A pragmatic analysis framework for the description of modality usage in academic English contexts

ELIA 2, 2001 213 A PRAGMATIC ANALYSIS FRAMEWORK FOR THE DESCRIPTION OF MODALITY USAGE IN ACADEMIC ENGLISH CONTEXTS Jordi Piqué (Universitat de València) Santiago Posteguillo (Universitat Jaume I, Castelló) J.Vicent Andreu-Besó (Universitat de València) In our pragmatic approach to research articles (RAs), we take a look at one of the main characteristics of academic writing: the use of modalized statements. Modalization, in the form of modal verbs, has been extensively studied by ESP researchers and practitioners. Most of their studies, however, have slightly touched upon the distinction between epistemic modality, which questions the certainty or probability of a statement, and deontic modality, which lays obligations or gives permission to the audience. This dichotomy is an important tool to describe disciplinary variations in academic and professional writing. We contend that different disciplines favor different types of modality. Results in this study indicate that health science RAs mostly use epistemic modality, whereas literary criticism RAs combine the use of both epistemic and deontic modality. 1. Introduction Within the pragmatic linguistic paradigm, genre analysis, as defined by Swales (1990) or Bhatia (1993), is a major approach in the current stateof-the-art study of English for Specific Purposes (ESP). In this line of research, much attention has been given to the definition of some specially relevant genres, such as abstracts (Cremmins, 1982; Lancaster, 1991; Posteguillo, 1996b). A number of studies have focused on RA structure (Hutchins, 1977; Stanley, 1984; Swales, 1990; Nwogu, 1997; Bhatia, 1993; Posteguillo, 1996a, 1999; Piqué & Andreu-Besó, 2000), on individual sections of the RA, such as the introduction (Swales, 1981; Dudley-Evans & Henderson, 1990; Piqué & Andreu-Besó, 1998); the results section (Brett, 1994); the discussion section (Hopkins & Dudley-Evans, 1988; Holmes, 1997) and including other discourse functions within the RA, such as the use of citations and references in RAs (Dubois, 1988; Lynch & McGrath, 1991), on RA argumentation (Hyland, 1990; Thompson, 1993), to name but a few. J. Piqué, S. Posteguillo y J. V. Andreu-Besó 214 Research articles --as the more deeply analyzed academic genre-- have been already been studied to unfold some of the most salient disciplinary variation to be detected. However, most of these studies have focused on whether the IMRD pattern is applicable to all disciplines as the representative macrostructure of the genre. We have also pursued this line of research in previous studies when, for instance, describing the semiapplicability of the introduction-method-results-discussion (IMRD) pattern to computer science research articles (Posteguillo, 1999), or the appropriateness of Swales’ (1990) CARS introduction pattern for medical and nursing research articles (Piqué & Andreu-Besó, 1998, 2000). It is true that these macrostructural descriptions represent a major feature of RAs as a genre, but there are other linguistic and communicative phenomena which differentiate the research article from other genres which have not been analyzed in such detail, especially having in mind disciplinary variation. Our aim in this paper is to analyze the language of three different academic and professional contexts (health sciences, journalese and literary criticism) in order to detect possible variations in the use of epistemic or deontic modality and to compare our results with those obtained by Simpson (1990). Our initial hypothesis is that different disciplines or professional settings favor different types of modality. 2. Corpus and analysis Mood and modality express the speaker’s attitude or opinion regarding the contents of the sentence or what the sentence proposition entails (Palmer, 1986: 21). He defines mood as realized by the verbal morphology, whereas modality appears as a linguistic feature generated by a variety of linguistic phenomena, as described by Downing and Locke (1992: 383-384), among which modal verbs play a special role. Both Palmer (1986: 18-19) and Downing and Locke (1992: 382) distinguish two main types of modality: epistemic and deontic. Epistemic modality, on the one hand, implies that the speaker assesses “the probability that the proposition is true in terms of the modal certainty, probability or possibility” (Downing & Locke, 1992: 382), as in It may be the case that ..., Results might change if certain conditions ..., or The ELIA 2, 2001 215 concert must be over. On the other hand, deontic modality means that the speaker “intervene[s] in the speech event by laying obligations or giving permission” (Downing & Locke, 1992: 382), as in One must look into this matter in detail..., or Shall we negotiate peace now? In our research we gathered three distinct corpora: first a corpus of health science RAs (with 30,086 words) made up of different RAs dealing with AIDS, cancer and surgery; a second corpus of newspaper articles (with 30,481 words) dealing with all types of topics downloaded from The Guardian; and a third corpus of literary criticism RAs (with 30,042 words) dealing with American and British literature from a selection of essays. The names of the authors, the titles of the articles --or headlines--, as well as the tables, pictures or figures have been eliminated. We are aware that corpus 2 is made of a genre (i.e. newspaper articles) other than the RA, but we believe in the relevance of comparing the use of epistemic and deontic modality across three distinct academic and professional settings --health sciences, journalese and literary criticism. Nevertheless, the differences --should there be any-- between corpus 1 (health sciences) and 3 (literary criticism) can only be explained in terms of disciplinary variations in the use of modality. In each corpus we have systematically considered the modal verbs can, could, may, might, must, ought, shall, should, will, would, and the two semimodal verbs dare, need, as defined by Downing and Locke (1992: 315316). 3. Results Tables 1, 2 and 3 below offer the results obtained in our investigation. The data from the health sciences corpus in presented in Table 1. On the left-hand column one may see the modal and semimodal verbs analyzed; on the second column we have the number of times which a particular verb has been used in that corpus; column three shows the frequency (f) of use of the modal verb with epistemic meaning; column four gives the relative percentage (%) of use of epistemic modality; column five shows the number of instances (f) in which the verb was used with deontic J. Piqué, S. Posteguillo y J. V. Andreu-Besó 216 meaning, and, finally, column six gives the corresponding relative percentage (%) of deontic usage. Tables 2 and 3 below are organized in the same way. Table 1. Epistemic and Deontic Modality in Corpus 1. 1-Health Science-RAs Epistemic Deontic Total Modals f % f % can 124 123 99.19 1 1.01 could 31 31 100 0 0 dare 0 0 0 may 125 125 100 0 0 might 20 20 100 0 0 must 17 15 88.23 2 2.27 need 30 28 93.33 2 2.14 ought 0 0 0 shall 0 0 0 should 49 41 83.67 8 9.56 will 30 30 100 0 0 would 24 24 100 0 0 Totals 450 437 97.11 13 2.97 According to these results, the use of deontic modality is almost non-existent in health sciences RAs representing only 2.97% of the total modal and semimodal verbs. It is also noticeable that many of the modals and semimodals show a 100% of epistemic usage. Such is the case of could, may, might, will and would, as illustrated in the following examples: (1) If all cells that harbored HIV could quickly be destroyed in this manner, the virus might be cleared from the body. (2) What's more, the immune system, in killing off infected cells, may play a significant role in its own destruction. (3) On the basis of these findings, researchers have developed theories on how HIV might indirectly destroy nerve tissue. ELIA 2, 2001 217 (4) This process will allow for the identification of those measures which hold the most promise of moving the field forward and of critical gaps that persist in the HIV HRQoL measurement arena. (5) Most clinicians would probably agree that this amount of difference would be a clinically significant difference. The results obtained in texts from The Guardian are shown in Table 2. Epistemic modality continues being the main type of verb used by journalists (91.20%), although a slight increase is present in regard to verbal expressions with a deontic orientation (8.80%). Table 2. Epistemic and Deontic Modality in Corpus 2. 2-Journalism—The Guardian EPISTEMIC DEONTIC Total Modals f % f % can 64 52 81.25 12 18.75 could 41 39 95.12 2 4.88 dare 00 0 may 22 21 95.45 1 4.55 might 10 10 100 0 0 must 22 14 63.64 8 36.36 need 19 19 100 0 0 ought 1 1 100 0 0 shall 00000 should 39 35 89.74 4 10.26 will 112 111 99.11 1 .89 would 79 71 89.87 8 10.13 Totals 409 373 91.20 36 8.80 The modal verb must, however, presents a particularly interesting change. While in health sciences this verb was used as an epistemic modalizing device (cf. Table 1), in journalism it combines epistemic (63.64%) with deontic (36.36%) meaning. We want to illustrate this use in the following two examples from corpus 2 in which the modal verb must is J. Piqué, S. Posteguillo y J. V. Andreu-Besó 218 presented in this double modalizing function: epistemic in example (6) and deontic in example (7). (6) As it is, the shower of Grammies and BPI awards that Telboy earned from his 1987 debut album now seems a long time ago, and he finds himself in what must be a slightly uncomfortable limbo. (7) 'What you did that evening in Mrs Longden's flat was a disgraceful exhibition of uncontrolled bad temper which split over into the commission of a crime for which you must be punished.' Corpus 3, from literary criticism essays, present the most striking differences. As shown in Table 3, the pattern of epistemic/deontic usage is reversed with modals must and shall, both of which are more frequently used in deontic modalizing expressions than in epistemic ones. It should be pointed out, however, that although the modal verb shall may not be particularly relevant, since it only appears in 7 instances, the modal verb must, with 58.62%, confirms the double function of this modal. ELIA 2, 2001 219 Table 3. Epistemic and Deontic Modality in Corpus 3. 3-Literary Criticism Essays EPISTEMIC DEONTIC Total Modals f % f % can 61 37 60.66 24 39.34 could 23 17 73.91 6 26.09 dare 00 0 may 26 20 76.92 6 23.08 might 29 27 93.10 2 6.89 must 29 12 41.38 17 58.62 need 7 5 71.43 2 28.57 ought 5 5 100 0 0 shall 7 1 14.29 6 85.71 should 10 8 80 2 20 will 55 37 67.27 18 32.73 would 36 31 86.11 5 13.89 Totals 288 200 69.44 88 30.56 The following examples illustrate once more this double function of must, in this case in the field of literary criticism: epistemic, in example (8), while deontic in (9). (8) The effort of accurate transcription involved must have been heroic. But exactly what the utility of the exercise was I, for one, have never understood. (9) If we want to understand America, you must realize that the greatest love is this homoerotic passion, which is in some ways as strong, though also as sublimated, as Dante's love for Beatrice. It should also be pointed out further differences in the use of modal verbs, especially in reference to corpus 3 (journalism) as compared to corpus 1 (health sciences). Five modal verbs –could, may, might, will, would– in corpus 1 showed a permanent use of epistemic modality. In corpus 3, however, these same five modals present several instances of deontic modalization, as can be seen in the following examples from journalism J. Piqué, S. Posteguillo y J. V. Andreu-Besó 220 texts: (10) There's a famous essay by Mark Twain, which is called 'The Literary Offenses of Fenimore Cooper", in which he points out that Cooper couldn't build a novel. His structures are pitiful; he couldn’t write dialogue. (11) It is, one may think, a relief that more records do not remain, since if they did, and if McKenzie transcribed them as assiduously as he does those for 1696-1712, his history of the press would run to 120 quarto volumes. (12) In a superficial and simplified way (which reaches, nevertheless, as far as we might wish to extend it, and with all of the nuances which are considered necessary), literature is a product of human activity ... (13) And this takes me back to my beginning which will also be my end. (14) But I would like to take this as far as possible. This progressive but evident increase in the use of deontic modality is shown in Figure 1 below which graphically summarizes the results shown in Tables 1, 2 and 3. In turn, epistemic modality use is close to 100% in corpus 1, it slightly diminishes in corpus 2, but there is a significant decrease in corpus 3. Figure 1. Epistemic vs Deontic Modality in the Three Corpora (in %). 0 20 40 60 80 100 120 Deontic 2.97 8.8 30.56 Epistemic 97.11 91.2 69.44 1-HS-RAs 3-Journalism 2-LitCrit ELIA 2, 2001 221 4. Discussion The results that have just been presented in this research confirm our first hypothesis in which we ventured that different disciplines favor different types of modality. This is especially so in the case of health sciences and literary criticism research articles. These two specialisms, as exemplified in corpora 1 and 3, quite similar in size and genre, represent two different uses of modal verbs in terms of epistemic and deontic modality. While in health sciences (corpus 1) seems to be practically the only acceptable option for the writer, in literary criticism both options seem to be equally resorted to by their authors. Their communicative purpose seems to dictate the use of a given modal verb through which they express their epistemic or deontic meaning. These findings open a whole new area of linguistic and sociolinguistic research and makes us pose the question whether the communicative purpose of an essay in literary criticism differ from the communicative purpose of a medical researcher/writer. This may very well be the case. Simpson (1990) noticed the tendency of literary critics to be highly assertive in their statements. He described instances of unmodalised assertions in the text he studied --The Great Tradition, by F. R. Leavis. Simpson quotes Leavis’ (1950: 1) opening statement as a typical example of unmodalized assertion: (15) The great English novelists are Jane Austen, George Eliot, Henry James and Joseph Conrad—to stop for the moment at that comparatively safe point in history. Simpson’s study, however, has the shortcoming of having analyzed only one text and, accordingly, the writing of only one literary critic. In addition, he does not provide specific quantitative data. Our analysis of the writing of different literary critics confirms that literary criticism reduces the use of modals in comparison to other academic disciplines. It also confirms that when literary critics do use modals they may do so in either epistemic or deontic assertions. This allows us to suggest that the combination of deontic