This book "supplies a vocabulary of English words and idiomatic phrases 'arranged … according to the ideas which they express'. The thesaurus, continually expanded and updated, has always remained in print, but this reissued first edition shows the impressive breadth of Roget's own knowledge and interests."
Date: Tue, 21 Jan 2003 16:16:04 +0100 From: Gerhard Leitner <firstname.lastname@example.org> Subject: Review: Corpus Linguistics, Westin (2002)
Westin, Ingrid, 2002. Language change in English newspaper editorial. Amsterdam: Rodopi. xvi+202 pp. ISBN 90-420-0863-6 (bound). Gerhard Leitner, Institut fuer Englische Philologie, Freie Universitaet Berlin "An editorial, in British journalism also referred to as a 'leading article or 'leader'", Westin says, "is a newspaper article expressing the opinion of the editor or publisher of the newspaper on some topical issue. A distinction is often made between personal editorials, which are by-lined with the writer's name, and institutional editorials, which are not." (p 7). Editorials thus are important parts of the dailies and have, obviously, been at the centre of interest in mass communications, text and discourse linguistics (e.g. Dijk 1998), in so-called Critical Discourse Analysis (Fairclough 1995), and, more recently, corpus linguistics. A historical analysis has not yet been published and Westin is, therefore, a first. She expects to find "a drift towards more informal styles ... during the period studied" (p 1). Such a drift has been identified in many studies on English worldwide and it would be interesting to see if, when and how it manifested itself in editorials. She also believes that "[T]he editorials [also] appear to have lost some narrative touch they had at the beginning of the century in favour of a style that is more matter-of-fact." (p 1). That would also be an interesting finding, if confirmed. If editorials became less narrative—whatever that means—reports clearly have become more so. So, one would notice an increasing cleavage between not unrelated print media text types. Her study divides into ten chapters. Westin begins with goals, expectations and a brief review of past research (chapter one) and ends with a summary that argues that the language of editorials is an agent in language change in general (chapter ten). A review of the shortcomings of available computer corpora, of methods used in the compilation of the "Corpus of English Newspaper Editorials" (CENE), and of analytical techniques is in chapter two. The remaining chapters fall into two parts. Chapters three to eight cover the details of her investigation, chapter nine re-integrates what was separated for analytical purposes and aims at a comprehensive account of the language of editorials. There is a sizeable bibliography and an appendix with all relevant statistical details. An index is missing. She looks at the language of editorials over almost the entire 20th century, hoping to identify change and continuity in what she calls up-market British newspapers editorials. The Times, The Guradian and The Daily Telegraph clearly fit that definition of 'up-market' and, moreover, appeared at and before the period she studies, which makes them truly comparable. Dividing the period into decades and using the principle of the 'constructed week', which means that a Monday is taken for January, a Tuesday in February, etc., she assembled a corpus of 864 editorials and a total of nearly 600,000 words. The data were tagged for automatic searches, which did not eliminate the need for manual editing. The quantities were normalized to 1,000 words each to eliminate imbalances in the corpus and then processed for decades, for all papers, for each one and submitted to validity and variance tests to permit valid inferences. The major linguistic background is Biber (1988) and Biber's and Finegan's other diachronic genre analyses. She selects 42 of Biber's features and uses his communicative-functional clusters, i.e. - Personal Involvement (PI), 17 features (chapter three) - Information Density (ID), 8 features (chapter four) - Narrative Discourse (ND), 6 features (chapter five) - Argumentative Discourse (AD), 5 features (chapter six) - Abstract Discourse (AbD), 6 features (chapter seven) - Explicit Reference (ED), 4 features (chapter eight) She added imperatives and questions and sentence length and subordination as expressions of PI and ID, respectively, and re-constituted the communicative clusters to some extent when she turns to an integrated analysis (chapter nine). Her analyses are detailed, with brief introductory sections, examples, normalized frequencies, validity data, tables and diagrams that mirror developments over 90 years (for each paper and the total) and comments that aim to provide reasons for change or continuity or else to argue other relevant matter. I will sum up her findings, ordered by statistical significance, either increases, decreases or continuity, and order them by communicative-functional clusters; I will add in brackets terminological and other changes Westin makes in chapter nine: (A) STATISTICAL INCREAESES - Information Density, i.e. Nouns, Attribute. Adject. (incl. Noun Phrases, NPs), Present participle clause WHIZ deletion (NPs), Word length, Type/token ratio - Information Density, i.e. Nominalization (reclassified from Explicit Reference) - Information Density, i.e. Infinitives (reclassified from Argumentative Discourse) - Informal Discourse, all reclassified from Personal Involvement, i.e. Present tense verbs, Not-negation, Questions, Imperatives, Contractions (B) STATISTICAL DECREAESES - Abstract Discourse, i.e. Agentless passives - Explicit Reference, i.e. who/which, Pied piping - Information Density, i.e. Prepositions (and Prep.'al Phrases) (NPs) - Sentence Complexity, all reclassified from Information Density, i.e. Sentence length, Subordination - Narrative Discourse, i.e. Past tense verbs, No-negation - Vagueness/uncertainty, all reclassified from Personal Involvement, i.e. Adverbial amplifiers, Private Verbs, First person pronouns, Pronoun it. (C) CONTINUITY - Abstract Discourse, i.e. 'Detached' present paticiple clauses, Past participle WHIZ deletion, Conjuncts, By-passives - Argumentative Discourse, reclassified from Abstract Discourse, i.e. 'Subordinators' with mutliple functions - Argumentative Discourse, i.e. Conditional subordination, Suasive verbs, 'Predictive' modals, 'Necessity' modals - Explicit Rerference, i.e. Relative that, Causative subordination - Narrative Discourse, i.e. 'Detached' past participle clauses, Perfect aspect verbs, Public verbs, Third person pronouns - Vagueness/uncertainty, reclassified from Personal Involvement, i.e. Demonstative pronouns, Possibility modals, Indefinite pronouns, General hedges - Informal Discourse, reclassified from Personal Involvement, i.e. Second person pronouns, Discourse particles, General emphatics Table 1: Features ordered by significance (and communicative-functional domains) This summary is based on data on all papers, though there are statistically valid patterns for some papers and/or periods. With nearly half the feature not changing at all, there is, obviously, a high degree of continuity, which, she says, is often "linked with sets of features that have either increased or decreased in use" (p 149). Continuity thus masks change elsewhere. Clearly, Information Density (ID) and Personal Involvement (PI) (5 features) are increasing in use. Two other features increase, one from Argumentative Discourse, the other from Explicit Reference, but both are reclassified as ID (see fourth column). As she replaces the term PI by that of Informal Discourse (for features 08 to 12), she can convincingly argue that "[C]onsequently, two conflicting linguistic paradigms are at work in the editorials: the aspiration for informality and the aspiration for information density and [a new term, GL] lexical specificity, the former probably the result of an adjustment to a new broader reading public, and the latter the result of an adjustment to the special 'house styles' that developed over the years." (p 164) That is well put, except that it remain unclear about the goals of house styles. Why should media write manuals that lead to odd results? One might think of such terms as addressee-orientedness and message-orientedness—a term that would refer to professional jargon (Leitner 1980)—and argue that both have been increasing. They lead to an unresolved conflict that is particularly acute in the structure on Noun Phrases, see features 04, 05 and 16 (chapter 4.3), where she notes a move from post-modification to premodification (cf. also Bell 1991). But she now has a problem with sentence length and sub-ordination, which had been treated as ID on the understanding that ID means 'high' density of information. But the drift to shorter sentences and less subordination creates less density and she feels compelled to separate the two out into a novel category, Sentence Complexity, which remains unexplained and which does not get rid of the conceptual problem either. As sentence complexity is decreasing, does that not signal a higher level of addressee-orientation, of informality? That would suggest that the notion of Personal Involvement would have to be reconsidered, which she does (see table 1, column 4). Her solutions, e.g. to reclassify demonstrative pronouns to Vagueness, do not appear convincing, though most of the other changes do. One could add comments here and there on findings that corroborate the trend she argued for. I will only add one further remark. She finds a shift towards informality for a number of features to occur during and after World War II, which may reflect an AmE influence earlier than is often assumed. I will proceed to an evaluation of what she says about differences between the three papers studied, which, while up-market, do not target the same readerships. She does find many differences between the papers, but most of them do not built up to statistically significant ones, she says (p 158). The ones that do can be summed up in these terms. The three papers differ regarding the way they exploit features that are allocated to either Informal Discourse, Information Density, Vagueness, Narrative Discourse and Argumentative Discourse. For instance, "The Times" stands out for its minor use of present tense, questions, and contractions (ID), while the "Daily Telegraph" does use present tense and question, though it has few contractions. "The Guardian", finally, use all three features. "The Guardian", incidentally, also turns out to be most 'argumentative' in that it has statistically more predictive modals, conditional and causative subordination, but few conjuncts. The other papers use of mix of these features, but are less argumentative. More interestingly still is the finding that "The Guardian" is also more 'narrative' in terms of 1st person pronoun use and past tense, and it is 'vaguer'—having more it and demonstrative pronouns. Westin summarizes her findings in this way: "When the features were compared across the newspapers analyzed, a clear distinction was noticed between The Times and the Guardian. The language of the Guardian is the most informal and the most narrative while that of The Times is the least so. The information density is the highest in The Times, and the lowest in the Guardian. In these respects the Daily Telegraph this takes an intermediate position. The editorials in the Guardian are the most argumentative." (p 161) That conclusion may well correlate with the social profiles of the papers' readerships, but would still have to be argued in detail for now and the past. In particular, it is peculiar that "The Guardian" should be both most argumentative and narrative. Has there been a style shift? It is interesting to add that the three papers do treat editorials as a fairly homogeneous type of text. A few closing remarks. Biber's diachronic genre analysis is the undisputed basis of her study, though Westin would have done well to be more critical of it. For instance, she maintains that the simple present is a reflection of PI or of Informality, that it serves to refer to what she calls 'topical issues'—the focal themes of editorials (see definition quoted above). Issues may be topical for a variety of reasons. "September 11" is a long time past and yet is clearly 'topical'. Do we have to expect the simple present? Linguists are generally agreed that the simple present has little to do with present time, some argue it is not even a tense, but an aspectual marker. Does it show PI, then? Certainly not. If anything does, it is the present progressive, which she lumps together with the simple present but should have been separated out. (Moreover, though she refers to the different uses of the simple present, she ignores that—for pragmatic, but indefensible reasons—in her analysis.) Westin's study is a sound contribution to historical genre analysis with many interesting facets. In relation to methods used and data analysis there is nothing that causes serious criticisms. What does is a certain amount of mechanical reproduction of a model on new data and unexplained changes that are made in the course of re-integration and interpretation. Another aspect is the fact that she has nothing to say in relation to other approaches to editorials—recall Critical Discourse Analysis, text linguistics and studies of ideology and opinion formation. Also, does corpus linguistics have anything to offer to such other research paradigms? Westin may have felt that to be beyond the intents of a doctoral dissertation but, at least, the reviewer can raise that question. Literature: Bell, Allan (1991) The language of news media. Oxford: Basil Blackwell. Biber, Douglas (1988) Variation across speech and writing. Cambridge: Cambridge University Press. Dijk, Teun A. van (1998) Opinions and ideologies in the press, in: Allan Bell, Peter Garrett, eds, Media discourse. Oxford: Blackwell. 21-63. Fairclough, Norman (1995) Critical discourse analysis: the critical study of language. London: Longman. Leitner, Gerhard (1980) BBC English and Deutsche Rundfunksprache. A comparative and historical analysis of the language on the radio", International Journal of the Sociology of Language 26, 76-100. ˘
ABOUT THE REVIEWER:
Gerhard Leitner's research interests and publications come from the fields of the sociolinguistics of mass media, varieties of English worldwide, Australia's language ecology and corpus linguistics; editor of GASt Newsletter, the newsletter of the (German) Association of Australian Studies (ISSN 1617-9900), co-editor, with Brian Taylor and Clemens Fritz, of Language in Australia and New Zealand. A bibliography from 1788 to 2001, CD-ROM. Berlin: Mouton de Gruyter (forthcoming, autumn 2003).