposes are accomplished through genres.
Register vs. genre
The concept of genre has been taken by linguists from the field of literary studies, where it refers to types of literary works (from poem, novel, short story, play, to sub-genres such as detective novel, romantic novel, spy novel and so on), and broadened to include texts, both written and spoken, that arise in a wide range of situations, including everyday transactions. In Martin's (1985: 250) words, genres are "are how things get done, when language is used to accomplish them".[2] His examples include lectures, seminars, recipes, manuals, service encounters and news broadcasts. His contention is that registers are realised through language, while genres are realised through registers (and registers are described in terms of configurations of Halliday's categories of field, tenor and mode) Both he and Ventola (1984) also refer to register and genre as different semiotic planes, with genre being the "content-plane" of register, while register is the "expression-plane" of genre. Eggins and Martin (1997) identify register with context of situation and genre with context of culture, which represent the two main dimensions within which texts vary. For Couture (1986), the two concepts are also distinct: registers operate at the lexico-grammatical level, while genres operate at the level of discourse structure.[3] Genres are probably best seen as structured, (potentially) whole texts that unfold in a particular way (a research or business report would be a good example), whereas registers represent the more general linguistic choices that are made in order to realise genres (for example, the language of scientific reporting or bureaucratic style). Thus genres have complementary registers and the communicative success of a text depends on an appropriate relationship to both systems.