ASSOCIATIONSUEDOISE DE LINGUISTIQUE APPLIQUEE (ASLA) The Swedish Association of Applied Linguistics

Text and Talk • ID Professional Contexts

Selected Papers from the International Conference "Discourse and the Professions" Uppsala, 26-29 August, 1992

Edited by Britt-Louise Gunnarsson ASLA Per Linell Bengt Nordberg ASLA:s skriftserie 6

ASSOCIATION SUEDOISE DE LINGUISTIQUE APPLIQUEE (ASLA) The Swedish Association of Applied Linguistics

Text and Talk in Prof essional Contexts

Selected Papers from the lnternational Conference "Discourse and the Professions" Uppsala, 26-29 August, 1992

Edited by

Britt-Louise Gunnarsson Per Linell Bengt Nordberg ASSOCIATION SUEDOISE DE LINGUISTIQUE APPLIQUEE (ASLA) The Swedish Association of Applied Linguistics ASLA is part of the international organisation AILA, which has members in more than thirty countries all over the world. The main aim of the association is to promote and disseminate information about linguistic research relating to practical language problems in society. ASLA's activities areas follows:

- arranging conferences, symposia and seminars, - publishing a newsletter for its members, - publishing symposium proceedings and other material, - distributing information and published material from AILA, - taking part in AILA's scientific commissions, committees and conferences. The newsletter "ASLA Information" is sent out three times a year. It gives details of literature, conferences etc. One issue a year also contains a section presenting ongoing research and developments in applied linguistics. A yearly symposium is arranged, with a theme which is of interest to researchers and practitioners. The proceedings of these autumn symposia are published in ASLA's publication series. The books in this series are yearbooks and they are distributed free of charge to ASLA members. It is possible to become a member of ASLA by paying the membership fee (SEK 175 from July 1, 1994) to the Swedish Association of Applied Linguistics, posta! giro 403286-8. Membership includes the yearbook in ASLA's publication series, ASLA Information, AILA Review and AILA News. Members can also attend ASLA symposia and buy back issues of ASLA's publication series at a discount. ASLA publications can be ordered from: ASLA, c/o FUMS, P.O. Box 1834, S-751 48 Uppsala, Sweden.

Text and Talk in Professional Contexts. Selected Papers from the lnternational Conference "Discourse and the Professions," Uppsala, 26-29 August, 1992. Britt-Louise Gunnarsson, Per Linell & Bengt Nordberg (eds.). ASLA, The Swedish Association of Applied Linguistics. Uppsala 1994. ISBN 91-87884-05-4 ISSN 1100-5629 © ASLA and the authors Preface

On 26-29 August 1992, 300 schalars from all over the world met in Uppsala, Sweden, for a conference on the theme of Discourse and the Professions. The conference, which was organized by Britt-Louise Gunnarsson, Per Linell and Bengt Nordberg, was held under the auspices of ASLA, the Swedish Associa­ tion for Applied Linguistics, and it was one in a series of annual conferences arranged by the Swedish affiliate. This, however, was the first international ASLA conference since 1981, when the 6th AILA World Conference was held in Lund, Sweden. The aim of the conference was to create an international forum for multidisciplinary and multifaceted explorations of the field of professional communication. Because of the ongoing differentiation and specialization of professions, more and more people are becoming aware of the importance of effective communication among organizations and individuals in science, bureaucracies, health care, business, government and administration, in short in practically every area of modern society. Language is, for obvious reasons, one of the most important tools of many professions. The ability to communicate, not only within one's own professional group, but also between different specialist groups and betweeen experts and laymen, is absoiutely vital if society is to function properly. In many professions, oral and written contact with the general public forms the core of professional activity. At the same time, the development of new communications technology has created new channels for professional interaction. Many people's professional work consists of an intricate interaction between human beings and advanced technical equipment and systems. This situation has given rise toa large and diversified body of research along different lines of thought and in different methodological traditions within e.g. text linguistics, the study of LSP, rhetoric, discourse analysis, interactional sociolinguistics, anthropology, sociology and the history of science. However, the general trend has been for these disci plines to pursue the study of profes­ sional language largely in theoretical and methodological isolation from each other. One of the aims of the conference was to contribute to an integration of these various research traditions. In accordance with these aims, the conference covered written as well as oral language, historical development as well as contemporary variation, communi- ii cation between experts as well as between experts and lay persons, communi­ cation within a language community as well as intercultural communication and interpretation. The conference was broad in scope and attracted scholars from different disciplines and fields, linguists, sociologists, psychologists, anthropologists and educationalists. The academic programme comprised four plenary lectures and 99 section papers. All in all, 217 scholars were active in one form or another

Uppsala, May 1994

Britt-Louise Gunnarsson PerLinell Bengt Nordberg Contents

Introductory address When two augurs meet, they smile ... 3 Stig Strö111/wl111

Professional Language across Time and Culture Competition and Discourse Community: Introductions from Nysvenska Studier 9 Kirstin M. Fredrickson & John M. Swales Robert Boyle's Views on the Language of Science 23 Mauri~io Gotti The Historical Development of the Register Barrier in Professional Language 37 Ger~tfrey Hughes Varieties of Psychological Discourse: Mental Reference in Professional and Everyday Language 49 Carl F. Graumann

Writing for Different Audiences and in Different Settings Generic Integrity in Professional Discourse 61 Vijay Bhatia Metadiscourse in Academic Texts 77 Minna-Riitta luukka Explicitness in Marketing and Linguistic Settings 89 Kjersti Flf)ttum Making Economics More Accessible? A Study of the Process of Rewriting an Economics Text fora Wider Readership I 05 Martin Hewings & Diane Houghton iv Discourse in a Technical-Directive Setting: The Case of Operating Instructions 125 Risto Hilt1111e11& Sirkku Nyman Writing in Organisations: The Technical Writer as an Example 137 Karin Mårdsjli

Talk and Interaction in Institutions Negotiating Genre and Power: Questions in Medical Discourse 149 Nancy Ainsworth-Vaughn Two Strategies for Clarification in Doctor-Patient Encounters 167 Ulla Melander Marttala The Social Worker as Moral Worker. Applying for Money- The Moral Encounter Between Social Workers and Clients 187 Lars-Christer Hyden The Treatment of Client Responsibility and Laughter in Encounters Between Social Workers and Clients 201 Ilmari Rostila Communication at the Work Place: Requesting in Asymmetrical Interaction Situations 219 Lenelis Kruse & Caja Thimm The Interaction in Conversations Between Aphasic Persons and Speech Therapists 237 Minna Laakso The Systems Analyst's Interview 257 Helen Tebble INTRODUCTORY ADDRESS

When two augurs meet, they smile ...

Stig Strömholm Vice-Chancellor of Uppsala University

"When two augurs meet, they smile." For more than two thousand years, this Roman saying has been used to indicate, and to characterize, the furtive complicity of humbugs. It must be admitted that for all its brevity, or rather by virtue of its incisive laconism, it describes brilliantly the brief encounter of two people sharing a secret of some importance and realising that partnership in knowledge: the quick exchange of glances, the passing smile, with its complex signal function: "I know, you know, we know that we both know ..." But is there really more to it? Are thc pejorative conclusions inevitable? I do not possess sufficient knowledge about the moral loftiness or baseness of Roman priesthoods in the days of the late Republic to venture any attempt at setting aside the verdict expressed by the time-honoured dictum about the smiling augurs-or rather haruspices, for the ironic phrase, which goes back on a repartee by the elder Cato, really refers to that particular branch of sooth­ sayers, as distinct from the older craft of augurium, based upon observation of the flight of birds. But even in the absence of such knowledge, which few are Iikely to possess today, it seems permissible to give the augurs' smile a more benevolent interpre­ tation than that proposed by Cato. and carried on through the centuries by Cicero. Is the smile of two experts meeting another inevitably dishonest? Does it necessarily express complicity in deceit? Speaking to this most distinguished gathering of experts-for we are all experts, in one field or another; otherwise we would not be admitted to this conference-I realise I am pleading before a partial court when I propose that the augurs' smile might at least be a perfectly acceptable, honest, and even reassuring expression of the satisfaction of experts, who realize when they meet, that they have something in common; the command of a language, of a system of meaningful signals inaccessible to others. The emphasis, in their smile, would be on the one hand, on partnership, on the pleasure of sharing something with a fellow human being. on the other hand on the secrct as such, the exalted. strenuously acquired art of interpreting the signs of the Gods. If that is so, deceit where is thy sting. humbug where is thy victory? Speaking as an expert among experts, I dare submit to you that the Catonian dictum on the smile of the augurs is nothing but a particularly nasty explosion of the jealousy of the vulgar against us experts, against us who command 4 vocabularies and systems of signs framed for the purposes of our art, or craft, or secret-not in order to create obscurity or to deceive, but with an intention to achieve a maximum of clarity and precision in our communication. That purpose can only be realized with a maximum effort of linguistic creativity, logical strictness and structural inventiveness-qualities which the vulgar detest. Hence the Roman word on the smiling augurs. However early the augurs and the haruspices were in the field, as inventors of the specialized professional discourse, inaccessible to others, which isat the roat of their mutual smile, others were before them ... The idea, and the creation, of a set of specialized signs, invented for the purpose of conveying specialized knowledge with that maximum of clarity and precision which experts require, is of course closely connected with the concept of interpretation. In fäet, interpretation, as an art, or a craft, is likely to be even more ancient than the invention of special kinds of discourse and special systems of signs for special professional purposes ... For already at a very early date, it was felt that even everyday language, even the vernacular, the discourse of the vulgar, may convey messages that are not immediately accessi ble and that consequently call fora specific intellectual procedure before their true meaning is revealed. There are, mainly, three fields in which bodies of hermeneutic precepts claiming to be scientific canons for the treatment of texts were developed already in Antiquity, and although the connections between these three doctrines are far from sufficiently investigated, it is hardly subject to doubt that such interrelations, with mutual taking and giving, existed at an early date. The first reasonably well-known framers of a theory of interpretation which claimed to be a coherent "doctrine", or at least an arsenal of elaborate and generally accepted maxims of construction, were those Greek philosophers and grammarians who made use of their analytical acumen in the study of early literature, above all of the Homeric poems. The ideas and solutions developed by these Hellenistic men of science and of letters were adopted, to a !arge extent, by those Jewish rabbis in the diaspora who tried to make the books of the Old Testament the foundation of a theology sufficiently coherent and rational to be acceptable to an educated Hellenistic public. Both directly, in the writings of the philosophically and philologically trained Fathers of the Church, and indirectly, through the intermediary of Jewish writers, and also in connection with the influx of Arabian science in the Middle Ages, this Hellenistic tradition came to cxcrcise a decisive influence upon the second great centre of hermeneutic discussion and system-building: early Christian theology. Indeed, the Hellenistic ars interpre­ tandi became once more a source of inspiration both to philologists and theologians. A new wave of ancient influence rose with the Renascence humanists, such as Flacius in Protestant theology and Bude in classical philology. 5 The third centre of early reflection and construction of theories and solutions is to be found among classical Roman lawyers. That on a general level these lawyers were subject to a strong influence from Hellenistic philosophy and grammar would seem to be beyond reasonable doubt. The precise strength, scope, and consequences of that influence are, however, the subject of contro­ versy. That the art of rhetoric, where problems of interpretation were also treated, was known to classical Roman lawyers would seem to be certain-some questions are much debated among the scholars- but at the same time, the usually very laconic maxims of interpretation which are to be found in the pre­ served legal material have an unmistakably original and independent character. Why was it precisely in the fields of classical philology, theology and law that artes interpretandi-systems aiming at completeness and claiming to be "sciences" in the older sense of a great mass of knowledge, coherence, consis­ tency, and rationality-experienced this early birth and growth? Any reflected and elaborated view of the process of linguistic interpretation presupposes, consciously or unconsciously, certain foundations in the form of at least some ideas concerning a theory of knowledge, logic, psychology, linguistic theory in general, and semantics. These foundations, which were very largely-at least before the introduction of new psychological and linguistic theories in the 17th and 18th centuries-a mixture of ancient speculations and generally accepted common-sense observations, must have been common not only to those theoreticians and practitioners who elaborated hermeneutic systems in the three fields just referred to, but also to the learned world at !arge. Theologians, classical philologists, and lawyers possessed no specialized knowledge of their own in these general matters. However, it was to take a very long time before efforts were made to develop a general theory of hermeneutics without special reference to these three sciences. Traditionally, hermeneutics as such was regarded-to the extent it was discussed at all-as part of theoretical philosophy, or to be more precise, of the logic of probability, but problems of interpretation could also be treated within the growing science of literature, ars critica. Attempts to formulate hermeneutic precepts of a general scope can be noted in the Renascence, inter alia in the writings of Melanchton, but it was not until the middle of the 18th century that the construction of more general theories commenced outside the three traditional branches of science. The theo­ ries which then emerged must be characterized as imprecise and groping and­ as might be expected-were heavily dependent upon the maxims of, in parti­ cular, classical philology. This remained true until well into the 19th century, when Schleiermacher, who could draw upon his theological learning, and a growing number of other scholars with various backgrounds began to lay the foundations of modern hermeneutics. The most obvious observation, when looking for the reasons why theologians, lawyers and students of ancient philology were the first to elaborate complete and coherent sets of hermeneutic principles, is certainly that these three cate- 6 gories of scholars all had to deal with texts which were considered to be author­ itative. That this was the case with Roman law and with the Scriptures is a fäet that needs no comment. It was also very largely true, however, of the Homeric poems and, further on in the course of development, ancient literature as a whole. With some simplification it can be contended that the scholar's con­ ception of his task in relation to these three groups of texts was for a very long time poles apart from what is likely to be considered today the foremost duty of modern critical science: it was not to search for, and demonstrate, contradictions and ambiguities-it was to make loose ends meet, to understand texts which were by definition "true", and to eliminate whatever could be called obscure or ambiguous. The uncontested point of departure was the firm authority of the text; all flaws were laid at the door of the ignorant or impercipient interpreter. Thus normative doctrines of interpretation were bom first, and were elabo­ rated earliest, among those who had to deal with authoritative texts-linguistic utterances which for one reason or another bad to be coherent. These doctrines express the endeavour to rationalize, but also, if possible, to formalize, and thereby to legitimate intellectual operations which, by virtue of the majesty of the text and the possible consequences of the interpretation, were conceived of as being particularly exalted and important. The needs that have called forth, in almost all contemporary fields of science and scholarship, both specialized systems of signs, terms, and formulae, and, concomitantly, equally specialized doctrines of interpretation, are of a different character. They do not reflect the kind of externa! authority which were at the root of ancient hermeneutics as a way out of the dilemmas of understanding and putting into practice authoritative messages. They reflect intellectual authority: The need for specialized linguistic vehicles for communication in a growing number of dwindling areas of knowledge. Be it far from me to trespass upon the land you are going to explore in the following days. Let me only hint at one single problem of a highly general character. It would, in my view, be foolish to question the necessity of profes­ sional discourse developing into more and more highly specialized systems of symbols, signs, and terms. We all know that this is an inevitable development. The best explanation has been given in a short and brilliant formula from the beginning of this century: Une science est une !angue bienfaite. But it would be equally foolish to disregard the moral problem this gives rise to: how, and by whom, is the responsibility for conveying the true message of all this professional knowledge to the common man, our Lord- le menu peuple, notre Seigneur, to quote a Medieval French expression-to be carried out? The importance of that responsibility, and of its proper fulfilment, is beyond doubt. For, as a highly competent and experienced observer says, St. Paul in 1 Cor. 15:24, "If therefore the whole church be come together into one place, and all speak with tongues, and there come in those that are unlearned or unbelievers, will they not say that we are mad?" PROFESSION AL LANGUAGE ACROSS TIME AND CULTURE

Competition and Discourse Community: Introductions from Nysvenska studier 1

Kirstin M. Fredrickson & John M. Swales

In this paper we report on a study of article introductions from recent issues of Nysvenska studier. The text analysis suggests that previous studies of article introductions in other languages and in other disciplinary areas have failed to recognize that competition for journal space among scholars in a particular area may be a variable rather than a given. We tentatively call this variable the hypo­ thesis of discourse community pressure and claim that the extent of this pressure has a number of consequences for the purpose and form of article introductions. As the hypothesis we are presenting is somewhat novel and fairly bold, we first provide a short intellectual history of recent work on research articles. During the last decade the scholarly article or research paper has become an area of intensive study-at least when written and published in English. Evidence of this can be seen in three fairly recent books: Bazerman's Shaping Written Knowledge (1988), Myers' Writing Biology (1990) and Swales' Genre Analysis (1990). Although all three volumes provide social constructionist accounts of the rhetorical forces that have shaped the English-language research article (RA) as we know it today, they differ in their points of departure. Bazerman approaches the research article from the perspective of its rhetorical evolution, Myers views the RA through the Jens of the "discoursal" school in the sociology of science, while Swales is motivated by the pressures placed upon 11011-nativespeakers of English as they make their way into increasingly Anglophone research worlds. Current work has continued to investigate selected disciplinary areas (Henderson & Dudley-Evans 1991, Melander 1991, Näslund 1991, Lindeberg 1992, Gunnarsson 1993), has further clarified genre differences between re­ search articles and more popular accounts (Nwogu 1990, 1991, Melander 1991, Näslund 1991, Gunnarsson 1993), or has expanded our understanding of the complex interplay among rhetorical, pragmatic and Iinguistic choices. This last group includes studies of politeness phenomena in scientific articles (Myers 1989), of subtleties in the seiection of reporting verbs (Thompson & Y e 1991 ), of the distribution and use of marked theme at sentential onsets (Gosden 1992), and of the juxtaposition of proposed and opposed claims (Hunston 1992, 1993).

1 We would like to thank Margaret Luebs for providing helpful comments. 10 Particular attention has been given to the introductory sections of research articles (see Swales 1990 fora review). Overall, this research shows the article introduction in English to be persuasive rather than descriptive, to be lexically rich, to be a distanced reconstruction of the original rationale for the study, to be complexly positioned with regard to tradition and innovation, and to manifest considerable negotiation between author, reader, knowledge, and claim. Attempts at capturing the broader textual movements within introductions include second-storying (Swales 1983), establishing newsworthiness (Huckin 1987), speaking with forked tongues (Jacoby 1987), mitigating face-threatening acts (Myers 1989), and creating a research space (Swales 1990). This last characterization (Swales' CARS mode!) is set up through preferred sequences and combinations of authorial moves and steps. A simplified version is presented below (adapted from Swales 1990: 141): Move 1. Establishing a territory Move 2. Establishing a niche (By one offour steps: A) counter-claiming, B) indicating a gap, C) question-raising, or D) continuing a tradition, where shifting from A through D supposedly indicates a steady weakening of the knowledge claim.) Move 3. Occupying the niche This mode! is interesting because the prevalent metaphor is unabashedly ecological: populations of researchers competing for visibility and resources are similar to populations of plants competing for light and nutrients. In other words, the CARS mode! assumes competition for research space among individual members, research groups or "schools" within any given academic discourse community. There is no doubt, of course, that strong competitive pressures can be shown to exist. They are incontrovertible in reports of funding struggles from major research agencies (e.g. Myers 1990) and have been vividly described in such influential monographs as Latour & Woolgar (1979), Knorr­ Cetina (1981), and Traweek (1988). However, it is equally true that these sociological and anthropological accounts have primarily been based on "big science" in the United States with all its consultancy and review processes, breaking stories, super-fast networks (Olsen 1988), cancerns about being "in the loop", and competition for influence, power, resources and prestige. In contrast, other academic discourse communities, possibly "kinder, gentler" and perhaps more prone to the support of others rather than Schadenfreude, have been under­ researched. We would therefore like to suggest that Swales' CARS mode! has been unduly influenced by these "big science" accounts. In consequence, we would like to argue that the concept of Discourse Community Pressure (DCP) should be treated not so much as a built-in assumption but rather as an important and 11 potentially illuminating variable. It would be both prudent and pertinent to recognize that there is no single highly-competitive ecological system in academia, but rather an interesting variety of ecological situations. More specifically, our working hypothesis is that, ceteris paribus, the greater the competition in a territory (as measurable by number of research papers per topic area, conference/journal acceptance rates, promotion criteria, percentage of funded proposals, etc.) the greater the rhetorical effort authors will have to expend in order to create research spaces for themselves. So, presumably the DCP will be high in very well-tilled areas (Shakespeare studies for example) or in those devoted to solving major human problems (AIDS research) and very low in, say, place name studies in Sri Lanka. We believe that the variable DCP hypothesis can be supported by a number of reasonable initial assumptions. First, the DCP is today likely to be higher in Anglophone scholarly contexts than in those operating in other languages, if only because of the well-attested phenomenon of non-native speakers of English continually trying to break into English-language journals with their "best" research (e.g. Najjar 1990). Second, the primary opportunity in an RA intro­ duction for creating a research space is offered at Move 2 (or in a series ofMove 2s) especially when the "indicating a gap" step-choice is made. It may be here, therefore, that we are most likely to perceive discourse community effects. Third, even in English, Move 2s may be a comparatively recent phenomenon. A close reading of Bazerman's (1988) study of the research article in physics shows, in the broadest terms, that RAs originally began with an unadorned Move 3; later Move 3 began to be prefaced by Move l; and only later still were Move 2s inserted tojuxtapose the old and the new. Similarly,journal articles in economics begin to use the CARS Mode) complete with Move 2s only in the 1960s or so (Dudley-Evans & Henderson 1990). In Swedish, journal articles in medicine show increased adherence to the CARS Mode), including the use of Move 2, after World War II, possibly due to greater Anglo-American influence (Gunnarsson 1990). The originating spark for this study was provided by Olle Josephson at a seminar at FUMS, Uppsala University, in April 1991. The gist of his obser­ vations are as follows:

This article introduction stuff is all very well, hut it doesn 't really apply here in Sweden. There are very few of us working on Modern Swedish; we all know what everybody is doing; there is more important data out there than we can ever hope to get to. As a result we are not competing with each other, and so we don't need to bother with creating a space. These remarks are intriguing. For one thing, they appear to point to an interesting case of very low discourse community pressure among those working on Modern Swedish. We might in consequence expect to find fewer of those elaborated moves and steps typical of more competitive environments. While any such finding would not itself offer confirmatory evidence for the 12 DCP hypothesis, it would at least suggest that part of any explanation of rhetorical variation might lie in discourse community social practices.

Materials The materials for the study consist of journal articles from six recent issues (1985-1990) of the annual journal Nysvenska studier (Studies in Modern Swedish). For reasons of comparability, certain articles were excluded from the study-reviews, very long articles (longer than 50 pages), and articles written in languages other than Swedish (there were two of these, one written in German and one in Norwegian). The remaining articles were included, yielding a total of 26 articles, representing the work of 22 different authors. (The articles are listed in the appendix.) The researchers mostly come from Departments of Nordic/Scandinavian Languages, hut some from linguistics or elsewhere. Most come from Sweden, hut several from Finland and one from Poland. Nysvenska studier publishes reports of research on the Swedish language. In terms of fields, the research can vary rather widely, from etymology to generative syntax to discourse analysis; as such, the authors often come from rather different theoretical backgrounds. However, as the journal is intended for modern Swedish, articles in runology and philology of older Swedish and Scandinavian tend not to be published here; such articles find their homes in other journals, such as Arkiv för nordisk filologi. The journal does not have a style sheet, and the forms chosen by the authors vary somewhat. The journal has a circulation of about 250 subscribers, many of which are libraries and institutions, so that the journal actually reaches considerably more than 250 individuals.2 Additionally, articles from Nysvenska studier are often assigned as required reading for academic courses, even undergraduate courses. Even so, it can certainly be said that the readership is considerably smaller than, for example, Language, which is distributed to each of the members of the Linguistic Society of America as well as to many Iibraries and institutions, nearly 7,000 subscribers in all.

Metbodology Each of the 26 articles was coded for moves according to the three moves of the C:reate a Research Space modet (CARS) for introductions (Swales 1990). Text parts that seemed not to fall into any of the moves were duly noted. To help

2 With the 1991 issue, the journal returned to its original title of Språk och stil (Language and Style). A style sheet and a referee system have been introduced, and the number of subscribers has increased by 45% to about 360. 13 ensure consistency of coding, each article was actually coded twice by the first author, with a six-week interval between the two codings, the second of which were then corroborated by the second author. Discrepancies were very few and were easily resolved. The articles were also classified as to sub-field area and the type of material used for data; for example, Iiterature only, narratives told in dialect, or a smattering of types of sources. The classifications used were taken from Josephson's (1989) study of the history of Nordic language research as reflected in Nysvenska studier and its predecessor Språk och stil over the twentieth century. It tums out that the article introductions do not show any detectable pattern in terms of either sub-field or material type; in other words, neither of these variables seems to influence the type of rhetoric chosen by the author. Therefore, this issue will not be discussed further.

Results The results for each text are summarized in Table 1. (See Appendix for a list of the articles in the corpus.) The S stands for story or story-like description and S' for modification of story, on which more later. A number in parentheses means that the move is embedded within the move denoted by the preceding number. As can be seen, all three moves are found within the corpus, and five texts exactly follow the 1-2-3 sequence predicted by the CARS mode!. Another nine follow a variant of the CARS mode! in that they have all three moves in the expected order, but have one or more moves repeated (for example text no. 23). These two groups of "normal" introductions constitute 14 of the 26 texts, or just over half, a rather low fraction in comparison to the other results reported in the literature (e.g. Swales & Naljar 1987). A text which follows the CARS mode! is no. 18, the second half of which is given below (with an English translation added). As might be expected, Move 2 shows that there are questions which are unanswered, and Move 3 offers an answer and indicates the content of the present article. 14 Table I. Summary of results Text number Moves 1 S-S'-1-3-2-1-3 2 1-3 3 s 4 1-2-3 5 1-3 6 1-2-3 7 l(S)-1-2-3 8 3-1 9 1-3 10 1-2&3 11 no introduction 12 l(S)-2&3 13 1-2-3 14 1-3-1-3-1-3 15 2-1-2-3 16 l(S)-3-2&3-3 17 S-S'-2 18 1-2-3 19 1-3 20 no introduction 21 1-3(2) 22 1-3-2-3-1 23 1-2-1-3 24 S-1-2 25 1-2-3 26 1(2)-1-2-3

(1) Standard moves (text no. 18)

Move 2 Riktig svenska (1973) lämnar åtskilliga frågor obesvarade både när det gäller förledsbestämningen som sådan och när det gäller sättet att behandla denna.

[Riktig svenska (1973) leaves many questions unanswered both conceming modifiers of first elements of compound nouns as such and conceming the way to treat them.]

Move 3 Eftersom båda problemkomplexen intresserar mig (se vidare 1.3) ger jag nu först en sammanfattning av de viktigaste inläggen i debatten kring förleds bestämning.

[Because both problem complexes interest me (see 1.3), I will first give a summary of the mast important contributions to the debate on modifiers of first elements in compound nouns.]

(From: Söderberg, Barbro, 1987: Svensk språkvård och ofödda lammskinn. Om förledsbestämning: ett språkvårdsfall. Nysvenska Studier, 67:89-122.) 15 Table 1 also shows that many of the texts deviate from the CARS Model. Twelve of the 26 texts (46%) are missing one or more moves, either entirely or, in the case of text no. 21, found only embedded within another move. Of these 12 texts, two actually have no discernible, independent introduction at all, and one (text no. 3) has an introduction which falls outside the CARS move system; that is, its introduction does not display any of the three moves. Nine texts (35%, quite a high percentage) do not contain a Move 2, Establish­ ing a niche. An example of an introduction lacking Move 2 is text no. 5, found in (2) below. This introduction is quite short and is reproduced in its entirety.

(2) Missing Move 2 (text no. 5)

Huvudregeln vid verben brinna och bränna i svenska är att brinna, brann, brunnit med stark böjning används intransitivt, I.ex. i meningen Elden brann friskt, medan däremot bränna, brände, bränt med svag böjning är transitivt, exv. Jag brände löv i trädgården. Från denna regel finns en avvikelse. Ibland används nämligen former av bränna i intransitiv betydelse, såsom i Ljusen bränner stilla. Denna avvikande användning av bränna är även känd från den svenska litteraturen och skall här något beskrivas. (Wiktorsson, Per-Axel, 1990: Det intransitiva verbet bränna i svensk litteratur. Nysvenska studier, 70: 143-158.)

[The main ru!e with the verbs brinna ['to bum'] and bränna ['to bum'] in Swedish is that brinna, brann, brunnit ['bum, bumed, burned'] with irregular conjugation is used intransitively, e.g. in the sentence Elden brann friskt ['The fire burned fierce!y'], while bränna, brände, bränt ['burn, bumed, burned'] with regu!ar conjugation is transitive, e.g. Jag brände löv i trädgården ['I burned leaves in the garden']. There is same deviation from this rule, namely that forms of bränna are sometimes used intransitively, as in Ljusen bränner stilla ['The lights burn quietly']. This deviant use of bränna is also known from the Swedish literature and shall be described somewhat here.] As is typical of the introductions missing Move 2, there is in fäet an opportunity where a Move 2 could have been developed in this introduction hut was not. Where the author says that there is a deviation from the normal rule governing the usage of brinna and bränna, he could have gone on to say either that this deviation has not been noted before, or that this deviation has been noted in passing hut that there is an inadequate understanding of where and why this deviation occurs. One of the more striking features of the corpus is that seven articles feature an opening story oi a story-like description of the phenomenon to be studied in the article. In some cases, for example text no. 7, the story is a variation of Move 1, Establishing a territory. In other cases, however, the story fulfi!ls some function that lies outside of the CARS mode]. An example of such a story is found in text no. 17, reproduced as (3) below. 16 (3) Story (text no. 17)3 I januari 1988 kunde man på många tidningars löpsedlar och förstasidor läsa rubriker som »33-åringen till nya förhör». Ingen av oss som våren 1986 följde dramatiken efter Palmemordet fann rubriksättningen märkvärdig. Den majoritet av tidningar som den gången valde iltt inte gå ut med mannens namn och foto hjälptes i stället åt med att göra »33-åringen» till ett begrepp - ja nästan ett egennamn. När mannen nu två år senare på nytt dök upp i rubriker och spalter blev detta bekymmersamt eftersom han inte längre var.33 år utan 34. Det vållade antagligen en del huvudbry på landets tidningsredak­ tioner. Man löste problemet på olika sätt. Några framhärdade i att peka ut honom som 33-åringen; andra döpte om honom till 34-åringen - fast fortfarande i bestämd form; åter andra tog sig ur konflikten genom att hitta på ett nytt namn, I.ex. Cafemannen. Den sista lösningen vittnar om framsynthet. Är det nämligen någonting som dagspressens rubriksättare bör ha lärt av den här historien skulle det väl vara att inte benämna det som ännu lever och åldras med hjälp av en så obeständig egenskap som åldern eller graden av aktualitet i själva namngivningsögonblicket. Lever och åldras gör på sätt och vis också språk ... (Thelander, Mats, 1987: Nu eller alltid? Perspektiv på 110 års nusvenska. Nysvenska studier ,67:5-14.) [In January 1988, one could read headlines on both the front and inside pages of newspapers, such as "The 33-year-old interrogated again". None of us who in the spring of 1986 followed the drama after the murder of [Prime Minister Olof] Palme found this type of headline remarkable. The majority of newspapers which at this time chose not to come out with the man's name and photo were instead aided in identification by making "the 33-year-old" into a concept-yes even a proper name. When the man again showed up in headlines and stories 2 years later, it was troublesome since he was no longer 33 but 34. Presumably this caused quite a few headaches among the country's newspaper editors. They solved the problem in various ways. Some persisted in referring to him as the 33-year-old; others resolutely rebaptized him as the 34-year-old-though still in definite form; still others look themselves out of the conflict by finding a new name, e.g. the Cafe man. The latter solution attests to foresightedness. If there is something which the daily headline­ writers should have leamed from this story, it is not to name something still living and aging with the help of such an impermanent property as age or degree of currency at the moment of naming. Languages also live and age in their own way ...] While many of the texts lack Move 2 and several of the texts feature an additional "move" (a story), it should be noted that it is not the case that Move 2 has simply been replaced by the story option. Of the seven introductions with a story or story-like part, only one is missing Move 2, and that introduction does not display any of the CARS moves. Also, when there is a story, it always comes as the first part of the introduction whereas Move 2 generally comes in the middle of the introduction, after Move 1. Rather it seems that the story serves a different function, one which is not incompatible with that of Move 1.

3 We are aware that this particular article started off life as an inaugural lecture given to a university. However, it surely cannot be true that all the articles featuring stories have a similar spoken provenance. 17 Discussion The !arge number of articles lacking a Move 2 and the presence of several stories raise a number of interesting questions: 1. Can these phenomena be associated with the disciplinary norms of the language sciences? 2. Can they be associated with the fäet that Swedish is not a major scholarly language (Baldauf & Jernudd 1987)?

3. If neither of the above applies, what alternative explanation could there be? One possibility, of course, is that this is just the way language researchers write. Most of the other studies of article introductions have used material from science, engineering or medicine. It is possible that language scholars simply use a different rhetoric from that of scientists. To test this possibility we examined all 18 articles from the 1987 volume of Language and found that 16 of the 18 followed the CARS mode! rather closely, one fronted Move 3, and one was anomalous. None of the articles featured any kind of story. It thus looks as though it is not language specialists in general, but rather the sub-group writing in Swedish-and perhaps in socially comparable languages-who use opening rhetorics of the sort found here. In order to assess the second hypothesis, we have re-examined the literature to check the occurrence of Move 2s and to see whether alternative Move Is are used in other languages. Forty of the 48 English-language article introductions originally analyzed by Swales (1981) contained a clearly identifiable Move 2 (83%) with relatively Iittle variation across the main disciplinary areas. In contrast, Lopez (1982) found only 12 instances of Move 2 in a corpus of 21 articles from Latin America published in Spanish (53%), the articles being again drawn from a range of fields. Najjar (1990) investigated 48 article introductions in Arabic all taken from the agricultural sciences. Of these, he classified 27 as broadly fitting the CARS mode!, 18 as having a technical report problem­ solution format, and three remained unclassified. Of the 27, only 13 had a Move 2, on which he makes the following observations:

More interestingly at this point, it can be seen that none of the gap-statements in our corpus apparently offers an evaluation of the previous research. There are no challenges to the work of others, hut rather a simple indication of the scarcity or absence of local research on the present author's topic or purpose. The closest that any of the present statements comes to an evaluation is the vague statement in B9 (see above) which claims that "our present knowledge (i.e. in Egypt) about [ ...] is not totally c!ear" (pp. 104-5). The final directly relevant study we know of is Taylor & Tingguan's (1991) analysis of science papers written in English by Anglo-American researchers 18 (1 I introductions), in English by Chinese researchers (10 introductions) and in Chinese by Chinese researchers (again 10 texts). Here the results are very different. This is because the main differences among the three sub-groups lie not in Move 2, hut elsewhere, in the handling of the previous literature, as the Chinese schalars-in English and even more so in Chinese-tended to omit or truncate this step. Although Taylor & Tingguan admit that this finding might simply be ascribed to lack of bibliographic resources, they also offer a number of interesting speculations:

There might well be an answer in terms of preferred rhetorical structures. This is that Chinese scholars find it less acceptable to identify by name and to summarize the work of others whom they will then proceed to "expose", as it were, in the following move [... ] , which focuses on the gaps or shortcomings of previous research. (p. 331.) The available literature on article introductions in Spanish, Arabic and Chinese thus seems to suggest some avoidance of face-threatening acts by, in the first two cases, downplaying opportunities for establishing-rhetorically at least­ gaps in current knowledge, and in the last by refraining from directly citing others whose work may be characterized as incomplete in the subsequent Move 2. What we have not seen so far are any alternative new strategies to replace or supplement elements which might be judged to be missing or atrophied from the perspective of the Anglophone schema. However, the Swedish data do present such a strategy, the opening story, thus leading to the question of why severai members of this particular discourse community feel the need to provide the other members with entertaining stories. The answer, we suggest, may well lie in the small size of the group. The stories may entice the students who read the articles for ciass to join the group. Additionally, if a particular article appeals only to specialists in one of the several sub-fields represented by the journal, the number of readers who are already full-fledged members of the discourse community will likely be very small. We therefore suspect that the function of the approachable and attention­ getting opening stories is to create an opportunity for authors to attract (at least initially) a readership spanning several sub-fields. Language specialists writing in Swedish may not face competition in terms of research space, hut they do face competition of another sort-competition for readership. The implications of this study for our wider understanding of the forces that shape academic discourse are quite substantial. First; it looks as though the CARS mode!, or its variants, is a construct perhaps over-influenced by corpora derived from highly competitive Anglophone science, technology and medicine. Second, the study has consequences for the field of contrastive rhetoric. Over the last decade this field has rightly moved beyond stylistics and textual rhetoric to include the investigation of such areas as educational and cultural traditions, as in Mauranen's (1993) exploration of the use of metadiscourse in English and Finnish economics articles. Our study suggests that discourse community size 19 may also be relevant, both in terms of a variable need to create a research space and in terms of a variable one to capture .a readership. Broader characterizations depicting language status and established rhetorical norms thus appear to be substantially moderated by the micro-situation of author X writing on topic Y for journal Z. Nor is this study without its practical implications. As authors or potential authors, we all may need to understand that changing size of discourse community may affect our rhetorical practices just as much as switching from one community to another. And if that is the case, it surely follows that differential effects of differing community pressures should in some small way be worked through in our training and teaching materials.

References Baldauf, Richard Jr. & Jernudd, Björn H., 1987: Academic communication in a foreign language: The example ofpsychology. ARAL, 10:98-117. Bazerman, Charles, 1988: Shaping Written Knowledge: The Genre and Activity oj the Experimental Article in Science. Madison, WI, University of Wisconsin Press. Dudley-Evans, Tony & Henderson, Willie, 1990: The organisation of article introductions: Evidence of change in economics writing. In Tony Dudley­ Evans & Willie Henderson (eds.), The Language oj Economics: The Analysis oj Economics Discourse. (ELT Documents 134.) London, Modern English Publications in association with The British Council, pp 67-78. Gosden, Hugh, 1992: Discourse functions of marked theme in scientific research articles. Englishjor Specific Purposes, 11:207-224. Gunnarsson, Britt-Louise, 1990: Makrotematiska och pragmatiska mönster i medicinska artiklar. In Svenskans beskrivning, 17, ed. by Erik Andersson & Marketta Sundman. Åbo, Åbo Academy Press, pp 97-114. Gunnarsson, Britt-Louise, 1993: Pragmatic and macrothematic patterns in science and popular science: A diachronic study of articles from three fields. In M. Ghadessy (ed.), Register Analysis: Theory and Practice. London, Pinter Publishers, pp 165-179. Henderson, Willie & Dudley-Evans, Tony, 1991: The new rhetoric and discourse analysis: The case of economics. Paper presented at the Inter­ national Seminar on the New Rhetoric and Discourse Analysis: The Case of Economics, Birmingham, England, April 1991. Huckin, Thomas, 1987: Surprise value in scientific discourse. Paper presented at CCCC, Atlanta, GA, March 1987. Hunston, Susan, 1992: Professional conflict: The management of disagreement in academic discourse. Paper presented at the International Conference on Discourse and the Professions, Uppsala, Sweden, August 1992. 20 Hunston, Susan, 1993: Evaluation and ideology in scientific writing. In M. Ghadessy (ed.), Register Analysis: Theory and Practice. London, Pinter Publishers, pp 57-73. Jacoby, Sally, 1987: References to other researchers in literary research articles. ELR Journal, 1:33-78. Josephson, Olle, 1989: Från Ruben G:son Berg till Svenskans beskrivning. Språk och stil och Nysvenska studier 1901-1975. Nysvenska studier, 69:91- 116. Knorr-Cetina, Karen, 1981: The Manujacture of Knowledge. Oxford, Pergamon. Latour, B. & Woolgar, S., 1979: Laboratory Life: The Social Construction oj Scientific Facts. Beverly Hills, CA, Sage. Lindeberg, Ann-Charlotte, 1992: Article introductions in economic sciences-a pilot study. Paper presented at the International Conference on Discourse and the Professions, Uppsala, Sweden, August 1992. Lopez, G. S., 1982: Article lntroductions in Spanish: A Study in Comparative Rhetoric. M.Sc. Dissertation, Language Studies Unit, University of Aston at Birmingham. Mauranen, Anna, 1993: Contrastive ESP rhetoric: Metatext in Finnish-English economics texts. Englishfor Specific Purposes, 12:3-22. Melander, Björn, 1991: Innehållsmönster i svenska facktexter. (Skrifter utgivna av Institutionen för nordiska språk vid Uppsala universitet, 28.) Uppsala, Sweden, Uppsala University. Myers, Greg, 1989: The pragmatics of politeness in scientific articles. Applied Linguistics, 10:1-35. Myers, Greg, 1990: Writing Biology: Texts in the Social Construction oj Scientific Knowledge. Madison, WI, University of Wisconsin Press. Najjar, Hazem, 1990: Arabic as a Research Language: The Case oj the Agricultural Sciences. Unpublished Ph.D. dissertation. University of Michi­ gan, Ann Arbor .. Nwogu, Kevin Ngozi, 1990: Discourse Variation in Medical Texts: Schema, Theme and Cohesion in Projessional and Journalistic Accounts. (Mono­ graphs in Systemic Linguistics, 2.) Nottingham, England, Department of English Studies, University of Nottingham. Nwogu, Kevin Ngozi, 1991: Structure of science popularizations: A genre­ analysis approach to the schema of popularized medical texts. English for Specific Purposes, 10: 111-124. Näslund, Harry, 1991: Referens och koherens i svenska facktexter. (Skrifter utgivna av Institutionen för nordiska språk vid Uppsala universitet, 29.) Uppsala, Sweden, Uppsala University. Olsen, Leslie A., 1988: Computer-based proposal creation in the American scientific community: Collaboration practices and system design require­ ments. Proceedings for the I 1th International Symposium an Computers af the University, Vol. I. Dubrovnik, Yugoslavia, pp 1-11. 21 Swales, John, 1981: Aspects oj Article lntroductions. Birmingham, England, Language Studies Unit, University of Aston at Birmingham. Swales, John, 1983: Developing materials for writing scholarly introductions. In R. R. Jordan (ed.), Case Studies in ELT. London, Collins, pp 188-200. Swales, John, 1990: Genre Analysis: English in Academic and Research Settings. Cambridge, Cambridge University Press. Swales, John & Najjar, Hazem, 1987: The writing of research article intro­ ductions. Written Communication, 4:175-192. Taylor, Gordon & Tingguan, Chen, 1991: Linguistic, cultural and subcultural issues in contrastive discourse analysis: Anglo-American and Chinese scien­ tific texts. Applied Linguistics, 12:319-336. Thompson, Geoff & Ye, Yiyun, 1991: Evaluation in the reporting verbs used in academic papers. Applied Linguistics, 12:365-382. Traweek, Sharon, 1988: Beamtimes and Lifetimes: The World oj High Energy Physicists. Cambridge, MA, Harvard University Press. 22 Appendix

1. Adelswärd, Viveka, 1990: Vems är poängen? En empirisk studie av den personliga berättelsen i institutionella samtal. Nysvenska studier, 70:71-94. 2. Lange, Sven, 1990: Om ordval och synonymik i äldre och yngre nysvenska texter. Nysvenska studier, 70:95-114. 3. Salberger, Evert, 1990: Spunnit turkiskt garn. Ett ställe i J. A. Wadmans Dryckesvisa. Nysvenska studier, 70: I 15-122. 4. Widmark, Gun, 1990: Bra - lånord med historia. Nysvenska studier, 70: 123-142. 5. Wiktorsson, Per-Axel, 1990: Det intransitiva verbet bränna i svensk litteratur. Nysvenska studier, 70: 143-158. 6. Josephson, Olle, 1989: Från Ruben G:son Berg till Svenskans beskrivning. Språk och stil och Nysvenska studier 1901-1975. Nysvenska studier, 69:91-116. 7. Sigurd, Bengt, 1989: Om »annan», »till», »också» och andra tilläggsmarkörer. Nysvenska studier, 69: 117-126. 8. Törnudd-Jalovaara, Elisabet, 1989: Skall och svenskans tre modus. Nysvenska studier, 69:127-148. 9. Hansson, Eva, 1989: Mäss ik a we litn 'medan jag har varit liten'. Om ett avvikande bruk av perfektum i Orsamålet. Nysvenska studier, 69: 149-172. 10. Josephson, Olle, 1989: Än en gång på stubben. Nysvenska studier, 69: 181-186. 11. Ekbo, Sven, 1988: Korta ordstudier. Nysvenska studier, 68:5-18. 12. Nyholm, Leif, 1988: »Ödmiukast til Protocollet». Ett stilistiskt experiment med Labov och Quintilianus. Nysvenska studier, 68: 19-40. 13. Olsson, Magnus, 1988: Om årtalsbenämningar. Nysvenska studier, 68:41-56. 14. Strzelecka, Elzbieta, 1988: Svenska rumsliga verbpartiklar och deras polska motsvarig­ heter. Nysvenska studier, 68:57-74. 15. Wijk-Andersson, Elsie, 1988: Individuella reflexioner om generella betydelser. Artikel­ användningen i några texter med generell referens. Nysvenska studier, 68:75-86. 16. Cederberg, Ann, 1988: Läsbarhet och begriplighet på kultursidan. En analys av sex artiklar på kultursidan i Svenska Dagbladet. Nysvenska studier, 68:87-122. 17. Thelander, Mats, 1987: Nu eller alltid? Perspektiv på 110 års nusvenska. Nysvenska studier, 67:5-14. 18. Söderberg, Barbro, 1987: Svensk språkvård och ofödda lammskinn. Om förledsbestäm• ning: ett språkvårdsfall. Nysvenska Studier, 67:89-122. 19. Larsson, Kent, 1987: Kellgrens Nytt Försök til orimmad Vers. Nysvenska studier, 67: 123-138. 20. Jonsson, Hans, 1987: Växlingen mellan normbestämd och neutral innebörd hos vissa adjektiv. Nysvenska studier, 67: 139-154. 21. Gunnarsson, Britt-Louise, 1987: Textmönster i vår äldsta vetenskapliga tidskriftsprosa. En analys av medicinska rön i Vetenskapsakademiens Handlingar 1750-1769. Nysvenska studier, 67: 155-180. 22. Thelander, Mats, 1986: Norm, bruk och variation i en svensk 1700-talsgrammatik. Nysvenska studier, 66:43-88. 23. Platzack, Christer, 1986: Inverterad objekt med infinitiv. Nysvenska studier, 66:89-112. 24. Sigurd, Bengt, 1986: Söka, finna, träffa på. Nysvenska studier, 66: 113-124. 25. Thelander, Mats, 1985: Inte - lite(t) - mycket. Skriftspråklig variation i historisk belysning. Nysvenska studier, 65:57-76. 26. Randquist, Madeleine G., 1985: Stil, stiltermer och stilistik. Ett bidrag till diskussionen om begreppet »stil». Nysvenska studier, 65: 159-180. Robert Boyle's Views on the Language of Science

MaurizioGotti

This paper will take into consideration Robert Boyle's attitude towards the language to be used for scientific purposes and will identify both the relationship that such a language should maintain with the principles of scientific research and the linguistic features that should implement such criteria. These linguistic aspects were central to Boyle's scientific activity, and his great interest in the problems of language and style is often manifest in his works. Boyle 's interest in linguistic matters derives from the consideration - common to most scientists of that period 1-that the English language appears to be inadequate for the expression of specialized concepts as it is not sufficiently equipped with suitable vocabulary and because its style is deemed not to match the epistemic and methodological criteria adopted in scientific research. In particular, what is considered especially evident is the inadequacy of several words used in the English language to express specialized concepts due to the vagueness of their referents, which makes them particularly ambiguous when employed in scientific publications. This is mainly due to the fäet that several traditional concepts have gone through a process of redefinition and reconside­ ration as a consequence of the rapid, continual developments in the various scientific fields. The use of such words, therefore, is to be avoided and discouraged, as it may lead interlocutors to confusion and difficulty of comprehension. Moreover, scientists consider the existence of such words a danger, as they may lead experts to reciprocal misunderstanding and unnecessary conflict. This opinion is shared by Boyle, who believes that even words commonly used by English speakers should be reconsidered or avoided when dealing with the concepts they stand for in specialized contexts. The following, for example, is Boyle's comment on the use of the word nature and others considered equally ambiguous:

I On the considerations of 16th and 17th-century British scientists on the inadequacy of the English language lo be used as a suitable tool for specialist publications and on the proposals to improve such a tool cf., among others, Sprat (1667), Birch (1756), Christensen ( 1946), Jones ( 195 I, I% I), and Viekers ( I 985). 24 On this occasion, I must not forbear to lake notice, that the unskilful use of terms of far less extent and importance, and also less ambiguous, than the word nature is, has been, and still is, no small impediment to the progress of sound philosophy. (A Free lnquiry info the Vulgarly Received Nation oj Nature 5:247,2) Boyle is conscious of the fäet that the change of signata for the same signans is a consequence of the continuous readjustment within the scientific field due to new principles and recent discoveries, and the remedy that he suggests consists in the·coining of new terms providing stricter delimitation of meaning:

And I confess I could heartily wish, that philosophers, and other leamed men (whom the rest in lime would follow) would by common (though perhaps tacit) consent, introduce some more significant and less ambiguous terms and expressions in the room of the too licentiously abused word nature, and the forms of speech, that depend on it; or would, at least, decline the use of it as much as conveniently they can; and where they think they must employ it, would add a word or two, to declare in what clear and determinate sense they use it. (A Free lnquiry info the Vulgarly Received Nation oj Nature 5: 168.) As can be scen from this quotation, Boyle identifies an oligarchy of "philosophers and other learned men" required to reformulate the contents of the vocabulary that scientific discourse needs, in order to provide meanings and forms that all specialists would then adopt. The solution is therefore to be found in the creation of "more significant and less ambiguous terms and expressions" that would replace the "too licentiously abused" ones. In order to avoid writing in a "canting" language, 3 it is also very important for the single writer to specify the "clear and determinate sense" in which he makes use of his terms. In this way the use of the language would be rendered objective and unambiguous. Another suggestion put forth by Boyle is "to alter words, that they may better fit the nature of things" (The Origin oj Forms and Qualities 3:41). However, not all the words needed for specialist expression could be derived from existing vocabulary, due to ·the lack of scientific terms in the English language. One of the quickest ways to increase lexis4 was the coinage of words derived from Latin. Robert Boyle himself recurred to this process of word formation, as he confirms in The Sceptical Chymist, where he makes his persona Carneades explain the recourse to such a process when reporting the arguments of the alchemist Sennertus:

2 All the quotations from Boyle's works mentioned in this paper (unless stated otherwise) are taken from Boyle (1772/1%5). 3 The adjective canting is used by Boyle as a synonym of ambiguous or obscure; indeed, in the preface to The Sceptical Chymist he writes that he will not take other arguments into consideration "if any impertinent person [... ] shall write against them in a canting way, I mean, shall express himself in ambiguous or obscure terms, or argue from experiments, not intelligibly enough delivered". (The Sceptical Chymist 1:461.) 4 For a more detailed analysis of the processes followed in the growth of the scientific vocabulary of the English language in the 17th century cf. Gotti ( 1992). 25

I therefore thought fit to retain the language, wherein the author proposes it, that I might also retain the propriety of some Latin termes, to which I do not readily remember any thatfully answer in English. (The Sceptical Chymist 1:551.) Some of the latinate words that Boyle uses in this text are: acetous (l :532), aqueous (1:479), calcination (1:543), colliquated (1:479), comminution (1:478), congelation (1:490), distillation (1:537), dulcification (1:543), factitious (1 :479), fasion (l :543), gradation (1:489), homogeneal (1:478), incinerated (1:486), injlammability (1:532), mistion (1:504), oleagenous (1:479), opacous (1 :490), porosity (l :546), sequestration (I :489), terrestrial (I :486), tincture (1 :537), unctuous (l :532). As well as borrowing foreign words, to refer to newly-discovered elements or phenomena the specialist often recurred to the coinage of completely new terms to indicate them. However, when redefining the properties of already-studied compounds or fäets, the scientist would frequently continue to employ existing vocabulary in spite of the fäet that the use of an old term to refer to a new concept would create a mismatch between the existing signans and its new signatum. Boyle himself followed this practice. In fäet, when reporting his pneumatic discoveries in New Experiments he specifies the concept of a totally void space by making use of the existing term vacuum, which was commonly employed by the English-speaking community to refer to the amount of air present in an exhausted receiver.5 However, he tried to avoid problems of ambiguity by specifying the context of reference to which his terms should be attributed.6 For example, in using the term vacuum, to which many of his contemporaries continued to attribute its traditional metaphysical characteristics, Boyle was careful to specify the physio­ logical meaning which he conferred to it:

By which I here declare once for all, that I understand nota space, wherein there is no body al all, hut such as is either altogether, or almost totally devoid of air. (New Experiments 1: JO.) Despite his comments on and examples of the correct use of words, Boyle was also aware that often the ambiguity in the interpretation of a term derived from the inconsistency in its use on the part of the scientist and from his idiosyncratic behaviour. This bad habit is clearly pointed out by Boyle:

(Writers) do not always mean the same things by the same terms, hut some imploy them in one sense, others in another, and sometimes the same writer uses them in very different senses. (The Origin oj Forms and Qualities 3:6.)

5 For an analysis of Boyle's language in dealing with his pneumatic experiments cf. Shapin (1984) and Shapin & Schaffer (1985). 6 The clarity and appropriacy of Boyle's use of terminology is confirmed by William Petty, who believes that he "understand(s) the true use and signification of words, whereby to register and compute your own conceptions." (Quoted by Jones 1951: 149 n. 18.) 26 In spite of this criticism, Boyle himself is not completely free from blame. In fäet, in his works he sometimes makes an inconsistent use of terminology, as in New Experiments, where he refers to the heaviness of the cover of the empty vessel as "spring of the externa! air", "force of the interna! expanded air and that of the atmosphere" or "pressure". In the same text the term pressure is often used alternately withprotrusion (e.g. 1:11) or to refer both to spring and weight (e.g. 1:18, 76). Boyle is conscious of this inconsistent behaviour, but justifies it by stating that in using the various terms different people commonly employ to refer to the same concept the writer can be more easily understood by a wide reading public:

And this put me in mind of intimating, that whereas, on the contrary, I sometimes imployed variety of terms and phrases to express the same thing, I

7 For the identification of Boyle as the author of this anonymous essay cf. Rowbottom ( 1950). The text of the essay was reprinted in Annals oj Science XVII, 1961, pp. 165-173 edited by R.E.W. Maddison. 27 (1:460). Moreover, "they deliver their hypotheses as darkly as their processes; and it is almost as impossible for any sober man to find their meaning, as it is for them to find their elixir" (1:470-1). In the same work Boyle points out another strict relationship, establishing an equation between the author's clarity of language and his honesty of behaviour. As an example of this questionable behaviour he once again quotes alchemists, whom he accuses of using obscure language to mask the inaccuracies and unreliability of their principles:

(They) write thus darkly, not because they think their notions too precious to be explained, but because they fear, that if they were explained, men would discern, that they are far from being precious. (1:521.) Boyle's conclusion is that "when they pretend to teach the general principles of natura! philosophers, this equivocal way of writing is not to be endured" (1:521). The solution he puts forth is that "judicious men, skilled in chymical affairs, shall once agree to write clearly and plainly of them, and thereby keep men from being stunned, as it were, or imposed upon by dark or empty words" (1:522).

Boyle and universal languages The dissatisfaction with the existing language for the expression of scientific concepts and the opinion that the English tongue is inadequate for specialist communication led many 17th-century men of science to take up a radical position and put forth the suggestion that rather than accepting and coping with the language as it actually is or working towards its improvement and the enlargement of its lexis, specialists could solve the problem of accurate and efficient communication in the scientific environment only by the creation of a completely new code of expression, free of all the faults that existing language possesses and endowed instead with all the features that its functions require. The features of this new language are inspired by Bacon, and imply the adoption of "real characters, not nominal; characters [ ...] which represent neither letters nor words, hut things and notions" (De augmentis VI, i). This use of "real characters" would not only have the advantage of being more precise, hut could also be employed at an international leve!. In fact, as Bacon continues, "insomuch that a number of nations whose languages are altogether different, hut who agree in the use of such characters (which are more widely received among them), communicate with each other in writing; to such an cxtent indeed that any book written in characters of this kind can be read off by each nation in their own language." 28 In agreement with Bacon's proposal is Wilkins' elaboration of a universal language, and other similar projects suggested by Delgarno and other members of the Royal Society.8 Boyle is in favour of the elaboration of this new code, which he sees as a very useful tool for international communication, and in a letter to Hartlib writes:

If the design of the Real Character lake effect, it will in good part make amends to mankind for what their pride !ost them at the Tower of Babel. And truly, since our arithmetical characters are understood by all nations of Europe the same way, though every several people express that comprehension with its own particular language, I conceive no impossibility, that opposes the doing that in words, that we see already done in numbers. (1 :XXXVII-XXXVIII.) The advantage of this new tool is particularly identified in the strict reduction of the function of words to their mere denotational value, in which the relationship between res and verbum is established on a one-to-one basis. This ideal situation is envisaged in the analogy to mathematical plainness, in which words stand as mere unambiguous symbols of items of reality. 9 Another great advantage of this new universal code is the compactness of its elements, which represents another of the main pragmatic criteria characterizing specialized languages. 10 This highly-appreciated feature is pointed out by Boyle himself when commenting on Delgarno's project:

(His Universal Character) will certainly bee far more stenographical than any short writing. (Quoted by Turnbull 1952: 119.) In spite of the advantages that such a new code would guarantee, Boyle remains however somewhat sceptical about the implementation of the projects put forth for its realization. In fäet, when he praises the validity of the various schemes, he also expresses doubts on their adoption on a !arge scale, and, while commenting on Delgarno's proposal, writes:

8 On the subject of universal languages cf., among others, Slaughter (1982). 9 This need for a new scientific language, quite different from verbal language and mainly based on mathematical principles is often pointed out by scientists of the 16th and 17th centuries. Cf., for example, Galileo:

Philosophy is written in this grand book, the universe, which stands continually open to our gaze. But the book cannot be understood unless one first learns to comprehend the language and read the letters in which it is composed. It is written in the language of mathematics and its characters are triangles, circles and other geometric figures without which it is humanly impossible to understand a single word of it. (Galilei 1623/1957:237-238.)

10 For an analysis of the main pragmatic criteria of specialized languages and their relationship with the surface manifestations of such languages cf., among others, Gotti (1991). 29 Whether it will bee or how far a Real character to all nations that cannot bee yet determined. (Quoted by Turnbull 1952:119.) The same attitude is evident as regards Wilkins' scheme; he appreciates the project, but is sceptical about its applicability:

The Universal Character Dr. Wilkins has now brought to that perfection, that hee is pleased to promise mee some lime next week to shew it to Dr. Ward and mee together, that it may bee jointly considered by vs all, and if hee be as wary in his Affirmatives about it as he vses lo bee about other things, I make no doubt but it will bee found extraordinarily ingenious, though I confesse I yet distrust the practicablenes of it in divers cases. (Quoted by Salmon 1988:164.) · One of the problems that he perceives is the vasthess of the lexis that the various branches of science require as a result of their great and rapid evolution. In fäet, when considering the idea of a universal language, he admits that such a code can be a useful tool "only if the dictionary [ ...] do not over-swell and disease it of a tympany" (l:XXXIX). However, the main obstacle that he comes across in t·he creation of a universal language is that there is still no clarity in the identification of the basic qualities on which real characters should be based. In his studies on corpuscular philosophy Boyle has come to the conclusion that there are two kinds of qualities: primary (referring to the size, motion and shape of particles) which are imperceptible, and secondary (based on the primary ones) which are, on the contrary, observable. Nonetheless, as Boyle, himself admits, there is not yet a definite existing taxonomy based on this distinction:

I confess, that I have not yet, either in Aristotle or any other writer, met with any genuine and sufficient diagnostic and boundary for the discriminating and limiting the species of things; or lo speak more plainly, I have not found that any naturalist has laid clown a detenninate number and sort of qualities or other attributes, which is sufficient and necessary to constitute all portions of matter endowed with them, distinct kinds of natura! bodies: and therefore I observe that most commonly men look upon these as distinct species of bodies that have had the luck lo have distinct names found out for them, though perhaps divers of them differ much less from one another than other bodies, which (because they have been huddled up under one name) have been looked upon as but one sort of bodies. (The Origin oj Forms and Qualities 3:50.) The interest with which Boyle follows all the projects connected with universal languages is therefore closely linked to his corpuscular studies, and also his appreciative comments on the various schemes put forth by various members of the Royal Society reflect his approval of what he considers mainly a so'rt of scientific experimentation, which. can help to bring same order and improve­ ment to the unsatisfactory state of an important part of the scientific world. Boyle, however, is very doubtful about the realization of those projects, and is convinced that a more concrete solution could be found in the systematization of 30 the linguistic tool already available to men of science and in the improvement of their style of expression.

The improvement of Scientists' style The suggestions that Boyle gives as regards improvements to the style of writing of men of science are strictly connected to his views concerning their method of research and code of behaviour. In particular, Boyle is convinced that writing scientific papers implies absolute honesty on the part of the writer and rigorous fäithfulness to the results of his experimentation. The use of language is therefore strictly connected to the scientist's approach to reality in general, in the sense that as the specialist does not take anything for granted but reconsiders each single phenomenon according to an objective viewpoint, also the way in which he expresses his activities will reflect a prudent and objective method. Facts must be reported as clearly as they are seen to happen, and if they are perceived as unclear or inexplicable, caution should be used by the scientist in describing them. Boyle is deeply convinced of these principles, not only following them in his writing activity, but also pointing them out explicitly by means of metatextual comments. For example, in A Proemial Essay he writes:

Perhaps you will wonder, Pyrophilus, that in almost every one of the following essays I should speak so doubtingly, and use so often, perhaps, it seems, it is not improbab/e, and such other expressions, as argue a diffidence of the truth of the opinions I incline to, and that I should be so shy of laying clown principles, and sometimes of so much as venturing at explications. But I must freely confess to you, Pyrophilus, that having met with many things, of which I could give myself no one probable cause, and same things, of which several causes may be assigned so differing as not to agree in any thing, unless in their being all of them probable enough; I have often found such difficulties in searching inta the cause and manner of things, and I am so sensible of my own disability to surmount those difficulties, that I dare speak confidently and positively of very few things, except of matters of fäet. And when I venture to deliver any thing, by way of opinion, I should, if it were not for mere shame, speak yet more diffidently than I have been wont to do. (1:307.) Thus the scientist should speak with certainty only of "matters of fäet", that is, of those fäets which have been proved true by abundant evidence, and can therefore be considered truly valid and able to obtain universal assent. The writing of a report on one's experiments is, indeed, conceived as a way in which thc scicntist can reproduce the various steps of his research experience and obtain a universal consensus on the conclusions he has reached. This assent­ creative function assigned by Boyle to his essays has been perceived very clearly by Shapin (1984:484) when analyzing his pneumatic writings:

The literary reporting of air-pump performances provided an experience that was said to be essential to the propagation of the material technology or even to be a valid 31

substitute for direct witness, In studying Boyle's Iiterary technology we are not, therefore, talking about something which is merely a 'report' of what was done elsewhere; we are dealing with a most important form of experience and the means for extending and validating experience. Another important principle followed by Boyle in the writing of his essays was to reproduce in that particular activity the same distinction correctly adopted by the scientist between the setting out of the fäets observed and his considerations on them. In fäet, Boyle suggests leaving "a conspicuous interval" (New Experiments 1:2) on the page between the two textual parts (report of experi­ mental findings and reflections on them) so as to show the methodological procedure adopted while underlining the rhetorical and pragmatic difference between those two parts. This objective approach to the analysis of scientific phenomena can also be found in Boyle's habit of examining fäets and opinions critically and severely, while at the same time respecting the people whose views and conclusions are being considered. On this subject he writes:

For I love to speak of persons with civility, though of things with freedom. (A Proemial Essay 1:312.) According to Boyle, this impersonal way of analyzing other scientists' conclusions and attacking wrong or inconsistent ideas allows the argumentative paper to be accepted more readily and highlights its persuasive strength, thus increasing its perlocutionary effect, which consists in convincing the reader of the validity of the writer's thesis. The advantages that avoiding an ad hominem way of arguing and adopting a more objective style provide in terms of raising the interlocutors' consent are clearly perceived by Boyle:

And as for the (very much too common) practice of many, who write, as if they thought railing at a man 's person, or wrangling about his words, necessary to the confutation of his opinions; besides that I think such a quarrelsome and injurious way of writing does very much misbecome both a philosopher and a Christian, methinks it is as unwise, as it is provoking. For if I civilly endeavor to reason a man out of his opinions, I make myself hut one work to do, namely, lo convince his understanding; hut, if in a bitter or exasperating way I oppose his errors, I increase the difficulties I would surmount, and have as well his affections against me as his judgment: and it is very uneasy to make a proselyte of him, that is not only a dissenter from us, hut an enemy to us. (A Proemia/ &say I :312.) This adoption of a more objective way of arguing is closely linked to Boyle's code of behaviour, in which the policy of a scientist cannot be distinguished from that of a gentleman. Indeed in his writings he often claims that it is fundamental for a man of science to keep his style as good as the current standards of polite conversation require. For example, in The Sceptical Chymist he claims that his language aims at the "due decorum" of gentlemanly style, and points out that "in a book written by a gentleman, and wherein only gentlemen 32 are introduc'd as speakers" it is important to show "that a man may be a champion for truth, without being an enemy to civility" (The Sceptical Chymist 1:462). Boyle is fully aware that a scientist's style 11 plays a relevant role in the re­ inforcement of the perlocutionary strength of his writings and in the acceptance of his theories, and sees a strong interrelation between method of expression and success in communication. This interrelation is guaranteed in particular by the writer's adoption of a plain and unadorned style, considered particularly func­ tional to the language of science. Metaphors and literary embellishments are seen as obstacles to a clear perception of the fäets reported and to a correct understanding of the scientific principles outlined in the paper. The adoption of such devices can hamper the reader's comprehension of the text, just as painting "the eye-glasses of a telescope" (A Proemial Essay 1:304) makes it impossible for a scientist to observe the universe. 12 This opinion of Boyle's proved to be very influential on contemporary and subsequent men of science and was later codified by the Royal Society. In fäet, article IV of Chapter V of the Statutes of the Royal Society (1728) reads:

In all Reports of Experiments to be brought into the Society, the Matter of Fact shall be barely stated, without any Prefaces, Apologies, or Rhetorical Flourishes, and entered so into the Register-Book, by order of the Society. Ho,vever, Boyle himself roade some use of figurative devices in his writings, a practice for which he has been accused of inconsistency (cf. Jones 1951: 145). 13

11 Boyle hasa complex view of the term style, which "comprehends not only the phraseology, the tropes and figures, made use of by a writer, but his method, his lofty or humble character (as orators speak) his pathetical or languid, his close or incoherent way of writing, and in a word, almost the whole manner of an authors expressing himself." (Same Considerations Touching the Style oj the Holy Scriptures 2:256.)

12 In pointing out these features of plainness and simplicity of style Boyle is also following Bacon's teachings:

Never cite an author except in a matter of doubtful credit: never introduce a controversy unless in a matter of great moment. And for all that cancerns ornaments of speech, similitudes, treasury of eloquence, and such like emptinesse, let it be utterly dismissed. Also Jet all those things which are admitted be themselves set down briefly and concisely, so that they may be nothing less than words. (Bacon 1620/1968, 4:254.) 13 In the use of figurative language, however, Boyle was not alone, as Markley (1988:46) confirms:

It seems safe to argue that Sprat's call for "Mathematicall plainess" had little effect on the stylistic practice of his contemporaries. No one stopped using metaphors, syntactical forms did not become simpler, and an authoritative language devoted to objective description did not emerge. What did occur, however, was that Sprat's critical rhetoric gave a new, quasi-official impetus to efforts to identify stylistic propriety with the 33 Boyle defends hjmself not only by highlighting the need to guarantee "decorum" in his writings, hut also by pointing out the argumentative value of such devices. For example, in .the Preface to the Second Part of his Christian Virtuosa he defends his use of analogy and metaphor by saying that:

Apposite comparisons do not o'rt!ygive light, but strength, to the passages they belong to, since they are not always bare pictures and resemblances, but a kind of arguments; being oftentimes, if I may so call !hem, analogous instances, which do declare the nature, or way of operating, of the thing they relate to, and by !hat means do in a sort prove, !hat, as it is possible, so it is not improbable, that the thing may be such as it is represented. (The Christian Virtuosa 5:511.) Moreover, also the criterion of conciseness is subordinate to the higher principle of clarity of exposition, which is considered the most important in terms of perlocutionary value. In fäet, also the principle of the avoidance of verbosity so often maintained by Boyle, Sprat and the other members of the Royal Society is neglected whenever the author runs the risk of being too brief and therefore unclear. This hierarchy of criteria is confirmed by the following quotation, in which Boyle apologizes for the impression of verbosity that some parts of his essays might arouse:

I have knowingly and purposely transgressed the laws of oratory in one particular, namely, in making sometimes my periods or parentheses over-long: for when I could not within the compass of a regular period comprise what I thought requisite to be delivered at once, I chose rather to neglect the precepts of rhetoricians, than the mention of those things, which I thought pertinent to my subject, and useful to you, my reader. (A Proemial Essay 1:305.)

Conclusions The analysis carried out in this paper has pointed out that like his fellow­ members of the Royal Society Boyle is very concemed about the problem of finding a language which is suitable for the expression of specialized texts. The faults that he finds in the English language are those hampering the possibility of directness of reference and clarity of interpretation. The problem of guaranteeing successful communication of scientific experience stands out as fundamental from an examination of Boyle's works. Another important aspect which is frequently taken into consideration in his writings is the problem of style, which should faithfully follow and clearly reflect the research method that the scientist has adopted, and which should be made to fulfil the pragmatic purpose that he has set out to achieve in writing his paper. This pragmatic purpose is clearly identifiable in the perlocutionary need

fiction of a dispassionate, gentlemanly style. Sprat and Glenvill, in effect, defined the terms in which debates about style would be cast for the next half-century. 34 to convince the reader of the reliability of the research being reported and of the validity of the researcher's conclusions. In his effort to achieve this perlocutionary effect, Boyle is very keen to predict the exegetic work of the reader and adopt all measures which favour this convincing strategy. It is this need to lead the reader carefully towards the attainment of the writer's goal which at times makes Boyle adopt also those devices.-such as analogies, metaphors, parenthetic expressions and reformu­ lations-which on principle he was so eager to discourage. In this way the features of the language are carefully made to match the argumentative purposes of the writer and to guarantee successful communication between the writer and the reader.

References Bacon, Francis, 1620: Novum Organum. In James S. Spedding, Robert Leslie Ellis & Douglas Denon Heath (eds), The Works oj Francis Bacon. 14 vols. London, Longmans & Co, 1857-74; rpt. New York, 1968. Birch, Thomas, 1756: History oj the Royal Society oj London. 4 vols. London, Millar. Boyle, Robert, 1772: The Works. Thomas Birch (ed.). 6 vols. London; rpt. Hildesheim, Georg Olms, 1965. Christensen, F., 1946: John Wilkins and the Royal Society's reform of prose style. Modern Language Quarterly, VII: 179-187 and VII:279-290. Galilei, Galileo, 1623: Il Saggiatore. (The Assayer, Discoveries and Opinions oj Galileo, trans. Stillman Drake. New York, 1957.) Gotti, Maurizio, 1991: I linguaggi specialistici: caratteristiche linguistiche e criteri pragmatici. Florence, La Nuova Italia. Gotti, Maurizio, 1992: The development of a scientific language in the 17th century. In C. Nocera Avila, N. Pantaleo & D. Pezzini (eds), Early Modern English: Trends, Forms and Text. Schena, Fasano (Br,-Italy), pp 319-343. Jones, Richard Foster, 1951: The Seventeenth Century. Stanford, Stanford University Press. Jones, Richard Fo~ter, 1961: Ancients and Moderns. J,.,ondon, Dover Publi­ cations. Markley, Robert, 1988: Two-Edg'd Weapons. Oxford, Clarendon Press. Rowbottom, M.E., 1950: The earliest published writing of Robert Boyle. Annats oj Science, VI:376-89. Salmon, Vivian, 1988: The Study oj Language in 17th-Century Engiand. Amsterdam, Benjamins. · Shapin, Steven, 1984: Pump and circumstance: Robert Boyle's literary technology. Social Studies oj Science, XIV:477-520. Shapin, Steven & Schaffer, Simon, 1985: Leviathan and the Air-Pump. Princeton, Princeton University Press. 35 Slaughter, M.M., 1982: Universal Languages and Scientific Taxonomy in the Seventeenth Century. Cambridge, Cambridge University Press. Sprat, Thomas, 1667: A History oj the Royal Society oj London for the lmproving oj Natura! Science. London, Martyn; rpt. London, Routledge & Kegan Paul, 1959. Turnbull, G.H., 1952: Samuel Hartlib's influence on the early history of the Royal Society. Notes and Records oj the Royal Society oj London, IX: 101- 130. Vickers, Brian, 1985: The Royal Society and english prose style: A reassess­ ment. In B. Vickers & N.S. Struever, Rhetoric and the Pursuit ojTruth. Los Angeles, William Andrews Clark Memorial Library, University of California, pp 3-76.

The Historical Development of the Register Barrier in Professional Language

Geoffrey Hughes

"All professions are conspiracies against the laity." This shrewd observation is made in the extended polemic which passes fora Preface to George Bernard Shaw's play, The Doctor's Dilemma written in 1906 (1971:72). Shaw's cynical remark contains a linguistic truth familiar to us all and supplies one motive for these deliberations, namely that professional status and social distance between practitioner and layperson are notoriously maintained by linguistic barriers and by obfuscation. The ramparts of professional jargon, legal and medical espe­ cially, have been powerfully satirised and undermined over the centuries by such notable authors as Chaucer, Shakespeare, Ben Jonson and Dickens. 1 However, the alienation of the laity does not derive solely from cynical motives, but also from the division of register in the English lexis, whereby native terms (which are generally transparent and readily comprehensible) have increasingly been displaced in professional language by opaque classical borrowings. This paper will employ a diachronic perspective and a lexicoiogical methodology to focus on the development of medical, legal and psychological vocabularies. A brief examination of medieval medical texts, such as the English translations of Lanfrank's Cirurgery (c. 1400) and The Cyrurgie oj Guy de Chauliac (c. 1425), demonstrates the point that core words now regarded as coarse and grossiy impolite were previously acceptable as medical terminoiogy. "In wymmen," we read in Lanfrank, "pe necke of the bladdre is schort & is maad fast to the cunte" (p. 172). We encounter in Wycliffe's contemporary translation of the Bible the graphic term arse-ropes for intestines, a word whose coinage half a century did not drive the variant term arseguttes out of use until the seventeenth century. Arse was of course, sufficiently acceptable in former times to be used in Shakespeare and Dr Johnson, while guts was empioyed without embarrassment by Sir Philip Sidney in his translation of the Psalms in 1580,2

1 In his General Prologue ta the Canterbury Tales, Chaucer presents both the Man of Law and the Doctor in terms of their own opaque jargon as specious opportunists. Ben Jonson used similar technique and presentation in Volpone and The Alchemist, while in Bleak House Dickens satirises the law with devastating effect as cumbrous, inefficient and heartless. 2 Sidney has "My eyes, my guts, yea my soule, grief doth wast" for Psalm 31, v. 5, where the Authorised Version (1600) prefers belly and Prayer Book body. 38 Several native terms of basic physicality have become obsolete. When we read that "[the urine] passith out by the 3erde" (1379) or of the "cancre of the mannes yerde" (1425), we are reminded (after a momentary confusion) that yard was a basic word for 'penis' in medieval times. Guy de Chauliac has sections on the "passiouns of the yerds" and defines priapisme as the "vnwilful stondynge of the 3erde." The main rival in medieval times was tarse, an older term deriving from Anglo-Saxon teors. Both terms had become obsolete by the nine­ teenth century, having been supplanted by classical penis and phallus. Indeed, the definition of penis in its first recorded use in 1693 (in the translation of the second edition of Blancard's Physical Dictionary) is an interesting combination of the old and the new: "Penis, the Yard, made up two nervous bodies, the Channel, Nut, Skin and Foreskin." In addition to the use of native nut for the glans penis, the old word for testicle was stone, recorded from the twelfth century, and carrying the usage marker in the O.E.D. of "Obsolete, except in vulgar use, e.g. stone-horse for 'stallion' and stone-priesf for 'a lascivious cleric' ." Similarly, the old word for 'copulate' was swive, recorded from Chaucer until it suffered a Victorian demise, shortly after this piece of quasi­ medical advice recorded in 1890: "Don't bathe on a full stomach: nor swive." Reference to the sexual organs was, of course, problematised by the fäet that the term sex (recorded from the fourteenth century) tended to be used in a generalised way (e.g. "the fair sex", also the gentle, soft, weak, second and even "the women sex"), rather than in the modern physiological and genital fashion, until the seventeenth century. Hence the sexual organs are referred to by Guy de Chauliac as "the membres of generacioun" (p. 526) or "/Je membres generatifes" (p. 527); hermajrodicia (as he calls it) is defined as "/Je nature of double kynde" (p. 529). Here we have an interesting use of two broad terms, nature and kynde, which creates ambiguity. Both Kynde and another highly generalised term, shape, had from Anglo-Saxon times been used more specifically to refer to the sexual organs: mlfric's version of the drunkenness of Noah in Genesis ix 23 runs: "Sem and Iafeth [... ] beheledon heora frederes 3ecynd"; the previous verse has the term 3esceapu. These senses became obsolete in the course of the late Middle Ages, a late instance being "Couer thy shap with some fly-flap", an injunction found in Skelton (1529). Other bodily functions are described in everyday language: Guy de Chauliac writes of "the crafte of pissynge" (p. 525), while Lanfrank offers the following prescription: "If he [the patient] schite oones a day, helpe him faerto [ ...] with clisterie" (p. 12). All in all, it is a valid generalisation to claim that all the available "four-letter" words were used in medical contexts up to about 1500. As table l3 illustrates, opaque classical nomenclature started to become current in this professional area, as in many others, in the sixteenth century.

3 The

Spleen c.1300 Cranium 1543 Palate 1382 Wyc!iffe Tendon 1543 Stomach 14th C, Vulva 1548-77 Vein 14thc. Larynx 1578 Banister Sperm 1386 Chaucer Pancreas 1578 Banister Genitals 1390Gower Scapula 1578 Banister Cell 1393 Gower Ske/etan 1578 Banister Artery 1398 Trevisa Scrotum 1597 Guillemeau Colan 1398 Trevisa Phallus 1613 Diarrhoea 1398 Trevisa Clavicle 1615 Crooke Oesophagus 1398 Trevisa Clitoris 1615 Crooke Semen 1398 Trevisa Coccyx 1615 Crooke Thorax c. 1400 Lanfrank Epiglottis 1615Crooke Trachaea c. 1400 Lanfrank Pelvis 1615 Crooke Uvula c. 1400 Lanfrank Radius 1615 Crooke Virus c. 1400 Lanfrank Uterus 1615 Crooke Testicle c. 1425 Vertebra 1615Crooke lntestine 1425 Guy de Chauliac Faeces 1639 Embryo 1477 Digit 1644 Muscle 1533 Anus 1658 Nerve 1538 Sternum 1667 Abdomen 1541 Copeland Saliva 1676 Cartilage 1541 Copeland Patella 1693 Blancard Cavitv 1541 Copeland Penis 1693 Blancard Fraciure 1541 Copeland Phalanges 1693 Blancard Rectum 1541 Copeland Pharynx 1693 Blancard Ulna 1541 Copeland Tibiå 1693 Blancard

Terms like abdomen, cartilage, cavity, fracture, rectum and ulna appeared in 1541 in Robert Copeland's The Questyonary of Chirugery; cranium and tendon appeared in 1543, larynx, pancreas, scapula, and skeleton appeared in John Banister's The History of Man (1578) and scrotum in 1597. These were followed by clavicle, clitoris, coccyx, epiglottis, pelvis, radius, uterus and vertebra, first recorded in Helkiah Crooke's Mikrokosmographia, A Description of the Body of Man ( 1615). As one might imagine, there was an uneasy period of lexical cohabitation between the native and the classical elements as they contrived to divide the territory: this rivalry is epitomised in Robert Copeland's definition under ulna: "the arme is deuided in thre great partyes. One is called vina, the other lyte( arme." (A charmingly unexpected example of the excluded middle.) As we have seen, radius emerged only in 1615, and the humerus arrived lexicographically nearly a whole century later in 1706, in John Kersey's updating of Edward Phillips' The New World oj English Words, where it is tentatively defined as "the shoulder; the shoulder-bone or first bone of the arm." Clearly, by the sixteenth century we have moved into a quite different lexical world from that of Guy de Chauliac's translator, who wrote of the brayne pan, the rigge bane, the forked bane, the spawde bane, the canell bane and the rollynge bane (pp. 40 20-21), as well as the priue stones (p. 511) for testicles and coldnesse for frigidity (p. 527). The point of the modern separation of registers according toa principle of decorum is most clearly illustrated in sexual terminology, on which C. S. Lewis made the trenchant observation that "As soon as you deal with it [i.e. sex] explicitly, you are forced to choose between the language of the nursery, the gutter and the anatomy dass". 4 Historically what has occurred is that the old terminology has been driven out of accepted usage, either into obsolescence, the farmyard or the gutter. In the whole area of the genitalia there are no neutral terms like arm, legor head. One must concede that a specific technical terminology was needed for certain physiological areas, for two reasons. Firstly, native terms like ankle, knee, shoulder, elbow and hip tend to be general: classical terminology effi­ ciently defines by breaking down these joints into their component parts. Secondly, there is the factor of semantic instability. As the term womb, for example, made its extraordinary transition from generalised 'stomach' to exclu­ sively feminine 'uterus', so cod, meaning a 'bag' or 'covering' made an oppo­ site shift in the male sex toan ambiguous area incorporating both 'scrotum' and 'testicles', subsequently surviving only in cod-piece. Chaucer's Pardoner uses the old senses in one of his hysterical outbursts: "O wombe O bely, 0 stynkyng is thi cod" (Pardoner's Tale 206). When we encounter Dr Johnson's definition of the pelvis as being "the lower part of the belly", we realise the imprecision of the terminology of this area. Bowel (from Later Latin botellus, "pudding or sausage") moved from its early sense of 'intestine' or 'gut' in the fourteenth century to the broad plural sense of the 'seat of compassionate and tender feelings' or the feelings themselves, a sense which lasted through to the nine­ teenth. Furthermore, the term nerves has itself undergone a major shift from its early physical sense of 'muscle' or 'sinew' in the sixteenth century, via the neurological sense. of the nervous fibres connecting muscles and brain from the seventeenth century, to its more subtle modern psychological meaning, which emerges in the nineteenth. In concert, and less excusably, professional titles changed: the ancient native term leech gave way to the Latin doctor, which is in tum being replaced ( espe­ cially in the United States) by Greek physician and therapist. Virtually all new fields of medicine style themselves in Greek terminology. Hence psychiatry, gynaecology, orthodontics, otorhinolaryngology, gerontology, paediatrics, oncology and so on. The only branch of medicine with an Anglo-Saxon name is midwifery, steadily being ousted by obstetrics. The problem of the register barrier is partially solved by translating the alien and opaque into the familiar.

4 Quoted by one of Lewis's students, Kenneth Tynan, 1975: The Sound oj Two Hands Clapping. London, Cape, p. 154. 41 Turning now to the language ofthe law, in which the base term is, of course, a Scandinavian word, one is especially struck by the fäet that the law was originally not in the hands of a profession at all. The ancient tradition is epito­ mised by the role of the Icelandic lawman, who recited the whole law from memory before the annual session of the Althing. Thereafter matters of dipute were settled by the parties concerned. The earliest extant Anglo-Saxon legis­ Iation is found in the Laws oj lne, King oj Wessex, promulgated between 688 and 695; these have a continuous and complementary relationship with those of Alfred (c. 885-899). These early laws are, unlike their successors, models of native simplicity and clarity:

lf anyone steals so that his wife and children do not know, Jet him pay sixty shillings fine. If he should steal with the knowledge of his family, then they should all go into slavery. Aten year old boy may be [considered] an accomplice toa theft. (Laws oj lne §7,5) However, legal terminology acquired an alien register from the time when the Norman overlords used their standard practice of defining tliemselves and all aspects of their rule in their own terms. From the point of view of the native majority of the population, this meant that a foreign nomenclature was hence­ forth used to define offences which had previously been denoted by native terms. Thus the direct terminiology used in the Anglo-Saxon Iaws, afabryce ('oathbreaking'), brernette ('burning'), stale ('stealing'), gejeohte ('fighting') were supplanted by perjury, arson, ·larcency, assault etc, which were opaque or ambiguous or both. The point of obfuscation is nicely made in one of the earliest instances of malaproprism, which occurs in the fourteenth century poem, Piers Plowman, where a character evidently thinks that restitution means precisely its opposite, namely 'robbery' (Skeat 1886:238). An interesting exception lies in what we now call "capita! crime": both murder and manslaughter are Germanic in origin, and have maintained the crucial distinction between deliberate and unintentional kill.ing. Generally, however, there consequently deve]oped parallel vocabularies of broad, general native words and specific alien technical terms: wrong-doing vis a vis jelonylcrime/delict; and theft vis a vis larceny. Further­ more, there is often confusion caused by competing definitions between the technical and the general senses in libel visa vis slander; assault visa vis bat­ tery and rape; damage and damages; guardian in its various senses, and so on. Thereafter came the infiltration of law Latin in the form of numerous phrases like habeas corpus, primajacie, quid pro quo, inter alia, inflagrante dilecto, in jee simple. Culprit represents an odd case of a word being invented for legal purposes and then abandoned.6 Critical comments on the voluminousness and

5 The Anglo-Saxon Laws are conveniently collected in D. Whitelock (ed.), 1955: English Historical Documents vol. I. London, Eyre & Spottiswoode. 6 Culprit, the result of the fortuitous or ignorant running together of two terms from Law French, was first used in 1678. 42 unintelligibility of legal language have been de rigueur for centuries, as David Mellinkoff shows in his classic study, The Language oj the Law (1963). He devotes two whole chapters to the necessity of making the law clearer and briefer: they are entitled, simply and appropriately, "Shorter" and "More Intelligible",7 Terms for mental disturbance and madness, set out in table 2, have tradition­ ally been lacking in precision, mainly as a consequence of euphemism. The older terms were characterised by great semantic breadth, whether classical or native in origin: they included natura!, simple, silly, the vapours, and a variety of terms involving the nerves, e.g. nervous conditionlcollapse etc. A surprising­ ly early example of a clearly foreign term is mania, recorded in Lanfrank: only in the seventeenth century is it joined by a first influx of similar classical terms such as hysterical, melancholia, obsession and psychology itself (in Blancard). These drive out the old words like wood, woodness. The second wave of classi - cal borrowings comes in the second half of the nineteenth century, which yields psychosis, neurotic, paranoia,psychotic, psychopathic and psychotherapeutic. Two processes are notable in the evolution of the nomenclature: one is the degree of back-formation evidenced in several key terms. Neurotic, for example, is coined as a medical term in 1661 in the sense of a drug having a marked effect on the nervous system, especially a bracing one; the pathological sense of the adjective is defined by Webster in 1847 as "a disease having its seat in the nerves"; neurosis itself emerged only in 1871 when Huxley made the distinction between psychosis and neurosis. Freud further refined the meaning by identi­ fying anxiety neurosis in 1918. Similarly, hysterical precedes hysteria by nearly two centuries, while psychopathic precedes psychopath by nearly half a century. It is as if the discipline proceeded from identifying the symptoms toan analysis of the given condition. The second trend is specialisation, classically evident in obsession, a term which would appe'ar to encapsulate in its semantic development the history of psychiatric symptomatology and its interpretation. lts original meaning, recorded from 1513, is "a siege, or the action of besieging". This introduces the basic and enduring paradigm of the mental condition being conceived of in terms of attack and defence, resilience and breakdown, security and anxiety. A century later, from 1605, emerged the sense of "the hostile action of the devil or an evil spirit in besetting any one." By the end of the century, demonic possession had been replaced by a more rational explanation: ''The action of any influence, notion or 'fixed idea' which persistently assails or vexes, esp. so as to discompose the mind."

7 Mellinkoff cites hundreds of castigating and cynical comments on the voluminousness, indistinctness and unintelligibility of legal language. 43 Table 2. Word field ofterms within psychology arranged historically

Traditional Specialisation Coinage mania 1400 maniac adj 1604 Cawdrey hysterical 1615 hysteric (adj) 1657 neurotic 1661 obsession 1680 melancholia 1683 psychology 1693 Blancard neurosis 1776-84 hysteria 1801 psychosis 1847 psychopathic 1847 neurotic 1847 paranoia 1857 paranoic 1857 neurosis 1871 Huxley inhibition 1876 William James psychomotor 1878 katatonia 1883 psychopath 1885 psychotherapeutic 1890 blocking 1890 William James psychotic 1890 dementiapraxox 1891 trauma 1894 transference 1895 Freud abreaction 1895 Freud psychoana(vsis 18%Freud psychosexual 1897 manic 1902 psychotherapy I 904 complex 1907 Jung rationalization 1908 defence mechanism 1909 libido 1909 Freud Freud anxiety 1909 Freud repression 1909 Freud projection 1909 Jung ego 1910 Freud regression 1910 Freud sublimation 1910 Freud obsessive I 9 I 1 schizophrenia 1912 schizophrenic 1912 displacement 1913 Freud fixation 1913 Freud id 1917 Freud angstl922 Freud superego 1924 Freud schizoid 1925 block 1931 manic depression/ve 1958 44 There were, however, "false friends" in the existing terminology, notably the signal case of the "etymological fallacy" which problematised the diagnosis of hysteria. Since the term is derived from Greek hystera, 'a womb', many early practitioners assumed that it was an exclusively feminine condition. The feminine association was reinforced by the sense of mother in the sense of 'womb', specially in the phrases fits oj the mother (from c. 1400) and the rising oj the mother (from c. 1527) to mean 'hysteria'. Thus in 1615 Crooke's Body oj Man carries the definition: "Hysterical women, that is, such as are in fitts of the mother." Freud had the greatest difficulty in persuading the more obdurate ofhis colleagues that hysteria could present itself in males. There is also a notable difference in register between the first wave of modern psychological enquiry, epitomised by William James, and that of Freud and Jung in the second. James tends to give specialised meanings to established words, such as blocking (later block), inhibition, and obsession. Huxley, in the same era, similarly extended the use of neurosis and psychosis from the provenance of pathology to psychology. Freud predominately, and to a lcsser extent, Jung, sought a more scientific and precise terminology which would be more exact because it was unshaped by semantic history, and would be within the ambit of their own legislation. As Malcolm Bowie observes, "They retreated from desire in the act of writing about it" (1991: 199). They either resuscitated or redefined classical terms, just as William James had done in redefining trauma as an unhealed psychic wound. Freud added specialised meanings to anxiety, ego, transjerence, regression, repression, displacement, fixation and sublimation (originally a term from alchemy first recorded in Gower c. 1390). 8 Similarly, Jung used complex in specialised fashion, famously of Oedipus and Electra, 9 as well as extending the meaning of projection. Freud went further, coining psyc!wanalysis in 1896, abreaction in 1895, libido in 1909, angst in 1922. 10 His famous tripartite division of the personality becomes the more interesting when one considers the evolution of the key terms: ego was given its Freudian specialisation in 1910; id was formulated in 1917 in the course of Freud's correspondence with Groddeck, and superego was coined in 1924.11The components of Freud's final duality in

8 Several of these straight translations of German originals, e.g. abreaction from abreagieren, transference, from Ubertragung and repression from Verdrängung. 9 Freud had clearly divined the symbolic significance of the Oedipus myth in the 1900 edition of The Interpretation oj Dreams (Standard Edition IV pp. 261-264), but Jung seems to have been the first to use the formulations Oedipus-complex and Electra-complex (in 1913 in Transactions of the International Congress of Medicine). 10 Libido is not strictly Freud's coinage, since as A. A. Brill (the translator) notes (1938:553): "The single adequate or fitting word of the German language, 'Lust', unfortunately has many meanings and signifies the sensation of need as well as that of satisfaction. 'Libido' is the motor force of sexual life." 11 The Translator's Note on this point runs: "To translate the German es, which means 'it' and thus implies the impersonality of the mind apart from its ego, the Latin id has been selected." (J. Riviere et al, Freud's Col/ected Papers (II xii, 1924) p. 254.) 45 the personality were given such classical names as eras (first used by Freud in 1922) and thanatos, though it should be noted that Freud's own terms were todestrieb ('death instinct')12 and todeswunsch ('death-wish'). The predilection for arcane classical vocabulary was an extension of the process well established in medicine; it was the natura! koine of members of the intelligentsia, and was also probably designed to give the discipline the status which it lacked. One should recall that the earliest sense of psychotherapeutic (1890) was "the treatment of disease by 'psychic', i.e. hypnotic influence". Associations with mesmerism and occultism were common in the 1880s. (Before coining psychoanalysis in 1896, Freud had himself used the formula­ tions psychische analyse.) In this context, one can understand the coining of impressive formations like katatonia in 1883 and dementia praxox in 1891, this last by Pick, though much associated with Kraepelin (1899), later superseded by schizophrenia. As in medicine, status was achieved, but at the expense of clarity. Unsurprisingly, the term was subsequently (c. 1944) retranslated back into layman's English as "split personality." However, the fundamental problems of definition have not been resolved. The whole foundation of the discipline is essentially problematic, because neither patient nor therapist operates within the same clear parameters of evidence, symptoms, conditions and definitions as is the case in medicine: indeed, the evidence of the patient may be misleading or manipulative, as Freud recognised in his use of the term transjerence. This aspect has become a major focus in the work of Lacan and the neo-Freudian revisionists. The modern history of the discipline has been one of endless schism between "druggists" and analysts, biologists and linguists. Lacan's dogmatic assertion that "The unconscious is structured like a language" evoked Laplanche's riposte "What linguistics? What language? And what Lacanianism?" (1989:41). Lacan "eroti­ cises the language of 'theory"', indulging self-consciously in provocative word­ play, fabricating the evidence which Freud and his followers had detected as a key to the unconscious: hence his quasi-sexual use of jouissance and conneries (Bowie 1991:199, 148, 151). More seriously, there seem to be no commonly agreed-on definitions of the major conditions. Above all, an essentially prob­ lematic aspect of the "talking cure" is the crucially important area of ambiguity inherent in the use and interpretation of language in the process of analysis. The enterprise of Freud and Jung, though worthy, has not been entirely successful, for another reason: the speech community has borrowed their terminology wholesale and used it with the unscientific latitude which the founding fathers sought to avoid. "Terms not so long ago confined to specialists are handled familiarly by the laity: moran, inferiority complex, mental age, [... ] paranoid de/usions, psychopaths." That comment was made in the New

12 Though trieb is usually translated as instinct in English, the meaning is closer to the cognate term drive. 46 Republic in 1927. Since then the trend has accelerated. Furthermore, terms intended to be neutral have become universally unsympathetic. There is mani­ festly a fundamental contrast in the acceptance of the taboos which still surround the naming of terminal illnesses and the practice of callous labelling by means of the names of mental conditions. The interesting new factor here is the growth of Politically Correct language: its emphases have so far been of discriminatory terms concerning race, gender and "abnormality" of physique: it remains to be seen whether it will have the effect of censoring the sti:gmatising usage of psychological conditions. Professions traditionally take to themselves the powers of definition and self­ styling. We have seen same examples of their appropriation of high register titles, such as therapist, gerontologist, gynaecologist and so on. A spectacularly cynical example occurred in the year 1360, when royal purveyors, notorious for their corruption, were restyled by statute achatours.13 However, the exploited majority does have its recourse in the slower vengeance of semantic change: the word cheat derives from the dishonest practices of the medieval escheators; quacksalver, quack, mountebank, charlatqn and sawbones aH reflect critical views of unprofessional medical practitioners, mast of thern from the sixteenth century; pettyfogging and its variants have maintained the layman's view of the law from a similar distance in time; the contemptuously belittling term head­ shrinker is recorded from 1950, while psychobabble, expressing disdain for psychological jargon, has been in existence since 1976. The problems of definition in these professional areas essentially replicate the mode! of the "conflict of speech economies" posited by G. K. Zipf in 1949. Specialists prefer a wide range of precise terms; the public, contrariwise, prefers a small range of broad terms. It is hard to see what can be done about the split in the terminology. Despite calls for "openness", one cannot imagine modern medical practitioners reverting to the direct demotic terms of medieval times. · These three examples or nomenclature represent, however, differing modets of interaction between the professions and the speech community across the register barrier. The older professions of medicine and law have managed, with the benefit of time, tradition and status, to maintain powers ofdefinition, and of obfuscation, within their fields. While they may not have gained the total trust of the public, · they have ensconsed themselves as part of a professional establishment. Such claims cannot be made for psychiatry. Despite a virtual epidemic of psychoanalysis, especially among the Åmerican bourgeoisie, there still seems to be a lack of sympathy for mental conditions, perhaps because they represent an affront to the western ideology of success, and have yet to gain the status of socially acceptable disorders. More seriously, there remains profound scepticism about the efficacy of psychiatric treatment. Consequently, this mast

13 According to Act 36 of Edward III (1360): "[ ... ] le heignous noun de purveour soit chaunge & nome achatour." 47 recent discipline has yet to acquire the professional status of law and medicine -despite the illuminating contributions of magisterial figures-and has seen its terminology largely demystified and crudified.

References Banister, John, 1578: The Historie oj Man. Blancard's Physical Dictionary. 1693. Translation. Second edition. Bowie, Malcolm, 1991: Lacan. London, Fontana. Brill, A.A. (ed.), 1938: The Basic Writings oj Sigmund Freud. The Modern Library, New York, Random House. Copeland, Robert, 1541: The Questyonary o/Cirurgens. Crooke, Helkiah, 1615: Mikrokosmographia, A Description oj the Body oj Man. The Cyrurgie oj Guy de (;hauliac (c. 1363) ed. Margaret S. Ogden, 1971. E.E.T.S., no. 265. Gower, John, c. 1390: Confessio Amantis. Guillemeau's French Chirurgerye. 1597. Translation by "A.M." Lanfrank's "Science oj Cirurgie" (c. 1425), ed. Robert v. Fleischhacker E.E.T.S. Orig. Series 102 (1894). Laplanche, Jean, 1989: New Foundations for Psychoanalysis. Trans. David Macey, Blackwell, Oxford. Mellinkoff, David, 1963: The Language of the Law. Boston, Little, Brown. Shaw, George B., 1971: The Doctor's Dilemma. London, The Bodley Head. Skeat, W.W. (ed.), 1886: The Vision oj William Concerning Piers the Plowman. B Text. London, O.U.P. Trevisa, John of, 1398: trans. Bartholomeus de Glanvila de Proprietitabus Rerum. Wycliff, John, I 382: trans. The Holy Bible. Zipf, G.K., 1949: Human Behavior and the Principle oj Least Ejfort. Cambridge, MA, Harvard University Press.

Varieties of Psychological Discourse: Mental Ref erence in Prof essional and Everyday Language

CarlF. Graumann

1 Professional vs. everyday language If it is proper to state that psychology in this century has become part of our daily Iife this statement does not necessarily fil! the professional psychologist with pride. The psychology that we find in so many places in our everyday Iife is not the psychology being taught and studied at our universitfes and the subject of scientific research. Or is it? If it is, what has happened to it considering the dissimilarity between academic or scientific and everyday psychology? If it is not, are there two different psychologies with Iives of their own? As interesting as this issue may be for a social scientist of science and, mainly, of psychology, I suggest a more modest leve) of analysis by concentrating on varieties of psychological discourse. The problem is still big enough and, as we shall see, there is hardly any research from which to start. It is true that there has been some research on the relationship between professional language and the vernacular, hut Iinguistic experts in the field, like Oksaar (1989) and Steinig (1981), see Iittle development, and mainly for the language of psychology there is, after a soiitary monograph of 1959 (Mandler & Kessen), no relevant research at all. Only with respect to psycho-analysis there is, beside some scattered reflections and interpretive comments, the influential study of how Freud's basic theory became vulgarized in terms of a "social representation" (Moscovici 1976). Although we do not have a consensual definition of professional language (cf. Fluck 1976; Oksaar 1989), scientists seem to agree that vernacular discourse is more ambiguous, variable, and situational, hence, more heterogeneous than scientific discourse (Mandler & Kessen 1959; Fluck 1976). 50 2 The psychological in scientific and in everyday discourse

2.1 Traditional key-terms Historically, modern psychology has developed 'as the scientific ·study of consciousness or experience, and of behaviour. Accordingly, the language of this discipline refers mainly to mental states and processes and to patterns of behaviour. But this is not unique to the science of psychology (if there is such a unitary enterprise). A language of mind and of conduct had been an essential and !arge part of philosophy since classical antiquity. Equally, talking about other people's and one's own mental states and patterns of conduct has always been a constituent element of everyday interpersonal communication, certainly in our culture. Hence, we have at Ieast three varieties of psychological discourse: the traditional, hut still existing philosophical, the modern scientific, and the perpetual ordinary language variant of referring to mind and conduct. This situation is, however, complicated toan inextricable degree because of the interaction between the varieties. It is well known and firmly established (a) that concepts and theorems of philosophical psychology have, mainly in the nineteenth century, strongly and Iastingly influenced the establishment of a psychological science, (b) that, by way of popularization, key-terms and concepts of modern professional psychology (and psychoanalysis) have diffused into literature and, mainly channeled by the media, irtto everyday language, where (c) they mingled and blended with autochthonous mental and behavioural terms. But that is only half the truth. We partly know or have good reasons to hypothesize that many of these influences and dependencies are mutual: Philosophy has adopted and reflected upon concepts developed in psychological science. The latter has taken up lay-psychological terms and conceptions. And so-called lay-theories, concepts and accounts of both individual mind and of social conduct have become subjects of ordinary language, ethnographic and ethnomethodological studies. Instead of trying to disentangle this web of relationships I shall concentrate on one aspect of the relation between academic and everyday psychological discourse: the survival of the essentialist and moral meaning of the key-terms of philosophical psychology in so-called naive or folk-psychology, whereas modern psychological science, from its beginning, has endeavoured to get rid of this essentialist and moral heritage. For this purpose I shall propose two theses concerning the basically "ontological" and "deontological" usage of mental key­ tcrms. By this I mean that what we have referred toas psyche, soul, or mind is both an essential constituent and a governing principle of human life and conduct. (1) My first thesis, which I cannot fully substantiate in this paper (cf. Graumann in press), maintains that the establishment of psychology as a science, i.e. its emancipation from theology and metaphysics, involved the 51 necessary effort to overcome any essentialist and ethical interpretation of psychology's subject-matter, hut that this effort of "de-moralizing" the mental (during the transformation of the moral sciences to social or human sciences) has only partly been successful. The de-moralization of. psycholological discourse has been tried, on the one hand, by certain "semantic" avoidance techniques: The traditional nouns used to refer to psychology's subject matter, psyche, mind, Seele, were replaced by (less "substantive"?) nominalized adjectives: das Seelische, the mental, das Psychische. But even in this adjectival form, or totally replaced by the new key-terms of consciousness, experience, behaviour, or, nowadays, cognition (Graumann 1984; 1991), they have largely remained unidentified, at least undefined. Even if psychologists have restricted themselves to particularities, like sensation, perception, emotion, motivation, intelligence, thinking and memory, they still tend to use these terms as titles for research areas rather than as well-defined concepts or theoretical constructs. The notorious pseudo-operational "definition" of intelligence as that which intelligence tests measure may, without reservation, be generalized to psychology, defined as what psychologists do. Since what psychologists (as scientists) mean to do is the objective description, measurement and explanation of intersubjectively observable phenomena there is no room for ontological and ethical propositions. The other technique of exorcizing the metaphysical and theological spirits of the soul has been a preference for mechanical, i.e., guaranteed soulless, models-a Cartesian heritage, still preserved in today's dominance of the computational or computer modets of the mind, which characterizes the cognitive mainstream of psychology. So much about the attempts at objectification in psychological science. (2) My second thesis is closely connected with the first: It is exactly the ontological and deontological usage of mental terms that has been preserved in everyday psychological discourse, which I consider to be a largely and ultimately moral discourse. Referring to somebody else's or one's own mental states or dispositions does not only presuppose the reality or existence of such states, which sometimes are treated as entities. In many cases this reference to mental affairs is also an appeal toan accountable agency, something amenable to blame or praise. At Ieast, we may take from attribution research that the tendency of so-called observers of actions to attribute the causes of such actions to actors' dispositions (e.g., to personality traits) frequently amounts to holding actors responsible. In accordance with these two theses this chapter is not about the technical terminology of psychological science and its application in the practice of professional psychologists. But, in order to substantiate the above theses, it shall be demonstrated that and how the traditional (pre-scientific) conceptions of the human psyche are still reflected in the language by means of which we refer to 52 others' and our own mind, for only in everyday discourse the scientifically extinct species of psyche, soul, and mind have survived. A few examples from the German and English vernacular may do to illustrate this continuity. If in German we account for somebody's behaviour by saying "It is her psyche" or "es ist psychisch" ('it is psychical'), we refer to something sensitive, delicate, vulnerable, which resides invisibly in a body, "deep down inside". And if, in colloquial German, we refer to a person as a "Psychi", we discriminate him or her as a "freak". In English, the "psychical", lying outside the sphere of the physical and of physical science, has traditionally adopted a spiritualist meaning, but may still be an object of so-called "psychical research". More common than "psychical", however, is "mental" discourse. Besides the common usage as the adjective relating to mind, to state "it's mental" or even to speak of a "mental case", clearly refers to a deficiency or disorder in a psychiatric sense. Yet, a "mental patient" is not necessarily a person "sick in mind" (geisteskrank) as the term "mental disease" suggests. His or her behavior may mcrely impress others as odd, fantastic, unusual, or "deviant". More often than not, it seems to be difficult to decide whether the disorder we meet with is "mental" or "social". At least in some societies persons who are socially or politically "deviant" run the risk of ending up in a "mental home" (or "asylum"). On the other hand, we occasionally say "mental" only to indicate that there are no physical or visible signs or symptoms. That applies to making "mental notes" as well as to committing "mental cruelty". ·While many usages of "mind" are indispensable in everyday language, if we want to refer to the human capacities or functions of perceiving, intending, deciding, or remembering, the Germanic-English counterpart of the mind, the "soul", is a much more delicate and less frequent word, and so is German "Seele" and French "l'ame". While it has been banned from psychological science, it has been retained in religious, poetic, and erotic discourse; but it seems generally to be on the retreat in everyday talk. Still, it sometimes means something like the guiding principle and the center or the "heart" of things. Free speech, for example, can be the "soul of democracy"; a person may be "the soul of the department". If we call a person a dear or poor or "merry old" soul we mean Gemiif, the qualities that express and arouse sentiment rather than intellect. The oldest meaning of soul as a spiritual principle may be obsolescent in ordinary language, hut the recent Afro-American "Soul" is the most distinctive indication that the soul still is an elementary spiritual, moral, and expressive force. There seems to be a central tendency in the ordinary usage of psychological key-terms: Denoting something as "mental" in English ( "psychisch" in German) frequently means'something tender, frail, deficient, or even, as we have seen, sick. The state of mind of a healthy, attentive, and contented person 53 would not be referred to as "mental". But if a person looks glum, strained, and irritated, the cause might be "mental". In ordinary language, the "psychological" in general is frequently associated with the soft, delicate, weak, or vulnerable. Complementarily, if we occasionally label an executive's secretary a "good psychologist", we usually mean to say that she knows about the "weak sides", the foibles, and the vanities of her employer, and that she has learnt how to use her knowledge. In general, being a "good psychologist" in non-professional discourse implies the knowledge and the skill of how to influence or even to manipulate others. In this context, it is interesting to know that the ability to look through a person's "forefront", "facade", or "mask" is part of the social representation of a good psychologist to such a degree that people still believe that psychology students are trained in these skills at the university. Although academic psychology has long given up the claim to be the art of unmasking, of revealing the "true" character of a person, this is still a popular notion, going back to conceptions of psychology as once held by Friedrich Nietzsche and Sigmund Freud. Here we have one of the crucial points where, expressed by the same word "psychologist", the modern scientific and the popular notions diverge.

2.2 Mental terms: their referential and communicative functions So far we have dealt only with the fate of the headwords or key-terms of psychology. But our everyday discourse is interspersed, sometimes saturated, with a huge variety of mental terms. They are not, as in science, constitutive of, hut certainly essential for, the everyday universe of discourse. It would be difficult to find a representative language sample of interpersonal communication without a substantial part of mental terms. The mere fäet that our language is so rich in mental words and phrases may be taken as the cultural acknowledgment of and social belief in the conscious character of human subjectivity. Having commerce with others and oneself is having commerce with conscious beings. Without going into the philosophical controversy about the reality of mental states, it is, for the present purpose, sufficient to state that since people communicate effectively by using mental terms they mutually refer to something that they posit to be real (Searle 1969). However, when we make use of mental terms, do we always address our own or other people's intentions, feelings, memories, beliefs, attitudes, etc.? In many cases we indeed mean to refer our interlocutor to a mental state if we have the impression that it is relevant for the other one to know. The phrase "I love you" may (hut need not) be meant as the expression of an inner state, of a feeling that I wish to communicate and, hopefully, share. That, however, is not always the case. The utterance "I want to know what to expect", contains three mental terms; at least, "want", "know" and "expect" are also terms for mental processes. But this sentence does not necessarily nor normally refer the hearer to 54 three different mental states of the speaker, who may merely press for decision and be eager to get this intention across. Applying Karl Bi.ihler's "organon mode!" of signs (Bi.ihler 1982), the rule seems to be that one and the same mental term or phrase may serve referential as well as other communicative ("appellative") purposes, depending on (linguistic) context and (social) situation. The utterance, for instance, "I am afraid", if spaken by a child asked to fetch something from a dark basement room; may indeed refer the parent to, and be expressive of, the child's state of apprehension or fear. At the same time, the sentence communicates a tentative refusal to go. The utterance, however, "I am afraid not", if given as an answer to the question "Can you borrow me a hundred dollars?", isa more or less polite "No" or "Sorry", although occasionally it may also betray a person's fear of separating from his or her money. If the assumption is valid that we have to rely on the context and the situation to be sure that mental terms and phrases do carry mental reference, it is also safe to assume that the more personal a situation is the higher the probability that intcrlocutors mean to refer to their partner's, their own, or third persons' mental states because in such situations thoughts, feelings, preferences, aversions, memories, and dreams do become interesting. The prototype of a highly, or even exclusively, personal situation may be the case of two lovers mutually exploring one another's mind or soul-a way of becoming familiar and intimate with one another. There are, however, other situations in which this highly personal discourse is neither expected nor considered to be good taste. Nevertheless, we occasionally meet persons who, by means of a very personal language, try to inquire into another person's moods, feelings, sentiments, and motives when such information is not volunteered. It may be that some individuals are flattered if others seem to be interested in their mental life. But it will equally happen that such enticements into "self-disclostire" are resented as an indecent ot improper "invasion of privacy". Whichever obtains, those who like to intrude verbally into other persons' "inner spheres" as well as those who are embarrassed by such psychological curiosity, confirm the social reality of both the inside­ outside metaphor of mind and of a boundary between these two psychological "spheres" (cf. Kruse 1980). As we know from research on privacy and on self­ disclosure, communication across this boundary is socially regulated and so is interpersonal discourse. We should not hesitate to recognize such rulcs of discourse as basically moral rules. The socially amhiguous tendency to gain "insight" into the inner life of another person who is not an intimate, is certainly a case for differential psychology. But the recent spread of "psycho-talk" in everyday social life will probably have to be explained by two related public trends: the spread of clinical "psychologese'.' and the popularity of TV talkshows, for which personal or even intimate questions have become significant features. 55 3 Psychologizing: in interpersonal discourse We have now come close to what is called psychologizing in the narrow sense of the word. In order to understand this socially important as well as controversial technique, it may be helpful to start with a basic definition and an exemplification. The generally accepted meaning of psychologizing is the explanation or interpretation in psychological, especially psychodynamic terms of matters which in themselves are not, or not exclusively, of a psychological nature. One variant of psychologizing may be found within the sciences, viz., whenever historical, political, or social events, philosophical or Iinguistic concepts are accounted for in (if not reduced to) psychological terms. Such attempts to find a psychological foundation for non-psychological states of affairs and events are cases of psychologism. Historically, it has mainly been found (and criticized) in philosophy (logics), history, sociology, and Iinguistics, less so in modern psychology. This intra-scientific psychologizing will not be dealt with here. The focus of the following paragraph is on psychologizing in everyday life, i.e., on the usage of psychological concepts and arguments in contexts and situations where reference toa speaker's or writer's mental states is unwarranted or, at least, surprising. If, for example, in a serious, business-like discussion one of the participants, instead of responding to the proposition of an opponent's statement, reacts to the latter's mental makeup by referring (insinuatively or evasively) to the presumable affective or motivational state "behind" his adversary's argument, we speak of psychologizing ad hominem (cf. Graumann 1985). In general, psychologizing is the rhetorical device of not reacting directly to an act or to the propositional content of a speech-act, hut of referring to the alleged psychological "background" of this act, i.e. to the actor's conscious or unconscious mental states. Psychologizing ad hominem may serve different purposes. It may be the overwhelming interest in a person's mental life rather than his or her utterances. Psychologizing may be a technique of avoiding a critical or embarrassing issue. It may also be a case of verbal aggression if the motives or intentions hinted at are morally questionable or socially undesirable. Inversely, it may be used in defense if we refer to our own (or our friends') basically positive mental dispositions in accounting for criticized behavior (Graumann 1985). Since in interpersonal as well as in public situations the evaluative accounting for others' and one's own behavior seems to be one of the major functions of psychologizing, it is important to see how such accounting is achieved. Whenever we refer to a person's mental dispositions or traits we not only address features that we consider to be general and permanent. We also conceive of traits as essential features of a person, which tel1 us who the person "really" is. Stealing something easily makes a person a "thief'. Y et, once labelled a "thief', he will remain a "thief' even if he stops stealing. If a 56 subsequent act contradicts the attributed trait we still may argue: "The exception proves the rule". The disposition, although invisible, tends to become the more powerful psychological reality than a person's merely contingent behavior, which, furthermore, may be dissimulated. "Pinning" or "nailing" someone down toa socially undesirable or downright immoral trait or type by means of trait- or type-names is a common technique of prejudicial and discriminatory discourse (Graumann & Wintermantel 1989). The social power of psychologizing relies on this indeterminacy of the relation between mental dispositions and actual behavior, which is wide open for interpretation. Alleging, alluding, insinuating, deriding, suspecting are social and rhetorical techniques of interpersonal aggression aiming at questionable motives or traits. Inversely, we may appeal to !'positive" dispositions if we try to defend a controversial conduct. We may have done something stupid or inappropriate, but we need not take full responsibility since we can account for it by referring to pertinent mental states or traits. We may either excuse our misbehaviour by apologetic reference to a temporary state of mind, such as inattention, preoccupation with a problem, grief or stress. Or we try to justify what we did by relating our behaviour to underlying positive personality dispositions such as open-mindedness, good faith and kind-heartedness, which sometimes "make us do" inconsiderate things. Here, appealing to mental traits is like pleading for mitigating circumstances, as if one's character was a fixed given entity that others as well as we ourselves have to take into account (for an overview of theories and research on the "accountability of conduct" cf. Semin & Manstead 1983). The self-assertive proposition, "It's the way I am", which we occasionally present half-apologetically, half-defiantly, refers our interlocutor to such a last instance or ultima ratio. The presupposition in all psychologizing is that referring to mental dispositions is referring to the "true" character of a person.

4 Psychologizing: in public discourse A last word about psychologizing in everyday public discourse. From the fäet that psychologizing presupposes the existence of an individual (other or self) to whose mental states we refer it does not follow that psychologizing is restricted to interpersonal communication. It is equally practiced with respect to groups and social categories, which, in conclusion, shall be demonstrated with respect to public discourse. lf we are critical of a social movement or a political party we may, in a straightforward manner, argue against the program of the party or movement. We may, however,-and this is frequently done in political debates-use the rhetorical strategy of psychologizing. Y et, if we are up against a social group, such as a party or the Government, or against a social category, such as the Socialists, the Jews or the Feminists, we can only psychologize if we, first of all, 57 personify and personalize our targets. This is accomplished by either of two discursive strategies. A. We single out an individual who for us personifies, i.e. represents, the whole category or even embodies the spirit of a movement, and it is toward him or her that we can assume a given attitude, be it of admiration, of critique, or of contempt. We may, if we are confronted with a social issue or movement, concentrate our attitude toward that movement and its ideology on a person whom pars pro foto we consider to embody its significant features. Sometimes this figure is a socially accepted representative of the group, sometimes he or she is a personification ad hoc, i.e., for a given purpose: the urgently needed addressee of a critique or the target of a political or police measure, as when someone is alleged to be the leader of, or the moving force behind, a collective movement. Mainly, in so-called conspiracies it is essential to identify a person who, however innocent, is to be held responsible for a development that is considered harmful for the established order (Graumann & Moscovici 1987). This tendency toward personification does not only serve tp.e need of having someone or some group to blame. It is also endorsed and augmented by the modern media, mainly by television. The necessity to visualize political, economic and other social trends and events in a concrete audio-visual manner inevitably leads to the presentation of persons. They are either introduced as agents, victims, eye-witnesses, or they belong to the mixed category of experts. What they say and how they say it is often the only palpable concretization of an otherwise invisible and elusive occurrence. Once a person has been made a salient representative of a social category or current, one requirement of psychologization is met: We deal with an individual. Only now can we personalize, i.e., endow our personification of a cause or trend with personal, i.e., psychological qualities. The figure which stands for a supraindividual, e.g. social, state of affairs is furnished with mental dispositions, i.e., traits, habits, attitudes and motives and may thus become the target of psychologizing in political discourse. B. The other strategy is stereotyping, i.e., conceiving of (and, consequently, labelling) all members of a given or assumed group or category, as being of a more or less fixed mental makeup. This psychological outfit consists preferably of aset of traits that allows for no individuality. Mainly with regard to minority groups we find stereotypical statements like "All Jews (Freemasons, Feminists, gypsies, etc.) are ...(here usually follows a series of adjectives)". Each set of trait-names not only makes up a type, but implies a "psychology" of its own and enables the lay psychologist to psychologized in terms of his or her folk or."lay theory". Once again, this means that even in the rhetorics of the public domain we still refer to something which, although invisible and "abstract", due to its dispositional potency, is nevertheless ascribed a reality that for many is more trustworthy and reliable than manifest, but possibly dissimulative, behaviour. 58 For this is the socially shared representation and esteem in which "the mental", now as ever, is held in everyday discourse, not, however, in scientific psychology.

References Btihler, K., 1982: Semiotic Foundations oj Language Theory. (ed. Robert E Innis.) New York, Plenum. Fluck, H.R., 1976: Fachsprachen. Einfiihrung und Bibliographie. Mtinchen. Graumann, C.F., 1984: BewuBtsein und Verhalten. Gedanken zu Sprachspielen der Psychologie. In H. Lenk (ed.), Handlungstheorien interdisziplinär, vol. III, 2. Mtinchen, Fink, pp 547-573. Graumann, C.F., 1985: Psychologisieren als Angriff und Verteidigung. In Psychologie, Psychologisierung, Psychologismus. (Schriften der Carl­ Friedrich-von-Siemens-Stiftung, vol. 9.) Mtinchen, Oldenbourg, pp 133-149. Graumann, C.F., 1991: Wiederannäherung an Psychologie. In K. Grawe, R. Hänni, N. Semmer & F. Tschan (eds.), Uber die richtige Art, Psychologie zu betreiben. Göttingen, Hogrefe, pp 1-10. Graumann, C.F., (in press): Psyche and her descendants. In K.J. Gergen & C.F. Graumann (eds.), Historical Dimensions oj Psychological Discourse. New York, Cambridge University Press. Graumann, C.F. & Moscovici, S. (eds.), 1987: Changing Conceptions oj Conspiracy. New York, Springer-Verlag. Graumann, C.F. & Wintermantel, M., 1989: Discriminatory speech-acts: A functional approach. In D. Bar-Tal, C.F. Graumann, A.W. Kruglanski & W. Stroebe (eds.), Stereotyping and Prejudice: Changing Conceptions. New York, Springer-Verlag, pp 183-204. Kruse, L., 1980: Privatheit als Problem und Gegenstand der Psychologie. Bern, Huber. Mandler, G. & Kessen, W., 1959: The Language oj Psychology. New York, Wiley. Moscovici, S., 1976: La Psychanalyse: Son Image et son Public. (2nd ed.) Paris, Presses Universitaires de France. Oksaar, E., 1989: Alltagssprache, Fachsprache, Rechtssprache. 'Zeitschrift fiir Gesetzgebung,4:210-237. Searle, J.R., 1969: SpeechActs. Cambridge, Cambridge University Press. Semin, G.R. & Manstead, A.S.R., 1983: The Accountability oj Conduct. London, Academic Press. Steinig, W., 1981: Psychologische Fachsprache und Alltagskommunikation. In Th. Bungarten (ed.), Wissenschajtssprache. Mtinchen, Fink, pp 422--453. WRITING FOR DIFFERENT AUDIENCES AND IN DIFFERENT SETTINGS

Generic Integrity in Prof essional Discourse

Vijay K Bhatia

Legislative writing has acquired a certain degree of notoriety rarely equalled by any other variety of English. It has long been criticized for its obscure expressions and circumlocutions, long-winded involved constructions and tortuous syntax, apparently meaningless repetitions and archaisms. On the other hand, it has been emphasized again and again by the drafting community that the ultimate aim of legislation is to drive the judge to adopt the meaning which Parliament wants him/her to take and a majority of parliainentary counsels admit that it is almost impossible to achieve this objective using plain English. The competing demands imposed by the two sides appear to be almost irreconcilable. The inevitable result of all this has been a slow and rather painful pace of language reform in the past few decades. For the specialist community, the radical reforms requiring drastic simplification of legislative documents suggested by the plain language campaign amount to neutralisation of the generic integrity of legislative writing. I would like to suggest that it is possible to make legal texts more accessible to a wider range of intended readership without necessarily neutralising the generic integrity of the legislative genre. I would like to extend the notion of 'easification' as originally proposed in Bhatia (1983) to suggest that it is possible, at least to some extent, to meet the requirements of the drafting community, on the one hand, and those of the readers, on the other. Since easification procedures are meant to make a text more readable without in any way changing the essential nature of the text, there is relatively little chance of neutralising the generic integrity of legislative documents. In this paper I shall examine a few instances of legislative provisions in an attempt to unravel the mysteries underlying this complex artefact and to answer the following three questions: 1. Why are legislative provisions written the way they are? 2. What is it that makes them inaccessible to specialists and non­ specialists alike? 3. Is it possible to make them more readily accessible to specialist readers? If so, how? 62 To answer the first question and to appreciate the complexity of legislative statements, we need to have a better idea of the communicative purpose(s) these statements are meant to serve and the constraints that are imposed on the drafting of these provisions. Legislative writing is highly impersonal and decontextualized, in the sense that is illocutionary force holds independently of whoever is the speaker (originator) or the hearer (reader) of the document (see Bhatia 1982, Gunnarsson 1984-also). The general function of this writing is directive, to impose obligations and to confer rights. As legal draftsmen are well aware of the age­ old human capacity to wriggle out of obligations and to stretch rights to unexpected limits, in order to guard against such eventualities, they attempt to define their mode! world of obligations and rights, permissions and prohibitions as precisely, clearly and unambiguously as linguistic resources permit. Another factor which further complicates their task is the fäet that they deal with a universe of human behaviour, which is unrestricted, in the sense that it is impossible to predict exactly what may happen within it. Nevertheless, they attempt to refer to every conceivable contingency within their mode! world and this gives their writing its second key characteristic of being all-inclusive (see Bhatia 1982). Legislative writing differs significantly from most other varieties of English, not only in terms of the communicative purpose it is designed to fulfil, but also in the way it is created. In most other written varieties, the author is both the originator and the writer of what s/he creates, whereas in legislative provisions, the parliamentary draftsman is only the writer of the legislative act, which originates from the deliberations of a parliament in which s/he is never present. Similarly, in most varieties the reader and the recipient for whom the document is meant are the same person, whereas in the case of legislative provisions, the document is meant for ordinary citizens but the real readers are lawyers and judges, who are responsible for interpreting these provisions for ordinary citizens (see Bhatia 1982, Gunnarsson 1984, 1989 also). The result of this unique contextual factor is that the parliamentary draftsman finds his loyalties divided. On the one hand, s/he has to acknowledge his loyalty to the will of parliament and on the other hand s/he must use linguistic and discoursal strategies to help the intended readership towards what Candlin (1978) refers to as "the equalising of interpretative opportunity". In other words, s/he is required lo use linguistic resources and discoursal strategies to do justice to the intent of parliament and to facilitate comprehension of the unfolding text for ordinary readership at the same time. Reconciling the two is not always an easy task. His predicament is well summed up in the words of Caldwell, an experienced practitioner in the field .

... there's always the problem that at the end of the day there's a system of courts and judges who interpret what the draftsman has done. It is very difficult to box the judge firmly into a comer from which he cannot escape [ ... ] given enough time and given 63 enough length and complexity you can end up with precision but in practice there comes a point when you can't go on cramming detail after detail into a bin ... (Reported in Bhatia 1982:25.) So, it is this seemingly impossible task of achieving the dual characteristic of clarity, precision and unambiguity on the one hand, and all-inclusiveness on the other hand, that makes legislative provisions what they are. Moving on to the next question, legislative provisions display the use of several typical linguistic and discoursal strategies which are rarely used so predominantly in any other genre. They include unusually long sentences, an above-average use of complex prepositions instead of their simple counterparts, their highly nominal character and initial lengthy case-descriptions (see Bhatia, 1987, 1989, 1991, 1993). However, the most important characteristic of the legislative provision is the use of qualifications without which the provision will lose its essential nature. Most legislative provisions are extremely rich in qualificational insertions within their syntactic boundaries. Let me illustrate how this is done by taking a made-up provision based on a section from the Housing Act 1980.

[1) Where a secure tenant serves on the landlord a notice in writing claiming to exercise his right to buy the dwelling-house, and if the Iandlord refuses to admit the tenant's right to buy the dwelling-house, then, subject to the following provisions of this Section, the Secretary of State, if he thinks proper, may by means of a written notification make special regulations in pursuance of his powers under Section 15 of this Act for the purpose of enabling the tenant to exercise his right to buy the dwelling­ house (notwithstanding anything contained in Section 51 of the Land Registration Act 1925) within a period of six months from the

[lA]

Provisionary Clanse Qnalifications Legal Function

Where a secure lenant serves on ] the landlord a notice in writing PREPARATORY [ claiming to exercise his right to Case Description buy the dwelling-house,

[ ·· and if the landlord rcfuses to ] admit the tenant's right to buy PREPARATORY the dwelling-house, Specifying Conditions

then, subject lo the following REFERENTIAL provisions of this Section, J Defining Scope lhe Secrelary of Slate, J

lf hHhh,k.,,o,.,, ] PREPARATORY >[ Stating Contingency may ~/[ by m•= of • wd,,., »oUfioo

~~in pursuance of his powers under] REFERENTIAL Seclion 15 of this Act Textual Authority

for the purpose of enabling the ] OPERATIONAL tenant to exercise his right to Staling Purpose buy the dwelling-house

(notwithstanding anything ] REFERENTIJ\L contained in Seclion 51 of Defining Scope the land Registration Act 1925).

within a period of six months J OPERATION AL from the dale of such a rcfusal, Temporal Relations

provided that the dwelling-house,] PREPARATORY or any part of it, is not bcing Stating Exceptiolls ~r used for charitable purposes I. r wit~in thc meaning of the l REFERENTIAL 1.,'Charitable Purposes Act 1954'. J Terminological Explanation 65 The qualifications seem to provide the essential flesh to the main proposition without which the provision will be nothing more than a mere skeleton, of very Iittle legal significance. In the words of Caldwell, a very senior practising parliamentary counsel,

... if you extract the bare bones [... ] what you end up with is a proposition which is so untrue because the qualifications actually negative it all [... ] it's so far from the truth [ ...] it's like saying that all red-headed people are to be executed on Monday, hut when you actually read all the qualifications, you find that only one per cent of them are ... (Reported in Bhatia 1982:51.) It is not simply the presence of qualifications that makes legislative provisions extremely complex but more interestingly the way these qualifications are inserted within the syntax of the legislative sentence, It is understandable that if one needs to incorporate a variety of qualifications within a single sentence one would like to have as many syntactic positions to insert them as are possible. But a consideration that makes this task even more difficult is the fäet that if qualifications on the one hand make the main provisional clause more precise and clear, they can also promote ambiguity if they are not placed judiciously. That is the main reason why legal draftsmen try to insert qualifications right next to the word they are meant to qualify, even at the cost of making their legislative sentence inelegant, awkward or tortuous but never ambiguous, if they can help it. The result of all this effort is that these qualifications are inserted at various points where they create syntactic discontinuities rarely encountered in any other genre.

Discontinuous Noun Phrase

[2] A secure tenant has the right -

(a) if the dwelling-house isa house, ta acquire the freehold oj the dwe/ling-house; (b) if the dwelling-house isa flat, ta be granted a long /ease oj the dwel/ing-house

(Section 1 [ I] of the Housing Act, 1980, U.K.)

Discontinuous Binomial Phrase

[3) Where a secure tenant serves on the landlord a written notice claiming lo exercise the right lo buy, the landlord shall (unless the notice is withdrawn) serve on the tenant, withinjour weeh, ar, in a case fälling within subsection (2) below, eight weeks, either

(a) a written notice admitting the tenant's right; or (b) a written notice denying the tenant's right and stating the reasons why, in the opinion of the landlord, the tenant does not have the right to buy.

(Section I [5] of the Housing Act, 1980, U.K.) 66 Discontinuous Complex Prepositional Phrase

[4] Any power of the Secretary of State to make an order or regulation under this Act shall be exercisable by statutory instrument subject, except in the case of regulations under section 22( 1), 33(2), 52(3), 56(7) or paragraph 11 of Schedule 3 or an order under section 52(4), 60 or 153, to annulment in pursuance of a resolution of either House of Parliament.

(Section 151[1] of the Housing Act, 1980, U.K.) Discontinuous constituents with fairly long qualificational insertions like these and many others embedded within them add considerably to an already complex syntactic character of the legislative sentence and cause serious psycholinguistic problems in the processing of such provisions, especially in the case of non­ specialist readership (see Bhatia 1984). Legislative genre thus has a conventionalized communicative purpose which is largely reflected not simply in the way it is conventionally written and read by members of the legal community (see also Gunnarsson 1989), but also in the way some of the syntactic and discoursal resources are used in it. Legislative genre therefore has good reasons to be what it is, and one should try to understand it in its own terms rather than by imposing standards of ordinary expression from the outside. Many of the attempts to reform legislative writing in the Western world have largely been ineffective because of their failure to recognize the value of an ethnomethodological position that the legislative provision reflects a sphere of practical reasoning which needs to be understood in its own terms. However, one often gets a feeling, quite justifiably in many cases, that this concern on the part of the specialist community for clarity, precision, and unambiguity on the one hand, and all-inclusiveness on the other, has been taken rather too seriously and, perhaps too far. It is true that legislative writing hasa long and well-established tradition and the style of legal documents has become firmly standardized, with the inevitable result that legal draftsmen tend to become comfortable with tried, tested and time-honoured linguistic expressions and style of writing over a period of time. This becomes particularly significant where each subsequent generation of parliamentary draftsmen is trained by the preceding one while on the job. To be fair to the drafting community, one must admit that there have been some improvements in style that have taken place in the past few decades, especially in the way textual-mapping devices are used to reduce information load at a particular point in the provision (see Bhatia 1987) but such reforms have been few and far between and as a consequence have gone unnoticed by reformist lobbies. On the part of the specialist community, therefore, there isa need to show two kinds of concern: firstly, a need to show greater awareness of their loyalty to the real readers of legislative documents; secondly, a need to use linguistic resources more consistently, particularly in cases w here certain linguistic forms are traditionally associated with specific meanings (see Swales 67 & Bhatia 1983). This is especially desirable in the use of what specialists call 'proviso-clauses'. This practice will create fewer problems of interpretation for many of the readers, the specialists as well as the non-specialists. There is also a greater need to make more efforts to use long syntactic discontinuities more sparingly, perhaps as an exception rather than a rule. However, even if such practices are adopted, we still have the problem of making legislation accessible to specialists and non-specialists alike. This brings us to the third question we raised in the beginning of the paper. How can we make legislation more accessible to a wide variety of readers? I would like to suggest two possibilities to make legislation more readily accessible taking into account reader characteristics, their familiarity with the subject matter and range of cognitive abilities including their background knowledge, their purpose in reading the document, including the task constraints and the generic properties of legislative statements. One is to create simplified or plain English versions of legislative provisions. There are already pressure groups in various countries involved in producing the over-all cognitive structure of the legislative statement. They also tend to reveal any problematic gaps or ambiguities in the coverage of the legislative provision. These alternative forms are being used for information and education purposes and are also becoming increasingly popular in business, computers, public health, advertising and many other areas where ordinary members of public are meant to take decisions on the basis of given information which is of technical nature. All these alternative popular versions, linear as well as non-linear, are effective and in some cases excellent examples of text presentation. However, simplification of this kind is a special form of intra-lingual translation, the purpose of which is to make a given text simpler in terms of its content and/or form. This is usually done by bringing its content or linguistic form within the area of experience of a particular group of readers. It invariably involves a kind of tampering with the original linguistic input in order to produce either a simplified version or a simple account of the original text. Widdowson (1978:88) discusses the two terms in detail. Simplified versions, he points out, are passages derived from genuine instances of discourse by a process of lexical and syntactic substitution, and hence are regarded as alternative textualizations of a given authentic discourse. Simple accounts, on the other hand, are genuine instances of discourse designed to meet a communicative purpose, and do not represent alternative textualizations. They are genuine instances of discourse in their own right. However, in either case the authenticity of the original genre which the text represents is comprom1sed. Whether it is a simplified version or a simple account, the resulting discourse is less likely to be a genuine instance of the genre which the original text represents. In the case of simplified versions, the content and/or lexico­ grammatical resources are tampered with in order to bring the text within the linguistic competence of a specific readership. This certainly changes the 68 authentic nature of the genre. In the case of simple accounts, it is even more likely that the original text will lose its essential communicative purpose. It will certainly, as Widdowson (1979) claims, be a genuine instance of discourse but will not necessarily represent the same genre as the original. A typical instance of this phenomenon is found in simple accounts of scientific research reports, which are called popular accounts of scientific discoveries and are simplified for non-specialist science enthusiasts in magazines like the Popular Science or the Scientific American. They may be called simple accounts and considered genuine instances of discourse, which, of course, they are, but they are different from original authentic scientific research reports because they represent two different genres, in that, they serve two very different communicative purposes. The moment one simplifies an original to produce an alternative or a different version of the original to produce an alternative or a different version of the original, the generic integrity of the original text is likely to be completely !ost or at least neutralized toa certain extent (Coulthard 1977: 153). Simple or plain English versions may have !imitations when used as alternatives to conventional forms of legislative statements, particularly the ones which are meant for specialist use. None of these, for example, when used as an alternative to the legislative statement can be used to legislate, that is to negotiate justice on the court of law. In their non-linear alternative forms they will be difficult to quote in legal proceedings. As instruments of legislation, they have been and perhaps, for a long time to come. will continue to be rejected by the professional legal community as ineffective, inefficient, imprecise and devoid of legal content. The main reason for this lack of interest in so-called reformed writing is two-fold. Firstly, the rewritten versions often obscure the generic integrity of the original legislative statement, and secondly, these can serve the communicative purpose of informing and educating the lay audience but are poor instrument of regulating human behaviour. The only way one can make them work is by changing the legal system. So long as the main function of legislation is to "box the judge firmly in a corner" (Bhatia 1982), the legisla­ tive statements will continue to take the shape they conventionally do. Loss in generic integrity resulting from the simplification of an original text may have negative repercussions for language reform. If that is the case, is there an alternative which can make a particular text more accessible to its intended readership and yet maintain the generic integrity of the original? Bhatia (1983) introduces the notion of easification as an alternative to simplification. Easification. as he point out, attempts to make the text more accessible to the lcarner by using a variety of what he cal!s easification devices, the purpose of which is to guide the reader through the cognitive structurc of the text without making any drastic changes to the content or linguistic form of the text, thus maintaining its generic integrity. He suggests a wide range of easification devices to make an authentic text more easily accessible to the reader depending upon the nature of the genre which the text represents, and the purpose of 69 reading. Easification is not only a technique for text presentation, he points out, but also a learning strategy which helps the learner to simplify the text for himself or herself, depending upon his background knowledge of the subject matter and of the language. This concept of "easification" can be validly applied to make legislation more readily accessible to a larger group of specialist audience without neutralizing the generic integrity, of legislative statements. Whereas simplification procedures create popular versions of legislative documents in plain language for lay audience, easification procedures can be used to create authentic and yet more readily accessible legislative provisions for the specialist audience. The basic difference between the two sets of procedures is that easification procedures are "genre-preserving", whereas simplification procedures can be "genre neutralising", to say the least. In some extreme cases, they can be "genre­ destroying" as well. The use of these two procedures will give us two different versions for two different sets of readership, one, an authentic, easified and authoritative version for the specialist use; and the other, a popular or simple version for the lay audience for information and education. I have discussed the use of simplification procedures to create popular versions elsewhere (Bhatia 1993). In this paper, I would like to focus on some of the easification procedures which can be used to make legislative statements more readily accessible without sacrificing their essential generic integrity. Let me discuss the use of easification procedures, some of which are already being used but not very consistently. None of these are firmly established within the professional contexts.

Easification of legislative documents for specialist use [A] Clarification of cognitive structuring: As already discussed above, one major problem that legislative texts pose to a non-specialist reader is the depth and complexity of modification. Many of these modifications appear in those syntactic positions where they create discontinuities in the structure of the legislative statement. The syntactic structure of a typical legislative statement must be understood in terms of a complex interplay of qualifications and the main provisionary clause. The most serious obstacle to a clear understanding of these legislative documents, therefore, is the complexity of syntactic structure, which can be overcome by clarifying the cognitive structuring underlying the provisions. I shall now take up an example for the analysis of cognitive structuring and present another way of achieving the same clarity of cognitive structuring.

[5] Where the dwelling-house with respect to which the right to buy is exercised is a registered land, the Chief Land registrar shall, if so requested by the Secretary of State, supply him (on payment of the appropriate fee) with an office copy oj any document required by the Secretary of State for the purpose of executing a vesting order with 70 respect to the dwelling-house and sha/1 (notwithstanding section 112 of the Land Registration Act 1925) a/low any person authorized by the Secretary oj State to inspect and make copies oj and extracts from any register or document which is in the custody oj the Chiej Land Registrar and relates to the dwelling-house. (Section 24[5] of the Housing Act, 1980. U.K.) In this case the highlighted sections represent the main provisionary clause and the remaining sections give the attendant qualifications which make the provi­ sion operative. [B] Reducing information load: The second serious obstacle to the comprehension of legislative statements is the high density of information load at a particular point in the syntactic structure of the legislative sentence (see Bhatia 1987). The tension between simplicity and clarity on the one hand, and certainty of legal effect on the other, is very common in all legal systems. However, in some of them, particularly in countries of the Commonwealth, it is further complicated by the requirements of the parliamentary process. The demand for all-inclusiveness resulting in excessive elaboration comes not only from the Government and the instructing department but also from Parliament itself. However, the present-day legislative writing shows the use of a new device to reduce information load at a particular point in the expression of legislative content. I have referred to this as a "textual-mapping" device and it is being increasingly used in present-day legislation. Let me illustrate this notion by taking up an example from an earlier legislative writing to demonstrate how easification procedures can make it more accessible using textual-mapping devices.

Original version of Section 32 of the Housing Act, 1957 [6] A local authority may pay to any person displaced from a house to which a demolition order roade under this Part of this Act, or a closing order, applies, or which has been purchased by them under this Part of this Act, such reasonable allowance as they think fit towards his expenses in removing, and to any person carrying on any trade or business in any such house they may pay also such reasonable allowance as they think fit towards the loss which, in their opinion, he will sustain by reason of the disturbance of his trade or business consequent on his having to quit the house, and in estimating that loss they shall have regard to the period for which the premises occupied by him might reasonably have been expected to be availablc for the purposes of his trade or business and the availability of other premises suitable for that purpose. (Section 32 of the Housing Act, 1957:797) This is a relatively simple legislative provision although it is written in the traditional old style, entirely self-contained, packed into a single sentence. It looks very neat, precise, unambiguous and all-inclusive but is difficult to process because of its syntactic complexity and sentence length. Like most other legislative statements this also calls for the exercise of a cross word mentality to 71 ferret out its meaning. However, it is possible to break it into a number of sub­ sections without sacrificing its legal content, preciseness and all-inclusive quality, thus preserving its generic integrity.

Alternative version of section 32 using textual-mapping devices

[6A] (1) Where a house, to which a demolition order made under this Part of the Act, or a closing order, applies or which has been purchased by the local authority under this Part of the Act, the local authority may pay to any person specified in subsection (2) below such reasonable allowance as they think fit towards his expenses mention in subsection (3) below and the loss mentioned in subsection (4) below. (2) The person mentiöned in subsection (1) above is- (i) anybody who is displaced from a house mentioned in subsection 1 above, or (ii) anybody who is carrying on any trade or business in a house mentioned in sub­ section l(i) above.

(3) Expenses mentioned in subsection (1) above are the expenses incurred by any such person mentioned in subsection 2(i) above in removing as a result of displacement.

(4) The loss mention in subsection (1) above is that which, in the opinion of the local authority, any person specified in subsection (2)(ii) above will sustain by reason of the disturbance of his trade or business consequent on his having to quit the trade.

(5) In estimating the loss mentioned in subsection (4) above, the local authority shall have regard to the period for which the premises occupied by him might reasonably have been expected to be available for the purpose of his trade or business and the availability of other premises suitable for that purpose. Breaking a lengthy legislative statement into several subsections obviously makes the main provisionary statement more easily accessible to its readers by postponing the less important and more complicating legal details to subsequent subsection(s). It may add to the length of the provision hut it will certainly make it easier to construe and hence more readily accessible to a wider range of audience. [C] lndicating legislative intentions: The third easification device that can make legislative writing more readily accessible without endangering its generic integrity is the use of statements of purpose to explain and clarify the legislative intent at various levels, particularly in the case of complex contingencies. This will tell readers what the Act or any section of it is about. Conard (1985) aptly illustrates this point by looking at the Trading with the Enemy Act, of the U.S.A.

The first thing that any law-reader wants to know is what the law is about. Does it affect him or doesn't it? Most laws fait [ ...] to answer this question [ ... ] A typical example is the Trading with the Enemy Act, which in time of war suddenly applies to tens of thousands of Americans. But if it came in the mail to an American who was 72

about to send sorne rnoney to a foreign creditor, he would have to parse two or three pages before he found out whether he was likely to be affected by it.

The first thing that the citizen would notice is that the Act begins with section 2 (rnay be s. 1 got !ost, he thinks). Section 2, which has no subtitle, contains a long list of definitions, starting with a half-page definition of 'enerny'.

Sornewhere in the rniddle of the second page begins section 3, also with no subtitle. lf the citizen is still looking, he will see:

That it shall be unlawful: (a) For any person in the United Stares, except with a /icense oj the President, granted to such person, or to the enemy, or ally oj enemy, as provided in this Act, to trade, either direct(v or indirectly, with, to, or from, or for, or on account oj, or on behalf oj, or for the benefit to any other person ...

lf he keeps on for a few lines rnore, and understands what he reads, he will at last find out what the law is about. (Conardl985.) A better way to start a statute, he continues, could have been to begin like this: "This law applies to everyone in the United States who has any dealings with foreigners." It could then have proceeded to explain what foreigners and what dealings were specifically affected. The Renton Committee Report (1975:63) also seems to be clear about the importance of such statements of purpose when they recommend that "statements of purpose should be used when they are the most convenient method of delimiting or otherwise clarifying the scope and effect of legislation". Sometimes statements of purpose or intention are also given in the form of subheading(s) and/or additional notes, as in the following example from The Personal Injuries (Compensation Insurance) Act 1963, Government of India.

Compensation payable under the Act, by whom and how payable-

[7] (1) There shall, subject to such condition as rnay be specified in the Scherne, be payable by an ernployer in respect of personal injury sustained by a gainfully occupied person who is a workrnan to whorn this Act applies, cornpensation, in addition to any relief provided under the Personal Injuries (Ernergency Provisions) Act, 1962 (59 of 1%2), of the arnount and kind provided by section 7:

Provide that. ..

(2) ...

(3) ...

NOTES- This section irnposes on the ernployers a liability to pay cornpensation for 'personal injuries' to the workrnen and also provides for the circurnstances in which the ernployer will be discharged from this obligation and the Central Govemrnent would then be liable. 73 [D] Illustrating legislative issues: In the communication of legislative intentions there are always areas where a particular interpretation can be viewed as doubtful or difficult to interpret. Conard (1985) cites a good example of this from a bill in the recent Congress which provided,

[8] That hereafter, except as otherwise specially provided by the Act of Congress, no action of recovery of wages, penalties or other damages, actual or exemplary, pursuant to any other law of the United States shall be maintained in any court unless the same was commenced within one year after such cause of action accrued: ... He continues,

The only way a lawyer can know whether the one-year !imitation of this law applies is to search the United States Code from slem to slem. If he can't find a !imitation anywhere else, then this is it. A catch-all !imitation may be a good thing, just to cover the unknown and forgotten cases. But this wasn't that kind of a law. The draftsman knew precisely what actions he wanted to limit, and he listed them, with citation, in the Committee report. If he bad listed them in the statute, he would have saved hours of research for thousands of lawyers ... If the draftsman had wanted to enlighten bis reader, he would have written:

'The following actions must be brought within one year after the cause of action accrued: (I) Suits for treble damages based on infringement of a registered trademark (17 U.S.C., s25); (2) Suits based on infringement of copyrights (27 U.S.C., s25);' and so on down the list. (Conard 1985:77-78.) In cases like these, it is always helpful to add notes or additional illustrations as part of the access structure of the legislative statement. Illustrations of this kind definitely are in use but very sparingly. Let me take an example from The Indian Code ofCriminal Procedure, 1973.

[9] 212.(1) The charge shall contain such particulars as to the time and place of the alleged offence, and the person (if any) against whom, or the thing (if any) in respect of which, it was committed, as are reasonably sufficient to give the accused notice of the matter with which he is charged. (2) When the accused is charged with criminal breach of trust or dishonest misappropriation of money or other movable property, it shall be sufficient to specify the gross sum or, as the case may be, describe the movable property in respect of which the offence is alleged to have been committed, and the dates between which the offence is alleged to have been committed, without specifying particular items or exact dates, and the charge so framed shall be deemed to be a charge of one offence within the meaning of section 219. Provided that the lime included between the first and the last of such dates shall not exceed one year. 74 213. When the nature of the case is such that the particulars mentioned in sections 211 and 212 do not give the accused sufficient notice of the matter with which he is charged, the charge shall also contain such particulars of the manner in which the alleged offence was committed as will be sufficient for that purpose. [Illustrations]

(a) A is accused of theft of a certain article at a certain time and place. The charge need not set out manner in which the theft was effected.

(b) A is accused of cheating B at a given time and place. The charge must set out the manner in which A cheated B.

(c) A is accused of giving false evidence at a given time and place. The charge must set out that the portion of evidence given by A which is alleged to be false.

(d) [ ...]

(Sections 212 and 213 of the Code of Criminal Procedure, 1973, Govemment of lndia) Illustrations of this kind, whether real cases from past judgements or imaginary ones, do not interfere with or neutralize the generic integrity of the legislative statement, but they stand outside as additional aid to interpretation and thus help comprehension of complicating legal content. A variety of other easification procedures can be additionally devised to facilitate comprehension and interpretation of complex contingencies depending upon the nature of the genre, the kind of psycholinguistic difficulties it poses to aset of readers, and the purpose of the reading activity. An excellent example of the use of some of the easification procedures I have mentioned here may be found in the efforts of Law Reform Commission of Victoria in Australia, (see Eagleson 1988) who have quite successfully managed to combine the profes­ sional as well as social cancerns in the drafting of legal documents. So long as any of these easification strategies do not interfere with the generic integrity of the text, they should be acceptable to the drafting community. In plain language literature, however, we often come across examples which transform legislative statements into plain language quite effectively for non­ specialist readership, but for the drafting community these versions invariably neutralize the generic integrity of the legislative genre, thus giving the impres­ sion that the essential certainty of legal effect is either reduced or completely !ost. In conclusion, I would like to emphasize that ideally we need to have two different versions of legislative documents. One simplified in plain English with a lot of use of non-linear communicative strategies for lay audience. Its main function should be to inform and educate non-specialists, who do not need to be thoroughly well-versed in the intricacies of legal content but who do need to be aware of specific laws in general terms. The other version, authentic, though easified in a number of ways, and authoritative should be meant for specialist 75 use. Such a conventional version can be complex, detailed and often considered complete, and should be meant for specialists like lawyers, public officials and judges, whose job is to discuss and negotiate justice. It is possible to imagine a situation where the society and therefore the constitution accepts the dominant role of the judiciary in the negotiation of social justice, in stead of the parliament, as happens to be the case at present in mast of the countries of the Commonwealth. If such a situation arises, then the expression of legislative intentions can take the shape of general principles, thereby making the plain language as the obvious choice of legislative writers. As things are, the con­ ventional legislative statement and it plain English version serve two very different speech communities. The two versions serve different communicative purpose(s), they are meant for two different audiences, having very different background knowledge and different motivations for reading and understanding the document. The two versions, therefore, represent two different genres.

References Bhatia, Vijay K., 1982: An lnvestigation inta Format and Functional Char­ acteristics oj Qualifications in Legislative Writing and its Application ta English for Academic Legal Purposes. A Ph.D. thesis submitted to the University of Aston in Birmingham, U.K. Bhatia, Vijay K., 1983: Simplification v. easification: The case of legal texts. Applied Linguistics, 4,1:42-54. Bhatia, Vijay K., 1984: Syntactic discontinuity in legislative writing and its implications for academic legal purposes. In A.K. Pugh & J.M. Ulijn, (eds.), Reading for Professional Purposes-Studies and Practices in Native and Foreign Languages. London, Heinemann Educational Books, pp 90-96. Bhatia, Vijay K., 1986: Specialist discipline and the ESP curriculum. In M.L. Tickoo (ed.), Language Across the Curriculum. Singapore SEAMEO Regional Language Centre. Bhatia, Vijay K., 1987: Textual-mapping in British legislative writing. World Englishes 6,10:1-10. , Bhatia, Vijay K., 1989: Nativization of job application-A microethnographic study. A paper presented at the International Conference on English in South Asia, Islamabad, Pakistan, 4-8 January, 1989. Bhatia, Vijay K., 1991: Pragmatics of the use of nominals in academic and professional genres. A paper presented at the Fifth Annual International Confercncc on Pragmatics and Language Learning, (4-6) Aprii, 1991 at the University of Illinois at Urbana-Champaign, Urbana. Bhatia, Vijay K., 1993: Analysing Genre-Language Use in Professional Settings. (Applied Linguistics and Language Study Series. Edited by Christopher N. Candlin.) London, Longman. 76 Conard, Alfred F., 1985: A legislative text-New ways to write laws. Statute Law Review, pp 62-83. Coulthard, M., 1977: An Introduction to Discourse Analysis. London, Longman. Eagleson, Robert, 1988: Efficiency in legal drafting. In David St. Kelly (ed.) Essays on Legislative Drafting. The Adelaide Law Review Association, University of Adelaide, pp 13-27. Government of India, 1963: The Personal Injuries (Compensation lnsurance) Act 1963. Government of India, 1973: Indian Code of Criminal Procedure, 1973. Gunnarsson, Britt-Louise, 1984: Functional comprehensibility of legislative texts: Experiments with a Swedish act of Parliament. Text, 4-1/3:71-105. Gunnarsson, Britt-Louise, 1989: Text comprehensibility and the writing process. Written Communication, 6/1:86-107. H.M.S.O., 1980: The Housing Act, 1980. London, HMSO. Renton, David, 1975: The Preparation oj Legislation. A Report of a Committee Appointed by the Lord President of the Council, London, HMSO. Swales, John M. & Bhatia, Vijay K., 1983: An approach to the linguistic study of legal documents. Fachsprache, 5/3:98-108. Metadiscourse in Academic Texts

Minna-RiittaLuukka

Every communicative situation, whether spoken or written, is a form of interaction between the producer and the receiver of the text. Discourses and texts are not to be considered in factual terms only: they are more than propositions, or fäets to be conveyed. They also include elements that are used for interactional management and textual organization. In this article, these interactional features of academic texts are discussed. I call these features metadiscourse. First, I will briefly describe the description · system which is based on the analysis of my data (Luukka 1992). Secondly, I will discuss some possible applications of this system. The description system was developed for the Finnish language, but the functional character of the categories is applicable to other languages as well. My corpus consists of five papers that were delivered at a Finnish conference of applied linguistics, and of the articles based on these papers. Such data make the comparison of spoken and written text optimal, because the authors and the topics and even the audience are approximately the same. Thus, the only difference is in modality. This means that the possible differences found in the use of metadiscourse are likely to be result from the fäet that a spoken delivery differs considerably from writing as a process. The spoken corpus was recorded and transcribed. The papers were, as academic presentations usually are, planned in advance hut none of the speakers read his or her paper aloud. The amount of preparation varied: Two of the speakers had written a draft version of their paper and, more or less, followed it but they did not read it aloud. One of the speakers had only prepared a few transparencies and she used them to structure her spontaneous speech. On the whole, the presentations were quite informal and spontaneous.

The concept of metadiscourse The term metadiscourse was originally used to refer to 'writing about writing' or 'discourse about discourse' (Williams 1981). Recently, however, the concept of metadiscourse has been expanded. According to Crismore (1989), for example, metadiscourse has several semantic, social, psychological, commu­ nicative control and rhetorical functions. For instance, metadiscourse is used to 78 frame the propositional material, make the discourse cohesive and coherent, indicate attitudes, certainty and contact with readers. Metadiscourse is typically manifested in the use of connectives, modal expressions and personal pronouns. In previous studies, metadiscourse has usually been defined as non­ propositional or non-topical elements of texts (Vande Kopple 1985, 1988, Lautamatti 1978a, 1978b, Crismore 1984, Markkanen et al. 1993). According to these definitions, metadiscursive elements are not supposed to add meanings to the propositional information of texts. Thus, they could even be omitted from a text without changing its "actual meaning". This definition, however, seems problematic when a more holistic view is taken and texts are analyzed as communicative acts. The actual meaning of a text is not to be regarded as a list of propositions. Actually, the traditional concept of proposition may not be relevant at all, if we want to find out what is really said and meant in a text. Therefore, I do not define metadiscourse as "non-propositional elements of texts", but, instead, take a more functional approach. Metadiscourse is used for interaction management and textual organization. In Halliday's (1973) terms, metadiscursive elements fulfil textual and/or interpersonal functions of language (see also Vande Kopple 1988). Thus, metadiscourse enables authors to point out how they develop their arguments, and what kind of structure the text has. In addition, the attitude of the author both to the content of the text and to his/her audience is revealed through metadiscourse. The mearis of metadiscourse vary from one particular word, such as an adverb or a connective, to several words or whole utterances.

The description of metadiscourse

In the present system the three basic functions of metadiscourse are textual, interpersonal and contextual. By means of textual metadiscourse the author structures the texr, while interpersonal metadiscourse is used for signalling attitudes towards the content of the text or persons involved in communication situation. Finally, contextual metadiscourse offers the author means of com­ menting on the communicative situation or the text as a product. Textual metadiscourse is used for guiding the audience through the text by making explicit, how the text is organized, and by showing how different parts of the text are related to one another. First, these cues can be used to explicate the global structure, or superstructure of the text.

(1) In my paper I will first describe A, secondly deal with B, then clarify C and then conclude with D. Secondly, the author can, in a similar way, make explicit the structure of smaller parts of the text, the macrostructure. These structural signals are mainly textual connectors. 79 (2) A has many effects: firstly it causes B, secondly it increases C, and finally it decreases D. And finally, explicit topic shifts can also show the organization of a text, a speaker can, for example say: And now let's go 011 and discuss A. In written texts, topic shifts are often signalled by headings. The second type of metadiscourse is interpersonal. These elements are signs of the author's personality and make his/her presence explicit in a text. They can express the author's attitudes, evaluations, comments, and degree of certainty, hut they also imply the relationship between the author and the audience. Here, interpersonal metadiscourse is divided in two categories: 1) signals of attitudes towards the content of a text and 2) signals of attitudes towards persons involved in the discourse. With the signals of attitudes towards the content authors can make explicit the degree of their commitment to the content, that is, how certain or uncertain they are of what they are saying. I have called these indicators certainty markers and divided them into two categories: uptoners and downtoners. In addition, there are two subtypes certainty downtoners, i.e. reservations and restrictions, which differ in strength as well as in meaning. Firstly, reservations express most clearly the truth value of claims, and they are, broadly speaking, related to epistemic modality. Reservation devices include I) epistemic adverbs (eg. possibly, perhaps), 2) conditional and potential moods, which in Finnish are expressed with verb endings (eg. sanoisin =="I would say", tarvinnee =="He might need"), 3) modal auxiliaries (eg. can, may) and 4) factitive and unipersonal verbs (eg. I assume/suppose, it seems !hat). Compared to reservations, restrictions tone down certainty to a lesser extent because the writer considers his or her claim true, but restricts its truth value as regards time (eg. "X is often Y"), quantity (eg. In this particular case X is Y) or generalisability (eg. "In my view X is Y"). (3) It seems that in this case A may have an impact on B. I would say that these results suggest that A is often caused by B. On the other hand it is possible that... In addition, authors can also indicate their own attitudes towards the assumed importance or interest of the subject matter, and moreover, their personal feelings and evaluation of the fäets they present. I have called these eva!uation markers. (4) It's interesting to note that [... ] One very important fact is that [... ] unfortunately research on this problem is still lacking The second type of interpersonal metadiscourse involves signals of interactional attitudes. The function of these signals is to make the presence of 80 the author and the audience explicit or implicit in a text. I have classified the indicators of interactional attitudes in three groups:

1. Presence of author (I Involvement) These utterances highlight or efface the author in the text. (5) In my research l've found that [... ] Next, I will discuss ... The results of this study indicate that...

2. Presence of audience (Y ou Involvement) Utterances that highlight or efface the audience in the text. (6) As you all know, A causes B, but you may think that... If you look at this picture, you can see that ... In this picture one can see ...

3. Presence of a_uthorand audience (We Involvement) Utterances that highlight the audience in the text and create solidarity. (7) As we already know, A causes B Now let's move on and talka little about B In describing the interactional features of text, the focus is on how personal elements are either highlighted or effaced. Highlighting a person refers here to expressions by means of which the speaker or writer makes an explicit reference to persons, for example, by using a pronoun. Effacement refers to those instances where the speaker or writer chooses a less direct approach, that is, refers to persons but uses a circumlocution such as the passive or the pronoun we instead of /. Finally, the third main category of metadiscourse is contextual meta­ discourse. By contextual metadiscourse I refer to all utterances that somehow relate either to the communicative situation, to the text as a product or to the process of production. Authors use these expressions to comment on their preceding or coming actions, or to draw attention to the tables or figures they are presenting. Thus, contextual metadiscourse consists of comments about situations and materials. (8) l'll move this slide a little bit so you can see better ... Now let me see, what was I going to say ... These differences can be seen in picture 1. 81 As compared to the earlier studies of metadiscourse, the present system emphasises more the interactional function of metadiscourse, while the textual function is given less attention. Present categorization also provides tools for a more systematic analysis of the relationship between the author and the audience. Earlier systems of describing metadiscourse have been used for analysing written texts only, and this might be the reason why the personal and situational elements have not been regarded as relevant and interesting as in this study. The main types of metadiscourse are summarized in the following figure:

TEXTUAL MEf ADISCOURSE signals of superstructure signals of macrostructure signals of topic shifts

INTERPERSONAL MEf ADISCOURSE signals of propositional attitudes certainty markers uptoners downtoners evaluation markers signals of interactional attitudes presence of author (I) presence of audience (you) presence of author and audience (we)

CONTEXTUAL MEf ADISCOURSE comments about the situation and processes comments about the text as a product

Figure 1 . Functions of metadiscourse

Results: spoken and written metadiscourse This study was primarily qualitative, and my corpus was fairly small. Therefore, no claims are made about the nature of Finnish scientific writing styles or speaking styles. The results should not be considered as the wholc truth about Finnish academic discourse, hut merely as suggestions. In order to describe the texts in detail and to campare written and spoken texts, I have used siinple quantitative measures (eg. the amount of different types of metadiscourse). The main unit in the quantitative analysis was an utterance, which is virtually equal to a clause in a written text. The number of utterances in written and spoken texts was surprisingly similar: the spoken texts contained 2,081, and the written versions 2,190 utterances. 82 The total amount of metadiscourse in spaken and written texts was nearly equal: 44% of written utterances and 49% of spaken utterances contained metadiscursive elements. However, there were clear differences in the distri­ bution of the main types of metadiscourse. For example, the amount of textual and contextual metadiscourse was higher in spaken texts. On the other hand, the amount of interpersonal metadiscourse, as a whole, was quite similar.

0 10 20 30 40 50 60 %

Il textual tz:llnterpersonal lfficontextual

Figure 2. The amount of metadiscourse

Structuring the. text The analysis of textual metadiscourse did not produce very interesting results. The only noteworthy thing was the small amount of textual signals, especially in the written texts, where only 8% of the utterances contained some kind of textual signal. Thus authors did not appear to offer much guidance for their readers. In written texts, textual metadiscourse was mainly realized by signals of macrostructure. Other types were relatively rare. In spoken texts, the amount of textual signals was slightly larger (13%) and most of the signals were used to mark topic shifts (eg. "Let's move on and discuss X"). In particular, the signals of superstructure were more common in spaken texts. They were found in all spaken texts but only in two of the written ones. 83 Propositional attitudes In eontrast, the analysis of interpersonal metadiseourse was both revealing and interesting. Even though the total amount of interpersonal metadiseourse was quite similar in spoken and written texts, there were remarkable differenees in the use of subtypes. Signals of propositional attitudes were more frequent in written texts and signals of interaetional attitudes in spoken ones. This may imply that writers eoneentrate more upon fäets, while speakers are also eoneerned with persons.

% 100

~ effaced persons Hl highllghted persons ~ evaluatlon markers Il cerlalnty markers

0 written texts spoken texts

Figure 3. Interpersonal metadiseourse

As a whole, expressions revealing personal attitudes and interests did not seem to be typieal of aeademie texts. I found very few evaluation markers either in written or spoken texts, and most of them appeared to emphasize the importanee of some partieular fäet, sueh as: "There is still one important fäet to eonside ..." Most of the propositional attitude signals were eertainty markers, and their amount was slightly higher in the written texts. This seems to indieate that writers are somewhat more eoneerned about the eertainty of the fäets they are stating than speakers. If there was any doubt about the eertainty, writers were eautious and employed modality markers and hedges. Aetually, these markers could have been ealled "uneertainty markers" sinee very few utteranees signalled a high degree of eertainty (e.g. "l'm quite eertain that..." or "These results definitely indieate that..."). Instead, authors favoured hedging utterances sueh as: "It is possible that...", "These results seem to indieate that ..." Writers modified the degree of their eertainty mainly with reservations, especially with modal verbs and adverbs. Speakers used weaker modifiers sueh 84 as hedges, shields and approximators, for example, "These results in a way indicate that. ..", "A is B, at least quite often." It appears that spoken texts are vaguer than written ones in this respect. It is also noteworthy that speakers often used double or even triple modifiers in the same utterance. (9) It is possible that A is sort of B, at least quite often this is the case, at least according to my findings, in this study A seems to have some properties of B, and so maybe I could claim that...

Interaction The most obvious difference between spoken and written texts can be found in the interaction management, particularly in the use of interactional attitude signals. The human touch obviously was more explicit in spoken texts. On the whole, the texts appeared to be fairly "self-centred", for most personal references were to the author as I or inclusively to us. Personal involvement becomes explicit most clearly in the use of personal pronouns. Finnish has six grammatical persons, and they can be indicated with personal pronouns and personal endings of verbs, for example: Minä istun LI sit], Me istumme [We sit]. In Finnish, one can indicate a grammatical person with a construction where a personal pronoun is combined with a verbal suffix. It is also possible to omit the pronoun and use only the ending of a verb in the 1st and 2nd person forms. As a result, even if authors want to refer to themselves, they can choose between the more explicit form Minä käsittelen seuraavaksi X:ää [I shall discuss X] or the more implicit form Käsittelen seuraavaksi X:ää, [*shall discuss X]. In most of the cases found in spoken texts, speakers used the construction of a personal pronoun plus an ending, but, in contrast, writers never used the pronoun minä, 'I' .. Also, there was a tendency to starta spoken utterance with a personal element, as opposed to a _tendency to locate it in non-initial position, and thus make it less explicit in written texts. Authors have many ways to avoid an explicit self-presentation in a text. For example, they can use generic or unipersonal verbs (eg. modal verbs), or passive forms and nominals instead of active verbs. Obviously, the most frequently used means of effacing persons was the use of the passive. In my data, one of the writers even stated her own personal commitment to and interest in the topic with a passive expression: "In this study, one was interested in X." This could be called the ultimate implicitness! The most remarkable difference between spoken and written texts was the amount of passive and other effacing expressions. In spoken texts, only 9% of the expressions indicating author presence effaced the author, but in written texts the amount was 46%. This means that the speakers referred explicitly to 85 themselves in nine cases out of ten, whereas writers did that only every second time. Thus, the written texts were clearly more impersonal than spoken ones. The texts, both written and spoken, were not very interactive, as only 20% of all interaction indicators were osed to consider the audience. When the authors took their audience into consideration, they did it by using questions, either rhetorical or genuine questions, and by addressing their audience. In written texts, the interaction between the writer and the readers was on a very implicit leve!. For example, the audience was taken into account only with a few rhetorical questions. The contact to the audience was clearly more explicit and direct in spoken texts: speakers addressed their hearers, asked them to do something, and commented on their possible reactions. For example "of course you all know/think that...", "x is very important for you", "in this picture you can see x" were typical utterances in the spoken data. If the authors clearly wanted to avoid directly addressing their listeners, they tended to favour expressions that efface the audience, such as generic expressions or passive forms instead using the pronoun you ("in this picture one can see ..."). However, an author can also refer to us, that is to himself/herself and the audience and by doing so perhaps create solidarity. On the other hand, it is possible, that this kind of we-talk makes the audience feel uncomfortable, as it might remind them of a mother-child or a teacher-pupil discourse. The pronoun we can refer to either the author and the present audience or to the author and some other persons ('me and my colleagues'). This Iatter use does not refer to the interactional relationship between the author and the audience, and it is primarily osed to efface the author. Therefore, I have taken into account only those cases where we refers to the author and the audience, or to people in general. Again, spoken texts were found more interactional as speakers referred to us more often than writers. In this, they referred to assumed mutual knowledge or shared experiences. These assumptions, however, were not necessarily correct. (10) As there has been no research, we do not know how the Finnish morphology is learned and therefore we are not able to ... Authors can indicate their solidarity also by involving the audience in the actual presentation process. This is shown in the next case, in which the speaker does not use the first person singular, but plural we instead. He does not say that be is not going to talk about vocabulary because it would lead him to trouble. In~tead he says: "Now if we would start to talk about vocabulary, it would lead us to difficulties ..." Or, instead of saying that he is going to show a transparency, he involves the audience in this act as well and says: "If we put these results obtained by Toivanen alongside mine ..." 86

%

18:leffaced yau ~yau lmwe ~effaced I 1111

written texts spaken texts

Figure 4. Author and audience

%

100 rr.=~A"?--x-x-;,c-r-r .. ,'5?'__,_ ~-.x."'5?__ ,_r_ 5("_ ><, :-i. 3 ..77

18:leffaced ~ highlighted

0 L.>C.'-LL.LLLLL<'-LL..

Figure 5. Visibility of persons

To sum up, the orientation of interaction was quite different in spoken and written texts. If we look at the total amount of the relationship indicators, spoken and written texts look very similar (figure 4). About 14% of all utter­ ances contained some interactional indicator. However, there are remarkable differences to be found when we look at the distribution of different types of indicators between spoken and written texts. Speakers refered to themselves more directly, addressed their audience more freely, and showed solidarity more than writers did. In comparison, writers were more implicit, tended to efface themselves, and addressed their readers far less than speakers did. 87 The difference between written and spoken texts is quite ciear (figure 5). 92% of the interactional utterances found in spoken texts were used to highlight the persons involved, while only 8% were used to efface them. But in written text, 66% of these utterances highlighted and as much as 34% effaced the persons involved in the interaction.

Conclusions One of the aims of the present study was to develop tools for the analysis of metadiscourse, as it appears in authentic material. In addition, I also wanted to find out the descriptive force of the system with respect to academic discourse. To sum up, my findings suggest that spoken and written conference papers differ from each other in various ways. For example, the textual structure was more implicit in written papers. Speakers seemed to realize that their Iisteners need cues about the structure of discourse to be able to follow the talk. In written texts the need for guidance does not seem to be as important. In addition, written texts were clearly more fact-oriented whereas speakers were also concerned with their audience. Speakers were more personal: they used evaiuation and opinion markers more frequently and more freely. Writers, on the other hand, were more concerned about the degree of certainty and used (un)certainty markers. Speaking and writing are different as processes and this is also evident in different outcomes. For example, statements were vaguer in spoken texts, speakers favoured hedges and double-hedges, even though the talks were to some extent at least preplanned. Nevertheless, the most remarkable differences were found in the interpersonal leve! of discourse. Written texts were clearly more impersonal and monological than spoken ones. When reading their paper, speakers referred directly to themselves, addressed their audience, asked questions, and created solidarity by referring to us more often. Academic texts are generally characterised as impersonal, informative and abstract. The choice of style is, however, governed by the norms and customs of scientific writing. Thus, when academic styie is under study, it is important to pay attention to the distinctive characteristics of the particular scientific community which the texts are reflecting. The norms held by the academic community can be used to explain, for example, why an author resorts toan impersonal style of writing. Only objective truths are regarded as scientific, and the author's explicit self-presentation in the text could be interpreted as a face threatening act. The author's resorting to impersonal expressions might thus even be regarded as a negative politeness strategy. It seems that scientists observe, for instance, the following implicit "rules": do not push yourself to the foreground, refer to your own research only in subordinate clauses, do not say I too often. Regard your audience as experts, and refer to common knowledge. Emphasise your solidarity with your listeners. 88 Do not present other people's ideas as your own, but tel1 openly where they have come from. If your criticise other researchers, do it in a polite and implicit manner. Do not present your findings as truths, if there are any doubts. It is always safer to be a Iittle uncertain. The concept of metadiscourse, in spite of its fuzzy nature, seems to be useful in the analysis of texts. The functional description developed for this study revealed interesting features of academic texts that were worth studying, and clear differences between spoken and written forms of texts were found. Consequently, the system could be of further use in the analysis of other types of text as well. Thus, future areas of application could be found in, for example, the analysis of metadiscourse in popularisation, in textbooks, and newspapers.

References Crismore, Avon, 1984: The rhetoric of textbooks: metadiscourse. Journal oj Curriculum Studies, 16. Crismore, Avon, 1989: Talking with Readers: Metadiscourse as Rhetorical Act. New York, Peter Lang. Halliday, M.A.K. 1973: Explorations in the Functions oj Language. New York, Elsevier North-Holland. Lautamatti, Liisa, 1978a: Observations on the development of the topic in simplified discourse. In Viljo Kohonen & Nils-Erik Enkvist (eds.), Text Linguistics, Cognitive Learning and Language Teaching. Suomen ovelletun kielitieteen yhdistyksen (AFinLA) julkaisuja Nr 22. Turku. Lautamatti, Liisa, 1978b: Some observations on cohesion and coherence in simplified texts. In Jan-Ola Östman (ed.), Semantics and Cohesion. (Med­ delanden från stiftelsens för Åbo Akademi Forskningsinstitut, Nr 41, Reports on Text Linguistics.) Åbo. Luukka, Minna-Riitta, 1992: Akateemista Metadiskurssia. Tieteel/isten Tekstien Tekstuaalisia, Interpersonaalisia ja Kontekstuaalisia Piirteitä. (Reports from the Language Centre for Finnish Universities, N:o 46.) University of Jyväskylä. Markkanen, Raija, Steffesen, Margaret S. & Crismore, Avon 1993: Quantitative contrastive study of metadiscourse: Problems in design and analysis of data. In Jacek Fisiak (ed.), Papers and Studies in Contrastive Linguistics, 28:137- 152. Adam Mickiewicz University, Poznan. Vande Kopple, William, 1985: Some exploratory discourse on metadiscourse. Coilege Composition and Communication, 36:82-93. Vande Kopple, William, 1988: Metadiscourse and the recall of modality markers. Visible Language, 22:232-271. Williams, Joseph M. 1981: Style: Ten Lessons in Clarity and Grace. Glenview, IL, Scott, Foresman. Explicitness in Marketing and Linguistic Settings

Kjersti Fl~ttum

The communicative value of explicit structure in text is obvious. Signalling of the structure by means of various types of marker facilitates considerably the reading and understanding process. However, it is well known that all texts are not structured in an explicit manner. The degree of explicit structure may depend on several factors, like the setting in which the text is produced, the author's overall intention (or global speech act) in writing the text and the text form (or presentation form) which is chosen for the particuladext. In this paper these factors will be considered in relation to the explicit structure of some selected texts from marketing and linguistic settings. Text structure can be divided at least into two types: Iinear structure (the progress from one text sequence to another) and hierarchical structure (the organizing of text sequences of different rank in relation to what is important and less- or non-important information). Hierarchization of information is a domain that clearly deserves more attention, in a theoretical perspective (linguistic, text linguistic and text thematic) as well as in a more practical perspective (text production and comprehension, manual and machine-based abstracting, summarizing and indexing; see Andersen 1990, Le Roux 1990 and Fl~ttum 1990, 1991, 1992a, 1992b, 1992c, 1992d). The general aim of my work in this domain is to map potential explicit markers of textual structure, linear as well as hierarchical, and to describe them linguistically. A specific aim, related to the hierarchical structure, is to contribute to the clarification of what can be considered important or non­ important information in a text. The presence of markers and their linguistic properties will be of capita! importance in this matter. For example, there is reason to believe that markers which signal linear structure, for example, like premierement 'firstly', deuxiemement 'secondly' d'abord 'first', ensuite 'next', en.fin'finally', can also signal hierarchical structure in certain contexts. The aim of this paper is to describe the use of markers in general, and the use of markers signailing hierarchical structure in particular. In the analysis of a few texts, some characteristics of texts written in the two very different settings of marketing and linguistics (see below) will be exposed. The analyses will be limited to the following two points: 1) explicit marking of structure, in particular hierarchical structure, and 2) in the case of explicitness, types of marker used. Finally, some reflections on further work in this domain will be presented. 90 1 Text material The text material consists of four examples of French written texts or text sets, two from each of the settings of marketing and linguistics. There is no well­ founded reason for the choice of such heterogenous settings. But in studying texts from these settings, for teaching purposes as well as for research purposes, the following intuitive ideas have emerged: the structure of texts written in a markyting setting, especially advertisements, is often not explicit, while the structure of texts written in a linguistic setting is generally explicit. These intuitive ideas will be examined in relation to the following texts (except for text 3, the texts are presented in the appendix):

Text 1: "Portugal" (Text set) This text consists of four independent parts picked out from a !arge brochure on Portugal attached to the French magazine L'Express, no. 2098, of September 19th 1991. These parts are advertisements published by the national tourist office of Portugal (text la), by an airport association (text lb), by a hotet chain (text le) and by a bank (text ld).

Text 2: "Transport" Transport is an advertisement text published in the magazine Le Nouvel Economiste by a transport company, (also presented in the text book Frenchfor Business (Le Goff 1986)).

Text 3: "Adverbes" (Text set not included in the appendix) The "Adverbes"-text corresponds to different articles in Langue jranr;aise, "Classification des adverbes", number 88, December 1990, edited by Henning Nl!'Slke.These academic texts will be referred toas text 3 (with exact reference to the page numbers when cited).

Text 4: "Ortographe" This polemical text isa contribution to the discussion on orthographic reforms in French, written by Robert Gueho and published in the journal Le jranr;ais dans le monde, 299, 1991. Before undertaking the analyses of the texts, some of their more general characteristics will be considered. Some of the central concepts used in the analyses will also be defined here. 91 The overall intention (global speech act) of texts 1, 2 and 4 is that of conviction or persuasion: to make the reader believe in (and eventually act in accordance with) the content presented in the text. (The difference between conviction and persuasion will not be discussed here.) The general intention of text 3 is that of assertion: to present fäets as true and thereby increase the knowledge of the reader. (See Adam 1985 fora text classi­ fication based on global intention.) These characteristics correspond to those commonly attributed to the text genres to be analysed here: persuasion is the normal intention of advertisements (texts 1 and 2) and of polemical texts (text 4); assertion is at least one of the normal intentions of academic texts (text 3). It is commonly accepted that the intention can be realized in different text forms. (See Lundquist 1990: 18 for the concept of "formes de representation" and Nielsen & N~lke 1991 for the relation intention-form). The text forms represented here are the following: the descriptive form, text 1 (characterized by the presence of stative verbs in durative aspectual forms and the enumeration of properties related to the persons or objects introduced), the argumentative form, texts 2 and 4, ( characterized by the presence of argumentative expressions like connectors of the type but, consequently, which make explicit the relation between the fäets and ideas presented), and the expository-explicative form, text 3 (characterized by an "objective" or "neutral" and exact presentation of fäets possibly with an explanation of (the relation between) these fäets).

2 Explicit marking of structure

2.1 Texts from marketing settings Texts la-ld, advertisements, are all characterized by a structure of enumeration. They list and describe the assets of Portugal as a holiday country, the assets of its airports, of a hotel chain and of a bank company, respectively. In general, these are enumerations without any structural marking, except for the conjunction et 'and' used in some cases. The enumerations contain a series of specific terms as in text la: Le Portugal. Un fäscinant survol de huit siecles d'histoire de 1'art. Orfevrerie et verrerie. Ceramique et broderie ... (Text la.) [Portugal. A fäscinating flight through eight centuries of the history of art. Silversmith's art and glassware. Ceramics and embroidery ...] 92 The specifications are sometimes introduced by typical expressions of descriptive writing like c'est 'it is' and nous sommes 'we are': Pour certains, c'est l'arrivee et I'emotion des retrouvailles, pour d'autres, le frisson du depart ... (Text lb.) [For some, it is the arrival and the emotion of reunion, for others, the thrill of departure ...] The specifications are often summarized by general expressions of the type tout 'all, everything' and rien 'nothing': Dans les aeroports ANA EP, tout est pense pour le confort ... Rien n'est laisse au hasard, ... (Text lb.) [At the airports, everything has been thought of for your comfort ... Nothing is left to chance, ...] Thus, there is semantic-lexical structuring in generic and specific terms, but there is no explicit interna! structuring in the enumeration of the specific terms (like the French equivalents of first, second, and so on). Although the non-marked enumeration is a general characteristic of these texts, there are some exceptions. In text ld, for example, there is an explicit hierarchization signalled by the following expressions: nous sommes le no 1 dans: ..., et nous sommes de plus en plus actifs dans: ... En outre, notre Succursale ... (text ld.) [we are number 1 in: ... , and we are more and more active in: ... More­ over, our branch ...] Text 2 is different from the ones characterized above. There are also distinct descriptive parts, like A Roissy, nous sommes le seul transporteur de fret aerien ... (text 2.) [At Roissy, we are the only air freight carrier ...] but on the whole, this text is clearly more argumentative in its form. The argumentative form appears in expressions like for example Tout cela conduit a... 'All that leads to ... ' and Tout ceci signifie que ... 'All this means that... '. These expressions also signal what is the most important information in the text: We are the best in transport. In addition there are pragmatic connectors of different kinds. Although the explicitness of structure is clearer in text 2 than in texts la-ld (at least la-le), the two sets of texts have several common features. As in texts la-le, there are in text 2 enumerations of the assets of the transport firm, without an interna! hierarchization between the different assets enumerated. Both in texts la-le and 2, all the assets are presented as equally good, or 93 important, by means of general expressions like tout. In text 2, for example, this is marked by Tout ceci signifie que ... Jn conclusion, the examined sets of texts (texts 1 and 2) are produced in the same setting: marketing, or, to be more precise, advertisement. They express the same intention: both intend to persuade the reader that all the products or services presented are the best. When it comes to text form, there are notable differences. Texts la-ld are dominated by descriptive characteristics, while text 2 is much more argumentative. As for the explicitness of textual structuring there are also some differences to notice. In text 1 there is little explicitness, in general. In text 2 there are several hierarchical markers, but these markers do not rank the different enumerated assets. One could ask why there is no such ranking in the advertisement texts. An answer based on intuition could be that the intention is to present everything (and as much as possible) as equally important.

2.2 Texts from linguistic settings In text 3 (different articles in linguistics), the information structure is very clear. (It must be stressed that these are clearly academic texts.) First there is the system of subtitles and the numbering of different parts which both mark the linear structure explicitly. Examples of subtitles are Jntroduction (p. 12), Conclusion (p. 26), La classification traditionnelle (p. 60). Second, there isa strong dominance of markers within the introduction and conclusion parts, also important for the structure of the text as a whole. Example from an introduction: Notre etude se bornera a ..., Elle prendra en compte ... , Nous effec­ tuerons ensuite ... (p. 28.) [Our study will be limited to ... , It will take into consideration ... , Then we will carry out ...] Example from a conclusion: Resumons, brievement, les resultats auxquels nous sommes arrives ... (p. 78.) [Let us summarize, briefly, the results we have obtained ...] This sentence is followed by a listing of three numbered points. This type of explicitness is characteristic of relatively long texts of the expository-explicative form, especially the academic ones. It is the linear structure of the text which is signalled by the explicit markers. There are few examples of explicit markers of hierarchical structure, at least on the super­ structural leve!. In the conclusion of some articles, however, one can find a great number of pragmatic connectors, which signal hierarchical structure. (For a well-founded genre-approach to texts in academic and research settings, see Swales 1990.) 94 In text 4, a polemical contribution to the discussion of orthographic reform, the structure is also quite explicit. First, the linear structure is clear from the division into paragraphs; second, there are markers signalling the linear organization of the text information: C'est le premier reproche ... Et le second reproche: ... (text 4.) lThis is the first reproach ... And the second reproach: ...] Another example is the author's enumeration of les principesfondamentaux 'the fundamental principles' behind the orthographic reform that is discussed. He says that he wants to formulate them, and in the formulation he numbers them explicitly: 1, 2, 3 and 4. This text is more argumentative in its form than the articles in text 3. The argumentative form manifests itself by the relatively clear hierarchical structure in the text, signalled by several pragmatic connectors. Thus we can again point at the correspondence between argumentative form and explicit marking of hierarchical structure. In conclusion, texts 3 and 4 are produced in the same setting: linguistics. They differ when it comes to intention: in text 3 (linguistic articles) the intention is to assert or explain linguistic matters while in text 4 the most important intention is to persuade the reader of a particular view (although this text also has elements of exposition in it). When it comes to text form, there are also clear differences. The articles in text 3 are dominated by the expository-explicative form, while text 4 is more argumentative. As for the explicitness of textual structure in general, the texts taken from the linguistic setting are fairly similar. All the texts are structured quite explicitly. However, when it comes to markers of hierarchical structure, there is a difference between 3 and 4: text 4, in a clear­ ly argumentative form, has a more explicit hierarchical structure than text 3.

3 Types of marker used The markers used to make explicit the information structure in the analyzed texts are of various types. Within the scope of this paper, only some examples can be mentioned.

3.1 Texts from marketing settings As already mentioned, in texts la-ld, of the descriptive form, there are, on the whole, very few markers. The conjunction et 'and' can be mentioned as an example. In text 2, of the argumentative form, there are several markers, especially of hierarchical structure. There are pragmatic connectors like en fait 'in fäet', car 'for', aussi 'also' (cf. the Geneva-school 's studies of such connectors in a 95 hierarchical perspective in the journal Cahiers de linguistique franr;aise and in Roulet et al. 1987). There are also non-conventional expressions with lexical components indicating hierarchy, like the following (from text 2): Ces quelques reflexions n'ont pas pour but de ... [fhe aim of these reflections is not to ...] Nous voulons simplement mettre en evidence ... [We just want to make clear ...] Tout cela conduit a... , [All that leads to ...] Tout ceci signifie que ... [All this means that...] The two last expressions can easily be paraphrased by the concluding donc 'so', signalling that the information in question is of concluding, and superordinate, value. So, the presence of hierarchical markers is obvious, but the lack of markers ranking the different assets in advertisement enumerations should also be noticed.

3.2 Texts from linguistic settings In text 3, consisting of academic articles of the expository-explicative form, there are numerous markers, especially of the type signalling Iinear structure. The markers are either conventional, like d'une part - d'autre part (p. 90) (on the one hand - on the other hand) and other enumeration markers, or they are non-conventional, often in the form of whole sentences, as in the following: Cet article concerne ... Nous commencerons par ... , puis nous expose­ rons ... , avant de conclure par ... (p. 51.) [fhis article deals with ... We shall start with ... , then we shall explain/put forward ... , before concluding by ...] There are of course several markers of metatextual deixis also, contributing to the explicit structure of the text (fora presentation of such markers, see Evensen 1990). In some cases one finds hierarchical markers. When present, the markers are often of the pragmatic connector type, accumulated in the conclusion, as in the following example taken from the concluding last paragraph of an article (my italics): 96 Il peut alars sembler que ... Je pense neanmoins... En effet... On peut donc esperer que ... Or, seules des recherches ulterieures dans cette voie nous permettront d'enjuger ... (p.29.) [So it might seem that... I never­ theless think ... lnfact ... Consequentlywe can hope that... But only further research in this direction will allow us to judge what just has been presented ...] In this passage, there is in fäet one marker in each sentence. In text 4, of the argumentative form, there are, as mentioned above, both linear and hierarchical markers, conventional as well as non-conventional. A typical example of markers of linear structure is the following: le premier reproche and le second reproche 'the first/second reproach'. These expressions, which can be called denominative enumerators, consist of an ordinal number anda noun describing the enumerated segments. In certain contexts, they may signal not only linear structure, but also hierarchical structure (see Fll')ttum 1992b). In text 4 there are also quite a number of markers of hierarchical structure, conventional markers like pragmatic connectors (mais 'but', en effet 'in fäet', certes 'certainly') and non-conventional markers like Concluons ces developpements ... [Let us conclude this presentation ...], je[ ...] souhaite vivement, mais ... [I keenly desire [... ], but...] and Ce que j 'en retirerais personnellement? [What I would gain from it personally?] (text 4).

4 Discussion After these analyses of the degree of explicit structure in texts 1-4, some comments should be made on the intuitive ideas about marketing and linguistic settings presented in the introduction: non-explicit marking of structure in marketing setting and explicit marking of structure in linguistic setting. These ideas are confirmed only to a certain extent. The analyses have shown that the setting in itself is not a sufficient criterion for distinguishing between non-explicit and explicit structure. Although the texts produced in linguistic setlings are considerably more explicit as to the marking of structure, we have also seen examples of explicitness in texts produced in marketing settings. Of course there are also the fäctor of academic/non-academic texts and the length of the texts to take into consideration. For example, it is not very likely that a short text contains a great number of structure markers. When it comes to the intention (or the global speech act), this does not seem to be a satisfäctory criterion either. For example, the persuasive intention is 97 realized in texts with explicit structure marking (texts 2 and 4) as well as in texts with non explicit structure marking (texts la-le). It is the criterion of text form, which, problematic as it may be, seems to be the most powetful in this context. However, it must be stressed that the material is very limited and does not allow any kind of generalization. The texts produced in the expository or the argumentative form are both characterized by explicit structure marking. But when it comes to markers of hierarchical structure, these seem to be most frequent in the argumentative form, in entire texts of this form or in sequences of texts where this form dominates. In conclusion, I shall present some reflections as to further work in the domain of explicit marking of structure, and in particular the hierarchical textual structure. The argumentative text form seems to be important for the further study of markers of hierarchy, and the expository (and toa certain extent the descriptive) text form might be important for the study of markers of Iinear structure, and their capacity of marking hierarchical structure. Some of the questions which should be studied in morc detail are the following: Is there a hierarchical relation between the segments announced Iinearly as in the introduction of a linguistic article cited above: Nous commencerons par ... , puis nous exposerons ..., avant de conclure par ... (text 3, p. 51)? And is there a ranking in the enumeration of different segments made explicit by ordinals like first, second and third? In fäet, this type of explicit structure is found not only in the introductions, but also in other part&-ofthe texts: ... je me contenterai de trois remarques. La premiere est [... ] ce sera ma deuxieme remarque [ ...] Ma troisieme et derniere remarque ... (text 3, p.10.) [.. .Ishall confine myself to three remarks. The first one is[ ...] that will be my second remark [... ] My third and final remark ...] If there is no ranking, why do text authors, in some contexts, say explicitly that the order is of no importance? Finally, I want to stress polyphony as an interesting domain which deserves to be studied in relation to structural markers. To summarize very briefly, the fundamental idea of the polyphonical theory, as developed by Ducrot and, in particular, N~lke (Ducrot 1984, N~lke 1989 and 1990), is that one and the same utterance may consist of a number of "discourses" which express different points of view in relation to the subject. In other words, the speaker may present several points of view, or "voices". Thus, it is vital to the interpretation of the utterance and of the entire text that one can determine the relations between the different points of view and the author. The three types of relations mentioned by N~lke (1989) are: adherence, acceptance and dissociation. 98 In order to relate the polyphonic aspect to the question of hierarchization, I have proposed the following hypothesis: The point of view which the author adheres to is the strongest or the most important and the one the author dissociates himself/herself from is the weakest or the least important. The relation of acceptance occupies amid-position (when the author accepts a point of view without taking any responsibility for it) (Fl~ttum 1992c). One example of a polyphonic marker is paraft-il 'it seems/apparently', taken from text 4. The expression paraft-il can indicate that the author dissociates himself from or at least does not take any responsibility for the point of view related to this paraft-il. In addition to the syntactic negation, the canonical marker of polyphony, other examples are pragmatic connectors like certes, mais, etc. Polyphonic analyses seem to be fruitful in the interpretation of structural markers and of their capacity to signal hierarchical structure and thus information value.

References Adam, Jean-Michel, 1985: Quels types de textes? Le Franc;ais dans le Monde, 192:39-43. Andersen, 0ivin, 1990: Lingvistikk og dokumentasjonsbaserte Tesaurus- konstruksjoner. Dr.art.-thesis, Bergen, University of Bergen. Cahiers de linguistiquejranc;aise, 1980-1990, Geneve, no 1-no l l. Ducrot, Oswald, 1984: Le Dire et le Dit. Paris, Minuit. Evensen, Lars Sigfred, 1990: Pointers to superstructure in student writing. In U. Connor & A.M. Johns (eds.), Coherence in Writing. Alexandria, TESOL, pp 169-183. FI~ttum, Kjersti, 1990: La Nature du Resume scolaire, Analyse formelle et informative. Oslo/Paris, Solum Forlag/Didier Erudition. FI~ttum, Kjersti, 1991: Pointers to value of text information. Paper presented at Nordtext Workshop, Copenhagen 1991. Fl~ttum, Kjersti, 1992a: La hierarchisation d'information comme activite resumante. In M. Charolles & A. Petitjean (ed.), Le Resume de Texte. Paris, Klincksieck, pp 91-102. Fl~ttum, Kjersti, 1992b: Marqueurs de structure textuelle. Paper presented at the International Congress of Linguists, Quebec 1992. Fl~ttum, Kjersti, 1992c: Polyphonic structure. In A.-C. Lindeberg, N.E. Enkvist, K. Wikberg (eds), Noråic Research un Text and Discourse. Åbo, Åbo Academy Press, pp 161-172. FI~ttum, Kjersti, 1992d: Suppression d'information dans l'activite resumante. In Actas do XIX Congreso Internacional de Lingiiistica e Filoloxia Romanicas. Coruna, Pedro Barrie de la Maza, pp 405-418. L'Express, 1991, no 2098. 99 Le Fran9ais dans le Monde, 1991, no 299. Le Goff, Claude, 1986: Frenchfor Business, Le Fran9ais des Affaires. Paris, Hatier International. Le Roux, Dominique, 1990: Automatisation de l'Activite resumante: Essai de Typologie. Paris, EDF. Lundquist, Lita, 1990: L'Analyse textuelle. Methode, Exercices, 2nd edition. Copenhagen, Handelsh!,'ljskolensForlag, Nyt Nordisk Forlag A.Busck. Nielsen, Anne Ellerup & N!,'llke, Henning, 1991: Persuasion disguised as description. In I. Zagor (ed.), Speech Acts: Fiction or Rea/ity? Ljubljana, pp 89-105. N!lllke, Henning, 1989: Polyfoni. En sprogteoretisk indfr;ring. ARK 48. Copen­ hagen, Copenhagen Business School. N!lllke, Henning, 1990: Formes et emplois des enonces negatifs: polyphonie et syntaxe de "ne ...pas". Revue Romane, 25, pp 223-239. N!lllke, Henning (ed.), 1990: Langue jran9aise, Classification des adverbes. Paris, Larousse, no 88. Roulet, Eddy et al., 1987: L'Articulation du Discours en Fran9ais contempo­ rain. Berne, Peter Lang. Swales, John, 1990: Genre Analysis. English in Academic and Research Settings. Cambridge, Cambridge University Press. APPENDIX TEXT 1: PORTUGAL lA Et si vous deveniez poete apres vous etre plongedans tant de culture.

'L,'.nrluyaf~f,P"P,,(,., 6ruisJan/e encorrt r/4 r;,r,, .C,., uilfap.s t>OW p

homme.s fUi .Jon/ a./Ns d /4 d'icouu,,,r/e cfha.rmonteJ 6/anche.s, 6/4ue, el ocNJ mimo1N. ~ encor"ft, ce, calhilra/4,,

clumonch.Qc~/d ,ue (.K)UJ iret, t.ew ruif,nl ~/J.Ier å r/<1u/re1 r:on/inenl,. ce.J chilre-1, cel/e umver.11/l, limo,n.s

atJ.J.J1;de dicouwr/e.J e/1 dlcouwrlu. E, 7briuyaf 'lin /å,cinan/,u,...,/ duf,nd d

coeur du 'lirluyaf Ge coeur fUi 6a/ OrJ?urerie el ~rrer~. CJram/yue el ,ouche ancrenne el tefkmenl uit,,an/e,

d,vu la podrme de charve '7brlugai.J. 6rod,rie. 7lrilau:x marlefi, el 6oisdorh. iitfinimenl 9iiniN'use de JeJ 111nJ

Ea •&ud:,J,•, 'lin 6rm J. nc.,lafyi<. Ee 7brluyal CJel/efu:.r.urianc, el dr JeJ parl0,1. f:J..rut mi/amorpho.u Qln JOujJkde milancohe. archtlec/ura/4. Ei9ne1 J/w.nuil,ne ou i., pala.iJ en ou6er7ei. Jmmuab/4

7/k,,-, '{"andpar/-,.,x,u,f '2uand 7bm6aftne. C/lne mal/r,,e tfUI va cfu 7-brlu5al. Uu, y e/e,. Co y est, 1/fJUJ

INl•wu.s d la renconlre J'e,p/wancf.em priromon ou roroco, cfu ni<>--d"Q.J.J,cjJ,n-e i/e, poil,. confin.s de fCurope el defeun lrbonf d far/ nouveou. C! ces chdl,ou:z •I Gu Ll//k.Jenl,fre1 proliyie.s par la ctlade/k, de, ':.kmp/,usel cfu Jl(oure, ~,..,,...... •.,..,':::::211 G,,n,nis..uon JK'oml,Je ek, monumenh /a,//,;, cl';nr6ranlob/4 7ra.n,I. t-"'--'• '-""'111~'-All. AiJloriyue.J. Gt//e prem.riM c,;& du CJ.s ~OJalfUV azurtt..J fUi -~ OJ7,c,rYr1/,ona/ ✓u ':kur1Jxwdu 'RrluJ«f monde d wcahon mondia.l,, Joule fu,mn,nl pdais el 1117n•urw.1 el ,-ui :::l..nw ltJu$tt /29 o, /OJO kwrfk1.

1B

Ce sont des mllllers de passagers que n6S aeroports accuelllent chaque jour. Pour certalns., c'est l'arrlvee et !'emotion des retrouvall/es, pou; d'autres le frlssondu depart. Dans les aeroports ANA EP,tout est pense pour le confort et le blen-etre de ces passagers: des termlnaux modernes offrant restaurants, bars, boutiques duty- free, banques et servicesde locatlon de vortures. Rlen n'est lalsse au hasard, et chaque jour nous leur a nzl prouvons et leur dlsons:

'Nous sommes heureux de vous accuellllr" a,zroporf-cn,r nau1r9a~So a(zr1ra,tz.p. · 1 lC

UN WEEK-ENDAU PORTUGAL, UNE DELICIEUSEGORGEE DE PORTUGAL.

Si vous ftes dans un des hötels Meridien du Portugal p0ur affaires... restez avec nous pour le week-end. Nous avans tout prevu.. vos golfs serent juste comme vous les aimez, les

musees vous.devoileront leun tresors 1 les restaurants typiques et les chan• teun de Fado seront au rendez-~ous, .C'esl vrai. un..wee.k.,.end.dansun hötel Meridien å Porto ou å Lisbonne, c'est plus qu'un week-end. - Informations et reservations a Paris au: (1) 40.68.34.20 el dans les agences Air France et agences de voyage, Plus de 50 HOTEl.S MERIDIEN dans le monde, Compagnons de Voyage d'AIR FRANCE. ::&; MERJDIEN PORTUGAL

1D LEMONDE EST NOTRE OBJECTIF A Il CAlXAGEiAL DE DOOSITOS,nous dooooru Il pri!!mxt • llOO cli!!prtf!roic<. Ainli, au Portugal,now SOOllilCilen!ldani: - • la opcntioos flJWldl,cs et mucht de capiuux -k finlncancnt de projeu t lDO)'ell mdelitres -le martht des thange,

A CAIXA GER.All.. ~ DE DEPOSBTOS

DUPOR'JUGAL POUR LE MONDE usso,. RIO DE.,W,BH) SÅOPAWl PMIS MAOrnAOffSHOI"' CAIXAGEAAL DE OEPÖSITOS 8AHCOf1'.MRO 1M ctPO'OOSSÅOHJ.A.D !SPI OEPOSITOS "'-'a .ln!On(I Jo.5,t 6t A.m!sia.19 A' o>ORT\JG-'l BS.~1.1.-aa, .,.,,,_ •~.,, 135111113"7 -'S 81 '"""~lc'.121•2SJll121 """"~1111358101 7!:.016P.va5-fll.A.t-.0: •.U,O(JAA,PQf!TUGAl. •~,,u 11621 Af(j(NTP ~,;;1~;AOPSSl telax:25069.lGfPM fNf~ (1)4,()69~00 ;...L 1)5111911)'()20 ~...,tu· 1,351!!1J J.c.6 7(33 ~•• l',ZlJHJJ ltcfu::(lHJ63176 ~6:)jl72CAJ)'..A.! ':t1«7<621CASSf{P r~u::!ll&Oi00309 ~t"!'IJt ,)51! 1911)J:)92 TEXT 2: TRANSPORT

Lemonde est plein d'hommes d'affaires qui s 'imaginent que tout est reso/u lorsqu'ils ont conclu la vente,el fait signer le bon de commande. En fait, c'est .souventfå que commencent les vraispro­ bt3mes.~ dans les affaires, l'important c'est le paie· ment,pas le bon de commande.El le paiementne vient qu'apres la livraison. Plus il faudrade tempspour livrer,plus long sera le delai de reg/ement.Avec toutesses incidencessur la tresore­ rie de /'entreprise. Ges guelgues reflexions n'ont fjfcs pour but de vaus apprendrevatre metier. Nous vauons s1mp/ementmet· tre en evidencele fait qu'il ex,ste,entre le depart el i am• vee de la commande,un laps de temps inevitablemais qui peut etre plus ou moins long, selon /'efficacite el Ja rapidite des services responsablesde /'expedition.De plus, tous les efforts auront ete faits en pure perte si la marchandisearrive endommagee. Toutcela conduit å vaus adresser å nous. Notre societe a ete creeeå Pa.risen 1959.Aussi. apres ving/ ans d' expe­ rience, nous savons nous y retrouverdans le labyrinthe de la paperasserie. A Roissy,nous sommes le seul transporteurde fret (•) aerien å posseder nos propres bätiments. Plus de 3 000 m1, /'installationla plus moderneen Europe.Nous avons des Agences dans les 26 premieres vi/les de France et dans le monde entier.Nous avansune equipe completed' expertstraducteurs interpretes qui parten/ tou· tes les /angues,ou presque... Et, bien sur. le fran9ais ! C'est pour cela que nouspouvons resoudre en quelques instants vas problemes internationaux. Nouspossedons un dessystemes de suividu fretsur ordi· nateurle plus sophisliquequi soit. Il nouspermet de /oca· fiser dans le monde enlier les expeditionsen l'espace de quelquesminutes. Aussi, s'il se produit un encombre­ ment du traficlntemational, nous powons derouter/'expe­ dition pour eviter cet inconvenient. Si cela se reve/e impossible,Ja clientete sera avertie. Toutceci signifie que notre entreprisepeut apporteraux hommesd'affaires franyais, ou qu'ils soient, un service mondiai de fret serien de porte å porte, et cela avec un degre de qualite inimitable. Textepublicitaire paru dans : "Le Nouvel Economiste" TEXT 4: ORTHOGRAPHE ORTHOGRAPHE: OUVRONS LE DEBAT-

Druin dr Pltillrm parn dan:.rLire. nov. /990

qui rcdouble une autre marquc. audiblc cellc~IA(les. des. etc.) ct qui scrait suffisante en principc. Unc bonnc reronnene doit pas viser ?I LE POINT DE VUE D'UN ENSEIGNANT en maintcnant une l1iquette, une fice//e: ttant donn6 d'åutre part la accroi1rc la rcdondance. mai~ plutöt A la ~duirc. Or. on fcrit ct on .. Les enfonts et les ~,rangers scron1 contcnts:· a dit M. Bemard preponderance 6crnsante des finales -tlle, -elit, c'est cette graphie continucra ~ Ccrire : les prie-Dieu. les trompe-r ub~tres. dcs voycllcs pn­ la dcmandc de r Acad~mic fram;aise ! Or. la tt:fonne muhiplie les priorit6 de 1'invariabilit6. Et cenc priorite vaudrait 6galemcnt pour rasite~ ou dcs accents circonflexes (niveau a) : cllc s'attaquc 11Ja mais. les cependant. nous sen m~me toute une listc d'cxceptions A les accords r6gressifs du type c' elai(en)1 les ~nfan1s. pour les noms i.:onjugaison des vcrbes en -eler. -eterct au pluricl des noms compo­ l"accord du panicipe passt dcs verbes pronominaux ... Il s'agil lä. il et adjcctifs empruntCs. rapponCs. abrcges. etc. ,ts (nivcau b). mais en mati~re d"accord (nivcau c) clle nous octroie est vrai, de confirmer /'invariohilirl de ces panicipes. Ccla Conclu9ns ces dCvcloppcmdnts, trop longs pcut-C1re: loin d'Ctre op­ en 1out cl pour toulrinvariabilitl! de luissl + lnfinitif. m ·am~ne A mon quatri~mc principe : je dirais mieux pcutMern:ccne pos6 ll une r6formc de l"orthographc. je la souhaite vivcmcnt. mais c·cst le premier reprochc que je ferais 11·1a tt:forme : e//e s· occupe fois poslulat : il est lil!. plus que les trois autn:s. aux particularir~s plu!eoricntee ver.; le grammiltical el fondee surdes principcs ou pos­ tn•p ,Ju dictionnaire. pas ossez de la flrammaire. Et le second rc­ de I' orthographe fran~aise. tulats bien rentchis. applitju6s avec toutc la cohCrcncc possible ... ~e : 11ncherrhe en ,•ain les principes fondamentaux qui au­ 4. Il faut donner systlmatiquement la priori1i a la Kraphic courre Ce guc j"cn rctircrais pcrsdnnellement? Je n~aurais pas davantage rufrnt d,1 prhider d son llahoralinn. J"ai la t~~ril~ de vouloir les Tur la graphie longue, d /' invariahle sur le variah/e. Le but n •est pas d'6tudiants. non; mais je pourrnis mieux mc concentn:rsur ressen­ funnulc·r: de raccourcir les textes par souci d'6conomie. mais de se pencher ticl : leur apprendre 11s·cxprimer en rrn~ais avec dant. avec tlC­ I. Tnme nom•elle rjRlc doit accraftre la cohlrencc glohale du sys­ sur le phenom~ne de la redondance orthographique : le s du pluriel, gance mCme pcut-~trc ... Mc.-.coll~gucs de France ont-ils. a vrui Wme ,)T(hnxrophiqut: on ne peut pa., proposer f ltiquht.je ji,·n,. par exemple. qui n ·a gCn~ralement pas de corrcspondant A I "oral ct dirc. un tout autre objcctif? Robert Gueho 1~"111.ll l'RAN(,"AIS l>ANS l.l( MOND~ W ?lit UnivcrsitC de Tr~vcs

Making Economics More Accessible? A Study of the Process of Rewriting an Economics Text for a Wider Readership 1

Martin Hewings & Diane Houghton

The objective of this paper is to present a case study of the composing processes and products of a professional economist writing for two different audiences. The first text studied is an academic paper for presentation to an expert audience. The second is a version of the paper rewritten by the author for a highly-educated hut non-expert audience, and submitted as an entry for an international prize. Through analysis of the subject's texts and his methods of composing them, we aimed to identify his perceptions of the different audi­ ences, how these perceptions affect his texts, and how the perceptions are reflected in the texts. This study considers audience accommodation in professional writing-in this case economics-from the perspective of discourse analysis, and is a contribution to a recent hut growing body of literature in this area. Swales (1990) and Berkenkotter & Huckin (1993) have developed general criteria for identifying genres and the discourse communities who make use of them. In well-defined genres, like the academic paper in a specialist area, the discourse analyst can make same fairly reliable inferences from the text about the writer's intentions in regard to the audience. Studies by Fahnestock and Myers have examined popularisation in the sciences. Fahnestock (1986) suggests that the results of scientific observation undergo a change of genre in popularisations; information changes its status, as a result of change in the rhetorical function of the text, from a focus on fäet and cause to one of val ue and action. Myers ( I 990) links textual content and eon text in his investigation of distinctions between academic writing in the sciences and its more popular counterpart. He is concerned primarily with evidence in text of the social nature of scientific discourse, arguing that different narratives are provided when different audiences are perceived, and that these narratives convey different views of the work of the scientist. The narrative of science preserves the scientist's objectivity, and presents progress as cumulative, and

1 As can be seen in this paper, Professor Maxwell Fry has given us invaluable assistance in conducting this research. We wish gratefully to acknowledge our debt to him. 106 the narrative of nature seeks to present science essentially as an accumulation of fäets about the topic of the research. Interest in professional economics writing has developed comparatively recently, mainly under the inspiration of Donald McCloskey. McCloskey (1986) suggests to economists that their work is persuasive not because of conformity to scientific principles, but because of its rhetoric, studies of which he locates within Iiterary criticism. KJamer (1990) and Swales (1993) campare the rhetoric of eoonomics textbooks with the discursive practices of professional economists. Goodwin (1988) identifies three roles for the economist, those of "philosopher" (roughly corresponding to the role of academic economist), "priest" (the role of practitioner) and the "hired gun" (the role of the economist interpreting economic theory for lay audiences for persuasive purposes) in relation to a rhetorical analysis of text types. Other research has looked at the expectations of economists' performance in areas outside their main discipline, such as in political debate on economics issues (Galbraith 1988); or at how economists and their discursive practices appear to members of other discourse communities, such as students of politics (Keohane 1988) and journalists (Warsh 1988). Given the lack of an established theory of audience accommodation in writing about economics, our study is largely empirical and exploratory. The procedures of the case study were as follows: (1) The academic paper (Version A) and the competition entry paper (Version B) were compared, changes noted, and motivations for the changes established where possible. (See Section 2.) (2) A parallel study was made of two comparable papers by prizewinners. (Section 3.) (3) The approach to audience revealed by the work of the prizewinning writers was then compared with the approach adopted by our subject; and some general principles for the study of rewriting for different audiences suggested. (Section 4.) Our supposition in this research is that many of the changes made in rewriting the academic paper were in response to the author's perception of two distinct sets of target reader, having different purposes for reading the papers and different degrees of understanding of the subject. Findings suggest that there are two different approaches to audience accommodation when rewriting for a different audience. The first is rephrasing, where the revision process is text­ baseå, and the second is reconceptualisation, where the starting point is the perceived needs of the audience. 107 1 Background The subject of our study, Maxwell Fry, is an economist of international renown. His present position is that of Tokai Bank Professor of International Finance at the University of Birmingham, England. His particular research interests lie in the relationship between financial development and economic growth in developing countries. The inspiration for the present case study was a paper he wrote specifically as an entry for the prestigious Amex Review Prize. This prize is sponsored annually by the American Express Company in honour of Robert Marjolin, first head of the Organization for Economic Cooperation and Development (OECD) and vice-president of the European Commission. The winning articles appear annually in an edited volume of papers. The articles must focus on policy, but beyond this there is scope for a wide variety of economic topics to be pursued. Fry's topic for his entry was the relationship between foreign debt and the current account deficit in selected developing countries. Too much foreign debt would stimulate capita! flight, and should be avoided by governments. Fry's entry to the competition was hased on an earlier version of the paper, which he had prepared for an academic seminar at the OECD Development Centre in May 1991, although based on work conducted for the Intemational Monetary Fund (IMF). Discussions with Fry indicated that the original paper was targeted at specialised economists at the IMF, OECD and, more widely, academic eco­ nomists interested in the area of macroeconomics and foreign debt. For the AMEX entry he said that he assumed "an intelligent lay audience" such as journalists who write economics columns in newspapers, with a background in economics generally, but not specialists in a particular field.

2 Analysis of Fry's papers: Versions A and B compared In this section we present forma! differences between the original version of the paper written for the OECD (Version A) and the version submitted for the AMEX Review Prize (Version B), and go on to consider what motivated the author to make the changes observed. There were two stages in this part of the research. First, förmal differences between A and B were isolated. The papers were first compared in terms of their overall organisation, and then detailed differences were identified. A provisional categorization of these detailed changes was then made which sought to relate the changes to their apparent motivation, Second, interviews took place between the author of the papers and the researchers in order to gain further insights into the process of rewriting the paper and to try to establish the validity of the proposed categorization of the changes. 108 2.1 Differences in overall organisation A major influence on length was clearly the requirement of the Amex Review prize that papers submitted should be no more than 4,000 words. Certain cuts in A, then, were necessary to meet this condition. The main focus of the original version is the presentation of an econometric mode) of current account and monetary policy behaviour in developing countries. This model comes in the main _bodyof A. In B, however, while there are references to the model in the main text, the model itself, with certain technical equations and regressions, is placed in an appendix. This movement of text into an appendix and a reduction in the amount of detail provided goes some way towards both reducing word length in the main text and accommodating the perceived requirements of different readerships. This was clearly stated by the author during the interview: "You want models for academic stuff [... ] you want to get away from models for [... ] policy-oriented laymen." The process of rewriting was not, however, simply one of cutting Version A. A requirement of papers submitted for the Amex Review prize is that they should be "policy relevant". As Version A did not have this focus, the author added a short section on policy implications in his abstract and a more substantial section (19 lines) to the conclusion of B. Other, mainly new, sections were necessitated by the deletion or removal to the appendix of the more technical parts of the text. In particular, a new section was added referring to the current account equation reported in the appendix and what it indicates, anda new section at the beginning of the appendix summarizing the method of calculating the current account ratio. In general, however, the broad organisational changes in the paper are mainly concerned with deleting sections from A. There is also some reordering, particularly the removal of text into the appendix in B, and a few short additional sections in B. In terms of organisation, we can characterise B as a pruned version of A rather than a substantially reordered version.

2.2 Detailed changes Following the isolation of detailed forma! differences between the two papers, a provisional categorization of the changes was devised. This sought to relate these forma! differences to the apparent motivation for the changes, Clearly, an attempt to establish motivation could only be speculative at this stage. The validity of the categorization was tested by discussing the changes with the author. It was not feasible to talk about every change, so examples from each of the categories were identified, a total of about 40. On the basis of these discussions the categorization was modified slightly. The final version is pre­ sented in figure 1 below. 109

MOTIVATION FOR CHANGE CHANGE 1 Difference in 1.1 Purpose in 1.1.1 More explicit statement of perceived reading paper objectives readership 1.1.2 Rephrase to make more policy-oriented 1.1.3 Delete a section

1.2 Amountof 1.2.1 Rephrase to include less information or information delete a section necessary to include

1.3 Sophistication of 1.3.l Grammatical restructuring readers; ability to 1.3.2 More explicit understand signalling of concepts or relationship between elements of text explanations 1.3.3 Change terms used 1.3.4 Restatement to reduce "complexity" 1.3.5 Paraphrase quotation 1.3.6 Divide text different- ly (paragraphs)

2 Improvements 2.1 Style on version A

2.2 Contents

Figure 1. Categorization of changes- Versions A and B compared

As figure 1 suggests, it was considered that most of the detailed changes were made in response to differences in the perceived readership for the two papers. The audiences were seen as likely to have broadly different purposes in reading the paper. Some of the details of A were thought to be of less relevance or interest to the intended audience for B and consequently deleted. In addition, changes were made in response to the relative sophisiicaiion of the readerships in dealing with technical econometric models and texts. Examples from each category are presented, together with a commentary'and, where appropriate, a report of remarks made by the author. 110 More explicit statement of objectives (Change category 1.1 .1) A more explicit statement of objectives was largely confined to the title of the paper and to subheadings. The title, for example, was changed from (1) A: Current Account Imbalances and Foreign Debt Dynamics in 26 Developing Countries to B: Debts and Deficits in Developing Countries: Same Stabilising and Destabilising Effects of Foreign Debt Accumulation on Current Account Deficits. In both titles a relationship is signalled by and: in the first, the topic of the paper is the relationship between current account imbalances and foreign debt dynamics; in the second, the topic is said to be the more general relationship between debts and deficits. The first part of the second title was, according to the author, intended to be a "much snappier, journalistic title", reflecting the perception of a more general readership. The second title also goes on to state explicitly that same of the factors or effects that influence this relationship will be presented in the paper. Presumably, for the policy-maker an explicit promise of information about potentially manipulable variables is more enticing than a more vague statement that a relationship is to be investigated.

Rephrase to make more policy-oriented (1.1.2) As well as the major additions to B noted in 2.1, the author makes a number of other smaller changes which reflect his wish to direct his comments primarily to policy makers rather than academic economists (not that the two groups are nec­ essarily exclusive). Consider, for example, the following extracts. (My italics.) (2) A: Policy cancern over the balance of payments is a cancern not so much. over the current account imbalance but rather over the accumulation of net foreign claims. B: Policy cancern over the balance of payments should not focus on a current account imbalance per se but rather on the accumulation of an unmanageable amount of foreign liabilities. The need for a policy message leads the author to switch from a positive statement in the original version to a normative statement in the Amex paper. For the specialist audience, the author writes that "Policy cancern [... ] is not so much over [ ...] but rather over ..." (a positive, testable statement); while for the Amex audience it is rephrased as "Policy cancern [... l should not focus on [ ...] but rather on ... " (a normative statement based on an untestable value judge­ ment). This change represents, according to the author in our interviews with him, "a big stepping stone in economics": a step from "what is" to "what the author would like it to be." 111 Delete a section (1.1.3) We noted that major deletions from A in writing B were mainly related to the removal of text to the appendix and the trimming down of text connected with the econometric mode) of current account and monetary policy behaviour in developing countries. Other less substantial deletions are also obviously related to audience. For example, complimentary reference to the IMF (for whom the original paper was written) is removed from B: (3) A: Ptompted by reports from such international organisations as the International Monetary Fund, the Western press has bombarded its readers ... B: Over the past five or six years, the Western press has bombarded its readers ...

Rephrase to include less infonnation or delete information (1.2.1) Deletions are made in order to remove some of the technicalities and detail found in the IMF paper. For example, campare (4) A: While the cumulated current account deficit and small amounts oj government debt do generate stabilising forces, possibly ... with B: While the cumulated current account deficit does generate stabi- lising forces, possi bly ... Here, one of the more minor factors generating stabilising forces is deleted. The reason for this deletion, as stated by the author in our interviews, is that he did not want to "bombard the reader with anything more than absolutely necessary." In this case he "wanted to [... ] just say what the main things are [ ...] and forget about the trivials." A number of similar examples are found in which such things as conditions, peripheral information or implications are deleted.

Grammatical restructuring (1.3.1) In categorising the changes, a number of instances of grammatical restructuring or reformulation were noted, and it was postulated that these could be explained as an attempt to make things easier for the less specialised readers. Most of these involved making sentences shorter; for example, in (5) A: The accumulation of debt resulting from the current account deficit in year O raises the foreign saving function to Sf1, producing an effective cost of foreign borrowing of r1 in year I. 112 B: The accumulation of debt resulting from the current account deficit in year O raises the foreign saving function to Sf 1• This change produces an effective cost offoreign borrowing ofr, in year 1. In discussion, the author stated that he considered shorter sentences to be more appropriate for less sophisticated readers.

More explicit signalling of relationship between elements oftext (1.3.2) There were frequent additions in B involving the more explicit signalling of the relationship between parts of the text. For example, in discussing the extracts (6) A: ... and real interest rates which act as stabilisers ... B: ... and interest rates. These feedback mechanisms act as stabi- Iisers... the author suggested that he considered it necessary to inform (or remind) readers of B that (real) interest rates are "feedback mechanisms" as well as acting as stabilisers, while for the more specialized readership this knowledge could be assumed.

Change tenns used (1.3.3) More of the differences between the two papers were placed in this category than in any other. Some of these were related to a description of the mode! used and graphical representations. Comparing papers A and B, variables became factors, steady states became positions, more volatile became fluctuatedfar more over time, junctions became lines, and shijts (in these Iines) became movements. In our discussions with him, the author commented particularly on one of these changes, from junction to fine in relation to a graph. He pointed out thatfunction is taken to be a "shorthand", describing "what influences the line" rather than the line itself. Strictly speaking, then, fine is not synonymous with junction, but chosen because it is something more concrete than an unseen junction. "A technical audience", he added, "would find savings fine a bit funny." At other times, he judged that in order to make terms used in paper A accessible to a less specialized audience, an explanation or some further clarification would be needed. Rather than do this, however, the auihor decided to modify the terminology used. Thus, for example, real exchange rates be­ comes exchange rates, real interest rates becomes interest rates, net foreign claims becomes foreign debt, and the domestic real shadow interest rate becomes the domestic real interest rate. 113 Restatement to reduce "compkxity" (1.3.4) This category is something of a ragbag including stretches of text longer than the word or phrase which are reworked to reduce some of their apparent complexity. It is frequently difficult, however, to say whether some of these changes are made specifically to accommoda.te the perceived potential problems of the new audience, or simply to improve style. For example, in (7) A: ... the three developing countries exhibited far greater current account volatility than did the three industrial countries. B: ... current accounts in the three developing countries exhibited much greater volatility than they did in the three industrial countries. the nominal group current account volatility is changed to current accounts (exhibited) volatility.

Rephrase to make more explicit (1.3.5) It was observed in category 1.3.3 that in B the author sometimes uses other terms or deletes parts of terms in order to avoid having to include what is seen as being a necessary explanation. At other times he chooses to add a short explanation or clarification. For example, in discussing

(8) A: At the effective cost oj joreign borrowing r 0, domestic investment l 0

exceeds national saving Sn0 • B: The ejjective cost oj joreign borrowing is also the effective

domestic real interest rate. At an interest rate oj r0 , domestic

investment I0 exceeds national saving Sn0 • the author stated that it was "sensible to make explicit [the equivalence between the effective cost oj borrowing and the ejfective domestic real interest rate] for the Amex readership," presumably because knowledge of the equivalence could not be assumed.

Different paragraph breaks (1.3.6) There are two instances in which an additional paragraph break is included in paper B. The author provided no motivation for this other than his impression that a paragraph break makes comprehension easier.

Improvements in style (2.1) The major difference between the categorisation presented in figure 1 and our earlier preliminary classification was that a number of changes which we considered bad been made in response to differences in perceived readership 114 were said by the author to be improvements in style that he would probably have made irrespective of assumed audience. For example the change in (9) A: ... current account or net stock of these countries' foreign assets/ liabilities, ... B: ... current account, which is the net stock of these countries' foreign assets or liabilities ... was said to have been made to remove the ambiguity of or rather than to make more obvious the relationship between current account and net stock for a less specialist audience. Also, examples of personalisation, as in (10) A: The modet developed here from saving and investment behaviour attempts to capture ... B: My current account modet attempts to capture ... were originally thought to indicate a more "interactive" stance prompted by the assumption of a less sophisticated audience. Discussion with the author suggested, however, that while this was a conscious style change, it was not motivated by audience perception, but by the feeling that the personal form "read better".

Improvements in content (2.2) There is just one type of change classified under this heading: updating of the cited literature.

2.3 Summary We have seen that the majority of changes made in rewriting Version A as Version B were in response to the different purposes of the two papers and their intended readership, as perceived by the author. The most obvious effects are the reduction in length and the inclusion of policy implications in Version B. The audience for Version B was assumed by the author to have either less interest in more technical detail or less ability to comprehend it. The main organisational change made in response to this was to remove some of the more technical information to an appendix. Many of the small-scale changes were also made to accommodate this by reducing complexity: terminoiogy was changed or explained, and detail deleted. In addition, the author considered that the audience for Version B would be less skilled at reading technical economics texts; less skilled in the sense that they would either have little experience of reading such texts, or would have less of a background in economics generally. Attempts to accommodate to this 115 perception of audience included a more explicit statement of objectives, a more overt signalling of the relationship between parts of texts, grammatical restructuring, and the splitting of longer sentences into two.

3 Analysis of a comparable prize winning text and a related paper The article by De Grauwe & Vansanten, Speculative dynamics and chaos in the foreign exchange market, won second prize in the AMEX Review prize competition for 1991. Like Fry's competition paper, there was a more academic paper on which it was based, Deterministic chaos in the foreign exchange market, written as a discussion paper for the Centre for Economic Research, in January 1990. We selected these two papers for comparison with Fry's work, because both Fry and de Grauwe & Vansanten had developed theoretical models as the basis for their papers, whereas Brainard, the first prizewinner, bad written a heavily policy-orientated paper that did not require any mathematics or modelling as part of the main text or appendices. Paul de Grauwe is Professor of Economics at the University of Leuven in Belgium, and also a Research Fellow of the Centre of Economic Policy Research in London, as well as the Centre for European Policy Studies in Brussels. Kris Vansanten is an economist with the Bank of Brussels Lambert in Brussels. Both their papers report attempts to mode! the foreign exchange market using a mode! based on recent chaos theory, under the assumption that there are two classes of speculator, "chartists" who follow exchange rate movement and extrapolate from events, and "fundamentalists" who compute and base their calculations on the equilibrium rate. What are the most immediately perceivable differences between the two papers (which we will label CEPR and AMEX for ease of references)? Length is the first observable difference. CEPR is about 6,650 words long, and AMEX about 3,450 words. However, AMEX is not a straightforwardly shortened version of CEPR. In general, comparable sections in AMEX are considerably shorter than in CEPR, but AMEX has both a longer introduction and a longer conclusion. There are also two short sections of CEPR which are omitted from AMEX. Another noticeable difference between the papers is the removal of all equations and mathematics from the body of AMEX. Of the original 18 equations in the CEPR text, some, but not all, are transferred to the appendices in AMEX: six cquations relating tu chaos theory are omitted altogether. The number of figures and tables is also very much down in AMEX, to eight (plus two tables in the appendices) from an original total of 25. Also, section eight of CEPR has six figures and a flow diagram, and none of these are transferred to the comparable section in AMEX (section six). 116 Closer scrutiny of the two papers reveals that the AMEX paper is much stronger on policy implications than CEPR. AMEX has its own section on policy, the comparable section in CEPR focusing on the models used to describe government intervention, rather than making recommendations (reflected in the absence of equations and figures from the AMEX section). The differences are also identified in a comparison of the two sets of references. CEPR contains a !arge number of references to chaos theory, and to areas such as statistics, econometrics and areas of science not shown in AMEX, where AMEX has additional references to intervention and sterilisation, and chartists and fundamentalists. Before analyzing the texts to attempt to establish differences between them which, in our view, derive from different perceived audiences, it is useful to make some general observations about the papers. They follow accepted conventions for reporting research, using variations of the scientific method of reporting: Introduction + Method + Results + Discussion. Each signals clearly what is going on at key stages in the text; for example: (11) CEPR: In this paper an attempt is made to introduce some concepts from chaos theory in exchange rate models. It will be shown that relatively simple models are capable of mimicking real-life exchange rate movements. (p. 2) AMEX: In this section we presenta simple partial equilibrium mode! of the foreign exchange market and describe the non-linear features of the mode! that are essential to produce the chaotic results. (p. 81) The two examples can also act as illustrations of the highly interactive nature of both texts. The writers are continually intervening between the reader and the underlying subject matter of the text to signal, to interpret, to evaluate, to convey their attitude to the truth, to refer the reader to diagrams, illustrations or real-world behaviour. There is a tendency in some of the social sciences for nominal groups to become very dense, for example, "thereby inhibiting any long-run steady-state equilibrium from emerging" (from Fry's papers). While de Grauwe & Vansanten do not entirely avoid the use of complex nominal groups, they generally limit the nouns to two per group, such as parameter values, expectations formation, chartist behaviour and money stocks. Their longer nominal groups tend to be related to items which are generally known in the everyday business world, such as juture exchange-rate movement and exchange­ rate paths. Surprisingly, there are fewer of the latter type in CEPR than in AMEX. Our experience of reading other prizewinning articles suggests that comprehension is helped where the relationship between several nouns in the same nominal group is identified in some way, such as through prepositions. 117 One would need to reword Fry's statement above to show this: "thereby inhibiting any equilibrium from emerging in the steady state in the long-run." We now turn to the writers' perceptions of audience as we infer them on the evidence of the texts. A useful way of establishing this is to compare sections having similar functions in the two articles. Analysis of the papers reveals that both at the leve! of function and wording, direct comparison is very difficult. Despite overall similarities of organisational structure, the two papers are essentially different. An example of this is given for illustration. For this purpose we have chosen the beginning of section 3 from both articles.

(12) CEPR:

3 The exchange rate mode!

(A) IOur main goal in this section is to use a mode! of the exchange rate which has a wide degree of acceptance in the profession and which produces an aperiodic dynamic behaviour of the exchange rate under certain parameter values. 2The danger with non-linear models is that they introduce ad-hoc non-linearities. 3In order to avoid this danger we will introduce non-linearities in the mode! where there is a strong foundation from economic theory to do so.

(B) IThe mode! consists offour equations. 2Qne equation explains capita! flows, the second orre the trade account; the third orre specifies the expectations formation, the final orre is the equilibrium condition in the foreign exchange market. 3Two non­ linearities are introduced, orre in the trade account equation where we assume a J­ curve effect. 4A second orre is the expectations formation equation where it will be assumed that agents take into account both the past behaviour of the exchange rate and its "fundamental" value when forecasting the future exchange rate.

AMIDC:

Exchange rates and deterministic chaos

(A) lJn this section we present a simple partial equilibrium mode! of the foreign exchange market and describe the non-linear features of the mode! that are essential to produce the chaotic results. 2The mode! consists of four equations. 3One equation explains the trade account, the second orre explains capita! flow via an international portfolio allocation mode!, the third orre describes the expectations formation and the last orre is the equilibrium condition in the foreign exchange market.

(B) 1The trade account is determined by the present and past movements of the exchange rate. 2We assumc that a J-curve phenomenon is present, i.e. a depreciation first tends to worsen the trade account before it improves it. 3This introduces the first non-linearity in the mode!. Functionally there are considerable differences between the two excerpts, even though they cover substantially the same topics. Paragraph A of CEPR is designed to persuade academic readers of the text that the writers are members of an academic profession who work within agreed parameters (using a mode! that "has a wide degree of acceptance in the profession"). They recognise that 118 the mode!, despite its acceptance, has !imitations which opponents might focus on, so de Grauwe & Vansanten counter possible objections from the start, by stating what they propose to do to solve the problem, working within the generally agreed parameters of the discipline ("we will introduce non-linearities [... ] where there isa strong foundation from economic theory to do so"). The first paragraph of AMEX by contrast is broadly descriptive. It is a summary of what their mode! consists of. The only possible element of required justification for the reader is the comment that they will describe non-linearities that are "essential to produce the chaotic results." This could be interpreted as an attempt to play down the difficulties of the text for the non-expert reader; difficult, technical aspects will only be dealt with when absolutely necessary. When we come to the introduction of the "J-curve effect", we notice that it is the AMEX audience that is considered to need an explanation. The CEPR reader is assumed to be familiar with the J-curve effect, whereas the AMEX reader requires a very simple explanation ("a depreciation tends to worsen the trade account before it improves it"). The writers are continually evaluating the information they are dealing with, and the claims they wish to make; making decisions about what the two different readerships need to know, what they may not know and that the writers must explain for them, what can be omitted because it is too difficult, or of less importance, who they must persuade, whose objections they must counter. Turning from the functional to the förmal differences in comparable parts of the two texts, and returning to the beginnings of section 3 (example 12 above), it becomes clear that there is very little, if any, copying between the two papers. Each excerpt contains the sentence "The mode! consists of four equations", hut beyond this there is little verbal similarity. Items are described in different orders, and synonyms are used to differentiate similar wording. This is typical of the differences between the two papers in general. It is very difficult to identify detailed se.ctions as identical in the way we

4 Comparison of the work of Fry with that of de .Grauwe & Vansanten The study of the way in which de Grauwe & Vansanten manage the process of audience accommodation in their different papers allows Us to make some statements about how Fry copes with the adjustment required in the presentation of his work. Comparing the two processes, it is clear that for Fry, the process of accommodation is text-based, looking at his föst text, cutting, editing, simpli­ fying, inserting, and altering words and structures. Most of this takes place in front of the word-processor screen, the original text gradually turning into a second, which then develops into a third, and so on. This process we refer toas rephrasing. In contrast, de Grauwe & V ansanten are focused more on their academic work, reconceptualising it in a different way fora different audience. There are other differences toa. De Grauwe & Vansanten make frequent interventions in their texts to signal to the reader how they intend him/her to comprehend the text. Their approach is interactive. With Fry, however, the texts are much more autonomous; that is, he makes a series of statements which are in general to be taken as true, without indicating what is expected of the reader. This is not to say that he is not personal; he gives his views ("when the World Bank asked me to analyse [ ... ] I was convinced that ... "). But he is not interactive. It is difficult to identify from the text itself any explicit sense of a triadic relationship between reader, writer and the text to be understood. He indicates his attitude to the truth by use of modality ("the level of this time of foreign debt [... ] may itself affect the position of both the saving and investment Iines"). He is thus signalling to an academic audience. But he does not signal to the non-expert reader what is to be made of his statements, and this produces some very dense and difficult prose, even in Version B. For example: (13) B: The reaction of the monetary authorities to a buildup of govern­ ment and government-guaranteed foreign debt is quite different. Instead of reining in domestic credit expansion, the monetary authorities expand domestic credit. In this case, therefore, the current account is worsened by an increase in DETY: (1) directly in 120 the current account equation (actually by the DETY2); (2) indirectly through its effect on monetary policy. Both behavioural and policy induced reactions to increased government and government-guaranteed foreign debt are destabilising, thereby inhibiting any long-run steady-state equilibrium from emerging. Apart from the references to DETY, which the non-expert reader may have difficulty with despite its being defined earlier, what is being described here is comprehensible to the norr-expert reader with a Iittle effort. What is missing is any guidance as to the relative importance for the reader of the statements made. Should the reader find this passage difficult and decide not to try to fully establish what it means, there are no clues as to whether something vital is being missed or not. The density is compounded by Fry's use of complex nominal groups, where the relationship between nouns in the same group is not always spelt out. Some examples are: "a buildup of government and government­ guaranteed foreign debt", "behavioural and policy-induced reactions to increased government and government-guaranteed foreign debt", and "long-run steady-state equilibrium". In order to read these groups correctly, possible ambiguities have to be disentangled. For example, in the first there could be a build-up of government (i.e. increased governmental intervention) plus a build-up offoreign debt which has been guaranteed by government. This is not what is intended, and one would assume the reader knows enough not to misinterpret, but this is an example of a way of writing that by its density adds to the mental-processing load required if the reader is to decode correctly. It is interesting to speculate on influences that have helped to make Fry's writing the way it is. One possibility is his method of composing. He is heavily dependent on the use of the word processor, even to the extent that he typeset his own 1988 academic book of some 522 pages. Where this skill on the one hand would enormously increase his ability to produce efficiently and at speed, it would, on the other hand, encourage the process of text-based shifting of sections of work from one source to another, and of concentration on fairly small-scale changes to his texts. In fäet, almost identical sections can be found in several different papers. It is not being suggested that any of the variants are inappropriate in the textual contexts in which they occur. What is being illustrated is a tendency which, if taken up extensively, will encourage a rephrasing, rather than a reconceptualisation approach when using existing text to rewrite for different audiences. For all kinds of reasons it is far easier to press a few buttons on a keyboard to transfer text than it is to start from scratch. 121 5 Conclusion In this paper we have presented evidence for two processes in the rewriting of specialised papers in economics for a lower-level scholarly audience. These processes we have termed rephrasing and reconceptualisation. In rephrasing, the writer might be thought of as responding to the question: "How can the original text be rewritten to accommodate to a different readership?" On a )arge scale, the result is to make additions and deletions and to reorder certain material. On a small scale, there are changes in terminology, grammatical restructuring, and more explicit marking of the relationship between sentences. The questions preceding reconceptualisation would seem to be something like: "Given the different readership, what material is it approp­ riate to include, and how is it best to organise this material in such a way as to make it accessible?" The result is a substantially different paper, with major changes to the organisational structure, and emphasis given to different parts of the material. Our evidence suggests that reconceptualisation is more likely to produce a new product which is appropriate for this audience. · Clearly, rephrasing and reconceptualisation, as we have characterised them, are not discrete processes. A more explicit marking of the relationship between sections of the text, for example, could be as much a part of rephrasing as reconceptualisation, while an assessment of audience may in fäet result in only minor, mechanical changes toa particular section of the original text. Generally, however, in the two sets of papers investigated here, one or other of the processes seems to predominate. A further feature which seems to distinguish the de Grauwe & Vansanten and Fry submissions to the AMEX Review prize is the degree of interaction with the reader. De Grauwe & Vansanten, we have argued, carefully guide the reader through the text, referring him/her to visual material in the text and to real-world behaviour outside it. Help is provided in interpreting and evaluating content. In Fry's writing, however, there is much less evidence of such assistance, often leaving the non-expert reader independently to assess the relevance of statements. The !imitations of a case study such as this are obvious. We have focused on three authors whose writing processes may or may not be representative of those of economists more generally. Our inferences about audience accommodation by de Grauwe & Vansanten are based on only two written products, without direct access to information about the processes which shaped them. However, while the processes of rephrasing and reconceptualisation as we have represen­ ted them here may well need to be refined with further investigation, we would argue that the distinction made between them hasa general validity. 122 References

Primary sources

De Grauwe, P.C. & Vansanten, K.J.L., 1990: Deterministic chaos in the foreign exchange market. CEPR Discussion Paper, No. 370. De Grauwe, P.C. & Vansanten, K.J.L., 1991: Speculative dynamics and chaos in the foreign exchange market. In R.O. O'Brien & S. Hewin, Finance and the lnternational Economy: 4, The AMEX Bank Review Prize Essays. Oxford, Oxford University Press, 79-103. Fry, M.J., 1988: Money, Interest and Banking in Economic Development. Baltimore (Maryland), John Hopkins University Press. Fry, M.J., 1991a: Current account imbalances and foreign debt dynamics'in 26 developing countries. Paper prepared fora seminar at the OECD Develop- 1 ment Centre, 29 May 1991. Fry, M.J., 1991b: Debts and deficits in developing countries: some stabilising and destabilising effects of foreign debt accumulation on current account deficits. Paper submitted to the AMEX Bank Review prize, 1991.

Secondary sources

Berkenkotter, C. & Huckin, T.N., 1993: Rethinking genre from a sociocognitive perspective. Written Communication, 10, 4: 475-509. Fahnestock, J., 1986: Accommodating science: The rhetorical life of scientific fäets. Written Communication, 3, 3: 275-296. Galbraith, J.K., 1988: The grammar of political economy. In A. Klamer, D.M. McCloskey & R.M. Solow (eds.), The Consequences oj Economic Rhetoric. Cambridge, Cambridge University Press, pp 221-239. Goodwin, C.D., 1988: The heterogeneity of the economists' discourse: philo­ sopher, priest, and hired gun. In A. Klamer, D.M. McC!oskey & R.M. Solow (eds.), The Consequences oj Economic Rhetoric. Cambridge, Cambridge University Press, pp 207-220. Keohane, R.O., 1986: The rhetoric of economics as viewed by a student of politics. In A. Klamer, D.M. McCloskey & R.M. Solow (eds.), The Con­ sequences oj Economic Rhetoric. Cambridge, Cambridge University Press, pp 240-246. Klamer, A., 1990: The textbook presentation of economic discourse. In W.J. Samuels (ed.), Economics as Discourse: An Analysis oj the Language oj Economists. Boston, Dordrecht and London, Kluwer, pp 129-165. McC!oskey, D.N., 1986: The Rhetoric oj Economics. Brighton, Sussex, Har­ vester Press. 123 Myers, G., 1990: Writing Biology: Texts in the Social Construction oj Scientific Knowledge. Madison, Wisconsin, University of Wisconsin Press. Swales, J.M., 1990: Genre Analysis: English in Academic Research Settings. Cambridge, Cambridge University Press. Swales, J.M., 1993: The paradox of value: six treatments in search of the reader. In W. Henderson, T. Dudley-Evans & R. Backhouse (eds.), Economics and Language. London, Routledge, pp 223-239. Warsh, D., 1986: "Yellow rain" and "supply-side economics": Some rhetoric that failed. In A. Klamer, D.M. McCloskey & R.M. Solow (eds.), The Consequences oj Economic Rhetoric. Cambridge, Cambridge University Press, pp 247-264.

Discourse in a Technical-Directive Setting: The Case of Operating Instructions

Risto Hiltun~n & Sirkku Nyman

Operating instructions are texts that everybody is familiar with. As texts they are short-lived; they are forgotten as soon as the operation itself is mastered. In fäet, we tend to become aware of the existence of instructions only when they do not work and something goes wrong with the operation. This can be a frus­ trating experience, and it is therefore in the interest of both the manufacturer and the user to try to avoid this kind of situation. Consequently, every effort should be made to ensure that the instructions are (a) informative enough and (b) linguistically as clear and simple as possible. Because of these constraints, instructions have developed a particular format: that of a schedule. The information which schedules order and classify varies from simple lists of items to elaborate sets of rules with graphs and tables (Sager, Dungworth & McDonald 1980: 118). It is also characteristic of schedules that the information they transmit is more commonly addressed to occupational and professional groups than individuals. As far as the linguistic simplicity of the instructions is concerned, it will naturally vary according to the kind of equipment and the levet of expertise expected from the reader, and some instructions are more helpful than others, but generally speaking, the message of a schedule tends to be self-contained. This is because schedules are seldom read like a regular text, in a single process, but are rather referred to selectively and at different times, and because the receiver of the information is expected to be competent enough to understand the text. Following Sager, Dungworth & McDonald (1980: 164) we define instructions as specifications of behaviour in certain situations, which are not necessarily strict orders or imperatives for action, but indicate the ways the actions are to be carried out if the recipient decides to undertake them. Operating instructions are designed to provide practical, portable instructions that can be understood independently of any other text, since they are likely to be used in situations where they are the only texts available. Operating instructions are directive texts. The receiver of a message is directed to do something, and if he acts as expected, the communication has been successful. The sender of a directive message is superior to the receiver in knowledge, and because of his prestige he can use distinct linguistic forms in making his purpose clear: requests, orders, instructions, prohibitions, exhorta- 126 tians, and warnings. Directive intentions are more common in technology texts than in science texts because science, according to Sager, Dungworth & McDonald (1980:27), is concerned with the extension of knowledge, whereas technology is concerned with the application of knowledge. Operating instructions usually divide into three parts: instructional, informa­ tional, and storage (Trimble & Trimble 1978:81). Instructional information is met in chapters on operation and maintenance; informational function is present in the introduction and theory; and storage consists of specifications and a parts list. (For further discussion, see Nyman 1991:23ff.) Reading operating instructions will soon indicate that nouns, complex noun phrases, prepositional phrases, adjectival modifiers, and a high type/token ratio, are common in such texts. These characteristics can, according to Biber (1988: 104), be associated with a high informational focus and careful inte­ gration in a text. Noun phrases, whether simple or complex, are the primary bearers of referential meaning; and a high frequency indicates high density of information. Prepositional phrases serve to further integrate the information into the text. A high type/token ratio also indicates a high density of information. Attributive adjectives are used to elaborate the information conveyed by nominals. They also help to pack information into relatively few words and structures. The features which are very rare in operating instructions include private verbs (e.g. think,feel), first and second person pronouns, wh-questions, emphat­ ics, amplifiers, and sentence relatives. These are associated with an involved, non-informational focus, interactive or affective purposes. What is typical of operating instructions, and any instructive material, on the other hand, is that they use either present passive forms or imperative forms. Attributive adjectives are used throughout to specify the particular referent intended. Procedural discourse differs from expository discourse in that it is event-driven and concrete rather than conceptual and abstract.

1 Three central aspects of operating instructions This paper discusses three central aspects of operating instructions from a functional point of view. First, the functions of the directive speech act and their realizations are described in section 2. Second, the active and the passive voice are dealt with as an example of syntactic realization of certain text functions. Finally, iconicism is presented as an example of a crucial device for text structuring.1

1 In the preparation of this paper, Risto Hiltunen has been mainly responsible for the functions of the directive speech act and their realizations, and Sirkku Nyman has been responsible for describing the active and the passive voice and iconicism. The authors would like to thank Dr Keith Battarbee for kindly revising the English of the article. 127 These aspects are present in most operating instructions and contribute significantly towards the pragmatic goal of the text: a successful description of the action to be carried out by the addressee. The data of the study consist of authentic operating instructions (under the heading Operation in the relevant manuals) of some highly sophisticated medical equipment for (a) oxygen monitoring systems, (b) blood pressure monitoring, and (c) an intra-aortic balloon pump. (For details, see the list of Sources.) These particular data were chosen for three reasons. The first reason was to study texts that are concerned with fairly sophisticated equipment so as to reveal as many of the characteristics of the instructional sections of the texts as possible. The second reason was to find instructions for the kind of equipment which the user simply must get right and where he cannot afford to make mistakes. Consequently, we hypothesized, such instructions must be maximally watertight. The third reason was to examine instructions that are written for a particular professional group with a fairly high leve! of technic.al expertise.

2 Directives The operating instructions represent a text type where the directive is regularly realized through the imperative. Of the other possibilities, it is only the modal declarative (with should) that is frequent This provides a possibility for more indirect way of expressing directivity.

2.1 Typology The typology of operating instructions can be presented in the form of a flowchart. In the present data it takes the form of either of the following:

mtl ~ 1 1 i i 2 2 i ...,. i 3 i I,(' 3 etc. 4 etc.

Figure 1. Types of operating instructions 128 Type 1 is the straightforward, simple, type, where one direction is followed by another (cf. instructions how to operate the telephone). Type 2 has "loops", or digressions, in between the commands, which typically either give extra information or recommend a particular procedure for carrying out the action, as in the examples below (all emphases in italics in the examples are our own). The first example gives an idea of how extra information can be presented in a loop. (1) Remove the protective cover, dilate the umbilical artery and advance the catheter gently into the descending aorta in the same manner as an umbilical artery catheter. Depth marks af 5, 10, 15 and 20 cm from the catheter tip are provided ta aid insertion. A typical recommendation looks as follows: (2) The position oj the tip should be determined by the physician and can be checked radiographically since the catheter section and in particular the electrode, are radiopaque. Additional information is usually expressed in the form of a statement, but even conditional and negative clauses may have this function. The recommendations characteristically take the form of X should be done, but here too there may be other possibilities, e.g. negatives, as in example (3): (3) Treatment of the umbilical area with antibiotic aerosols is not recommended other than possibly at the time of insertion.

2.2 Functions of directives Directive statements are usually positive (do X), but they may also be negative (do not do X). The negated imperative is a prohibition; it is emphatic and functionally often a caution or warning. It typically specifies a preceding positive command, as in the following: (4) Cut off remaining distal part of delivery tubing using scissors or leave in place. Do not cut the proximal part of the delivery tubing which contains the sensor. (5) PULL, do not tum. The direclive act is in most cases associatcd with the purpose or consequence of the action, which may or may not be expressed, i.e. it may be implicit or explicit. The purpose or consequence of the action is usually implicit; it is understood and taken for granted. This is probably true of instructions generally. The reasons for the implicitness may be threefold: pragmatic, contextual or linguistic. 129 Pragmatic reasons have to do with the reader's assumed le.ve! of competence and hence they are regarded as redundant. Sometimes this reason is actually stated; e.g. (6) The following instructions are written assuming that the operator is familiar with the preparation of extracorporeal bypass and are there­ fore only suggestions. It is possible to apply this procedure only when it is known exactly who is going to read the text. Contextual reasons have to do with the immediate physical context of presentation, e.g. information regarded as redundant because of accompanying illustrations. Linguistic reasons have to do with the meanings of the words used, and their collocations, i.e. lexical implication; e.g. (7) NOTE: do not touch or contaminate slightly protruding sterile sensor and surrounding area. The collocation of touch, contaminate and sterile is sufficient to suggest that the consequence of doing otherwise would be harrnful. The reasons for spelling out the purpose and consequence may be twofold: noteworthiness of the consequence, and icönicism. To deserve explicit mention, the risk involved in not following the instruction must be both significant and not self-evident, cf. (8) Once the sensor is in the desired position, tighten the rear screw assembly "A" to prevent jluid leakage. (9) Pull gently apart as far as the automatic stop. Do not force beyond the automatic stop or you may damage the delivery system. In the following example the element interrupting the flow of directives is noteworthy enough to be expressed explicitly: ( 10) Flush the Iine with heparinized solution. It is essential that the conti­ nuous flush device be utilized with a heparinized solution throughout the procedure to maintain adequate resolution oj the blood pressure waveform. Iconicism, on the other hand, is most evident in the step by step procedures where a particular function is equated with a particular effect, as in the cali­ bration of the device, cf. (11) Press [TEMP] to set patient temperature Press[<)[>] to adjust display to actual patient temperature Press [STOREJ[RECALL] to store current display value etc. 130 The aspect of implicitness vs. explicitness is an important consideration for those drafting instructions, and misjudgements in this respect may tum out to be costly. (For an example, see Shuy 1991.) Cautions and warnings are fairly common in the data, but it is not always clear why some statements are expressed as cautions and others warnings. The present material indicates, however, that there may be a functional pattern here. Most of the cautions refer to the functioning of the equipment, while references to consequences harrnful for the patient are rare. Warnings, on the other hand, are equally distributed in terms of the harm caused to the equipment and the patient or the user. Cautions usually take the form verijy/ensure!assure!that ... or, more indirectly, care!caution should be taken to ... Warnings, on the other hand, show a definite preference for the pattern do not+verb (Nyman 1991 :73- 74). More comprehensive data might reveal whether warnings are merely emphatic cautions or whether, in fäet, there is a difference in the sense that cautions are equipment-oriented and warnings person-oriented, as our results would seem to suggest; cf. the examples below: (12) CAUTION: Use of presoaking fluids other than isotonic saline may result in damage to the oxygen sensor (13) WARNING: Do not use interna! triggering when the patient gener­ ates a cardiac cycle.

3 Active versus passive A noticeable syntactic feature in instruction manuals is the alternation between the active and passive voice of the verb. What functions do the actives and passives have in different contexts?

3.1 The active The active forms comprised 56% of all finite verb forms in our texts, imperatives excluded. They were frequent in both the descriptive sections of the text and the actual operating instructions. They are used to describe what the system comprises, e.g. (14) One lumen of the catheter contains the connecting wires from the electrode to the monitor. Then they are used to describe what the equipment itself performs, e.g. (15) The blood gas analyzer measures oxygen tension at 37 °C, whereas the in vivo system measures at the temperature of the patient. They are also used to describe the result of a particular function, e.g. 131 (16) The monitor will show "Act" on the P02 display until the sensor is activated (normally 10-20 minutes), when an uncalibrated reading will appear on the display.

3.2 The passive In all other cases the passive is used. In the analysis of the passive we compared our examples with the findings of Tarone et al. (1981), hut they could be applied to our material only partly, since their texts are scientific, while our texts are directive. For example, their generalization about the quasi-passive use of the first person plural pronoun does not apply at all to our texts. This is only to be expected on the basis of the directive function of the instructions. However, as in the data of Tarone et al., the use of the passive to describe established, standard procedures was very frequent, as examples (17) and (18) indicate. (17) The sample should be analyzed immediately after it has been taken from the patient, and gas-tight syringes must be used. (18) The blood sample should be stored in ice water if it has to be trans­ ported to the analyzer or if it cannot be analyzed immediately. Sometimes the procedures described were not actually "standard", hut rather issued as recommendations to be followed, e.g. (19) The catheter may be secured with a suture at the exit of the umbilicus. In (19) the choice of compliance is left to the user, hut with the understanding that he will presumably adopt the procedure. The category of established or standard procedures can also be said to include the cases where the passive is used to describe things on which the user has no influence, i.e. what the manufacturer or the device provides for the user, e.g. (20) A hold state is indicated by two decimal points and a zero displayed in the mean pressure window. (21) Tum monitor on and observe that all elements of the front panel have an illuminated display and that the audible alarm is operational. If the BATT LOW indicator is illuminated connect the monitor to • the A/C supply and switch ON at the rear panel. Under established procedures one must also count the instances where the passive is used as an indirect instruction or a command. This was the case with the modal constructions in particular. 132 (22) The suture should be tied loosely to avoid constriction of the sam­ pling lumen and in a purse string or similar form to permit easy removal prior to withdrawal of the catheter. In several instances the use of the passive is dictated by textual factors, but,, as Tarone et al. point out, these constitute a generalization that follows from general language. Passive verb forms are used, for example, for textual organi­ zation and emphasis. (23) An empty 2 ml syringe is first used to clear blood and infusion solution from the catheter. A heparinized syringe is used to draw out the volume fora blood gas sample; the contents of the 2 ml syringe used to clear the line are replaced slowly via the catheter. In examples such as (23), the action is the most important thing; the agent is less important. It is not surprising that the agentless passive is far more frequent than the agentful one. When the agentful passive is used, its use is also motivated by textual reasons such as topicalization, e.g.

(24) Deflation is controlled by the setting of the DEFLATION POINT (2-10) control so that it occurs prior to the next systolic upturn.

4 Iconicism Of the iconic patterns it is the syntactic one that is most frequent in the instructions. The same syntactic arrangement is often used throughout the text, e.g. (25) Check that cylinder is Medical "E" size Check that plastic washer is present Check that cylinder is Medical "D" size. (26) Prior to operating the equipment Prior to assisting a patient Prior to initiating IABP These examples show that syntactic iconicism often co-occurs with lexical iconicism. Both devices contribute to maximizing clarity and minimizing ambiguity. Another type of iconicism which is characteristic of instructions is so-called experiential iconicism (Enkvist 1981: 97), meaning that in natura! order (ordo naturalis), text and discourse have the same arrangement as things in the universe of discourse, while in artificial order ( ordo artificialis) this arrange­ ment has been somehow disrupted. 133 Operating instructions normally follow the natura! order. However, even in these instructions there were a few deviations from the natura! order, usually fully marked, however. The most important motives for the deviations are: thematic factors, relative importance of a point, "aftyrthoughts", or a possible error. For thematic factors (including emphasis), see (26) and the examples below: (27) Loose.n the rear clamp by partially unscrewing section "A". (28) Wait for at least 30 minutes after insertion for sensor tö stabilize before starting calibration. Relative importance of point is the motive used in such cases where deviation is possible, as the point is not crucial for the main issue, e.g. (29) Ensure that the oxygen electrode at the tip of the catheter is immersed in the fluid. Flush the catheter to ensure that no air remains in the lumen. Presoaking the electode in sterile salinefor 15 minutes is desirable to accelerate activation in vivo. (30) It is advisable to use the same Bayonet Cap with each In-Line Sensor as was used du ring calibration procedure for greater accuracy. In "afterthoughts", the writer adds a comment to the instruction, e.g. (31) Immediately draw blood gas sample and send to lab for analysis. [ ...] A three way stopcock should be added to the infusion circuit between the infusion set and the catheter. In some cases, the possibility of error is created, where the writer does not consider the possibility that the reader will not read the whole sentence to the end, e.g. (32) All traces of cold sterilising fluids must be evaporated from the surfaces of the In-Line Sensor prior to use. (33) To provide an area for the patient name record the alarm limits as set and set the date. Then press the START button ajter the mode selector has been set in the desired printer format. Other deviations are caused by the order of things being seif-evident, e.g. you cannot add paper to the printer until you have opened the door. Here, too, emphasis is given to that operation which is considered to be more important. (34) The chart paper may be replenished ajter opening the side door. 134 Conclusion Operating instructions are normally highly standardized, stylistically frozen texts. They have an obvious pragmatic goal: to instruct the reader about the performance of the operation. Here we have explored in a micro-level analysis how the overt simplicity of such texts is turned into a complex set of interacting parameters. By giving some examples of the discourse properties of individual parts of the message, we have indicated that the procedure applied is a helpful way of looking at the relationship in operating instructions between the encoded contextual information (about the equipment, the. addressee, and the physical context) and the cues for recovering such information available for the reader.

References

Primary sources Cardiomet l000tm. PaO2/PvO2 (Extracorporeal). Operation Manual. 1983. High Wycombe, Bucks, England; Orange Medical Instruments Ltd. Continucath lO00tm. Intravascular Oxygen Monitoring System. Operating Instructions. 1983. High Wycombe, Bucks, England; Orange Medical Instruments Ltd. Datascope Accutorr tm JA and 2A. Non-invasive Blood Pressure Monitors. Operating lnstructions. 1984. Paramus, NJ., USA; Datascope Corp. Datascope System 83. lntra-Aortic Balloon Pump System. Operating lnstructions. 1983. Paramus, NJ., USA; Datascope Corp. Neocath lO00tm. Umbilical Artery Oxygen Monitoring System. Operating lnstructions. High Wycombe, Bucks, England; Orange Medical Instruments Ltd.

Secondary sources

Biber, Douglas, 1988: Variations across Speech and Writing. Cambridge, Cambridge University Press. Enkvist, Nils Erik, 1981: Experiential iconicism in text strategy. Text, 1-1: 97- 111. Nyman, Sirkku, 1991: Understanding Operating Instructions: A Study oj the Language from a Psycholinguistic Point of View. Unpublished Master's Thesis. Department of English, University of Turku. Sager, Juan, Dungworth, David & McDonald, Peter F., 1980: English Special Languages. Principles and Practice in Science and Technology. Wiesbaden, Brandstetter Verlag. 135 Shuy, Roger W., 1990: Warning labels: Ianguage, law, and comprehensibility. American Speech, 65/4: 291-303. Tarone, Elaine et al., 1981: On the use of the passive in two astrophysics journal papers. The ESP Journal, 1: 123-140. Trimble, Mary Todd & Trimble, Louis, 1978: The development of EFL materials for occupational English: the technical manual. In Mary Todd Trimble, Louis Trimble & Karl Drobnik (eds.), Englishfor Specific Purposes: Science and Technology. Corvallis, Oregon State University, pp 74-132.

Writing in Organisations: The Technical Writer as an Example

Karin Mårdsjö

What are the reasons for the similarities and differences between cornrnunicative genres within a given society at a given tirne, and what are the reasons for the sirnilarities and differences of cornrnunicative genres in their historical transformations and in cross­ cultural cornparison? The quotation presents a central issue of Thomas Luckmann's article Communicative genres as an elementary form of the social transmission of knowledge (Luckmann 1986: 11). In the article he discusses how the shaping of texts, in content and form, often aim at organising, finding routines for and maybe completely standardising the solutions of recurring communicative problems. Luckmann claims that these recurring solutions, forms of texts, are very dependent on their varying social contexts, in history, in cultures, in organisations. In this discussion, he introduces the notion of "communicative genre", a middle leve! between linguistic code and social organisation. Those two levels in a communication process have been the objects of research, but not to the same extent the meeting point in between, i.e. the shaping of the text and reasons behind the choices at that leve!. Luckmann "borrows" the term genre here, not in order to discuss very stable linguistic patterns, but in order to be able to describe relatively routinised solutions of communicative problems within a given society-or part of the society-at a certain point of time. Luckmann discusses how these solutions of recurring communicative prob­ lems (texts, speeches, etc.) find their form due to two kinds of structures: inner and outer structures. With inner he means existing genres and texts patterns, both on a macro and a micro leve!. With outer he means social relations, to the audience, to people around, and the interaction with the "environment" in a wider sense.

1 The aim of the article In this paper, the objective at hand is the shaping of texts-and the reasons structuring those processes-in organisational settings. The specific setting in this case is the technical writer communicating the use of new or relatively new 138 technology to broad audiences. 1 This situation is naturally related to matters concerning text production in other organisational contexts. Research concerning "institutional discourse" in a wide sense is of course of relevance in these matters, for example Agar (1985). Issues concerning the institutional impact on communication, particularly in a bureaucratic setting, is summarized in Cedersund and Säljö (1993) in an article on bureacratic discourse, conversational space and the concept of voice. The communicative setting in this case, is, however, of a somewhat different kind. This paper deals with communication in a commercial setting; it is not a case of communicating within the bureaucratic sphere, within the context of official language. The restrictions and ambitions are of another kind. In the paper, I will introduce the specific communicative setting for this kind of discourse and present some main results from a study of technical manuals as means of communication (Mårdsjö 1992a). The study included text analysis and a field study with technical writers.

2 Communicative setting The communicative situation in the study I discuss here is the linguistic introduction of technology, more specifically technical manuals. The text sample included in the analysis was manuals for microwave cooking and word processing-technologies that are widely used and really need good and comprehensible manuals. The role of the technical writer as he or she is creating manuals is to work as an intermediator between the technology producer and the technology user. She is, in the words of Erving Goffman (1983:144), the "author" or maybe "animator" to the "principal's" (Goffman's quotation marks) ambitions towards the user, and the technology producer is the principal in this case. We could also discuss the role of the technical writer as the role a translator, "translating" the use of more or less difficult technology to wide audiences. Their work could be described as parallel to the work of those who write about legal matters for the lay audience (Gunnarsson 1982) or create curricula for school text books (Selander 1988), thus transforming the complex values of the society into comprehensible texts. The user and the producer have somewhat different ambitions concerning the manuals and the technology. These ambitions could be summarised in the following way: both want the technology to function, but the producer also finds it important that the manuals are read, whereas the user mainly wants to use the

1 With "relatively new technology" I mean technology that has been used by a small number of enthusiasts, but is now being diffused to larger user groups. In the terms of Everett M. Rogers (1983), the stage of diffusion has gone from the "early adopters" to the "late majority" and the "laggards". 139 technology, not to read. To some extent, the different parties' hopes and ambitions meet, but concerning major aspects they don't. This is somewhat paradoxical, as both producer and user want the things, the technical devices or technical systems, to work. But the user wants to start using the new equipment right away, whereas the producer wants the users to optimize their use, i.e. to fully develop and utilize the possibilities of the hair dryers, micro wave ovens, lawn mmvers etc. Manuals are written with the intention that the full possibilities ofthe use should be made clear to the users-and those may not even cast a glance at it, which will quite possible lead to disappointment on the part of the users. This is what Carroll & Rosson (1987) describe as the "paradox of the active user"; the users run the risk of making mistakes through the process of "trial and error" rather than making the effort to read. Earlier research on this kinds of communicative settings has, for natura[ reasons, focused on the instructional aspect, and stressed the issues of comprehensibility. This kind of research is certainly motivated as technical manuals are such important tools in a modern society. The research of John Carroll and his colleagues at the IBM Thomas J Watson Research Center is an important part of this work. They have studied the way people learn to use different kinds of computer programmes, how people learn actively and creatively and how these efforts can be supported by better manuals. This research has lead to this group's advocating what they call a "Minimalistic Manual"; a manual that is brief and to the point. The results of this work is summarized and discussed in Carroll (1990), and developed and commented by van der Meij (1992). My research has not been that normative. I have used an analytical and descriptive approach in order to study the aims of manuals, as tools and as commercial instruments for the producers. In this work, I have widened the analysis to regard other aspects than the purely instructional. Manuals are tools for the users; they provide instructions and descriptions (background information) for the correct use of different devices-from toasters to nuclear power plants. Technical manuals work as tools for the technology producers, as well. They are not only aimed at creating instruction and description; they also aim at creating motivation for use. Wrongly or little used technology is poor advertisement for the products.

3 The field study In a study including text analysis and a field study with interviews, I have analysed technical manuals a a communicative genre, or maybe several communicative genres (Mårdsjö 1992a). The main objective of the study was­ as mentioned above- to find answers to the following questions: why technical writers write the way they do when they are introducing new or relatively new technology and what organisational demands ( outer structures) and personal 140 ambitions or genres (inner structures) compete-or co-operate-in the writing process. The companies included in the study were five Swedish companies, two companies (for word processing manuals) and three test kitchens (for micro wave cooking manuals). Fourteen persons were interviewed: the technical writers of my text analysis sample and some of their closest colleagues. Before my visit to the companies, I had analysed their manuals in order to understand the inherent aims towards the users. The manuals were all of a rather high quality; they were printed and illustrated and relatively wide-spread. The interviews were semi-structured. During the interviews, I tried to clarify and get to know their writing situations on three levels: what general ambitions they have with their manuals (those analysed and others of their production), restrictions and driving forces in the technical writers' work and the different organisational models for their manual production.

4 Results

4.1 General ambitions During the interviews, the technical writers described the general ambitions of their work in terms that very much resembled the results from my text analysis. They meant-to a varying extent-that the aims of the texts were threefold. The first, and most obvious, was to give instructions, such as recipes for micro wave cooking and "lessons" for word processing.2 But this was not all. As I had suspected, there were other ambitions, as well. The second aim was to describe the technology by giving an overview of the possibilities and by explaining the functionality of the technology. The latter was more usual in the micro wave case than in the word processing case. The word processing users were obviously regarded as more acquainted with the technology than the micro wave cookers. The writers also used verbal means to describe the function of the technology, by making analogies between daylight and microwaves, between radio waves and micro waves. The third, and on the surface least conspicuous aim mentioned, was to create motivation for use. The following figure illustrates how this is done in a subtle way:

2 The chapters were called lessons and they were structured according to traditional pedagogic models: easier tasks first and more complex later on, recurring tasks in the beginning and less recurrent towards the end, etc. 141

--- ~ Z\l\!\!\N\JWWN Figure 1. "Function ofMicro Waves" (From: Laga mat i mikrovågsugn [1989:8].)

The picture is descriptive on a surface leve), hut the inherent aim is of a motivational kind. The readers are supposed to be less frightened of the technology in question because of this picture. The methods for creating this is the use of analogies. The person responsible for putting this picture in a manual described that: the main purpose of the picture was to create a feeling of safety: "The oven is put between safe and well-known devices." This was, however, described as a way of introducing the technology which was Iinked toa specific stage of the technology diffusion. This book was written in the late nineteen eighties, and the person in charge of the editing told me that it was not going to be included in the next version. The analogies mentioned above can also be interpreted as having motivational ambitions, as well as the instructive parts. The lessons for word processing were specifically described as being designed for motivational purposes; the interviewed writers emphasised that it is imperative to get the users quickly started-or they will loose patience and interest in the technology.

4.2 Restrictions and driving forces Let us now return to the question of why the technical writers write the way they do. The answers mentioned in the previous section-concerning the aims of the manuals-were relatively homogenous and very much in accordance with my text analysis. The answers concerning resirictions and driving forces were, however, more varying and also added other explanatory factors, of which some were related to the "outer" structure, to the organisational demands. This heterogeneity in the result is of course related to the fäet that the general ambitions for the manuals are relatively easily interpreted from the design and content of the texts, whereas the factors related to the communicative situation are not that easily extracted from the texts. They are more hidden in the process and surfaced through the interviews. 142 What I in fäet had expected as answers to this interpretative question were matters related to three different aspects: The dependence of genres, the image of and relation to the users (these matters are discussed more thoroughly in Mårdsjö 1992b) and the relation to the technology. My expectations concerning these aspects were confirmed, and those factors were mentioned as (more or less) important in the writing process. Thus, the first one is what Luckmann calls an "inner" structure, and the second and third "outer" structures. The dependence of stable genres varied between the technologies. The recipe-parts of the micro wave oven manuals are very much designed like ordinary recipes, with a distinctive graphical form. In the word processing manuals there are, however-apart from the "lesson" model-no such stan­ dardised ways. The persons interviewed in that field claimed that the non­ standardisation is due to competitive factors. The companies want the manuals to differ for competitive reasons, rather than to develop a standard format, that might be more convenient for the readers.3 The issue of the users was, however, not very controversial. The technical writers in the two fields gave expression to the view that they write for an average user with no or low previous knowledge of the technology in question. They do not consciously write for "local experts", hut for the "end users".4 Their opinions concerning feed-back were also rather similar. Their users phone when they want to complain or ask for advice. They also ask a lot of questions at computer fairs or cooking shows in department stores. Their questions and complaints had, in some cases, given cause to revisions and additions of some of the texts in the manuals in later versions. The writers,

3 This is well in accordance with international experience. The standardisation document IEEE Standard for Software User Documentation (1988) has not been as widely spread as the committee responsible had hoped; the !arge companies use their own standards. Information about the standard and the diffusion comes from Richard A. Ries, who was involved in that project. 4 Allwood, C.-M. (1991): Människa-datorinteraktion. Lund: Studentlitteratur. In this book, Allwood discusses the roles of "local experts" as intermediators and informal teachers. 143 however, in both cases, describe the role towards the technology as instrumental and wanted to describe the technology as a tool. This is, however, not the image presented in the books. The explanatory value of those three factors were, however, somewhat diminished by two other factors that were strongly emphasised: cultural cancern, "Swedishness" and a personal interest in the technology (technology function and technology development). Still another factor, legal restrictions, was merely mentioned. Every one of the interviewed persons, without exceptions, mentioned a deep national-cultural cancern as a major driving force. Let's call this "Swedish­ ness". It consisted of several elements, of which the linguistic was naturally very central. They did not want to translate texts but to write in Swedish for a Swedish audience. Most of them expressed strong criticism against poor translations and joint European efforts (for example micro wave recipe books). This concept also included a preference for an objective tone-Swedes are regarded as "allergic" to an enthusiastic "ad tone", like in "Congratulations to your choice of product." But there are also other elements in this strategy, which have to do not only with the language but with the relation to the world itself. In the micro wave manuals, they rather showed traditional Swedish dishes than foreign, and in the word processing manuals they wanted to use examples that were familiar to a Swedish audience. Moreover, some of the writers interviewed expressed a feeling of responsibiblity for the standard of the general education, which is a typical Swedish feature. 5 "We have to fil! in where the school has failed" was an expression of that, and in the manuals this ambition was manifested in the descriptive parts. The technology interest was also very strongly emphasized by the writers, in both fields. They mentioned a strong personal interest as a reason for choosing the profession, and they did not want to limit their work to "just writing". A very thorough knowledge of the technologies in question and opportunities for taking part in the development processes were mentioned as major professional ambitions. The writers gave me several examples of both. With a couple of exceptions, the technology interest was described with far more enthusiasm than personal stylistic ambitions. Legal aspects were also mentioned, but these were not considered to be of highest importance in the writing process. Same marketing laws and food regulations, etc. are part of the legal frame-work, but do not really influence the daily writing processes. The future co-operation with the EC was mentioned as a potential restrictive factor.

5 This has been analysed and discussed by Swedish ethnologists like Åke Daun. 144 5 Organisational models It is, naturally, hard to draw distinct lines between "personal" and "organisation­ al" factors in this communicative matter. In a broader sense, the organisational structures will of course dominate, as this writing is purely professional and not in any deeper sense personal; the ambitions of the technology producing companies will be the strongest determining factor. The manual is not the place for free discussions of the advantages or disadvantages of the technology, instead it isa vehicle for supporting the introduction of the technology. The impact of the organisation in a wider sense is difficult to trace directly in the texts. During the interviews, however, I identified some basic structures or "models" for the production of manuals: the negotiation mode!, the evolutionary mode! and the industrial-standardised mode!. The negotiation mode! was used when the technical writer in question had a strong counterpart in the writing process. In one case a publishing company was financially responsible for the project and a number recurring conflicts had arisen, for example how photographs should be used in the manual. The technical writer advocated the use of photographs, which is an expensive form of illustration, and the publishing company wanted to use a cheaper alternative for financial reasons. In the end there were, however, quite a few photographs, which the technical writer described as a success for the quality ambitions of the manual. In the negotiation mode!, such quality matters will be dealt with in an open way. The evolutionary mode! is relatively common in the production of manuals. New versions are added, new people are added in the revision process. It is not easy to identify "authors" to the separate parts of the manuals. This lack of a strong hand holding the project together might result in manuals that are not edited sufficiently. Unnecessary repetitions may occur, as well as lack of vital information. There may be positive effects, as well, as the texts will develop slowly, and new versions will empirically and gradually be tried out. If the feed­ back from the users is competently used in the revision process, the result might be fruitful for the users in the end. The industrial-standardised mode! was used in one of the software companies, making manuals both for word processing and other computer software applications. The technical writer I interviewed had-a few years previous to the interview-taken part in developing the standardised template. The standardisation included text structure, language use and the use of pictures. The editor of the manuals controlled that the different writers followed the prescribed template through close reading and discussions with the authors, not by formulating guide-lines in print. In this company, they had a feeling that their form was being used also by other software companies. In this way, an individual attempt at standardising might affect a genre. 145 6 Conclusions This paper has dealt with factors that determine the writing process in a specific professional setting. Is it the inner or outer structures that are the most important when these writers solve their specific communicative problems? The image that the writers presented during the interviews was that the outer structure, the relations to other actors, to the technology and to the values in society, was the more important of the two. There is, of course, a limited stylis­ tic choice for creating an instructive or a descriptive text, but the driving forces that were most explicitly expressed were those concerning the "environment" rather than those concerning existing genres. The writers' ambitions were to instruct, describe and create a motivational frame-work for the reading of manuals and for the use of technology. Other important objectives were to give a Swedish text toa Swedish audience and to develop a professionalism based on a thorough knowledge of the technology in question. Different kinds of restrictions were not mentioned to any greater extent. This is, of course, somehow dependent on the interview situation; no one wants to appear as a puppet. But this, I think, is not the whole explanation. Can the result presented above be described as representative and possible to generalize? During this field study, I began to regard these technical writers as a kind of elite, who expressed their independence in several ways: The personal ambi­ tions, "Swedishness" and interest in technologies, were driving forces as well as-and sometimes more than-company restrictions. Guidelines were hardly mentioned at all (the "industrial mode!" is an exception). Moreover, they could design their working spaces, and the layout of the manuals-and to some extent the technology itself. If the same kind of analysis was to be made on less ambitious manuals than those included, or at smaller companies with less emphasis on the quality of their manuals, the result might differ from this. One result would probably anyway be similar; the persons interviewed never forget that their ultimate aims lie within the technical companies' aim: To motivate for the use of technology.

References Agar, Michael, 1985: Institutional discourse. Text 5, 147-168. Allwood, Carl Martin, 1991: Människa-datorinteraktion. Lund, Studentlitte­ ratur. Carroll, John M., 1990: The Nurnberg Funnet. Cambridge, MA, The MIT Press. Carroll, John M. & Rosson, Mary Beth, 1987: Paradox of the active user. In John M. Carroll (ed.), Interfacing Thought. Cambridge, MA, The MIT Press, pp 80-111. 146 Cedersund, Elisabet & Säljö, Roger, 1993: Bureaucratic discourse, conver­ sational space, and the concept of voice. Semiotica 97-1/2, 79-101. Goffman, Erving, 1983: Forms oj Talk. Philadelphia, PA, University of Penn­ sylvania Press. Gunnarsson, Britt-Louise, 1982: Lagtexters begriplighet. Lund, LiberFörlag. IEEE Standard for Software User Documentation, 1988: New York, American ' National Standard & Institute of Electronic Engineers. ANSI/IEEE std 1063- 1987. Laga mat i mikrovågsugn, 1989: !CA-förlaget & Electroluxkoncernen. Västerås, ICA bokförlag. Luckmann, Thomas, 1986: Communicative genres as an elementary form of the social transmission of knowledge. In Bengt-Göran Martinsson (ed.), On Communication 4. (SIC 13.) Department of Communication Studies, Uni­ versity of Linköping, pp 1-17. Meij, Hans van der, 1992: A critical assessment of the minimalist approach to documentation. SIGDOC '92, 7-17. Mårdsjö, Karin, 1992a: Människa, text, teknik. Tekniska handböcker som kommunikationsmedel. (Linköping Studies in Arts and Science, No 82.) University of Linköping. Mårdsjö, Karin, 1992b: Technical writers' image of their audience; word processing and microwave oven manuals as an example. In Annette Grindstedt & Johannes Wagner (eds.), Communicationfor Specific Purposes. Ttibingen, Gunter Narr Verlag, pp 38--51. Rogers, Everett M., 1983: Diffusion of Innovations. New York, Free Press. Selander, Staffan, 1988: I..ii.robokskunskap. Lund, Studentlitteratur. TALK AND INTERACTION IN INSTITUTIONS

N egotiating Genre and Power: Questions in Medical Discourse 1

Nancy Ainsworth-Vaughn

In the United States it is widely believed that patients ask few questions and doctors ask many, in doctor-patient encounters: "most often, patients, particular­ ly clinic patients, are passive and powerless in interactions with providers and rarely ask questions ..." (Roter 1977:283); "in medical interviews, questioning is the principal device which physicians and other health professionals use to com­ municate with patients" (Dillon 1990a:54). These studies were based primarily on data from certain kinds of medical settings, such as free clinics; Roter, Hall & Katz (1988) found that 75% of the 61 studies of medical discourse they reviewed were based upon clinic data. But patients in clinics lack the cultural status which is associated with the reciprocity of full fees for services; these patients are relatively powerless. The treatment setting is not the only discourse feature on which we lack data. There have been no previous studies of medical discourse which balanced the gender dyads studied, although gender has been identified as a critical factor in medical encounters as well as in other settings. Diagnosis may make a differ­ ence, as well; but we have few data on this influence. In short, the present literature leaves us without a sense of the range of possible variation among medical encounters. As well as this distortion of our theories about the medical encounter, there are practical issues of great import; questioning, or the lack of it, may affect patients' physical health. Blood sugar in diabetics is lower, hypertension is lower, and other outcomes are better when patients are encouraged to question their doctors (Greenfield, Kaplan & Ware 1985; Kaplan, Greenfield & Ware 1989; Brody et al. 1989). There also may be an effect of questioning on compli­ ance, since a major study found that "there was increased compliance by parents [in a pediatric clinic] when doctors gave more in_formation than when doctors asked a greater proportion of questions" (Freemon et al. 1971). It is clear that there are both theoretical and applied reasons for the study of questions in medicai discourse. The aim of the present study is to examine

1 I wish to thank the patients and medical staff studied here, for their trust, generosity, and efforts in helping gather these audiotapes; the Michigan State University Foundation for an All-University Research Initiation Grant; and Mary Ann Crawford and Anne Hope, for their painstaking work in helping develop and apply analytical categories. 150 questions asked by patients and doctors in a private-practice setting, with attention to the influence of gender and diagnosis. Simple quantitative methods are used. These methods are set within a qualitative framework, so that definitions are based on discourse sequence and conversational inference as well as on linguistic form.

1 Previous research Dillon (1982, 1986, 1990b) reviewed the study of questions in more than a dozen different fields. One point upon which all researchers agreed is that questions are a very powerful discourse move. A question exerts power by requiring a response, setting the topic of the response, and choosing the next speaker, the questionee. It may elicit information, and information is power. Further, depending upon the context and content, a question may entail the expectation that the floor will be returned to the questioner. All of these actions exert significant control over the progress of the interaction and over the other participant. Questions have been seen as realizing social power in political (Bennett 1982), educational (Mehan 1979), and legal (Philips 1987, Walker 1987) settings. Like other discourse phenomena, questions are not unidimensional. They have multiple functions, and among those functions are some which serve to share power; the most noticeable of these is the function of involving another in participating in discourse (Goody 1978, Fishman 1983). However, the fäet that in American culture questions realize more power than they give away is suggested by studies of questioning in asymmetrical dyads (parent-child, teacher-student, doctor-patient). It is always the speaker with higher status who asks the most questions (Dillon 1982, 1986), and the imbalance in number is usually dramatic. Roter (1977), West (1984), Silverman (1987), Franke! (1990) and Biesecker & Biesecker (1990) have focused exclusively on questioning (with various restrictions) in the medical encounter. Methodologies, and consequently defini­ tions, differed considerably. In a large-scale study, Roter (1977) studied the effect on "direct questions" of an intervention in which a patient educator helped patients articulate their questions just before seeing the doctor. However, ber definitions were idiosyn­ cratic: she defined "direct" questions as those "raised clearly by patient initiative" (299), and excluded as "indirect" patients' questions which asked for clarification or amplification of a preceding comment by a doctor. These defini­ tions are not drawn from contemporary sociolinguistic and pragmatic analyses of questions. Another narrowly focused study is Frankel's (1990) research on 10 ambulatory care visits between adult patients and long-term practitioners in general interna! medicine. Fewer than 1 percent of the questions asked by 151 doctors and patients in the encounters studied by Franke! were "patient­ initiated" questions. Although an explicit definition of "patient-initiated" is not given, it appears that the utterance had to introduce new information in order to be "initiated" by one of the participants (pp. 240, 260). The data included "30 [... ] sequences initiated by patients; 17 ofthe 30 involved 'normal' troubles such as requests for clarification, information, etc., and are thus not included in the analysis" (p. 239). Again it is difficult to evaluate the quantitative data, to know what was counted. West (1984) used the adjacency pair definition of questions, excluding only repairs (when one party did not hear the other) and markers of surprise (Oh, really?). My data can be compared with hers because we used operational definitions of questions which were similar in mast ways, in spite of somewhat different theoretical orientations. 2 West studied questions in 21 encounters in a clinic staffed by residents, whose population was primarily drawn from lower socioeconomic strata. Fourteen of the physicians were men and 4 were women, all white; there were 9 male patients (5 white, 4 Black) and 11 female patients (6 white, 5 Black). West found 773 questions, of which 91 % (705) were asked by physicians. The clinic setting, imbalance in gender dyads, and the ethnic diversity of the patients may have affected West's totals. Silverman (1987), in a large-scale, semi-quantitative study of out-patient encounters in pediatric cardiology, gathered longitudinal data which provide examples of contextual influences on questioning. For instance, mean number of parents' questions was 1.14 per encounter for children with the least severe problems, 4.28 for those with moderate severity, and 2.9 for those parents whose children had the mast serious problems. Silverman pointed out the difficulties of categorizing utterances as questions and gave two very brief examples, but provided no theoretical motivation for his choices. Biesecker & Biesecker's (1990) findings supported Silverman as to the importance of situational variables. Biesecker & Biesecker studied correlates with the "information-seeking comments" of 106 patients (42 men and 64 women) in an out-patient clinic. Their work raises the issue of socioeconomic status; in contrast to Waitzkin (1985), they found a lack of correlation between SES and the number of patients' questions: "situational factors surrounding the doctor-patient interaction better explain patient information-seeking behaviors than do patient demographic variables, patient attitudes, or even the physician seen, especially for patients with interactions lasting at least 19 minutes" (1990:27). It is not clear whether Biesecker & Biesecker excluded requests for confirmation, action, and permission, which overlap with requests for information in many instances.

2 West generously contributed the time necessary to discuss her definitions and to analyze one of my transcripts, in order to maximize analytical comparability. 152 These studies all converge in agreeing that the balance of power in medical encounters is locally negotiated, and that questions are a crucial part of that negotiation.

2 Methodology and data This analysis is informed by concepts from interactional sociolinguistics and anthropological discourse analysis. I understand language users to construct texts in accord with their understanding of participants' cultural and social rights and obligations and their understanding of the event as an exemplar of a genre (Hymes 1972, Gumperz 1982, ten Have 1989). Situated discourse is then a major arena for the moment-to-moment realization of an individual 's social identity as well as having purposes specific to its context (e.g., medical, pedagogical). The 40 encounters studied here (Table 1) were chosen from a total of 88 audiotapes collected in two studies, to balance numbers of men and women, as both physicians and patients. There were 20 encounters with male patients, 20 with female patients; and 20 encounters with male physicians, 20 with female physicians. Eleven patients were male, 12 were female; 4 physicians were male, 4 female. All 31 participants were white. The setting was oncology in 28 of the 40 encounters. In an initial small-scale study, four women (LA, JB, EAA, KE) tape recorded their visits to seven physicians (including B, C, E, G, and Y studied here). Patients took tape recorders into their medical private-practice encounters over a period of months or years. Six tapes studied here (LADl, LAFl, JBO, EAAl, EAA2, and KE6) are from this study. Other encounters analyzed here (28 oncology, 6 hematology) were recorded in two oncology-hematology practices involving three physicians (R, Q, and W). A radio microphone was worn by the physician. Standard consent practices were followed. Ten of the eleven physicians approached agreed to participate. Approximately three-fourths of the patients approached agreed to participate; of those who refused, a disproportionate number were men. Only about half of the men who were approached agreed to participate. There were three stages of analysis. First, I went through the data testing possible criteria for the linguistic structure, referential content, and discourse function of possible questions. A research assistant and I then used the three types of criteria in independent analyses of the audiotapes/transcripts. The third step was a review of the analyzed transcripts with a second research assistant. Regardless of linguistic structure, at least two of the three analysts had to agree that an utterance met certain discourse criteria in order for it to be counted. At all stages of analysis, analysts listened to all 40 audiotapes while reading and marking written transcripts, replaying important segments repeatedly. 153 Table 1. Encounters

TaEe code Date Pt !lender Ph !lender Phisician Len!lth EAl 4/04/91 F F R 1,877 EA2 7/10/91 F F R 1,066 EAAl 9/28/89 F F G 2,095 EAA2 11/21/89 F F G 1,217 FBI* 4/17/91 F M w 1,760 FB2 5/02/91 F M w 2,906 FGl 10/16/91 F M Q 1,216 Fil 11/27/91 F F R 3,034 Fl2 1/23/92 F F R 4,236 HKI* 4/04/91 F F R 6,376 HK2 5/02/91 F F R 2,648 IJI* 4/17/91 F M Q 4,322 JBO 2/06/89 F F y 1,916 KE6 n.d. F M B 2,084 KJI 4/18/91 F F R 1,143 KJ2 5/22/91 F F R 1,556 LADl 10/30/89 F F C 4,130 LAFI 4/18/90 F M E 3,403 MSl 3/27/91 F M Q 3,163 MS2 5/22/91 F M Q 2,391 AG!* 4/17/91 M M w 839 AG2 5/08/91 M M w 1,864 BEl 5/22/91 M F R 819 BE2 7/17/91 M F R 1,090 BRl 3/27/91 M M Q 2,930 CRl 4/18/91 M F R 1,237 CR2 5/16/91 M F R 1,335 DWl 4/17/91 M M Q 1,319 DW2 5/01/91 M M Q 1,395 FRI 4/04/91 M F R 1,760 FR2 5/02/91 M F R 2,198 HEKl 5/08/91 M M w 723 HEK2 11/06/91 M M w 1,129 JRl n.d. M M Q 1,546 JR2 n.d. M M Q 2,239 LCl* 11/27/91 M F R 5,063 LC3 3/11/92 M F R 929 MPl* 3/27/91 M M Q 4,430 MP2 4/17/91 M M Q 2,031 PRl 3/27/91 M M Q 2,522 Total 89,937 *Initial visits

3 Defining questions West (1984) defines questions with the concept of adjacency pair in conver- sation analysis: the first member of an adjacency pair requires the production of the second. A question, in this view, is an utterance which requires an answer; if 154 an answer is not forthcoming, there is warrant to reinstate the question. This definition relies upon native speaker intuition as to the nature of questions and answers. Much of the preparation for counting questions in my data consisted of specifying the discourse criteria upon which intuition operates. Linguistic markers were seen as part of the information that speakers assess using discourse criteria. I was aided by West's (1984) analysis and Stenström's (1984) list of lexical, grammatical, and contextual criteria for utterances functioning as questions. Because ambiguity and polysemy are intrinsic parts of discourse processes (J. Brody 1991, Tannen 1993), there were some utterances which might have been interpreted as questions by one or the other of the participants hut which

Table 2. Linguistic and discourse indicators of possible question function

Linguistic markings: -WH When will the test results be back? [WH includes what, why, how, etc.] -SWH The test results will be back when? ["search" WH-question] -Yes/No Are the test results back? [inverted auxiliary] -TG The test results are back, aren't they? -QF The test results are back, right? [elliptical question; appended queryform] -PH The test results are back? [various intonation patterns, including rising intonation] Discourse/contextual criteria: -D The test results are back. (pause) [New topic, and addressee can be expected to have the information.] -DD I wonder whether the test results are back, [Addressee can be expected to have the information.] -QQ I have a question. It seems important to know the test results. [Addressee can be expected to have the information.] 155 Table 3. Patients' questions categorized by Iinguistic/discourse structure Taeecode NIO D DD PH* QF* QQ SWH* TG* WH* YN* AGl N AG2 N 1 2 EAAl N 1 1 3 EAA2 N 2 1 2 FBl N 1 1 FB2 N 1 1 1 4 HEKl N 2 HEK2 N 1 2 JBO N 1 2 3 KF.6 N 3 2 1 1 6 l.ADl N 2 2 1 6 l.AFl N I 3 5 BEl 0 1 1 2 BE2 0 1 1 BRl 0 1 2 2 CRl 0 2 2 CR2 0 2 DWl 0 3 1 3 DW2 0 PAl 0 2 6 PA2 0 3 FGl 0 1 3 Fil 0 2 1 2 2 Fl2 0 1 2 2 FRI 0 2 8 5 FR2 0 2 1 1 8 13 HKl 0 3 2 10 1 12 32 HK2 0 1 3 4 IJl 0 2 1 1 1 8 JRl 0 1 JR2 0 1 2 KJl 0 1 2 1 1 KJ2 0 1 3 LCl 0 2 5 10 11 LC3 0 1 4 MPl 0 4 1 2 2 MP2 0 1 2 MSl 0 1 1 3 3 4 MS2 0 3 PRl 0 1 3 17 8 57 7 3 1 9 75 147

* "Linguistically-marked" questions N =non-oncology patient; 0 =oncology patient 156 Table4. Physicians' questions categorized by linguistic/discourse structure TaEe code NIO D DD PH* QF* QQ SWH* TG* WH* YN* AGl N 1 1 AG2 N 1 1 3 EAAl N I 3 2 EAA2 N 1 1 2 9 FBl N 5 1 7 FB2 N 2 6 5 9 HEKl N 3 1 1 3 HEK2 N 2 1 2 JBO N 2 4 8 KF.6 N I 1 1 5 8 LADl N 2 12 1 7 12 LAFI N 9 4 9 12 BEl 0 1 2 BE2 0 1 1 2 BRl 0 10 2 2 7 3 CRl 0 3 1 1 CR2 0 8 1 1 DWl 0 1 1 2 DW2 0 3 2 1 EAl 0 7 2 6 EA2 0 2 1 5 5 FGl 0 4 5 5 3 Fil 0 5 2 7 Fl2 0 2 1 1 FRI 0 2 2 FR2 0 5 1 3 1 HKl 0 1 1 1 3 I HK2 0 2 1 3 4 4 Dl 0 2 1 7 5 JRl 0 1 1 1 JR2 0 1 1 6 KJl 0 1 2 1 3 KJ2 0 1 1 3 LCl 0 1 6 3 9 LC3 0 1 2 3 2 MPl 0 2 8 1 12 2 MP2 0 1 1 MSl 0 3 6 2 5 2 MS2 0 6 1 6 8 12 PRl 0 1 3 2 2 13 11 Total 50 3 128 20 0 7 8 133 165 * "Linguistically-marked" questions N = non-oncology patient; 0 = oncology patient About 10% of the utterances which functioned as questions were syntactically and phonologically marked as declaratives. These categories are potentially ambiguous as to discourse function: D, a declarative with fälling tone, has the most potential for ambiguity. 157 (1) D: I feel it now. That one's so Iittle (???) I'm not really sure that it's. Let me feel over here because this was a bad site lo start with. You didn 't find anything underarm. P: Nope I didn't find anything underarm. (BE2,3-9) This example is an A statement about a B event (cf. Stenström 1984:153), in the context of A's lack of certain knowledge and A's need to know. It translates to, "I am guessing that this is the case for you; confirm or disconfirm." These were differentiated from summaries. In deciding whether (1) is a question or a summary, we looked at tre context to see whether there was any previous mention of the topic. In this case, lumps under the arm had not previously been discussed; therefore, we considered this declarative sentence to function as a question, rather than as a summary/formulation used to close down a topic. There were other discourse criteria which identified questions, notably presence of a previous list of questions and pauses by A designed to allow B complete a sentence. There were many redundant cues from discourse. The following example provides a redundant combination of three discourse criteria: an A statement about a B event, presence of just previous questions, and A's pause to allow B to complete the sentence. A phonological cue (non-falling intonation) was present as well. (2) D: Have you lost any weight? P: Well I did, but I think I gained it back. I !ost about ten pounds. D: Your appetite's been (2 sec) P: Fine, I haven't !ost it or anything. (IJl, 6--1) Discourse redundancy such as this is parallel to morphological redundancy; both allow hearers greater opportunity to interpret speakers correctly.

4 Findings and discussion

4.1 Variation in genre Patients asked a high percentage of the questions in these encounters. Of the 838 questions doctors and patients asked one another, 61.3% (514) were asked by physicians and 38.7% (324) were asked by patients. This is in striking contrast with West's (1984) findings that patients asked only 9% of the questions. My method of counting differed from West's in only one minor way, and if we eliminate that difference we find that patients' percentage of questions actually goes up slightly, to 39 .1% of the total. These data establish the existence of a continuum between two genres for medical encounters, the interview and purposeful conversation. Medical encounters do not universally belong to the interview genre. The place of a specific encounter on this continuum will be negotiated by physician and patient, through discourse features such as first-naming, interruptions, floor- 158 holding, and questioning. Speakers have bases on which they ascribe to them­ selves and others the right to first-name, interrupt, hold the floor, and question. I have already described the possible influence of the medical setting, i.e., clinic vs. private practice, and patients' status in relation to whether they must use free or low-cost clinics.

4.2 Ethnicity Ethnicity may be a factor, as well. General ethnographies of African-American communication have had little to say about questions (Labov 1972, Kochman 1981, Smitherman 1986), but the differences they have found suggest that cultural differences in questioning on the part of patients cannot be ruled out. Physicians, too, may have changed behavior according to the race of the patient; Roter, Hall & Katz, in a summary of results of 38 quantitative studies of medical interaction in which ethnicity was reported, found that "non-whites received less information and less positive talk than whites" (1988: 112). West found several patterns in her data suggestive of the influence of ethnicity (e.g., more Black patients than white were first-named by the physician). Approxi­ mately half of West's (1984) patients, and an apparent majority of Roter, Hall & Katz's (1988) patients, were Black. If ethnic identity is one of the bases for negotiation, the fäet that all patients in the present study were white may be related to the high number of their questions. Two other bases for negotiation which may have influenced these data are gender and diagnosis.

4.3 Gender Gender differences noted in this study cannot serve as the basis for generalizations about physicians because of the small number of participants, especially in the female physician-male patient grouping, where we had only one physician. However, the data on patients are much more reliable. The data we do have on gender of physician clearly indicate the importance of further investigation of this factor in medical discourse. Female patients asked more questions (9.65/visit) than did male patients (6.55/visit); but this is qualified by the differences among gender dyads described below. From these counts it would appear that the genders of both the physician and the patient have significant effects upon question-asking. When the physician was a woman, male patients asked 10.9 questions per encounter; female patients asked 10.8 questions per encounter (see Table 5 for numbers of participants and encounters in these dyads). When the physician was a man and the patient a woman, the patient asked 8 questions per visit. Finally, the man-man dyad produced only 3.7 patient questions per visit. These are striking findings. 159 The percentages of physicians' questions in these same encounters are equally interesting. Overall, when the physician was a woman, 49.9% (216) of the 433 questions were asked by the physician, with different proportions for male and female patients (Table 5). In encounters with male physicians, 73.3% (297) of the 405 questions were asked by the physician.

Table 5. Percentage of questions by gender dyads PATIENTS PHYSICIANS Female Male 4Phy. 7Pat. 1 Phy. 4Pat. 12 Encounters 8 Encounters Female 53.6 46.4 43.5 56.5 4Phy. 6Pat. 2Phy. 7 Pat. 8 Encounters 12 Encounters Male 72.8 27.2 74.3 25.7 Data on gender of physicians in medical discourse are scarce. Gender of physician was reported in only one-fifth of the 61 quantitative studies of medical encounters reviewed by Roter, Hall & Katz (1988). In studies of medical discourse analysts have had access to very restricted types of gender dyads, due to the general difficulty of securing access to medical settings and to the larger number of men in the role of physician. Davis's (1988) attempt to discuss gender issues is handicapped by data which involve only female patient-male physician dyads. Pizzini (1991) found that Italian female physicians were more egalitarian than men in their use of directives (if a smaller number of directives from the physician implies a more equal balance of power), and they were less Iikely to use humor to stop patients from talking; but specifics about the number of interactions in each type of dyad are not given. Ainsworth-Vaughn (1992) found gender differences in physi­ cians' preferred ways of changing the topic among participants matched for gender and for severity of medical complaint. The contrast lay in the ratio between topic changing strategies which share power and those which allocate power to the speaker. For physicians who were women, this ratio was 5 power­ sharing strategies to 1 power-holding strategy. For male physicians, it was 1.4 power-sharing strategies to 1 power-holding strategy. West (1984) studied interruptions and questions in 21 medical encountcrs. There were 17 encounters involving 14 male white physicians, but only four encounters involving four female white physicians. Patients seen by the four female physicians included two Black women, one Black man, and one white man. 160 The percentages of interruptions in West's data, in general and over gender dyads, showed trends with general similarities to those of the percentages of questions in my data. For instance, West found physicians initiating 58.7% of the interruptions; compare physicians' 61.3% of the questions, in my data. However, the small number of female-female pairs, and varying ethnicity, in West's study do not provide a basis for reliable quantitative measures. In a qualitative examination of these encounters, West (1990) concluded that female physicians were more egalitarian in their issuance of directives than men. Aside from the research on gender and discourse features, patients' satis­ faction has been studied. Gender of physician was identified as a salient factor in patients' satisfaction with medical encounters (Linn, Cope & Leake 1984):

The most dramatic and consistent finding in the study was that both male and female patients evaluated care from female residents more favorably than care from male residents, especially on items pertaining to the art-of-care. (p. 966.) Arnold, Martin & Parker (1988) reported no differences between men and women physicians in therapeutic actions taken; the authors suggest that ways of communicating account for patients' greater satisfaction with female physicians. One possible source of the relatively !arge proportion of questions asked by patients in the present study is the fäet that here number of encounters with male and female physicians was equal. This suggestion must be heavily qualified by the possibility of individual differences in the small number of physicians stud­ ied. However, when these data are considered together with the various findings of gender differences in discourse in other settings and in medical settings, it seems unlikely that the gender differences found here are idiosyncratic.

4.4 Diagnosis, initial visit Table 6 lists six initial visits. In the two hematology initial visits, the patient was referred with no diagnosis; the purpose was to develop a diagnosis. AG had a tentative diagnosis of excessive cigarette smoking, which was affecting his red blood cell count. With this relatively non-threatening diagnosis, he asked no questions. FB received no diagnosis until a later visit; in her initial visit, at which the doctor recommended further tests, she asked only 2 questions. The other four initial visits involved patients who had been referred with an established diagnosis of cancer; so the purpose of the visit was to discuss what to do about the disease. These patients asked 14, 9, 28, and an astonishing 60 questions, an average of 27 questions per encounter. 161 Table 6. Number of patients' questions by length of transcript Tape code Length Questions HEKI 723 2 BEI 819 5 AGI* 839 0 LC3 929 5 EA2 1066 4 BE2 1090 2 HEK2 1129 5 KJl 1143 5 FGI 1216 4 EAA2 1217 5 CRI 1237 5 DWI 1319 7 CR2 1335 2 DW2 1395 0 JRl 1546 I KJ2 1556 4 FBl* 1760 2 FRI 1760 15 AG2 1864 3 EA! 1877 8 JBO 1916 6 MP2 2031 3 KE6 2084 13 EAAI 2095 6 FR2 2198 25 JR2 2239 3 MS2 2391 3 PRl 2522 4 HK2 2648 8 FB2 2906 7 BRI 2930 7 Al 3034 7 MSI 3163 12 LAFI 3403 9 LADl 4130 11 A2 4236 5 IJl* 4322 14 MPI* 4430 9 LCl* 5063 28 HKl* 6376 60 Total 89937 324

* Initial visit Overall, oncology patients asked an average of 11.57 questions per visit. Within oncology, 17 encounters were repeat visits by established patients with no pressing medical issues. These patients asked an average of 6.5 questions/ encounter, not so far removed from the non-oncology average of 5.75 questions/ encounter. 162 These figures suggest, once again, that medical discourse is fine-tuned to its context. Diagnosis, current condition, and the place of a visit in the physician­ patient interactional history all may make a difference.

4.5 Length Silverman (1987) and Biesecker & Biesecker (1990) found variables such as length and purpose of visit, diagnosis, and initial vs. repeat visits to be related to the amount of information patients sought. In my data these factors were confounded; initial visits with a diagnosis of cancer were also the longest visits (Table 6). If patients have a longer time with the physician, they have more opportunity to ask questions. But opportunity alone is not enough; in one study clinic patients had visits twice as long as private patients, hut they asked fewer questions than private patients (Roter, Hall & Katz 1988). This fäet calls into question the importance of length of visit as an explanation for questioning. Apparently, in order for patients to question, they need not only an opportunity to question, hut also some bases upon which they can negotiate an interactional status which gives them the right to question. Clinic patients lack one of those bases: the fundamental reciprocity of full fees for medical services.

5 Conclusion This paper is about language and power in the medical encounter. Who deter­ mines its topics and speakers, through questions? In the literature, the interview nature of some medical encounters, with patients asking fewer than 10% of the questions, has been overgeneralized to all medical encounters. By contrast, in the present study we find patients asking almost 40% of the questions. These data establish the existence of a continuum of variation in the discourse of medical encounters. The data show that some patients do develop a high degree of control over the discourse of their medical visits. Previous studies have primarily described one end of this continuum, the interview end. The place of a specific encounter on this continuum between interview and conver­ sation must be negotiated between physician and patient. Many patients in this study approached the conversational end of the continuum. The question becomes, what are the resources participants used in negotiating their rights to control over the medical encounter in general, and the right to question in particular? Among the possible answers, I have discussed and presented data on gender, length, diagnosis, current condition, and the place of the interview in a sequence of visits. I assume that the private-practice setting (though not quantifiable) also plays a role. Biesecker & Biesecker (1990) believe that patient demographic influ­ ences, such as socioeconomic class, are less influential than situational 163 variables, such as the medical setting. But situational variables cannot be neatly separated from demographic influences. Since the setting for this discourse was private practice, in which patients are either insured or paying their own way, patients' socioeconomic leve! can be assumed to be higher than in previous studies of clinic settings. In the private-practice setting patients are reciprocal partners in the cultural exchange, by virtue of being paying customers; this is seldom an overt topic of discourse, and yet it may have lessened social distance between physician and patient and thus pushed the balance toward greater freedom on the part of patients to ask questions. These are the first data on medical encounters in which gender dyads are balanced in number. The data also are unusual in including two interactions with most patients. However, the number of physicians who contributed !arge numbers of encounters to these data is small; the study suffers from the difficulty (general in analysis of medical discourse) of gaining access to medical sites. It also suffers from the relatively small number of women who are practicing physicians. By the year 2000, women will makeup one-third of all U.S. physicians (Epps 1991). As the number of female physicians grows, we will learn more about the relationship between gender and discourse. Diagnosis, gender, and initial vs. repeat visit all appeared to make a difference in the numbers of questions patients asked. On the average, oncology patients asked more questions than non-oncology patients; women patients asked more questions than men; all patients asked more questions when the physician was a woman; and initial visits, at least in oncology, contained more questions from patients than did repeat visits. This study has contributed data on the discourse continuum between medical encounters that are primarily interviews and those encounters in which physician and patient engage in purposeful conversation. These data are badly needed to provide a sense of the range of possible variation in the medical encounter. They also may provide a sense of the strengths, rather than the weaknesses, of patients as participants in medical discourse.

References Ainsworth-Vaughn, Nancy, 1992: Topic transitions in physician-patient inter­ views: Power, gender, and discourse change. Language in Society, 21 :409- 426. Arnold, Robert M., Martin, Steven C. & Parker, Ruth M., 1988: Taking care of patients-does it matter whether the physician isa woman? Western Journal oj Medicine, 149:729-733. Bennett, Adrian, 1982: Strategies and counterstrategies in the use of yes-no questions in discourse. In John J. Gumperz (ed.), Language and Social Jdentity. Cambridge, Cambridge University Press, pp 95-107. 164 Biesecker, Analee E. & Biesecker, Thomas D., 1990: Patient information-seek­ ing behaviors when communicating with doctors. Medical Care, 28: 19-28. Brody, D.S. et al., 1989: Patient perception of involvement in medical care: Relationship to illness, attitudes, and outcomes. Journal oj General Internat Medicine, 4:506-511. Brody, Jill, 1991: Indirection in the negotiation of self in everyday Tojolab'al women 's conversation. Journal oj Linguistic Anthropology, l :78-96. Davis, Kathy, 1988: Power Under the Microscope. Dordrecht, Foris Publica­ tions. Dillon, James T., 1982: The multidisciplinary study of questioning. Journal oj Educational Psychology, 74: 147-165. Dillon, James T., 1986: Questioning. In Owen Hargie (ed.), Handbook oj Communication Skilts. New York, New York University Press, pp 95-127. Dillon, James T., 1990a: Clinic questioning: Medicine. In J.T. Dillon, The Practice oj Questioning. New York, Routledge, pp 54-65. Dillon, James T ., 1990b: The Practice oj Questioning. New York, Routledge. Epps, Roselyn Payne, 1991: Commencement Address. East Lansing, Michigan, Michigan State University College of Human Medicine. Fishman, Pamela M., 1983: Interaction: The work women do. In Barrie Thorne, Cheris Kramarae, & Nancy Henley (eds.), Language, Gender, and Society. Cambridge, MA, Newbury House, pp 89-102. Franke!, Richard, 1990: Talking in interviews: A dispreference for patient­ initiated questions in physician-patient encounters. In George Psathas (ed.), Interaction Competence. Lanham MD. University Press of America, pp 231- 262. Freemon, Barbara et al., 1971: Gaps in doctor-patient communication: Doctor­ patient interaction analysis. Pediatric Research, 5:298-311. Goody, Esther N., 1978: Towards a theory of questions. In Esther N. Goody (ed.), Questions and Politeness. New York, Cambridge University Press, pp 17-43. Greenfield, Sheldon, Kaplan, Sherrie & Ware, J.E., 1985: Expanding patients' involvement in care: Effects on patient outcomes. Annats oj Internat Medicine, 102:520-528. Gumperz, John J., 1982: Discourse Strategies. Cambridge, Cambridge Univer­ sity Press. Have, Paul ten, 1989: The consultation as genre. In Brian Torode (ed.), Text and Talk as Social Practice. Dordrecht, Foris Publications, pp 115-135. Hymes, Dell, 1972: Modcls of the interaction of language and social life. In John J. Gumperz & Dell Hymes (eds.), Directions in Sociolinguistics. The Ethnography oj Communication. New York, Holt, Rinehart & Winston, pp 35-71. 165 Kaplan, Sherrie H., Greenfield, Sheldon S., & Ware, J.E., 1989: Assessing effects of physician-patient interaction on outcomes of chronic disease. Medical Care 27:110-127. Kochman, Thomas, 1981: Black and White Styles in Conjlict. Chicago, University of Chicago Press. Labov, William, 1972: Language in the Inner City, Philadelphia, University of Pennsylvania Press. Linn, Lawrence S., Cope, Dennis W. & Leake, Barbara, 1984: The effect of gender and training of residents on satisfaction ratings by patients. Journal oj Medical Education, 59:964-966. Mehan, Hugh, 1979: Learning Lessons. Cambridge, MA, Harvard University Press. Philips, Susan U., 1987: The social organization of questions and answers in courtroom discourse. In Leah Kedar (ed.), Power Through Discourse. Norwood, NJ, Ablex, pp 83-112. Pizzini, Franca, 1991: Communication hierarchies in humor: Gender differences in the obstetrical gynaecological setting. Discourse and Society, 2:477--488. Roter, Debra L., 1977: Patient participation in the patient-provider interaction: The effects of patient question asking on the quality of interaction, satis­ faction, and compliance. Health Education Monographs, 5:281-315. Roter, Debra L., Hall, Judith A., & Katz, Nancy R., 1988: Patient-physician communication: A descriptive summary of the Iiterature. Patient Education and Counseling, 12:99-119. Silverman, David, 1987: Communication and Medical Practice. London, Sage Publications. Smitherman, Geneva, 1986: Talkin and Testifyin. Detroit, Wayne State Univer­ sity Press. Stenström, Anna-Brita, 1984: Questions and Responses in English Conversa­ tion. (Lund Studies in English, 68.) Malmö, Liber Förlag-C W K Gleerup. Tannen, Deborah, 1993: The relativity of linguistic strategies: Rethinking power and solidarity in gender and dominance. In Deborah Tannen (ed.), Gender and Conversational Interaction. New York, Oxford University Press, pp 165- 188. Waitzkin, Howard, 1985: Information giving in medical care. Journal oj Health and Social Behavior, 26:81-101. Walker, Anne Graffam, 1987: Linguistic manipulation, power, and the legal setting. In Leah Kedar (ed.), Power Through Discourse. Norwood, NJ, Ablex, pp 57-82. West, Candace, 1984: Questions and answers between doctors and patients. In Candace West, Routine Complications: Troubles with Talk Between Doctors and Patients. Bloomington, IN, Indiana University Press, pp 71-96. 166 West, Candace, 1990: Not just doctor's orders: Directive-response sequences in patients' visits to women and men physicians. Discourse and Society, 1:85- 112. Two Strategies for Clarification in Doctor­ Patient Encounters

Ulla MelanderMarttala

Interviewer: What do you hope to get out of this visit? Patient: Well, I hope I will just get an answer, about what it is, because that's what I think is really tough, that it hurts but I don't know why.

(Extract from pre-interview.) A very important matter in most doctor-patient conversations is to identify the patient's problem, its nature, and what can be done about it. I The purpose for both doctor and patient is to collect and exchange information, evaluate it, to consider possible treatments and decide on the measures necessary to achieve the goal-to improve the patient's health. According to Agar (1985), the professional part in the medical encounter as an instance of institutional discourse, must accomplish three things: diagnos is, directive, and repor!. In doctor-patient encounters, diagnoses and directives are usually attained through questions, explanations, test and examination results, prescriptions and the like, generally calling for clarification. So one can say that reaching clarity and intersubjective understanding is also an important issue in doctor-patient conversations. The symptoms and problems of the patient, and their nature and origin, must be investigated and clarified, and development, prognosis and treatment have to be explained and clarified. To state that clarification is an aim for both doctors and patients may at times be too idealistic a view of doctor-patient conversations, in that clear communi­ cation need not always be preferred, nor unclear discourse be dissatisfying (Coupland et al. 1991, West & Franke! 1991). But the assumption that doctors and patients do want to clarify matters and reach an agreement (of some sort), though aims and outcomes sometimes may not always be in accordance or motives the same, will neverthelcss be adopted here. A matter of great importance in institutional discourse, and hence also in doctor-patient conversations is asymmetry and differences in perspectives stemming from differing knowledge, frames, social and cultural background,

1 The work reported here is a part of The conceptual worlds oj doctors and laymen, a research project which has been supported by a grant from The Commission for Social Research, Sweden, conducted at the Unit for Advanced Studies in Modern Swedish, Uppsala University. 168 interactional rights, vocabulary, etc. Also important is the personal relationship and treatment. Patients must be met with empathy and emotional reassurance as well as technical knowledge. Clarification can also be seen as belonging here, as well as to areas of more factual information exchanged (cf. Cicourel 1981, Fisher & Todd 1983, Franke! 1984, Mishler 1984, ten Have 1991, West & Franke! 1991, Drew & Heritage 1992; also Drew 1991, Linell & Luckmann 1991). Of special interest here is the use of technical vocabularies and expressions: the physician's use of medical terms and jargon, and the patient's lay expres­ sions. In both of these cases difficulties in understanding and thus also a demand for clarification may arise. Differences in vocabularies and the meaning of terms and concepts and their understanding have earlier often been studied outside of actual interactions, while other studies which have taken these problems in account, above all Meehan 1981 and West 1984, stress the importance of studying these phenomena in actual interactions, in authentic doctor-patient encounters. (See also Cicourel 1981, Shuy 1983, West & Franke! 1991; and, on the importance of analyzing discourse phenomena demonstrably relevant for the participants themselves, Linell & Luckmann 1991, Drew 1991, Drew & Heritage 1992.) Another area of interest related to both clarification and vocab­ ulary in doctor-patient conversations, is the meaning of important diagnostic concepts and distinctions of disease and symptoms that are of cancern for the doctor and the patient, and how they are managed and expressed in conversation, i.e. how they are investigated, described, clarified and understood. The whole conversation between doctor and patient in a medical consultation may be looked upon as a matter of clarification. The purpose is to understand and clarify the patient's problems and how they can be dealt with. In a sense, almost everything said in a doctor-patient encounter is a form of clarification or requests for clarification, at different levels. Clarification is thus found at global levets (global levels being the longer stretches of dialogue, or sequences treated at "higher" levels, e.g. stories, topics, as well as the whole conversation itself); and at local levets (i.e. single exchanges such as adjacency pairs and other contributions relating directly to "immediately preceding" and "anticipated next" contributions (Linell 1990, Linell & Luckmann 1991). As Linell (in press) discusses, understanding and misunderstanding are multi­ layered and multifaceted phenomena. Things can go wrong at different levels of these processes, so problems may involve hearing and attention, iopics or perspectives on topics. It is possible, then, to distinguish, among other things, thc following types of troub!e sources: identity of words; references and referential perspectives; meaning specifications; cognitivc, cmotive or conative attitudes towards things talked about; levels of intentionality; and frames and perspectives (Linell in press; cf. Schegloff 1987, Coupland et al. 1991). Clarification and requests for clarification are, however, not necessarily tied to misunderstanding and miscommunication. They are often part of any 169 communication, "good" or "bad", and can also be an essential part and the virtual goal of a conversation. This means that a conversation containing repairs and requests for clarification need not be a problematic one, or a conversation lacking such components unproblematic.2 Any conversation demonstrates all kinds of ambiguity, like non- and mis­ hearings, slips of the tongue, word search, problematic reference, lack of clarity and specificity, in addition negotiations of meaning. There also exists an organization of repair in conversation, in addition to other "ordinary" mechanisms or routines, (e.g. meta-comments, reformulations) that can be applied to these problems (Schegloff et al. 1977, Schegloff 1987, Coupland et al. 1991, Linell in press). Alongside these repair mechanisms, conversations contain features and routines that display the participants' understandings, and hence the possibility of recognizing problems and breakdowns in intersubjec­ tivity; e.g. sequential organization and feed-back (Schegloff 1982, 1984, Franke) 1984, Linell 1990, ten Have 1991, Schegloff 1992, Drew & Heritage 1992). It is self-evident that in order to elicit information one must ask questions. One key method of addressing clarification in doctor-patient conversations, that is to investigate the patient's problems, nature of the symptoms, etc., at a global leve), is naturally by topic introduction and development. For the most part, this is accomplished by way of questions. In doctor-patient interaction, as in many other forms of institutional discours, the importance of topic and questioning patterns in the obtainment of information is well known and have been shown in numerous studies. But these studies have also shown that topic control and questioning are in the hands of the professionals, and used as a means to control information and discourse and to fit the matters into the institutional frames (West 1983, Mishler 1984, Beckman & Frankel 1984, Agar 1985, Franke) 1990, Linell 1990, Drew & Heritage 1992). Another way of testing one's understanding of what is said or to negotiate meaning, is to request repair and clarification from one's conversational partner (Schegloff et al. 1977, Meehan 1981, Shuy 1983, West 1984). In a doctor­ patient encounter where the participants have differing background knowledge and perspectives, and may use special vocabularies or expressions, be they medical or lay, coupled with the complexity of content matter anda need to get correct and detailed information, as clear understanding is often very important, requests for clarification can be expected to be quite common, or at least play a key role in the clarification process. However, there also exist forces that may constrain the number of requests. One of these is politeness, or consideration of face, for examp!e a participant who wants to avoid giving the impression that his/her conversational partner is

2 cf. Linell (in press) on discussion of latent, covert and overt misunderstanding events, and treatment of such sequences. One notes, in particular, "that presence of repairs need not presuppose genuine MEs [misunderstanding events], and, of course, that many genuine MEs do not generate repair sequences." 170 ignorant or expressing him/herself unclearly. The difference in status and interactional dominance between a doctor and a patient in institutional inter­ action may also lead to hesitation on the patient's part to request clarification. Another counteracting factor is the desire to avoid too halting or tedious a conversation, where discourse is repeatedly interrupted and broken up by requests for clarification, instead of running smoothly. Instead, the participants may Jet utterances that are unclear or difficult to understand pass, hoping perhaps that everything will become clearer for them later on. A participant may also attempt a less obtrusive approach to requesting clarification, or try, in advance, to phrase his/her words in such a way that foreseeable problems do not arise (West 1984, Coupland et al. 1991, Linell in press). Self-repair and self­ correction, i.e. self-initiated repair and correction, may be another way of dealing with clarification (Schegloff et al. 1977). Other means of clarification, including different forms of self-clarifications, explanations, reformulations, and expansion and development of utterances, are also likely to be significant. Sequences of clarification, or expressions and sequences in need of clarification, may be difficult to identify. This is due, in part, to the fact that all communication involves partial and fragmentary understanding, and most contributions can be regarded as ways of showing one's interpretations, and attempts and efforts to clarify (Linell 1990, in press). However, there are also sequences in a conversation where clarifications and requests for clarification are noticed and marked-interactionally, formally and/or topically. There are, for example, sequences and tums in a conversation where the main purpose is not to give or elicit new information or introduce topics, or aspects of topics, hut where one of the participants requests some clarification in order to be able to answer or understand. The conversational repair organization is one means by which these problems may be addressed, and there are others (Schegloff et al. 1977, Meehan 1981, West 1984, Schegloff 1987, Drew 1991, Drew & Heritage 1992, Linell in press). The question I would like to put forward is then, what clarification signifies and how it can be accomplished in doctor-patient conversations. I also would want to know more about the usage and importance of medical terms and concepts, in addition to-in some cases-lay expressions, in actual conver­ sations. The content and distinctions of such vocabularies are important, both in the physician's professional knowledge and performance, and in the patient's understanding. Earlier studies that have investigated the usage of medical terms and clarification in doctor-patient conversations, have concentrated on repair organization. In this paper I will examine a number of strategies for clarification and requests for clarification in a corpus of doctor-patient encounters. The first step in my study was to investigate the usage of requests for repair for clarification. I have studied this in relation to what the problem sources are, in particular-whether or not they involve medical terms and concepts. Another issue dealt with is whether other methods of clarifying 171 medical terms and concepts exist. My second step was to explore how certain diagnostic concepts were handled and clarified in the conversations.

Two approaches In this article I will present two approaches to analyzing clarification in doctor­ patient conversations. I am particularly interested in how requests for clarifi­ cation are related to medical terms and concepts. In the first part of my study an approach to handling misunderstandings and requests for clarification in all conversation, viz. the organization of repair, has been applied. The use of requests for repair, i.e. "other-initiated repair", as a forma! method of clarifying, is looked at, to see what types are used, how they are distributed between participants, and above all what it is that needs to be clarified. In the second part of the study, I present an alternative method for investigating how medical terms and concepts are managed in the conversations. Here I have chosen to start from a number of important diagnostic concepts, and investigate how these are dealt with and clarified in the conversations. It may be noted that the "success" of a clarification, i.e. to what degree the clarification actually clarifies, is understood or misunderstanding is resolved, will not be dealt with (unless the "failure" of clarifying a matter leads to another request for clarification). I would also like to point out that cases which would formally constitute a request for clarification, and is treated as such by the participants themselves, may very well serve other purposes and functions than that of solely clarifying. I am, however, not addressing these issues in this paper. Although I do intend to consider conversational phenomena which are relevant to the participants themselves. The material consists of 15 doctor-patient conversations, of a total length of 420 minutes, and an average length of 28 minutes per conversation. The encounters comprise regular visits to two clinics of interna! medicine in Sweden, and all the patients suffer from some kind of rheumatic disease. Five of the consultations are new visits, i.e. first visits, where the doctor and the patient have not met before and are doing so to establish a diagnosis; three visits involve registration at a special rehabilitation center for rheumatic patients whose diagnosis has been established, hut where the participants have not met before; and seven cases are return visits where the diagnosis has been established and the doctor and patient have met before.

I. Requests for repair Earlier research on this subject, requests for clarification and repair in doctor­ patient conversations, has shown that while patient-initiated questions are often 172 dispreferred (West 1983, Franke! 1990), requests for repair are not (Shuy 1983, West 1984). It has also been demonstrated that requests for clarification and repair are one conversational device available to doctors and patients for handling difficulties generated by the use of medical terms and jargon (Meehan 1981, Shuy 1983, West 1984). I will deal here with explicitly requested clarifications, or "other-initiated repair". The common conversational device of self-clarification, especially self­ correction, i.e. "self-initiated self-repair", will not be looked at (Schegloff et al. 1977; see also Schegloff 1987, 1992). The analysis does not include what could be called "instances of corrections", where "simple" errors or mistakes in pronunciation, word choice, slips of the tongue, etc. have occurred and have been replaced (cf. Schegloff et al. 1977, Jefferson 1987, Schegloff 1992). Also excluded are sequences of "word search", in which the participants often cooperate to find "the right word". Nor will I study "requests for confirmation" or "markers of surprise" ("news marks") (Schegloff 1982, West 1984), or "formulations" (Heritage & Watson 1979; cf. "reformulations", Adelswärd et al. 1987). The markers and requirements used to identify requests for repair (cf. Schegloff et al. 1977, Meehan 1981, West 1984) are given below in Table 1.

Table 1. Criteria for requests for repair

1. Interrogative words and question phrases. What did you say?, What? What do you mean (by ... )?, What is ... ?3

2. Interrogative expression + candidate understanding; including question with a candidate interpretation, or suggestion with interrogative intonation Do you mean/You mean + suggestion af interpretation; Is that/That is ... + suggestion

3. Repetition of trouble source; full or partial followed by clarific~tion or 4. Longer pause and/or trouble markers lsuggestion (speech perturbations and incomplete utterances)

Combinations of criteria are also possible, the most common combination being one of the elements mentioned together with a pause. There are other additional criteria, the most important of these being that requests must be iocai, i.e. used in a local context and at a local leve!. That a request is positioned in a local contcxt almost ahvays means in the same sequence, as a part of an adjacency pair and as a conditionally relevant question type (or other utterance) (e.g. Schegloff et al. 1977, West 1984). An example of this is when asked a question,

3 Please note translations of Swedish Va sa du and Va are literal and where in English might be considered impolite, are very common and not considered impolite in Swedish. 173 the answerer produces a request for repair that topicalizes an item in the pre­ ceding question, before answering. The item then is clarified and subsequently answered. Also, the content of the requests must refer to a local leve! of the conversation, usually to a specific item or to the entire directly preceding utter­ ance. Thus a request for clarification that refers to the significance of discourse as a whole or all of the patient's problems or symptoms, e.g. a question like What could it be?, is not regarded as a local request for clarification. For requests of type 3 and 4 to be counted as requests for clarification, there should follow a clarification or some other reformulation from the participant being requested of clarification. 4 Further, the person whose utterance constitutes or contains the problem source must see to the repair ("other-initiated self-repair"). Some examples are given below.s Examples of requests for repair: Interrogative words and question phrases:

(1) D: But you have been able to work now / despite these problems P: Well it has been bard hut/ but l'm on a three months probationary period for ajob - D: What

(2) D: Do you have a mediset / ord- - P: What is what is that D: It's a boxa little red box with P: Nono Interrogative expression or question with candidate understanding:

(3) D: Yes you've x-rayed back and shoulders you said / is there anything else you've had x-rayed P: Hmh mhm ((shakes her head)) D: Nothing else P: Y ou mean x-rayed in general - D: Yes P: Well I broke my wrist / I was x-rayed then

4 Requests for clarification identificd by criteria 1 and 2, need not aiways be followed by a clarification, although in this material there is no such "unfulfilled" requests. 5 Some transcription conventions used are: I, Il, /Il pauses of different length, from approximately 0,5 (/) to 3 ,0 (///) seconds stumbling, stuttered talk pt clicking sound. Transcription in English translation has been simplified and standardized. 174 Repetition: (4) D: Y ou haven 't got any marks on your stomach Il -,. P: Marks on D: Like str- stretch marks striae let's see ifwe Longer pause:

(5) D: Well I so you work quite a lot then Il D: pt I Y ou have long working days P: Yeah Having illustrated the criteria mentioned, I will now show some results on the number and distribution of types of requests for repair, and, most importantly, trouble sources, by type.

Distribution of requests In the 15 doctor--:patientconversations studied, I found 45 examples of requests for repair. A total of 17 requests were initiated by doctors, and 28 by patients. Table 2 gives a summary of the results. I do not intend to comment on the individual figures or conversations in the table. I must, however, draw attention to the fäet that the number of requests for repair made by patients is greater than the number made by doctors; 62% were made by patients, and 38% by doctors. This differs from the findings of West 1984 (cf. also Shuy 1983), where the distribution of requests for repair was more equal, 59% for doctor-initiated requests for repair compared to 41 % for patient-initiated. West's study also demonstrates a reversed distribution where the doctors account for more requests than the patients. Table 2 also shows that there are some conversations in the material that contain no requests for repair, in two conversations no such requests for repair according to the criteria occurred. Here, we may assume that other techniques of requesting clarification have been used. It is also worth noting that approxi­ mately half of the total number of requests (24 out of 45) are made in four of the five "new" visits. (These four recordings also stand for about one third of the total recording time.) This perhaps was to be expected as these encounters were ones containing detailed and lengthy anamneses as well as diagnosis deliveries with careful explanations. It is, however, difficult to say whether the total number of requests for repair in the material is high or low. Nevertheless, self-clarifications on one hand, and confirming expressions, requests for confirmation and news marks on the other (cf. Schegloff et al. 1977, West 1984), are procedures one would expect to find to a larger extent. This is also the impression I get, and would like to investigate further. 175 Table 2. Requests for repair

Conver- Length Newor Total Numberof Numberof sation (min.) retum numberof doctor patient visit6 reguests reguests reguests

37 N 7 3 4 2 49 N 12 2 10 3 28 R 1 4 33 R 4 2 2 5 23 R 4 4 6 18 R 1 7 27 R 0 8 22 R 6 1 5 9 19 Nr 1 10 17 Nr 2 1 11 10 Nr 1 1 12 18 R 1 1 13 50 N 0 14 29 N 3 2 15 36 N 2 1 Total 45 17 28

Types of requests What types of requests, then, can be found? Table 3 shows how different types of requests for repair are distributed between doctors and patients.?

Table 3. Types of requests

Total Doctors Patients

1. "What did you say?", "What?" 9 6 3 2. Suggestion 9 5 4 3. Interrogative expression 8 2 6 4. Pause/hesitation 6 0 6 5. Repetition 5 2 3 6. Pause/hesitation + suggestion 5 2 3 7."You mean" + suggestion 3 0 3

6 Nr is used for registration encounters at the rehabilitation center. 7 In Table 3, requests identified according to criterion 2 have been divided into the three sub­ types "suggestion", "pause/hesitation + suggestion", and "You mean + suggestion". 176 The number of instances is very small but I want, nevertheless, to comment on this and point out some differences between the doctors and the patients. Perhaps the most striking difference is the comparatively high number of patient-initiated requests of the category "pause or hesitation" (not followed by a suggestion), 6 instances. There is no instance of this category for the doctors. This is also the case for the category "You mean + suggestion", where patient requests total 3, and the doctors reguests none at all. To account for the absence of the category "pause/hesitation" in the case of the doctors, one may take into consideration the interactional regulation of the conversation in institutional encounters, which among other things involves the professional usually taking responsibility for regulating the dialogue, its progress, etc. (e.g. Agar 1985, Linell 1990, ten Have 1991, Drew & Heritage 1992). Another point worth mentioning in connection with this, is that a request for clarification in the form of, for the most part, a pause or hesitation, can be regarded as less obtrusive and a more implicit way of requesting clarification in that this strategy does not formally mark the request.

What needs to be clarified? Another question is now what it is that needs to be clarified. Of special interest is naturally whether it is the use of medical terms and jargon, lay expressions or special concepts which is the trouble source. In table 4 I have compiled the results with respect to this.

Table 4. ltems requested to be clarified

A. Problems of hearing or attention Total Requesting person (a) "What did you say" preceding utterance repeated: 9 P:3,D:6 (b) Repetition: preceding utterance repeated: 3 P:0,D:3

B. Specific item requested to be clarified Total Requesting person (a) Medical or other technical term: 3 P:3 (b) Vague or unclear word: 5 P:5 (c) Unclear reference (incl. incompleted utterance): 5 P:3,D:2 (d) Identification of referent: 6 P:l,D:5 (e) Problem of remembering (point intime): 3 P:3 (f) Checking fäets, interpretation: 11 P:10, D:l

I have divided the material into two groups. One group consists of requests for repair that are treated in the discourse as problems of hearing or attention. 177 Almost all of these are initiated by the phrase What did you say or What, and the repair is generally made by repeating all or a major part of the preceding utterance, in an identical form or only slightly reformulated. This can be seen in Example 6 below. In another case, however, the repair is considerably more exact than the initial formulation (in the morning is changed to between eight and nine); and in yet another case the utterance totally reformulated, although with the same meaning, Example 7 below. There are also three instances where "repetition" is treated in the same way. In none of these cases does there occur any specific technical term or other problematic item in the preceding sequence that could account for comprehension problems.

Hearing or attention problems: (6) ((Blood pressure is taken.)) P: Is it low or -,. D: What

(7) P: Well I well I understand that it's all very diffuse and I D: What

Examples of items requested to be clarified

Medical terms: (8) D: Do you often have infections Il D: pt I Get colds P: No seldom Vague words: (9) D: You say that when you stand up it hurts Ido you feel weak as well Il or is it -,. P: What do you mean by weak like I'm going to fall or D: Well no but less strength in your musc!es Unclear reference: (10) D: But othenvise you don't need any other aids strips or anything I -,. P: What for D: Urine test strips 178 ldentification of referent: (11) D: And then you can use salicylic drugs whether it is Treo or Bamyl or P: I can do that / Bamyl S what's that like -,. D: Eh / well / that's the one that dissolves P: Yes that's right it's easier D: You think it's easier yes Problems of remembering:

(12) D: When did it start P: Well I was probably / I know that I was in to Emergency at least three times three years in a row there D: Three years in a row and when was that Il P: Hm -,. D: Was it in the seventies the eighties P: I guess it was the late seventies and early eighties Checking fäets and circumstances:

(13) D: Orudis and Naprosyn and that type are milder P: Mm D: So what I thi- _,. P: What d- y- s- Naprosyn is milder too D: Y es I Naprosyn and Orudis are such anti-inflammatory medicines

(14) D: And eh Il pt eh why l'm asking so many questions about the sun and that is because our experience is that usually in connection with or especially after sun tanning / we we get aggravated situations// and this is part of the tissue disorder that I suspect you may be suffering from P: Mhm D: That's why I asked about whether you get eczema from the sun if you get it in the summer there's a test that is called ANF which is an antibody / that you have Il and the amount you have of it can can be said to be divergent from the normal population Il P: And that has to do with the tissue D: That has to do with the tissue disord- it's called LED this disease P: I see Same comment is needed here. There do not appear to be many problems with medical terms or jargon. A low number of requests for clarification of medical terms may, generally speaking, be due toa reluctance on the part of the patient to request explanation of these words, a major reason being perceived dif­ ferences in status and dominance and a preference for avoiding obtrusiveness. It could also be that there are only a limited number of (unexplained) medical terms uscd in these conversations, which is how I interpret it. Another issue is what elements may be counted as medical terms. Are words like x-ray, cortisone, injection, sedimentation, infection medical terms? Words such as these are used in these conversations, not explained or requested to be explained, and they are used by both doctors and patients (as well as being found in "ordinary" language). There is also the question of intension depth. 179 Patients may be fami!iar with a broad meaning of a term, and use the words themselves, at the same time as their depth in understanding and background knowledge are not nearly as profound as the professional's. Vague or unclear words are slightly more problematic. All words categorized as "vague" in the material are uttered by the doctor and define important diag­ nostic concepts, at the same time as being everyday language words. Most difficulties, however, area result of unclear reference and identification of referent on one hand, and content of the discourse on the other, i.e. checking fäets and circumstances.

Implications One conclusion that can be drawn from the findings of my analysis, is that a little more than a quarter of the requests for repair concern more superficial problems such as hearing and attention. One also notices that medical terminology does not seem to be a major source of problems in the material, while lack of clarity and specificity, e.g. unclear reference and identification of referent, in quite a few repair requests is. As West & Franke! (1991:181) point out: "miscommunication in medical encounters can arise from talk that is non­ technical as well as technical in nature." The problems, however, seem to be quite easily resolved and, in these conversations, may represent minor problems of communication. It cannot, however, be overlooked that problems may some­ times arise from or be a sign of more profound differences in knowledge and perspectives. The devices for requesting clarifications that occur are standard conversational means for dealing with problems of hearing and understanding in any kind of discourse. I would also argue that this type of request for clarification is not overly frequent, as one might have imagined. One reason for this may be that there are other strategies for requesting clarification, in particular-self-clarification. As is discussed in Schegloff et al. 1977, self-initiated self-repair may be preferred as well as being a common means of clarification in conversation. The study by West (1984) illustrates the importance of requests for confirmation (which, in her material, were twice as frequent as requests for repair), which may be another method of requesting (and securing) clarification. Most important, however, for the purpose of this study, is the following question. If problems with medical terms and important medical concepts are not dealt with within the repair organization, i.e. as requests for repair, as the results would suggest, how are they managed? There may be other methods (in addition to the forma! repair organization) of requesting clarification. And more generally speaking, how are important medical terms and concepts clarified? One way of investigating this is to start with the important terms or concepts. Are they or are they not posing problems, and if so investigate how they are 180 dealt with, within the conversations. So, in the second part of this study, another approach to clarification has been adopted.

Il. Conceptual investigations In this part of my study I have chosen to proceed from a number of important diagnostic concepts regarding rheumatic diseases, to investigate how these are handled in the interactions, that is to say-if and how they are clarified. The concepts chosen are tiredness, strength (in hands), morning stiffness and rheumatism (in connection with heredity). Some preliminary results and observations of these clarification sequences will be presented. As the concepts are important parts in the professional agendas of diagnosing rheumatism, they are naturally introduced by the physicians first. But it should be noted that the words are also both vague and ordinary lay expressions. The concepts are most thoroughly discussed in the first visits, where a diagnosis is to be established, hut may also come up in return visits. Tiredness: discussed in 4 first visits. Strength: discussed in 3 return visits (physically examined in 2 return visits). Morning stiffness: discussed in 5 first visits, and in 3 return visits. Rheumatism: discussed in 5 first visits. These concepts are usually introduced and clarified through topic development clarification sequences at a more global leve!, and above all by questions. There are, however, also sequences in which a request for clarification of concept occurs at a more local leve!. In both cases, this occurs in something that we could call conceptual investigations. I would like to begin by looking at some examples of conceptual investigation concerning the concept 'morning stiff­ ness'. In the first example (15), the doctor introduces the concept with a question and continues the conceptual investigation with a second question about an aspect of the same, i.e. the duration of the patient's morning stiffness. 8 (15) D: This that you feel stiff and clumsy and that is it worst I at any special time of the day do you think or is it the same all the time I P: That I'm stiff that's m- mostly in the morning I think D: You are P: Yes _.. D: Mm how long are you stifffor I P: W ell maybe an hour

8 The duration of morning stiffness is an important diagnostic criterion of the rheumatic disease RA, rheumatoid arthritis. 181 In the next example (16a-b) the investigation becomes more complex and detailed; the topic is reintroduced in the conversation.

(16a) D: So Il pt I when you wake up in the morning do you feel anything specialare you stiff then P: Stiff D: Youare P: Yes -,. D: For how long Il P: Well Il an hour an hour anda half D: Anhour P: Yes The topic is then reintroduced about twelve minutes later (during the physical examination):

(16b) D: And this morning stiffness it I has it increased intime then P: Yes it has gotten like Il in three months it has gotten much worse D: Much worse

3 lines omitted. Pause 24 s while the doctor gets a blood pressure cuff.

D: So it just keeps getting worse and worse P: Yes Il D: The morning stiffness is it all over your body or P: No then it's Il my knees I ifI walk down stairs D: Mm 111I'm talking about morning stiffness I so it feels like it's in your knees P: Yes D: Not in the rest I I not your hands P: It aches like in my hands and up you know Il hut hut I've been I I mean I it's hard to explain Il like Il it's growing Il all the time you know in my hands and up into my shoulders D: But you you Il what I'm interested in is this morning stiffness then Ido you feel I you could say a time one anda half hours when when you are I -,. are you stiff all over or is it just the knees you mean or is it stiffness all over P: It's my whole body you know D: Okay I then we mean the same thing As a contrast, these examples can be compared to an example from a return visit where the patient has had rheumatism for several years and the doctor has no need of exploring the concept of morning stiffness, especially not its duration. (17) D: How far can you walk P: Wellat this time of the day it's hard to walk from the parking lotto up here D: Because of your feet P: Y es in the morning I only move around indoors D: You had already from the beginning a lot of trouble with this intense moming stiffness that lasted so long in the morning and now you've got it back again P: Yes I have The examples show that clarification can be found at different levels of conversation, and can be of different scope. At a more global leve), it is often accomplished by topic development and questioning. The examples also show 182 that one aim of doctor-patient conversations is to clarify matters, i.e. the patient's problems, symptoms the patient may be experiencing, the nature of them, and that certain conceptual distinctions are often important. At a more local level, there is another form of clarification operating. Here, requests for clarification are formally not marked as requests for repair, and identifying them is a problem. Tentatively, I would suggest that this form of conceptual investigation requests clarification in the form of specification and concretization of concepts at a more local level, i.e. immediately following a question-answer adjacency pair in which the concept was first introduced. This type of questioning and requesting clarification is mainly used by the doctors and seems to be used routinely in questioning and eliciting information from the patients about important symptoms or concepts. The doctor requests specifi­ cation without formally marking that a clarification is requested. This usually implies the doctor starting with a general question on an important subject, receiving an answer that is too vague or low in intension depth, after which he/she asks additional questions, dividing the concept at hand inta more precise and concrete parts or aspects. 9 Here I offer only a few examples of this strategy of conceptual investigation, 18-21 below, but I think the area is worth investigating more thoroughly.

(18) D: Do you think that you are tired / like in the daytime are you tired P: Y es I sleep a lot D: Y ou sleep a lot P: Y es in the evenings / I fall asleep right away --+ D: Y ou do / this tiredness is there any difference from how you felt a year ago for example do you think you're more tired now / or have you always been tired so to speak P: Well I think I have because I can't ((laughter)) it's hard to lie down and watch TV I just fall asleep / and I have never been like that before --> D: Y ou haven't / so you thin- has anyone commented on it P: My girl friend does yeah

(19) P: An aunt of mine well she had a bit of rheumatism that I know of --> D: Mhm did she have sort of twisted bony hands P: No ((shakes her head)) D: Nothing she had a little pain in her joints P: Yes

(20) D: How about the strength / in your hands P: Well it's nothing to boast about ((quiet laughter)) --> D: Mm how do you notice that P: I can't carry anything or pick anything up

(21) D: How is the strength in your hands I P: Well it isn't up to much ((quiet laughter)) I would say

9 This procedure, starting with a general and vague question and then dividing it up into more specified aspects and a use of concrete examples, is an interviewing technique recom­ mended in Swedish medical textbooks. 183

-+ D: Can you manage a frying pan I P: pt I but eh no lifting no lifting D: Not even a Teflon pan I P: Oh yes / yes I do but I was thinking of that D: Y ou were thinking of a cast iron pan P: Really heavy yes hut otherwise there's no problem D: Can you manage a Teflon pan with one hand P: Yes In these examples, the doctor is breaking up a concept, an important symptom (tiredness, rheumatism, strength), bringing up different aspects of the concept and making the concept more precise and concrete by giving and/or asking for examples. It would be easy to see this as a way of avoiding a request for clarification in a formulation such as What do you mean by ... ?. The information requested is divided, in advance, into smaller parts, and different, more concrete, specified aspects. In all probability the doctor already knows which concepts need to be clarified. Distinctions of the concepts that are relevant in these conceptual investigations can be seen as manifestations of the differences in knowledge, conceptions and perspectives between the professional and the layman.

Conclusions This study has shown that the conversational repair organization is operating and can be used for clarification in doctor-patient conversations. It has also shown that requests for repair often regard more superficial (viz. hearing and attention or reference) clarification problems. Medical terminology and concepts or deeper interpretation and understanding levels seem not to be dealt with to any great extent by this repair organization. We may, however, assume that these problems are more often handled through self-clarifications, which will be an interesting future line of investigation. Analysis of how a number of important concepts are dealt with in the conversations suggests that clarification of important diagnostic concepts is handled within topic development clarification sequences, by way of questions, in so-called "conceptual investigations". Repair can be seen as a narrower concept, often operating at local and more superficial levels, and as a part of any conversation. Clarification, as in the conceptual investigations, on the other hand, exists both at global and local levels, conccrning primarily deeper and more important elements. The requests for clarification found in a doctor-patient conversation may be due to differences in knowledge and perspectives between the doctor and the patient; such sequences in the discourse may be signs of differences in the "conceptual worlds" of doctors and laymen. At the same time one can say that the requests indicate that these conceptual differences are realized, dealt with, 184 and hopefully often overcome, allowing the participants to reach an understanding in their discourse. As stated above, a conversation with several requests for clarification is not necessarily a problematic conversation, or a conversation lacking requests for clarification not necessarily an unproblematic and smoothly running conversation where everything is understood. Requests for clarification are important for the participants in a conversation to gain clarity and understand each other. In the special type of conversation studied in this project, requests for clarification are also vital to the understanding of a patient's problems and thus, ultimately, to improving the patient's health.

References Adelswärd, Viveka et al., 1987: The unequal distribution of interactional space: Dominance and control in courtroom interaction. Text, 7:313-346. Agar, Michael, 1985: Institutional discourse. Text, 5:147-168. Beckman, Howard B. & Franke!, Richard M., 1984: The effect of physician behavior on the collection of data. Annats oj Interna! Medicine, 101 :692--696. Cicourel Aaron V., 1981: Language and medicine. In Charles A. Ferguson & Shirley Brice Heath (eds.), La.nguage in the USA. Cambridge, Cambridge University Press, pp 407-429. Coupland, Nikolas, Wiemann, John M. & Giles, Howard, 1991: Talk as "Problem" and communication as "Miscommunication": An Integrative Analysis. In Nikolas Coupland, Howard Giles & John M. Wiemann (eds.), "Miscommunication" and Problematic Talk. Newbury Park, Sage, pp 1-17. Drew, Paul, 1991: Asymmetries of knowledge in conversational interactions. In Ivana Markova & Klaus Foppa (eds.), Asymmetries in Dialogue. Hemel Hempstead, Harvester Press, pp 21-48. Drew, Paul & Heritage, John, 1992: Analyzing talk at work: an introduction. In Paul Drew & John Heritage (eds.), Talkat Work. lnteraction in Institutional Settings. (Studies in interactional sociolinguistics 8.) Cambridge, Cambridge University Press, pp 3--65. Fisher, Sue & Dundas Todd, Alexandra, (eds.), 1983: The Social Organization oj Doctor-Patient Communication. Washington D.C., Center for applied linguistics. Franke!, Richard M., 1984: From sentence to sequence: Understanding the medical encounter through microinteractional analysis. Discourse Processes, 7:135-170. Franke!, Richard M., 1990: Talking in interviews: a dispreference for patient­ initiated questions in doctor-patient encounters. In G. Psathas (ed.), lnter­ action Competence. Lanham MD, University Press of America, pp 231-262. 185 Have, Paul ten, 1991: Talk and institution: A reconsideration of the "asymme­ try" of doctor-patient interaction. In Deirdre Boden & Don H. Zimmerman (eds.), Talk and Social Structure. Studies in Ethnomethodology and Conversation Analysis. Cambridge, Polity Press, pp 138-163. Heritage, J.C. & Watson, D.R., 1979: Formulations as conversational objects. In George Psathas (ed.), Everyday language. Studies in Ethnomethodology. New York, Irvington Publishers, pp 123-162. Jefferson, Gail, 1987: On exposed and embedded correction in conversation. In Graham Button & John R.E. Lee (eds.), Talk and Social Organisation. (lntercommunication I.) Clevedon, Multilingual Matters, pp 86-100. Linell, Per, 1990: The power of dialogue dynamics. In I vana Markova & Klaus Foppa (eds.), The Dynamics oj Dialogue. New York, Harvester Press, pp 147-177. Linell, Per, in press: Troubles with mutualities: Towards a dialogical theory of misunderstanding and miscommunication. Forthcoming in Carl F. Graumann et al., (eds.), Mutualities in Dialogue. (Also available as "Arbetsrapport från Tema K", 1993:2. Linköping: Dept. of Communication Studies.) Linell, Per & Luckmann, Thomas, 1991: Asymmetries in dialogue: some con­ ceptual preliminaries. In Ivana Markova & Klaus Foppa (eds.), Asymmetries in Dialogue. Hemel Hempstead, Harvester Press, pp 1-20. Meehan, Albert J., 1981: Some conversational features of the use of medical terms by doctors and patients. In Paul Atkinson & Christian Heath (eds.), Medical Work: Realities and Routines. Famborough, Gower, pp 107-127. Mishler, Elliot G., 1984: The Discourse oj Medicine. Dialectics oj Medical Interviews. Norwood, ABLEX Publishing Corporation. Schegloff, Emanuel A., 1982: Discourse as an interactional achievement: Some uses of 'uh huh' and other things that come between sentences. In Deborah Tannen (ed.), Analyzing Discourse: Text and Talk. Georgetown University Round Table on Languages and Linguistics 1981. Washington, D.C., Georgetown University Press, pp 71-93. Schegloff, Emanuel A., 1984: On some questions and ambiguities in conversation. In Maxwell J. Atkinson & John Heritage (eds.), Structures oj Social Action. Studies in Conversation Analysis. Cambridge, Cambridge University Press, pp 28-52. Schegloff, Emanuel A., 1987: Some sources of misunderstanding in talk-in­ interaction. Linguistics, 25:201-218. Schegloff, Emanuel A., 1992: Repair after next tum: The last structurally provided defense of intersubjectivity in conversation. American Journal oj Sociology, 97: 1295-1345. Schegloff, Emanuel A., Jefferson, Gail & Sacks, Harvey, 1977: The preference for self-correction in the organization of repair in conversation. Language, 53:361-382. 186 Shuy Roger W., 1983: Three types of interference toan effective exchange of information in the medical interview. In Sue Fisher & Alexandra Dundas Todd (eds.), The Social Organization oj Doctor-Patient Communication. Washington D.C.: Center for applied linguistics, pp 189-202. West, Candace, 1983: "Ask me no questions" ... an analysis of queries and replies in physician-patient encounters. In Sue Fisher & Alexandra Dundas Todd (eds.), The Social Organization oj Doctor-Patient Communication. Washington D.C.: Center for applied linguistics, pp 75-106. · West, Candace, 1984: Mishearings, misgivings, and misunderstandings in physician-patient dialogues. Discourse Processes, 7:107-134. West, Candace & Franke!, Richard M., 1991: Miscommunication in medicine. In Nikolas Coupland, Howard Giles & John M. Wiemann (eds.), "Mis­ communication" and Problematic Talk. Newbury Park, Sage Publications, pp 166-194. The Social W ork er as Moral W orker Applying for Money-the Moral Encounter Between Social Workers and Clients

Lars-Christer Hyden

When faced with our own personal finances we are confronted with our own Iives; with our decisions, strengths and faults, priorities and omissions, and we encounter these in a way that resembles a cost-benefit analysis. In this respect our everyday household economy is a question of morality; not only in the sense of whether to steal money for beer from the food account, but also in the sense that our incomes, expenses, and deficits reflect and depict how we live, have been living, and would like to live. Our incomes are linked to our choice of profession; our expenses to our sense of what is of worth and what we consider necessary; and our deficits may tel1 us something about what we ought to do or they may compel us to shift priorities. In other words, the fäets of household economy obtain their meaning along a moral horizon: it is in relation to our moral conceptions that questions of personal finance are evaluated, and it is here that a specific amount of money becomes either a deficit, a burden, or a relief. Although most household economic fäets obtain their meaning from personal values, most societies have some kind of general conception of what the minimal economic level of existence is and what kind of responsibility the individual has for his or her own economic subsistence. In most Western societies different systems exist for guaranteeing citizens a minimal leve) of subsistence when their normal way of supporting themselves temporarily does not function due to for example unemployment or illness. Most of these systems are based on the principle of insurance: while you are earning money you must regularily pay a premium in order to be able to use it when times get tough. Most Western countries also have some sort of relief system for economic support when this social insurance fees system for some reason does not function for individuals. Persons who apply for this latter kind of social welfare compensation in most countries have to tum to a public social welfare bureau. In Sweden, these bureaus are run by the municipalities. In general, the eligibility of an applicant is decided by a professional social worker based on some kind of rules stating in broad terms which circumstances are seen as valid for receiving social welfare. In Sweden, as in probably most other Western countries, these rules are locally determined and do not for the most part have the status of legal rules. 188 The professional social worker can be characterized as a "street-level bureaucrat". This concept was introduced in the 1970's by Lipsky and referred to bureaucrats "who interact directly with citizens in the course of their jobs, and who have substantial discretion in the execution of their work" (Lipsky 1980:3). Of central concern to the "street-level bureaucrat" is classifying citizens in order to be able to apply existing and relevant rules to the individual cases. Starting with Lipsky's theory, Prottas (1979) writes that "the task that faces the street­ level bureaucrat is to match the relevant characteristics of the citizen to the eligibility rules of the agency and place the citizen into the category [ ...] The way the client is categorized determines how the bureaucracy will treat him or her" (Prottas 1979:4-5). The social worker who meets a client has to use a given set of rules and categories stating when and under what circumstances this particular client is eligible for social welfare benefits. And the social worker will have to present his or her decision as legitimate to supervisors and politicians. The client who meets a social worker is placed in a situation in which his or her economic choices and standards are scrutinized by the social worker: the client's ability to sustain him­ or herself, the choice of occupation, priorities of expenditures. Clients manage their economy in relation to their everyday context, possibilities, and choices, and assess and justify their economic doings in relation to mundane standards. These standards and the client's justifications reflect and are a part of the client's moral conduct in the sense pointed out in the beginning of this paper. The encounter between social worker and client takes place at the social welfare bureau, which conditions the förmal framework of the meeting. The client is compelled to make a förmal application to qualify as a legitimate applicant and then to be called to an encounter with a social worker in which his or her eligibility will be examined. During the actual meeting, the client must present his or her economic situation while referring to the everyday life context in a way that makes him or her appear eligible for social benefits. The social worker has to match the förmal eligibi.lity rules with the information provided by the client, not only to categorize the client hut also in order to account for the decision to supervisors and local politicians. The encounter between social worker and client is framed in a förmal setting­ the rules, procedures, and categories of the social welfare department -and is to a large extent concerned with the client's handling of his or her own personal economy. The situation is basically a moral one. As Adelswärd and Forstorp (1992) have argucd discourses can be defined as moral in at least two ways: the topic can be an explicit moral issue and the format of thc discourse can be asymmetric giving one of the parties certain moral rights over the other. The topic under discussion in social welfare encounters is ambiguous. In the everyday context of the client it concerns his or her way of handling money, life style, and priorities, in other words, the client's moral charactcr. (Cf. Hall, Sarangi 189 & Slembrouck forthc.) To the social worker the topic basically concerns the client's eligibility, that is how to apply the forma! rules. The format of the discussion is characterized by the social worker having the right to question the client and evaluating his or her handling of money. This makes it an asymmetric conversation (Linell & Luckmann 1991) giving the social worker certain moral rights in the situation how to talk and which perspective to use. The central question for the present study is in which way the social worker and client deal with these issues: how is a moral question processed in a forma! and institutional setting and by forma! means. This problem is discussed through an analysis of some conversations between social workers and clients at social welfare bureaus in the municipality of Stockholm. Initially the establishment of a conversational format is analyzed. In opening the conversation the social worker has to frame the conversation in a way that makes it acceptable to both parties present that the social worker not only has the right to decide the topics to be discussed and the conversational rules, but also to make inquiries concerning the client's private and personal matters. The investi­ gation proper of the economic situation of the client is analyzed in order to identify typical conversational sequences. These sequences are then analyzed further with respect to their content: the client's justification of the handling of the economic situation and the social worker's evaluation of these actions.

The data In 1986/87 a data corpus consisting of 20 encounters between social workers and clients was collected. All these encounters took place at social welfare bureaus in the municipality of Stockholm. The conversational interaction between the clients and social worker was recorded on a tape recorder and later transcribed. Fifteen of these 20 encounters concerned applications for cash public allowances and constitute the material in the present analysis. To become a social welfare client means going through a selection process, generally starting at the switchboard of the social welfare bureau. The telephonist at the switchboard makes a first screening of all incoming phone calls redirecting citizens who, according to the telephonisi, have called the wrong agency or for obvious reasons are not eligible for social welfare. The citizens passing this first obstacle are directed to the social worker who is on call. This social worker makes a preliminary investigation over the phone and those citizens found eligible at this stage are given a time for an appointment with asocial worker. Before their encounter the client is sent an application form by mail. The form consists of questions concerning date of birth, address, civil status, income, rent, and some other basic social and economic information considered relevant for the assess­ ment of eligibility for social welfare benefits. Just two encounters will be used as the basis of the following analysis. The first (# 204) involves a male person about 25 to 30 years of age visiting the social 190 welfare bureau for the first time and meeting a female social worker of about his age. The second encounter (# 106) involves a male client of about the same age as the first, but well acquainted with both the female social worker and the procedures at this social welfare bureau. The general structure of the 15 encounters has been analyzed and discussed in a previous paper (Hyden 1991) and have been subjected to various analysis by Fredin (1993) and also by Linell & Fredin (fortlicorning).

The opening: The making of the moral authority The opening of the encounter not only gives the participants the possibility to negotiate and elaborate on the reason for the meeting (cf. Schegloff & Sacks 1973), hut also to establish the conversational format: the social roles of the speakers, the rules of the conversation and the rights attributed to the speakers. (Cf. Schegloff 1991.)

Excerpt 1 (# 204)

01 S: let's see if you have filled in your application form fully 02 C: well yes what should I write [looks for his form] 03 S: hmm 04 C: ah here it is [finds the form in his bag] 05 S: you have not stated your address and present occupation 06 C: well you mean the period I was,working, 07 S: or the last three months 08 C: yes [... ] 09 S: then theperiod you have been unemplbyed·, 10 C: hmm [wrltes in theforrn:J 11 S: we do not ~Ui tfreernployers but we wanttcrl\e sure·tll.at',you aetually•have been working f'2 C:· sure This excerpt represents the first exchange in the encounter-excluding the initial greetings which were not recorded. In that sense it makes up the förmal start of the encouhter and identifies the social worker as 'the authorized starter' (Tumer 1972:393). But most important a formatfor the encounter is established: a topic and an agenda are identified and the speech rights are distributed. The application form and its oontents are introduced as botft,the prime topic and the agenda by the social 1workerr(line 1)) ,fh'e client responds by accepting this and asking w hat he should l writei (lliie 2). In doing so the social worker gives the form an importanfstatus: ihs the fäets stated in the form that will become the topics for discussion. The form contains information about the client's family situation, address,;occupation, incomes, rent, and so on. The social worker immediately introduces the topic of work and the time periods the client has been working. The social worker indicates that the relevant time period may be a three week period. 191 But most important, in this opening sequence the social worker positions herself as the forma! authority in the encounter: she has the right to open the meeting and she has the right to introduce topics (line l); she is the authority as to how to fill in the form (line 5 and 9); and she decides the rules on how to go about this. This fäet is elaborated on and stated more clearly in line 11 where the social worker introduces herself as a person invested with a special moral authority: She indicates that she is authorized to call the client's former employer, but she will refrain from doing so. The reason she gives is that she wants to be sure about the client's actual working period. She indicates the moral authority invested in her by a switch to a we form: "we do not call the employer but we want to know [ ...]". By this the social worker makes it clear that she is not speaking as a private person but as a representative of the social welfare administration and is invested by its authority to make investigations. The client by his more or less passive responses ("mmm", "yes", "sure") accepts and confirms the social worker's förmal and moral authority. Although this means of opening an encounter is the standard one (Hyden 1991; cf. Cedersund 1992), other ways could be adopted. In excerpt 2, the social worker starts the encounter by attacking the client (line 1, 5, 11 and 13) for seeming to be under the influence of drugs.

Excerpt 2 (# 106)

01 S: do you know what I thought when you entered the room 02 C: mmm 03 S: we have met once before in January 04 C: yes 05 S: I thought you looked high then and I asked you about that 06 C: yeah 07 S: then you said you were taking tab!ets and were on +sick leave 08 C: + yes I had been taking tranquilizers and 09 S: + and now I think 10 C: + sleeping pills 11 S: exactly and I was thinking the same thing when you entered the room now 12 C: no 13 S: I think that you once again seem to be high or intoxicated 14 C: no [laughs] I am not high 15 S: is this unusual am I the only one who says this is there anyone els= 16 C: =lately it's only you but I'm not healthy right now it was the same yesterday when I was applying for ajob I felt really weird and was shaking and had a bad stomach and last week I had fever 17 S: but I am the only one lately

In initiating the encounter by confronting the client, the social worker established the same kind of order as in excerpt i above: the social worker introduces a topic and establishes herself as a forma! and moral authority. By confronting the client for seeming to be under the influence of liquor (line 1 and 13), the social worker immediately focuses on the moral character of the client. The client tries to defend himself by denying the accusations and on line 16 by referring to everyday circumstances: he had not been feeling well lately and has problems 192 with his stomach and has even had a fever. In her reply the social worker does not accept the client's invitation to discuss his "problems" in relation to his everyday context, but clings on to her own line of reasoning, namely, her impression of him as being high. Although the social worker uses the first person pronoun-/-she twice makes references to contextual factors which defines her as an official person. In line 1 and 11 she refers to her office at the social welfare bureau in a way that indicates that to her, the client starts to exist the moment he enters the door to her office. On line 3 she refers to a previous encounter that had taken place some months before. By these references the social worker links the present situation to previous ones and to the fäet that the client only exists to the social worker in herofficial localities. In both these excerpts, the social worker quickly establishes herself not only as the förmal authority, that is, the person with the right to open, to introduce topics, and to evaluate answers, but also as the moral authority with the right to question the moral character of the client and his where abouts. In doing so the social worker draws upon the institutional context constituted by the social welfare bureau and the social worker's forma! and legal rights. In the next step of the conversation the social worker proceeds to the investigation proper, by turning to the information on the application form.

The investigation I: The form of the conversation The aim of the investigation is for the social worker to establish the legitimacy of the client's economic situation and his handling of it in relation to the given local rules for social welfare eligibility, and the trustworthiness of his answers. The questions of legitimacy generally concern the income situation of the client and the trustworthiness whether the information provided by the client could be regarded as correct or not. The collection of this information proceeds through a systematic discussion of the client's information on the application form. In this manner the topics given on the form determine which topics will be discussed. In that sense, the conversation has a high topical coherence, although the topics discussed may appear random and without inner coherence to an uninitiated observer. Two basic types of sequences can be identified which have different forms and places in the conversation. The first type is a simple question-answer loop. The second is elicited when the social worker wants to test the legitimacy or trustworthiness of the information provided by the client. The extract is taken from the beginning of an interview. 193 Excerpt 3 (# 204)

0 I S: let's go through the application are you unrnarried 02 C: mmm 03 S: do you live alone in yourapartment 04 C: yes 05 S: do you have a contract for the apartment 06 C: mmm yes I have 07 S: can you show me a payment slip 08 C: yes I have it here 09 S: well I can see that you haven'! paid the rent for November 10 C: no 11 S: l'd like a copy ofthe contractand then you can have it back 12 C: mmm 13 S: have you been to the employment agency In the question-answer loop, the social worker introduces a topic and the client generally gives an short answer. In excerpt 3 the social worker introduces the topic of civil status on line 1, on line 3 whether the client Iives alone in his apartment, and on Iine 5 if he has a contract. The social worker may follow up the answer by a new question and the client offers a new response (line 7-12). The question-answer loop comes to an end when the social worker either changes topics or shows a passive acceptance of the answer or answers provided by the client with an mm or yeah. When the social worker introduces a new topic a new question-answer loop begins. These conversational seque!}cesprimarily invoke basic social information, but even questions of this kind can be problematic and elicit sequences of the second kind with an elaborated response by the client.

Excerpt 4 (# 204)

01 S: m let's see do you have an ID card 02 C: I don't have one l've !ost my wallet 03 S: yeah 04 C: well so I don't have any ID at the moment 05 S: yeah 06 C: maybe this isa problem 07 S: yes it's troub[esome but we can start with your application 08 C: well yeah I've !ost my wallet 09 S: yeah 10 C: so I don't have any ID 11 S: have you reported it to the police 12 C: no [ ...] 13 S: well I think I have to ask my supervisor about it In excerpt 4 the social worker asks the client to identify himself by showing bis ID-card and the client tells the social worker he has !ost it. The client apparently senses that this may affect bis application and asks the social worker if this is a problem. The social worker confirms this and tries to introduce a new topic, but 194 the client repeats that he has lost his wallet. The social worker closes the topic by saying that she has to discuss it with her supervisor. In this sequence the social worker asks for an ID-card and implies that the client must prove that he is the person he says he is and not an impostor of some sort. The client has to show the social worker that he is a trustworthy person. When he cannot do this, he starts to excuse himself for the lack of ID-card by referring to the circumstances surrounding the loss of the card. The form of this sequence differs from the earlier one in that it starts with a question from the social worker which results in an relatively elaborate answer from the client. The social worker listens to his answer and confirms this by passive answer of the type yes and yeah. This second form of conversational sequence occurs whenever the social worker poses an open question or when the client attempts to clarify certain circumstances surrounding some fäets or when the social worker wants to question the quality of information or actions by the client. This takes place in situations when the client has to justify his behavior and certain circumstances related to his economy.

The investigation Il: The evaluation In his elaborate answers, the client not only tries to provide the information the social worker needs, but to present the information in a way that portrays himself as an eligible applicant. The client has to present himself not only as a trustworthy person; he also has to present his economic situation and his handling of it in a way that is justifiable. In doing this the client presents his actions by referring to everyday circumstances and by stating them as the best given in the situation. The social worker, on the other hand, is required to evaluate the client's story according to the standards of the eligibility rules.

Excerpt 5 (# 204)

01 S: why did you have to leave yourlastjob 02 C: I quit by myself 03 S: you quit yourself 04 C: mmm 05 S: yeah 06 C: because of my insccure income 07 S: yeah 08 C: my income varies between five thousand and ten thousand crowns a month 09 S: yeah 10 C: because I'm a freelancer and I work on commission 11 S: mmm 12 C: fixing lights and sound recording equipment and such things 13 S: mmm 14 C: driving car and fixing things and I can work fora week and then be free for[ ...] 15 S: yeah 195 16 C: then I work for two days and am free for two months then work a month and be free two days 17 S: yeah [ ...] 18 C: I must have another job which gives me a stable income each month 19 S: mm 20 C: 'cause I've got something's I have to pay for 21 S: yeah 22 C: in order to survi ve and now I never know if I 'Il earn one thousand crowns or live thousand [ ...] 23 S: well I think I'm beginning to understand now [cough] well it is like this if you quit ajob voluntarily we are usually very tough we usually [ ...] 24 S: have a benefitfree month [ ...] 25 S: since you didn't quit because you didn't like your job 26 C: no absolutely not it was a great job 27 S: mmm In excerpt 5, the social worker poses an open-ended question about the fäet that the client had stated that he was unemployed. The client responds by stating that he had quit his former job himself, and the social worker repeats the client's response and the client starts to describe his former job and the circumstances leading up to his quitting it (some of his explanations are left out of the excerpt). The client claims that he had a job he thought was "great" but that his work periods were random and his income insecure. As a consequence he had problems supporting himself. In order to stabilize his economy he quit his the job with the intention to find a new one. While the client is relating his story, the social worker supports him by interjecting listener support items (back chan­ nelling) of the type mmm and yeah. When the social worker finally sums up her impressions of the client's story, she does so by evaluating it according to the rules for social welfäre eligibility. She thus shifts to the first person plural, we, as to stress the fäet that she is not speaking on her own behalf, but as an official representative of the social welfäre administration. The social worker states the rules and describes the normal procedures in cases when a client quits by himself thus contributing herself to defining the moral status of the client. She sums up with an open-ended question, "for you didn't quit because you didn't like your job," and gets a negative response from the client. In his narrative, the client continually refers to his everyday life circumstances: where he was working, what he did, what he thought about it, and how he anticipated getting a new job. That is, he refers toan everyday context which he apparently supposes the social worker is familiar with and understands. By referring to these everyday contextual fäets, he attempts to present to the social worker the "moral logic" of his actions. He argues that he acted in the best possible manner given the circumstances: he presents his actions as deliberate choices made in order to avoid something less wanted, namely, a perpetually poor economic situation, which in its extension would lead to an economically 196 catastrophic situation. In this way, the client presents himself as a person acting in a morally responsible way, although he was to a certain degree forced by circumstances to act in the way he

Excerpt 6 (# 106)

01 S: what is the reason you are unemployed when you could get ajoh 02 C: what the hell I was at this clothing company the wage was seven and half thousand crowns [$1.200] a month do you think I could like ajoh like that 03 S: well you have to choose hetween living on social welfare benefits 04 C: no= 05 S= two thousand sixty crowns [$450] plus the rent for the apartment 06 C: there are other jobs I'm looking into so we have to wait and see 07 S: there is something important and fundamental which you yourself touched upon 08 C: yeah 09 S: it's this ahout rejecting ajoh offer 'cause you think seven anda half thousand is too small an amount [... ] 10 S: 'cause we can't accept that you tum downjoh offers or weed outjohs 11 C: no hut this rejecting well I didn't tum it down I went on an interview 12 S: no hut you're not out looking for certain kinds of johs 13 C: well I'm going toa new interview today In extract 6 the social worker asks the client why he is out of a job when it is rather easy to get a job, at least at the moment. The client answers by telling about a job he was offered hut declined because he thought the payment was too low. The social worker responds by pointing out the client's alternatives, either accepting the job or living on social welfare benefits which would give him much less to live on. She points out that the client is in no position to make a choice 197 between different jobs and that he according to the beneficiary rules cannot turn downjob offers. In doing this, the social worker changes from the first person/ to a we pronoun form, thereby indicating that she is speaking as a representative of the social welfare administration. The client mitigates his first statement by saying that he did not turn it down, since he just went in for an interview and that he has another one lined up for later the same day. The client presents his everyday moral standards when he tells the social worker that he will not accept a low salary. The social worker responds to this by stating the beneficiary rules, but in doing so this she criticizes the client's moral standards as being illegitimate. The client senses this criticism and immediately attempts to mitigate his position by reframing his job rejection as an initial interview and saying that he is going in for another interview the same day. That is, the client does not retain his own moral standards, but rather gives in to the social worker's moral position. He obviously does this so as not to jeopardize his eligibilityfor social welfare benefits; he has to accept whatever job he is offered. In other words, the social worker not only formulates a different moral standard, but she supports her moral standard by an indirect threat: if the client does not accept her moral standards he will not be eligible for social welfare.

Discussion If the fäets of household economy obtain their meaning in relation to the moral horizon of everyday life, the social bureau introduces a new horizon, that of the institutional and bureaucratic. In the encounter between social worker and client, both of these conceptions are put against each other as two different ways of evaluating the treatment of the private economy. To be asocial worker who examines applications for economic support and making decisions on these matters, is to act as a representative for the welfare administration; it is to match relevant rules to concrete cases and to formulate decisions which are formally accountable to superiors and local politicians. In this respect, the work of the social worker could be characterized and analyzed in forma! terms, as a question of matching "the relevant characteristics of the citizen to the eligibility rules of the agency" (Prottas 1979:4). To the social worker the moral topic primarily is defined as a forma) topic. A problem for the social worker, however, seems to be which are the "relevant characteristics". In a forma! perspective, this information concerns certain social and economic circumstances. But a doser analysis indicates that even thc most basic information, say about the client's confirmation of his or her identity, hasa moral aspect: the possibility that the client is not trustworthy and is an impostor. Information concerning current employment or unemployment implies a possibility that the client does not want to work and support himself. This means that the social worker has to question the moral character of the client. 198 By pursuing this forma! inquiry, the social worker is also conducting a moral search: the social worker has to determine the moral character of the client and the circumstances surrounding his handling of his own economic situation. In other words, the "relevant characteristics of the citizen" that the social worker matches with the beneficiary rules are also of a moral nature. In this respect, the social worker is not only asocial worker but also a moral worker. In order to define and defend his moral character the client has to pursue the economic matter as a moral topic. In the encounter the client refers to everyday circumstances to account for his economic situation and to justify it by applying everyday moral standards. The client also has to support his or her moral charac­ ter in more general terms: is he or she an impostor and fraudulent or is he trustworthy. In the encounter the client puts his moral standards and the moral quality of his actions related to his economic situation, on trial. The preferences concerning professional career, life style and so on, are normally accounted for by reference to the standards of life, or to the morality of everyday life. These standards and evaluations are further more normally part of the shared life world of family, relatives, and friends, although they generally also have an intelligibility for most other persons. These preferences and standards are evaluated against the institutional, forma! standards of the social welfare agency. This also means that the social worker in a conflict situation can use her institutional power: to decline the client's application for money in order to compel the client to conform to institutional standards. The specific character of the social worker as a moral worker is that he or she does not pursue his or her private moral cancerns-social workers use the pronoun we in discussing moral matters-but the moral norms of the social welfare agency. The norms pursued by the social worker to a !arge degree seem to converge with at least same of the moral standards of the clients. The reason for this is probably that they both draw on deep-rooted cultural and moral conceptions conceming the limits of individual rights and collective responsibility. Assumingly, this is also one of the reasons that conflicts between social workers and clients do not emerge more often then they in fäet do. Rather, shared moral standards seem to be an important condition for the smooth running of social bureau conversations.

References Adelswärd, Viveka & Forstorp, Per-Anders, 1992: Samtal om moral och samtal som moral. [Conversation about Moral and Conversation as Moral.] (Arbetsrapporter från Tema K 1992:9.) Linköping, Linköpings universitet. Cedersund, Elisabet, 1992: Talk, Text and lnstitutional Order. (Diss.) (Linköping Studies in Arts and Sciences, no. 78.) Linköping. 199 Fredin, Erik, 1993: Dialogen i socialt arbete. [fhe Dialogue in Social Work.] (Diss.) (Studies in Communications, no. 36.) Linköping, University of Linköping. Hall, Christopher, Sarangi, Srikant K. & Slembrouck, Stefaan, forth.: Moral construction in social work discourse. In Britt-Louise Gunnarsson, Per Linell & Bengt Nordberg (eds.), The Construction oj Projessional Discourse. London, Longman. Hyden, Lars-Christer, 1991: Moral och byråkrati. Om socialbidragsansökningar. [Morality and Bureaucracy. On Social Welfare Applications.] (Forsknings- och Utvecklingsbyrån, Stockholms socialförvaltning, Report 1991:7.) Stockholm. Linell, Per & Luckmann, Thomas, 1991: Asymmetries in dialogue: Some con­ ceptual preliminaries. In Ivana Markova & Klaus Foppa (eds.), Asymmetries in Dialogue. Hemel Hempstead, Harvester Wheatsheaf, pp 1-20. Linell, Per & Fredin, Erik, (forthc.): Negotiating terms in social welfare office talk. In A. Firth, (ed.), Negotiation in Workplaces: Discourse and Inter­ actional Perspectives. Lipsky, M., 1980: Street-Leve! Bureaucracy. Dilemmas oj the Individual in Public Services. New York, Russel Sage. Prottas, J.M., 1979: People-Processing. The Street-Levet Bureaucrat in Public Services Bureaucracies. Lexington, Lexington Books. Schegloff, Emanuel A., 1991: Reflections on talk and social structure. In Deirdre Boden & Don H. Zimmerman (eds.), Talk and Social Structure. Studies in Ethnomethodology and Conversation Analysis. Cambridge, Polity Press, pp 44-70. Schegloff, Emanuel A. & Sacks, Harvey, 1973: Opening up closings. Semiotica, 8:289-327. Turner, Roy, 1972: Some förmal properties of therapy talk. In David Sudnow (ed.),Studies in Social Interaction. New York, The Free press, pp 367-396.

The Treatment of Client Responsibility and Laughter in Encounters Between Social Workers and Clients

Ilmari Rostila

The aim of this paper is to describe some aspects of the treatment of client responsibility in the interaction between social workers and clients at municipal welfare offices as institutional interaction. According to Drew & Heritage (1992:22), institutional talk has three important characteristics: (1) goal orienta­ tions, (2) special and particular constraints and (3) particular inferential frameworks. Institutional interaction thus involves an orientation by at least one of the participants to some core goals, tasks or identities that are conventionally associated with that institution. It may often involve special and particular constraints on what one or both of the participants will treat as allowable contributions to the business at hand. Institutional talk may also be associated with inferential frameworks and procedures that are particular to specific institutional contexts. The social workers concerned were engaged in so-called integrated social work. In the traditional organization, social workers are divided into separate departments specializing in (I) child welfare, (2) social work with alcohol abusers and (3) income subsistence. In recent years, however, there has been a growing tendency to combine these organizations so that social workers are responsible for more than one sector at a time. The same worker, who usually has close contact with the local community, assumes responsibility for all three or at least income subsistence and social work with alcohol abusers. In most of the situations that I analysed, the social workers were responsible for all three sectors. In encounters between social workers and clients the participants seem to negotiate their way in a "bottom-up" fashion towards a sense of what the interaction will be about (Drew & Heritage 1992:23) so that one option is a relatively personal relationship (for the friendship mode! as a strategy in social work, see e.g. Stenson 1992). Often the participants have met before and will meet in the future as the clients have to renew their application from time to time in order to qualify for the benefits. In this way the encounters are conversations in-a-series; the way the participants leave each other, for instance, may be used as a resource in the next encounter (see Button 1991). One aspect of negotiating the character of the interaction cancerns the negotiation of the 202 "first topic", which is presented as having a special status in this conversation as the "first topic" or "reason for the conversation" (for these terms, see Schegloff & Sacks 1973). In most encounters it is the social worker who takes the initiative in order to get the discussion started on a certain topic. In some encounters he or she starts to administer economic support or try to get the client to give some special information or account. There is a clear similarity between these encounters and those studied by Hyden (see elsewhere in this volume). Sometimes the client takes the initiative to raise the financial issue first. Sometimes, however, the social worker will take the initiative to talk more freely as "phatic communion", or will start to negotiate with the client about the agenda so that the financial issue could be postponed or otherwise presented as not the "first topic". Thus, in contrast to clearly "monotopical" conversations (Schegloff & Sacks 1973), and the encounters investigated by Hyden, participants made use of their previous experience and knowledge about each other to invoke different activities and identities. Every discussion at the social welfare office necessarily touches upon the responsibility of clients for their situation. Laughter is often used when and where the client is either reproached or the discussion otherwise touches upon client responsibility. This finding provided an impetus to study in more depth the use of laughter in the treatment of responsibility as a special detail reflecting distinctive characteristics of certain instances of interaction. An analysis of those instances could provide a folier picture of the treatment of responsibility and of the moral work done by social workers (cf. Hyden in this volume). The analysis here is confined to those instances where the treatment of re­ sponsibility and laughter meet; other instances where the participants approach the responsibility of the client or laugh are not studied. My description does not seek to cover the use of laughter in social work nor the social significance of laughter in institutional discourse (cf. e.g. Adelswärd 1989). My material consists of tape-recorded discussions between social workers and their clients at social welfare offices. I have analysed the encounters of eight social workers with twenty clients, i.e. a total of 20 encounters. There are no encounters in the sample that deal explicitly with child welfare. In the few cases that did concern children, the clients refused to have the conversations tape­ recorded. Otherwise only a few clients did not want me to use the tape-recorder.

The treatment of responsibility in other contexts In conversations there always appear initial actions to which recipients are expected to do relevant next actions (Schegloff & Sacks 1973). These "adja­ cency pairs" have a first part and a second part (questions-answers, request­ granting/rejection, invitation-acceptance/rejection). If the first part of an adja­ cency pair is produced in a conversation, the recipient is expected to produce the 203 Iatter part of the respective pair. An accusation constitutes the first part to one such adjacency pair, and it requires a relevant response (Atkinson & Drew 1979:57). There isa range of relevant next actions: denials, counteraccusations, justifications or excuses, admission (i.e. acceptances of the charges) and apologies. However, denial is the preferred next action. It is produced in the form of a preferred-action tum shape (see Pomerantz 1984). The absence of a relevant response, a deliberate deviation from the relevance structure, is taken into account in the conversation. The speaker may repeat the first part, or there may follow a pause during which the speaker waits for a reply. The absence of the second part of the relevance structure may also be treated as a "noticeable absence", about which complaints may be made or from which various inferences may be drawn (Atkinson & Drew 1979:57). In ordinary day-to-day conversations, the identification of the blamed party in the process of allocating responsibility can be preceded by an account of an "unhappy incident" (Pomerantz 1978). The speech sequences in which partici­ pants attribute responsibility for "wrongdoings" are then divided into two parts: The first consists of a report of an "unhappy incident". This occasions a subsequent segment, where the aim is to identify the person responsible for that incident. In this sense the attribution of responsibility happens "next". The sequence begins with a statement that specifies the cause for allocating responsibility, hut it does not yet amount to an accusation. This gives the recipient an opportunity to take the initiative in the allocation of responsibility; in the words of Pomerantz, initiating a sequence in this way with a segment other than blaming allows for recipients to initiate attributing responsibility (by apologizing or making a confession). This can be a preferred action in relation to accusing the other party in the situation. In a study of cross-examination situations, it was found that the order in which the parties speak and the types of speaker tums are fixed: the speech exchange system is (partly) pre-specified (Atkinson & Drew 1979:65). In a cross-examination, each tum is either a question or an answer. In addition, it is prescribed that it is the examiner who asks the questions and the examined who does the answering. The examiner's questions amount to an accurate descriptive effort to find out what exactly happened in a certain incident and who were involved. This information provides the basis for the official charge. The examiner, in the sequence of questions and responses, produces descriptions of a situation and how these descriptions are used in producing a final evaluation of the witness 's activity (i bid.: 105-135). In courts of law, the actual accusation is thus preceded by fi nely discriminative "descriptive work" that is done within a sequence of questions and answers. The purpose of this descriptive work is to collect evidence that can help to prove that the examinee is responsible for the offence for which he or she will later be accused. The participants are well aware that the questions and responses can be implicative with regard to the official charge. 204 Talking about a difficulty and the responsiblity of the client The conversation in extract 1 takes place before the social worker announces her decision as to whether or not assistance is going to be granted; it is part of the investigation of the circumstances of the client in that context. The social worker's initiative and the client's responses are connected with the applying and getting the benefit and are more or less implicative with regard to it. In the sequence the social worker failed to bring the discussion to bear on the diffi­ culties that would have paved the way toa discussion of responsibility.

Extract 11

0 0 01 S is everything otherwise alright • 0 0 02 C ((no everything's)) (.) I haven't been 03 drinking for one and 2 half months (now). (.4) 04 s have you coped (.) on your own or have you been 05 using these antab[ us or anything], 06 C [NO: no I don't use] them-. 07 S 0 yees 0 (1.4) 08 S have you been to the AA or, 09 C nohhh 10 (1.6) 11 S Alice's (.) support has been enough then, 12 C ho ho .hhh (.4) 0 (1 haven't been drinking fora 0 13 month) ( ) (.6) 14 s ( ) (.4) 15 C ( ) (not now) 16 (4.0) 17 s .hhh how- (.) well (.) generally have you been 18 able to find work on YQl!! own or through the 19 manpower office. The social worker opens this episode with the familiar "how are you" question. The question gives the clierit an opportunity to upgrade himself: he hasn't been drinking for six weeks. This possibility helps to make this a legitimate topic of conversation. There is also another question that gives the client an opportunity to present himself in a favourable light (he has managed without taking medical drugs). The client answers the questions in the preferred action shape (Pomerantz 1984). However, the social worker is apparently expecting a continuation to these replies (the pauses in lines 7 and 10). The social worker has indicated that externa! support is extremely important in the effort to keep off the bottle. The clieni, on ihe other hand, has demon - strated that he is completely independent of externa! help. However, the social worker then says something which makes it extremely difficult for the client to hold on to his argument that he is independent of externa! help: the social worker presents a conclusion in which he praises the role of the client's wife

1 See for transcript notation Atkinson & Heritage ix-xvi. 205 (11). It is hard for the client to say that his wife has been of no help. However, the question could push the client towards a treatment of his shortcomings. The client responds to the social worker's conclusion by laughter (12), thus refusing to talk about the whole issue. The social worker reverts to the client's ability to cope independently, now in his capacity as ajob applicant. The treatment of responsibility and moral work can be more or less detached from the application of money and from the institutional context that is con­ stituted by the social welfare bureau and the social worker's forma! and legal rights. The conversation in extracts 2 and 3 took place after the financial issue had been settled; in that sense the activities in cases 2 and 3 are detached from the business of applying for money. Case 2, however, is taken from a rather monotopical, bureaucratic encounter after the social worker has refused to give assistance. The discussion about the client's responsibility is part of making the decision acceptable to the client. In case 3, the social worker and client have also had contacts and common experiences from earlier projects and group work in community work. These encounters differ clearly from each other in terms of how implicative criticism or accusation and responses to accusations can be with regard to the institutional task of applying for money. In the following sequence the discussion proceeded to deal with the client's difficulties (extract 2), but it is the social worker who formulates the trouble. In this extract the social worker accused the client in no unclear terms.

Extract 2 01 S wellat least there're some good sides here 02 like one is that you had the sense to leave 03 Helsinki intime, I mean it's really an 04 [expen]sive place= 05 C [(mm)] 06 S = >(been for you that)< .hhh you were there 07 a bit ((clears throat)) too, something of 08 a mil;:judgement (.) wasn't it? 09 (.6) 10 S I mean at least you shouldn't have got a 11 video from the first[:. er]: 12 C [m: l The social worker opens the episode with a preface in which she compliments the client for avoiding difficulties that could have seriously exacerbated her situation. However, the fäet that the client's decision (leaving Helsinki) is taken under consideration refers to the client as the subject in the problem situation. The social worker then tries to construct an unhappy incident out of the client's former life, touching upon the problems of her situation (3 and 4). The client produces a continuer when the social worker mentions the key word in the description of the problem situation: expensive. The social worker now suggests a more direct link between the problem situation and the client. 206 On the other hand, the client's responsibility has already been taken up in connection with her decision to leave Helsinki. As the social worker continues on the client's circumstances, taking them to imply a certain conclusion (the client should have left Helsinki even earlier than "intime"), the client's respon­ sibility is more or less explicitly stated. Now, the social worker refers to the client as a passive subject. She moves towards the passive voice as she talks about the mistake that has been made. The social worker combines, in a rather peculiar way, the unhappy incident and her accusation, referring to the client as the place where the mistake (misjudgement) happened. The social worker's statement resembles an assessment that requires a response by the client in the form of a second assessment. On the other hand, there are also some critical elements in the statement that would seem to require a response. At the same time, the social worker is beginning to moderate her conclusions. She starts out from a mistake, but this soon becomes a misjudge­ ment (this is more clear in the Finnish 2). The "in your case" is also a passive and neutral expression which helps to moderate the criticism. When she then re­ words her statement in the form of a question, the social worker mobilizes the resources of the adjacency pair question-answer and thus puts the client under increasing pressure to reply. The formulation of the assessment as a question leaves the conclusion open. The question format challenges the client to respond to the social worker's assessment by a comment assessment. However, the client refuses to reply; she neither acknowledges nor denies the conclusions. Now, the social worker moves from the passive voice (7-8) to the active voice in line 10 and thus steps up her criticism.

Extract 3 01 C (I begin to feel cold) (1.2) 02 S you mean outdoors. 03 C yeah (0.4) 04 S yeah well when you're inside with Ron it 05 feels like= 06 C =yeah [heh] ITT S [you're] indoors so much (0.4) 08 C suppose [that's it] 09 S [when you should] 10 C suppose you should get some exercise. 11 S exercisehh.

2 The translation of this kind of material is highly problematic, as is clearly evident in this example. In Finnish the tum of the social worker was as follows: 06 S: = >(sulle ilmeisesti ollu että)< .hhh olit 07 siä vähä mrh ((kakistaa kurkkuaan)) meni 08 virhe:arviointi (.) sun kohallas vai? 207 The client is complaining that she does not move about very much outdoors because she begins to feel cold. However, she is also displaying her troubles, in an innocent and modest way. This could lead to troubles-telling as an activity where participants orient themselves to each other as a troubles teller and a "properly aligned" troubles recipient as a set of relevant categories (see Jefferson 1980). In so-called service encounters, participants generally orient themselves to each other in a more non-personal and instrumental manner. This means that clients do not talk about their difficulties more than is necessary for purposes of seeking advice or other help (Jefferson & Lee 1981). This client dis­ plays her troubles to a greater extent than people in a service encounter would normally do, and thus displays an understanding of the relationship as more personal. Without aligning as a recipient in troubles-telling, the social worker provides a diagnosis that is fairly intimate (the feeling of cold is due to the fäet that the woman is indoors with the husband, who is mentioned by his Christian name). The client acknowledges the explanation as a preferred action and shows her approval of the social worker's intimate comment by soft laughter. The social worker repeats the explanation on top of the laughter, hut now the tone is more critical. The client admits the criticism in the shape of a dispreferred-action turn. The diagnosis and its admission by the client are used as an introduction to criticism.

Disengagement from accusation in a bureaucratic frame

Extract 2 (continued)

18 S I mean at least you shouldn't have got a 19 video from the first[:. er]: 20 C [m: l 21 S or hifi > (yes yes) it was a music centre 22 you said you'd bought< you shouldn't 23 have bought this ( .2) as a first. 24 C mm. 25 (2.0) 26 S it's often the most IMPORtant thing in life. 27 S y[ ehah it's] how you= 28 C [(it is)] 29 S =how you take i!, 30 C (it) is: (.) almost everyone's got hifi. (.4) 31 S ermm. 32 C everyone's always (.4) that's how it is. This exchange of words started with the social worker constructing, in a rather peculiar way, an "unhappy incident" on behalf of the client (see above). The client produces a relevant response (28-) to the accusation only after the social worker laughs towards the end of the sequence (27). Up until then, the client has remained silent or at best produced a continuer so that the turn has always been 208 the social worker's. The relevance of the accusation in the situation has been so modified that neutral responses have been possible. The participants have displayed to each other an understanding that this interaction happens between a professional and a client by the client refraining from any other kinds of tums than (minimal) answers or continuers to the questions or statements of the social worker (a multiple question-answer sequence as a characteristic feature of professional-client encounters, see e.g. Atkinson 1982: 111). These professional­ client roles are set in an encounter where the participants have so far dealt only with the client's financial problems according to the bureaucratic rules (see Hyden in this volume). Following the accusations, there is a long pause (25) that starts with the social worker refusing to take up the tum handed over by the client. The social worker allows a two-second pause, indicating that the accusation has created a rele­ vance for the client to respond either by admitting or by defending herself against the accusation. The client still refuses to respond, and the social worker continues to deal with the issue of responsibility on the client's behalf. She solves the problem by taking the bull by the horns and presenting an explanation on the behalf of the client (26). The social worker's statement (that this is "often the most important thing in life") at once provides an explanation and keeps the accusation alive. However, the accusation is now raised to a more general leve!. This is followed by laughter. This laughter can be taken more as an attempt to create an atmosphere of "normalcy", a time-out from the earlier business of applying for the benefit (cf. mutual laughter and time out for pleasantries in troubles-telling; Jefferson 1984:351), than one of "intimacy" (cf. above). Laughter provides a suitable environment in which the client can respond to the social worker's statement more like in an ordinary conversation. There are, however, certain features of impropriety in the social worker's statement which in an ordinary conversation could be interpreted as an invitation to intimacy. As an explanation, it remains an accusation and it is made on behalf of the client, as if an exposure of an opinion that the client shares on this issue. The client apparently treats the statement not as an invitation to intimacy but as a signal that the social worker has moved away from their actual business. Laughter serves here as a mediating element so that the issue can now be handled in freer, more normal terms. The client takes immediate advantage of the situation and starts to give a reply (28) that overlaps with the social worker's laughter. Here the client uses almost the same words as the social worker, indicating that an affiliation is also going on (cf. repetition in the production of reciprocal affiliation, see Tannen 1990). Now, the client begins to explain how common the problem is (30) and thus produces an explanation that she so far has not spontaneously given. The client is responsive to the social worker's "depends on how you look at it," but still does not make an unambiguous commitment to the fäet that the pressure to get a music centre should be ignored. Now, the social worker waits fora while before 209 taking up her tum, and then accepts the explanation as if it were an interesting opinion. The social worker is not expressing her agreement here, but listening to the client's opinion on the issue. The social worker takes the client's statement as a mediated explanation for the client's money problems as well. In terms of actions, the situation is organized around accusations and de­ fences. In terms of categories, we have asocial worker, a professional, an expert criticizing a client for things she has failed to do: the former is placed in the category of accuser, the latter in that of defender. In the activity of searching for a common solution, the relevant categories are citizens, people sharing certain problems and co-conversationalists. In this situation the allocation of responsibility is so managed that the social worker gives expression to the client's difficulties and touches upon the client's responsibility in this situation. Here the accusation has a different relevance than in ordinary conversation; the client's silence is actually adequate. The partici­ pants "exit" the situation by activating a different type of categorization. They move towards a more "intimate", more "normal" relationship through the acti­ vation of different types of membershi p categories. In providing an explanation, the social worker is partly disengaging herself from the accusation by providing an explanation on behalf of the client, even though there is some irony in the statement and the accusation continues. Start­ ing from the laughter, the client is constituted as a person who is trying to do her best, which means that the social worker and client are now on the same starting line. Following the laughter, the accusation translates into a problem that is jointly formulated by the social worker and client. At the same time a common membership categorization is activated in the effort to solve this problem. Laughter thus functions as a mediating element, a suitable environment for the movement from accusation toa common description of the problem. It is possible that in this particular environment-the social welfare office­ irrational consumption on the part of the client is interpreted as a cause of a specific personal problem. In denying or confessing her mistake, the client would therefore be entering upon a different act than in normal speech situations where responsibility is allocated. The social worker and client tum what seems to be irrational consumption into something that is more or less normal. The client avoids confronting the mistakes that have caused this problem situation until after the social worker has presented her special, non-personal formulation and the laughter; now the problems can be discussed at the general leve!. Following this "normalization", the client proceeds to talk about the mis­ takes she has made in more general terms, thus avoiding any reference to herself as she gives an explanation. 210 Disengagement from accusation in a partnership frame

Extract 3 (continued)

10 C suppose you should get some exercise. 11 S exercisehh. 12 C don't know I feel like I'm l!!,wayscold 13 (0.4) 14 S ,YQ!!should move around a bit:: (like there) 15 ((sniff?)) (0.2) move around damn it (0.2) 16 around. 17 (1.2) 18 S and- (0.4) you silly fools when I asked 0 (you) 0 19 over there to the centre but- (.) isn't 20 it more expensive: for you, (1.0) 21 C I: have been in such a bad shape that I haven't 22 been able to do (anything) ([ else ]) 23 S [mm:m] 24 (.2) 0 0 25 C (( else )) 26 (1.6) 27 s we 've really had fun I mean the last time 28 there were two women who had .hhh (.2) 29 this bailiff there and we carried out their 0 0 30 stuff- (.4) out of sight (dam-) , 31 (.2) 32 C ((h.hh)) 33 (.2) 34 s HA HA HA [HA HAJ(.) .hh ha [ha] .hh 35 C [(hhh)] [sha]me on .YQ!!, 36 s he heh .hh [((sniff))] 37 C [you're not] supposed to do that,= 38 S =whAT, 39 C but you shouldn't do that sort of thing .. hhh 40 S yes but I mean you must use a bit of common 41 sense= 42 S =you can't Jet [ every]thing 43 C [yeah] 44 S =[go for nothing. and it doesn't MEA]N anything= 45 C [no of course not.] 46 S =that sort of thing, .hhh 47 C well no::. (.2) 48 S in a half a million bankruptcy a (.2) 49 C no well you wouldn't think so he[h .hh] 50 S [a silly] old 51 s =typewrit[er]. 52 C [ye]ah:: The sequence started with the client's complaint and the social worker's not too serious criticism. It then proceeded to the social worker's critique that the client admitted. The social worker carries on with the advice and then relinquishes his tum (11). The repeated advice can be read as an accusation; a response by the client would be relevant. The client provides an explanation in the form of a complaint (12), which would make relevant an empathetic affiliation on the part of the 211 social worker. On the other hand the client is also in a way continuing to display her problem and telling her trouble. It also makes relevant a piece of advice and keeps the issue of responsibility alive. The social worker repeats his advice (in a more aggressive tone of voice on Iines 14-16). The client declines to respond. The repeated advice can be read as an accusation to which the client does not respond. On the other hand the client had already admitted same sort of advice. The next tum (18-20) by the social worker is again an accusation, but it isat once a reproach directed against the client. In this sense it can be read as a sanction. This aspect of the statement is evident in the way that the social worker addresses the client and his spouse as "you silly fools." The Finnish word pöperöt can be heard as a kind of kid name which softens the criticism. The social worker motivates his accusation by making it a complaint about the client who had disappointed the social worker as an organizer of community work. Towards the end of the turn the social worker stops his accusations and ends his turn with a question about the reasons for not coming. The client does not answer straight away (pause), but in the end gives an explanation. The social worker's response is minimal (23); he neither accepts nor rejects the explanation or apology. This can be seen as a way of avoiding representation of oneself as someone who accepts or rejects the client's explanations. Perhaps an acceptance would also imply a move away from the client's problems as shared troubles? Through his minimal response, the social worker is giving the client room to continue her talk as troubles-telling and at the same time showing his sensitivity to the troubles of the client. The client says (apparently) her last word (25). There then follows a pause (26) as the client does not carry on with her troubles. Next, the social worker tells a story (27-) about the client group, which can be read as a renewed invitation to take part in community work and to resume close relations as co-conversationalists. The story and its reception can be seen as an attempt to carry on the role of the social worker as a person who is trying to take part in resolving the problems (passivity, powerlessness) that the client has been displaying. The story describes an activity that is very nearly illegal. The social worker has taken part in that activity and allied himself with his clients against the authorities. This is something that the traditional social worker would categori­ cally refuse to do. The social worker uses a pause and a withheld sweaiword to indicate that this is a non-conventional act (30). There follows a relatively long pause (31-33) during which the client does not produce a clear response, but there is some indication of her hesitating to affiliate with the social worker's story (32). However, the client's disaffiliative response, her silence, can also be read as disaffiliation in relation to the described "improper" behaviour. 212 Jefferson, Sacks & Schegloff (1987) examined laughing together as a component of a particular type of sequence, "the expanded affiliative sequence". Frankness, crudeness, obscenity, etc. can be indices of relaxed, intimate inter­ action (ibid.:160). Participants can use this convention by introducing such talk as a programmatic action. By introducing such talk, a speaker may initiate a move into intimate interaction from a status he or she perceives as non-intimate so far. Jefferson et al. examined talk following impropriety as recipient treat­ ment of an invitation to intimacy (ibid.: 160-163). In many analysed sequences of ordinary conversation, affiliation could be achieved over interna! expansions that ran off in the following three steps: (1) impropriety followed by disatten­ tion, (2) disattention followed by laugh-appreciation and (3) laughter followed by affiliation. The affiliation occurred within an occasion of laughing together. Laughing together thus served as an environment in which the recipient might properly produce a lexical reference to the source of the laughter and thus become implicated in the sort of mentality which produces such talk. Laughing together can thus provide an environment which simultaneously urges and restricts affiliation so that it is both sought after and restricted in its domain. In the following tums the impropriety that appeared in the social worker's story and the disaffiliation that followed it start off the affiliative sequence. The social worker starts this by heavy laughter in his next tum (34). Following the laughter, the client produces a disapproving reply (35) that has the appearance of disaffiliation. Upon closer inspection, however, we see that the client's tum is in fäet a very elever transition from disaffiliation to affiliation. That is, the client's "reproach" is a preferred action in this situation. It is a response that the social worker is trying to invoke and succeeds in invoking with laughter. The client presents his reproach and disaffiliation in the preferred-action tum shape (Pomerantz 1984). The statement is made very emphatically, and it begins with a critical expression at the very moment that the social worker's laughter is ending so that the "reproach" becomes a seamless continuation to laughter. The social worker laughs again (36) and the client expresses her moral disapproval of this activity in the preferred-action shape (37). The client is speaking in a laughing tone of voice. Now, the social worker initiates a repair (38). He shouts "whAT," with an astonished emphasis at the end of the word. It is too loud to be a serious question; there is no doubt he has heard and under­ stood what the client said. In other words, the initiation of a repair has a very specific function here: it reflects the social worker's "ritualized disbelief' (Jefferson), Ritua!ized disbelief occurs in a situation where someone announces that, say, Mr Major has resigned and someone replies with an emphatic whAT! indicating that this was important news. In this case the ritualized disbelief can be translated as "are you (client) really accusing me (social worker)?". This deprives the social worker's counter-reproach of its real status; it is a quasi­ sanction just as the client's statement was a quasi-reproach. This gives the client 213 an opportunity to repeat his reproach (39), intensifying his participation in the joke. At the same time, the improper action of the social worker (and his group) is given a specific topical meaning in the discussion. The client's disapproval has a specific function that is related to the partici­ pants' membership categories. The client has produced a conception that is consistent with conventional moral standards. To counterbalance this, the social worker presents a serious comment in defence of his clients. The client has produced a traditional notion that the social worker refutes. By her (quasi-) criticism of the group, the client is criticizing the social worker for his deviation from conventional behaviour. She is building up a profile of this social worker as an unconventional, original social worker. We could assume that if the only response to the social worker consisted of a reproach directed at himself, then the purpose of this exchange of words would be too obvious. The focusing of the reproach on the social worker and his clients as a group also underlines the membership of the social worker of the client group-without this being underlined. The reproach and the criticism involve the risk that the social worker becomes associated with externa! control; the social worker isat risk of being associated with the accusing/professional category. This risk is eliminated by presenting the story in an amusing shape, with the social worker represented as being explicitly against the control mechanism. Moreover, the roles of the social worker and client are reversed. At the same time, the story repeats the invitation to join in the group. By accepting the invitation to intimacy, the client takes part in the mentality of the group activity and thus expresses his willingness to join in. As the renewed invitation is associated with an invitation based on unconventional methods, the client is not conforming to the will of a "controlling" social worker or consent­ ing to participate in a controlled activity. In receiving the story, the client participates in the client group's frame of mind. The client's positive attitude towards this activity is thus not only stated, hut represented in the details of the speech (laughing together and affiliation). The client's attitude towards the group has been verbalized and made available for future use. Here is an instance where the participants have aligned more clearly as a troubles teller anda troubles recipient and discuss the reasons and the responsi­ bility of the client for this trouble in a work-up phase of the trouble (fOi the work-up phase in troubles-telling, see Jefferson 1980).

Extract4 01 S how are they so (1.4) how've you been other- 02 things have been (.4) alright otherwise 03 C ye:p (.) though I did have these fits. (.) 04 S you did did you, 05 C well I still have these (1.) ((sniff)) hh (.4) 214 06 s oh I' -m sorry (.2) ((compassionate voice)) 07 C [yeah] 08 s [aren]'t you- don't you ha(.) medie- 09 don't you ha[ve medic]ation 10 C [n- n:-] 11 s =[or don't you remember to take] it or, 12 C [ye-ye:- (.) I: have] 13 C Ido remember. (.) 0 but 1° should take more 14 and I don't. (0.2) 15 s well m won't you, (('suffering' voice)) 16 C well because they make me feel tired so I: fool 17 0 them 0 (.2) 18 s (so) you I see: (.) you prefer to be your own 19 doctor and have these fits. =it can happen 20 anywhere you know. (.6) 21 C ((sometimes)) I had had a fit, (.) in 22 in xxxx there in the yard. (.6) 23 s at xxxx? only, 24 C yeah I collapsed there. (.) [clothes] on= 25 s [aha:::]= 26 C =[(but)] 27 s =[yes] but- (.) oh: dear dear oh:- 28 -erm- (.) that' s really dangerous you know 29 surely it's less of a botherifyou 30 feel a bit tired tha- (.4) than than than. 0 0 31 C (no it's) because I don't eat prop- (.) and 32 then suddenly if I don't eat (.) if I don't have 33 enough to eat. (.4) my feet won't hold.=I'm 34 like ajumping mouse (.4) 35 S? (mm)(.) 36 C you know it begins, (.) ((some strange sound)) 37 (.2) 0 0 38 s yea::h , ((mysteriously)) (.4) 39 C quite (.) strange movements. (.6) 40 s (yea::) (.4) well how- you or- =how many should 41 you take then, ( .4) 42 C aa: originally they prescribed three. (1.) then 43 once when there was a different doctor. so I 44 went to renew them. (l._)_ 45 C I was myself.=I didn't have the prescription 46 with me so I says they they'd prescribed two 47 (.4) so he wrote two (and). (.6) and 48 now I've been taking one.=except when I've felt 49 that something's wrong and 50 so I've (.4) taken another one at noon- 51 (.2) 52 C hh[h] ((could starta laughter)) 53 s [so] originally three [andhh]= 54 C [yeh] 55 s =ahfterh thathh .hhh (.) 56 s it isn't (.) no wonder [really] (.) 57 C? [((some kind of humming))] 58 s this is really a bit of a- a bit of a gamble, 59 C well they're always feeding me. ((continues in a 60 smiling tone of voice)) 215 The social worker's "how are you" question (1-2) provides a starting-point for the client's account of an "unhappy incident" (3), which then unfolds into a description of a problem. The client paves the way (5) and the social worker follows (6). This opens the doors to the subject of responsibility as the social worker moves on to propose a diagnosis (8-9). Line 8 is a dispreferred tum as it focuses on the actions of the client. It is interrupted and continued only as a softened, passive formulation ("have medication"). The client responds to the social worker's question which is contained in the diagnosis by assuming responsibility for the act (not taking his medicine) that caused the unhappy incident (13-14). The social worker tries to establish why the client is not taking his medicine (15). It is only after the explanation that the social worker criticizes the client for the action, although there is no denial of the explanation (18-20). The social worker does not contest the client's explanation for his activity but reproaches him for what he has done. The client does not respond to the criticism but has more details to tel1 about his fits (21-22, 24). It is only at this stage that the social worker asks the client to consider the validity of his expla­ nation (27-30). The client responds by pointing to the fäet that he hasn't been eating and describing the symptoms that this has caused (31-34). Thus the client evades the question conceming the grounds for his activity by introducing new information and (partly) unties the link between the medication and the fits. The social worker takes no stand on this but asks the client to give more details on his medication (40-41). Now, the client explains how he has "cheated" the doctors. The story (42-50) begins with an announcement of the förmal prescription. Note the emphasis on the word three and the pause that follows, which serves to underline the key word in the whole story: the number of pills prescribed by the doctors and the number of pills the patient has actually been taking have now been introduced as the hard core of the story. Then, the client explains how three becomes two (note again the pauses after this announcement). Finally, the client says he has actually been taking only one pill. Only after the possible laughter of the client (52 and see pause, 51), the social worker expresses her appreciation of this story and affiliates herself with the impropriety (53, 55). She repeats the plot of the story, starling to laugh precisely after the word three (the prescribed number of pills). The social worker wraps up the basic structure of the story but is unable to finish it; it is simply too funny. The client has succeeded in evoking the response he was looking for, which is seen in the affiliative statement (54) that overlaps with the social worker's laughter. The social worker then comments on the story by referring implicitly to the client's responsibility (for the fits) but at the same time appreciating the story affiliatively as an exposure (56). The statement reveals, in· the shape of a preferred-action tum, the reciplent's relationship to the story as an exposure, the significance of the story as a daring news (cf. ritualized disbelief) which 216 explains the fits. This is followed by a carefully worded reproach where the social worker criticizes the client for not taking the matter seriously enough (58). At this stage the social worker is still referring to the story in a humorous manner, hut she is already beginning to take some distance from it. In his reply (59) the client continues his stories of disobedience in the preferred-action tum shape, and does not respond to the social worker's criti­ cism. Apparently he does not take the social worker's comment on his taking the whole matter too light-heartedly as a reproach hut rather as an invitation to new exposures? The client's story of how he has cheated the doctors (improper behaviour), which is accepted, with some reservations, by the social worker, serves as an invitation to intimacy or an alliance ("partners in crime") and maintains the "normalcy of co-conversationalists". On the other hand, in relating this story the client has only responded to the social worker's question. The cheating of doctors is an invitation for the social worker to join in. The client laughs only modestly and apparently does not present his story as a definitive offer of intimacy. It is an option that the social worker partly seizes.

Concluding remarks In the course of dealing with clients' financial and other problems of everyday life, social workers and clients in social welfare offices come to deal with the issue of client responsibility. Client responsibility can be either linked up with financial support, or it can be detached from it. Referring to two cases where client responsibility was detached from financial support, this paper has examined how the social worker and the client moved back and forth between troubles-telling, allocation of responsibility and even accusing the client. The accusations, against which the client did not very much defend her- or himself, were jointly terminated through a more or less affiliative exit. This evasive management of accusations is well-suited to a partnership relation. Laughter was used in these activites before the speech sequences in which the participants "approached" each other. Laughter in the partnership-frame took place in the form of laughing together, which could be seen as a component of a particular type of sequence, "the expanded affiliative sequence" where speakers offer an invitation to their co-participant to produce talk.that constitutes a move towards more intimate interaction. It seems that social workers and clients in welfare offices may negotiate more or less personal and friendly relationships as special environments in which they can deal with client responsibility and accomplish the social worker's moral tasks. The participants use different professional-client categories. On the one hand, the "social worker" is a bureaucratic agent who deals with financial support. This bureaucratic agent does not actually provide help, but his or her 217 job is to establish what benefits the client is entitled to according to office rules. On the other hand, the "social worker" can display her- or himself as a more or less friendly agent, negotiating with the client and trying to find out if she or he wants to talk about some difficulties. This friendship strategy can raise questions about the overly ambiguous nature of social worker-client encounters (Baldock & Prior 1981, Drew & Heritage 1992:62), but it clearly functions as a special strategy for dealing with client responsibility.

References Adelswärd, Viveka, 1989: Laughter and dialogue: The social significance of laughter in institutional discourse. Nordic Journal oj Linguistics,12:107-136. Atkinson, J. Maxwell, 1982: Understanding formality: the categorization and production of 'förmal' interaction. British Journal oj Sociology, 33:86-117. Atkinson, J. Maxwell & Drew, Paul, 1979: Order in Court. London, Macmillan. Baldock, John & Prior, David, 1981: Social workers talking to clients: a study of verbal behaviour. British Journal oj Social Work, 11: 19-38. Button, Graham, 1991: Conversation-in-a-Series. In Deirdre Boden & Don H. Zimmerman (eds.), Talk and Social Structure. Studies in Ethnomethodology and Conversation Analysis. Cambridge, Polity Press, pp 251-277. Drew, Paul & Heritage, John, 1992: Analyzing talk at work: an introduction. In Paul Drew & John Heritage (eds.), Talkat Work. Cambridge, Cambridge University Press, pp 3-65. Hyden, Lars-Christer, 1992: The social worker as moral worker. Applying for money - the moral encounter between social workers and clients. Paper presented at the International Conference on Discourse and the Professions, Uppsala, Sweden, August 1992. Jefferson, Gail, 1980: Final Report to the (British) SSRC on the Analysis oj Conversations in which "Troubles" and "Anxieties" are Expressed. Report No. HR 4805/2. Jefferson, Gail, 1984: On the organisation of laughter in talk about troubles. In J. Maxwell Atkinson & John Heritage (eds.), Structures oj Social Action: Studies in Conversation Analysis. Cambridge, Cambridge University Press, pp 346-369. Jefferson, Gail & Lee, John R.E., 1981: The rejection of advice: managing the problematic convergence of a 'troubles-telling' and a 'service encounter'. Journal oj Pragmatics, 5:399-422. Jefferson, Gail, Sacks, Harvey & Schegloff, Emanuei A., 1987: Notes on laughter in the pursuit of intimacy. In Graham Button & John R.E. Lee (eds.), Talk and Social Organisation. Avon, Multilingual Matters, pp 152-205. Pomerantz, Anita, 1978: Attributions of responsibility: Blamings. Sociology, 12:115-121. 218 Pomerantz, Anita, 1984: Agreeing and disagreeing with assessments: some features of preferred/dispreferred tum shapes. In J. Maxwell Atkinson & John Heritage (eds.), Structures oj Social Action: Studies in Conversation Analys is. Cambridge, Cambridge University Press, pp 57-101. Schegloff, Emanuel, A. & Sacks, Harvey, 1973: Opening up closings. Semiotica, 8:289-327. Stenson, Kevin, 1992: The social work interview as governmental practice. Paper presented to the Intemational Conference on Discourse and the Profes­ sions, Uppsala, Sweden, August 1992. Tannen, Deborah, 1990: Talking Voices. Cambridge, Cambridge University Press. Communication at the Work Place: Requesting in Asymmetrical Interaction Situations 1

Lenelis Kruse & Caja Thimm

An interesting aspect in the psychological research of interpersonal behaviour is the question of expectancy formation and resulting behavioural actions. Since Merton formulated his concept of the "self-fulfilling prophecy" (1948), which was followed up by Rosenthal & Jacobson's (1968) teacher-expectancy effect (see also Jussim 1986), various studies have shown that the way people treat each other is widely determined by what they expect. All of these studies refer to the fäet that beliefs, hypotheses and expectations about a person or a future event may become fäets, that may be confirmed by the target of the hypothesis or prophecy when acting according to the partner's predictions. Most of this literature on interpersonal expectations and expectancy confirmation (e.g. Snyder 1984) or more recently, identity negotiation (e.g. Swann 1987) suggests that such partner-related hypotheses tend to be confirmed rather than disconfirmed. The explanation for these effects seems to be quite simple when Snyder (1981:300) states: "The behavior of other people is very much a product of our own actions toward them. How others present themselves to us is, in !arge measure, a product of how we first treat them." However, recent studies have shown that people also disconfirm certain hypotheses, especially when the hypotheses are socially undesirable for the target, contradict their self-image (Hilton & Darley 1985) or make them feel pressurised (Baumgardner & Brownee 1987). Partner hypotheses have to be regarded as complex phenomena. Apart from information about specific individuals with respect to age, ethnicity, personality traits, attitudes, group membership or status and forma! power, etc., which influence partner hypotheses, situational fäctors also contribute to the formation of expectations (Galtung 1959). It is obvious that partner hypotheses are largely communicated by language. However, most of the studies in this field have fäiled to examine this dimension, but focussed instead on non-verbal variables or more global communicative behaviours like praise and reprimand (Harris & Rosenthal 1985).

I The research reported in ~is article was made possible by· a grant from the German Science Foundation (DFG). 220 A research project in· the Heidelberg/Mannheim Collaborative Research Group "Speech and Situation," which investigates "social identity and partner hypotheses in conversations" has been mainly focussing on the aspect of the verbal communication of partner hypotheses. We want to know more. about communicative behaviours that occur in dyads in which a person A has hypotheses, assumptions, beliefs, expectations about characteristics or possible reactions, etc. of a partner B-hypotheses which B is not aware of.

1 Interaction strategies in asymmetrical interaction Asymmetrical interactions are part of everyday work situations and are therefore relevant and important to mast of the adult population at same time in their lives. Work-place interaction involves a specific type of talk which very often includes typical routine phrases connected with the type of work. Communi­ cation at the work-place is also largely determined by the specific status hierarchy and the sex of interacting partners (Fine et al. 1987). Ever since Brown & Levinson's (1978) study, status differences have been regarded as one of the central dimensions in asymmetrical interactions. Status asymmetry implies a difference in social power and consequently a different distribution of conversational rights (Erickson et al. 1978; Bradac & Wisegarver 1984; Holtgraves, Srull & Socall 1989). Recent research on power and asym­ metrical interaction has shown that two different aspects have to be considered: power and interactional dominance (Linell, Gustavsson & Juvonen 1988; Thimm 1990; Thimm & Kruse 1991). Being in power in asocial relationship is different from exhibiting dominant behaviour. Power may be defined as a potential for exercising influence over other people's actions, decisions and thoughts, whereas interactional dominance cancerns manifest properties of discourse (Linell et al. 1988). Different speech markers have been discussed in the light of power-related talk. For example, speech rate has been described as an important factor in the formation of impressions. Same studies showed faster speaking people to be more convincing and more competent and attractive (Street, Brady & Putnam 1983). Other relevant aspects of powerful/powerless talk are the amount of speech (Linell & Jönsson 1990) and the length and placing of pauses (O'Greene, Lindsey & Hawn 1990). An important influence on verbal behaviour is the type of situation. Herrmann (1982), Winterhoff-Spurk & Grabowski-Gellert (1987b) classify situations in terms of the legitimacy of the requesting speaker and the willingness or ability of the hearer who has to respond to the request: situations in which speakers feel their requests are legitimate and the hearers are willing to comply are called standard situations (SS). A variation to the standard situation is the risk of reactance, in short, a reactance situation (RS), which exists when the speaker considers her/his legitimacy to ask for something to be high, but is not so sure 221 about the hearer's willingness to comply, or even assumes that the hearer is reluctant to comply. Previous research (Dom-Mahler et al. 1989) has shown that speakers in SS (e.g. superiors vis-a-vis subordinates) used more indirect requests, interrogative or statement intonation accompanied by a half-smile, whereas in RS more direct requests (e.g. "please make some coffee now") will be used, which then, how­ ever, were softened by a full smile. Further results have shown that, in general, a more direct, explicit request, although presented in a friendly and polite manner (smiling, intonation) is most effective: targets of the request reported more positive emotional evaluations and were able to reproduce more elements of the request. The authors interpret the kind of speech behaviour which is either most effective in SS or in RS as a combination of both task- and employee-oriented behaviours. A situation in which a speaker might expect resistance from his/her partner could be a "face-threatening" situation (Goffman 1972). Empirical research has demonstrated that in particular speech which is more direct is considered "face­ threatening", while less forceful requests reflect the speaker's concern with the hearer's possible of face (Holtgraves et al. 1989). The importance of speech style in work-related asymmetrical interactions has been investigated in several studies. Steffen & Eagly (1985) found that high status persons were assumed to use a more direct and impolite style and were also thought more Iikely to gain compliance by using this style. Lower status subjects were more concerned with face-saving, and also perceived the style of their partner's talk as more direct and less polite. Both studies showed that softening and politeness strategies were directly related to status: the higher the status the more direct and less polite the style of talk was perceived to be. According to the politeness mode! of Brown & Levinson (1978), and research by Penman (1990), however, one would expect that higher status interactants should be less concerned with attending to the face of a lower status interactant. Another type of action to be considered in connection with communication in asymmetrical situations is compliance gaining. What, for example, happens when a boss expects resistance from an employee whom he believes to be reluctant to carry out his requests? How does this specific expectation about his partner affect his verbalisation of the request? Compliance-gaining strategies have been researched under various aspects (Marwell & Schmitt 1967; Miller et al. 1977; Barnicott & Roloff 1978). These authors describe specific types of speech acts which serve to achieve control over the relationship and/or the interaction. One result seems especially noteworthy: Barnicott & Roloff (1978) showed that prosocial strategies such as praise were used more often by so called "high Macchiavellians" to reach their goal than antisocial strategies such as direct commands. 222 2 The material: work-place interaction The material for this study consists of an experimental rolecplay, "The Miller file", focussing on communication between a boss and his/her subordinate in a work-place setting. The data were taken from audio- (and video-) tapes, produced by another team in our Collaborative Research Group, who were mainly interested in studying nonverbal concommitants of request behaviour (Winterhoff-Spurk et al. 1987a). For the experimental role-play subjects were asked to enact the following situation: they were to play a head of department who has spent the moming in a management meeting and now is retuming to his office, which he shares with his assistant, Ms Dom (a confederate). The subjects were told that they should inform her first about news from the meeting (there will be new fumiture for the office) and then tel1 her that the Miller file, which they have routinely already given to her, has to be completed today. It is one hour before Ms Dom would normally leave the office. The confederate was instructed to say very little, just enough to keep the conversation going. Two different situations were compared: the first (situation I) is meant to be a standard situation, while the second situation (situation Il) should correspond to a reactance prone situation. The subjects received the following instructions: Situation I: "It is easy to complete the file within the time given. Ms Dom knows the case. She has the file." Situation Il: "It is not quite certain whether the file can be completed within normal working hours or whether it requires overtime. Ms Dom knows the case. She has the file." 41 male students aged 24-39 served as subjects. They had previoulsy worked in a state employment office and were now attending a Civil Service College. 20 subjects played situation I, 21 played situation Il. Two types of speech pattems had to be produced: "giving information" (inform someone about new fumiture, i.e. something pleasant) and "requesting". In situation I this is a routine request, whereas in situation Il it is implied that the assistant might be forced to work overtime. Situation Il can be regarded as a more unpleasant task and potentially face-threatening for the "boss" who is doing the requesting. This is because any resistance by the clerk would require more decisive compliance-gaining moves, which include the risk of being not successful and being seen as authoritarian. The role-playing situation was thus characterised (1) by a difference in förmal power assigned to A and B (boss and subordinate) and (2) by differing hypotheses that A was given about B's willingness to carry out a specific task. This scenario is typical of work situations and the subjects felt that it was quite easy to enact their role. 223 We were seeking answers to the following guestions: - How is the dialogue constructed in the light of different beliefs about the willingness or unwillingness of the assistant? - Which speech markers associated with powerful talk prove to be significant for this type of work-place interaction? - What strategies does a superior use to achieve compliance from an assistant? - How are the different styles of talkjudged in terms of acceptance by externa! judges?

3 Methodological approach and results Three different methodological approaches were used. Firstly, guantitative variables, which had proved to be relevant for interaction situations with power differences (cf. Thimm & Kruse 1991), were analysed and evaluated statistically for significant differences between SS and RS. Secondly, conversational anal­ ysis was used for complex speech patterns which characterise strategic interaction. Thirdly, global ratings of externa! observers were carried out to gain an impression of the behaviour of the subjects. This threefold methodological approach is in accordance with our overall interdisciplinary approach involving social psychology and Iinguistics.

3.1 Speech markers Freguency counts were calculated with chi-sguare; comparison of means was calculated with t-tests for independent samples. For analysis the texts were divided into separate sections or topics (cate­ gorised independently by three coders) which were analysed for those speech markers which had proved relevant for asymmetrical interaction in previous research. The following sections were differentiated: - greeting - lead-up to the topics furniture and Miller file - topic furniture - topic Miller file (= reguest section) - evaluation, conclusion - remainder · These sections were analysed for the following guantitative categories: - overall speech rate - speech rate per section - amount of words per section and total - amount of time per section (in seconds) 224 - length of pauses before changing to another topic (pause 1 in seconds) - length of pauses within sections (pause 2 in seconds) - total length of pauses per section For this part of the analysis significant results were found in only two of the sections: the greeting and the evaluation phase. Greeting encompasses the first encounter of the interactants, the evaluation section means the concluding part of the dialogue. In some cases this was characterised by an evaluation of the situation such as, "I think that is all. Do you have any more questions?" In the standard situation the subjects used significantly more words to greet the assistant (z= -2.35, p<.02). In the evaluation section the speech rate yielded a significant difference: people in RS talk much faster than those in SS (z= -3.8, p<.001). Subjects in SS, on the other hand, use more time (z=3.06, p<.003) and more words (z=l.93, p<.06). These results are in accordance with other research findings. The increased speech rate can be taken as a sign of feeling inferior or threatened. Subjects engaging in this style of talk try to avoid interference or interruptions by their interaction partners. Yet in sum, these quantitative parameters did not produce clear-cut differences.

3.2 Requesting The analysis of the request section, however, revealed more results. Subjects varied in the way they realised the request to finish the file. For this reason, the request section of the texts was analysed using both qualitative and quantitative methods. An analysis of three different types of syntactical forms was carried out: statements, questions, and commands. Statements were utterances such as, 'Tonight the file is available," questions were, for example: "Could you com­ plete the file today?" and commands were short forms such as: "Complete the file today." In the standard situation we found 37 request utterances, in the reactance situation 62. The following distribution was obtained:

Table 1. Request utterances (in absolute numbers)

Svntactical form Standard Situation Reactance Situation

Statements 14 39 Questions 11 15 Commands 12 8 Total 37 62 225 In the reactance situation, subjects clearly preferred statements (2/3 of the utterances) to questions, while commands were rare (chi-square=7.36, p<.03). Hence the strongest and most direct forms of request are used less often in RS. The lack of command-like requests confirm the result of Barnicott & Roloff (1978), that antisocial moves, which might threaten the face of the other person, are not seen as very dysirable, because they would shed a negative Iight on the speaker her/himself. One interesting result was that the people in the reactance situation made more attempts to realise the request-they mentioned the "Miller file" repeatedly, with another topic in between. The topic "Miller file" was addressed a second time by 5 subjects in SS and by 13 subjects in RS (who also used more words and more time). It was addressed a third time by only one subject in SS, but still by 8 in RS. Although subjects in RS addressing the Miller file for a second time used significantly more words and more time, they altogether did not use more words and time for the request section, but just split it up into several pieces. This can be interpreted in the light of strategic interaction (see "chunking" below).

3.3 Strategic interaction Quantitative measures of the communication structure and a rather rough initial syntactical analysis of the request section yielded a number of interesting but still relatively weak indications of particular speech behaviour in standard and reactance-prone situations. Thus we considered it useful to apply analytical procedures derived from discourse analysis, focussing on interaction strategies that may be used in face-threatening situations. Strategic interaction can be regarded as central to reactance prone situations. We define strategy as a sequence of speech patterns serving the purpose of reaching the interaction goal of the speaker in a situation of actual or perceived reactance. We distinguish type of strategy from types of strategic moves which serve to realise the strategy in the context of the verbal interaction. The strategy itself is named according to the goal aimed for. The exact speech patterns, i.e. strategic moves, are analysed in relation to the strategy (Thimm 1990). To analyse strategic interaction in the context of asymmetrical work-related communication the following strategies and their concommitant strategic moves were analysed: · 226 Table 2. Analysed strategies and strategic moves Goals Strategy Strategic moves avoid, avoidance strategy delegation, changing prevent topic, vagueness, mentioning externa! sources, chunking maintain relationship-securing personal addressing, relationship, strategy confirming,making sure, secure the idiomatic phrasing interaction get person to cooperative strategy compliment, praise, cooperate offer compensation establish or power strategy order, threaten, confirm power mention hierarchy over others

3.3.1 Avoidance strategy Certain strategic moves help in a potentially face-threatening situation, either to reduce conflict or prevent it from arising at all. Avoidance of confrontation is one of the most effective strategies, because then nobody has to repair the damage. This was realised in three types of strategic moves: a. Delegation Avoidance is mostly realised by delegation of responsibility, i.e. the speaker distances himselffrom the problem and thus cannot be held responsible. Delegation is typically expressed by such phrases as "I was told" or "it was decided that. .. " Similar numbers of delegation moves occurred in both conditions: in 11 cases out of 20 in SS and in 13 out of 21 in RS. Delegation of responsibility obviously seems important to subjects in both conditions. However the way people delegate is quite different. One particular strategic move, for example, was only found in RS (3 out of 13): referring to an externa! authority, usually his own superior, e.g. "my boss told me", "I was approached by my boss." This implies that the partners are on an equal footing, both subordinate to another boss. Interestingly, 3 out of the 8 subjects in RS who did not delegate responsibility used the opposite tactic of making the task sound personally important to them. ("This is really important to me.") 227 b. Chunking Another way to get around a tricky situation is to approach it step by step and mentioning other topics in between. We called this approach "chunking". As mentioned above (requesting) the topic "Miller file" was addressed by significantly more subjects fora second and even a third time in RS than in SS. This means that they present the request in little pieces and thus try to avoid a confrontation. The topics they mention in between are pleasant topics, like the new furniture. This may be interpreted as a gradual approach to a potentially face-threatening request, a kind of softening or mitigating technique. c. Vagueness To make the request sound less direct and therefore less pressuring, subjunctive verb forms were used (in German there is a stylistic difference between the indicative and the subjunctive, the latter sounding less direct, more like "Could you maybe ... "). Although statistical evaluation did not show significant differences, a high standard deviation in RS (3.0705), indicated a wide variation within the RS group, which was later corroborated by other results. There seem to be two ways of coping with the situation. Subjects either delegate responsi­ bility, sounding vague and keeping the whole interaction short, or they do the opposite by sounding personal, engaging in dialogue, using a colloquial style or offering compensation, e.g. time off in Iieu.

3.3.2 The relationship-securing strategy To maintain personal contact and a certain leve! of intimacy interactants use various stategic moves (Cody, McLaughlin & Schneider 1981). Continuing an interaction always requires a certain degree of cooperation between the partners. It is also a prerequisite before the speaker can get the addressee to comply. In our material the following strategic moves to secure the relationship were found: a. Forms of address The choice of personal pronouns when addressing helps to define the relationship (Brown & Gilman 1960). Some of the "bosses" were trying to reduce the asymmetry: they used the personal form Du instead of the forma! form Sie (both meaning 'you') and sometimes even say we. Sie was used signifcantly more often in the RS (p<.02), indicating that the assistant was more often addressed directly. The we form was used mostly in the description of the task (e.g. "we have to finish the Miller file today"). Again, a high standard deviation (s=3.07) points to variation within the RS group: some employ a general strategy of making their communication sound personal and consequently use we very often, while others stay very forma! and businesslike, choosing to delegate responsibility. 228 Also the absence of addressing is noteworthy. While nota single subject in the RS failed to address the clerk, only 4 did so in the SS. b. Colloquial style Office talk is usually task-oriented and characterised by forma! speech. Yet previous research has also shown that shifting from a more forma! to a more colloquial style is a feature of powerful language (Scotton 1985). We wanted to investigate whether the "boss" engaged in purely forma! style (office talk) thus keeping himself at a distance from his assistant, or whether he tried to sound personal and easy-going by using a colloquial style. Analysis of the degree of colloquiality revealed interesting differences. In RS 37 colloquial phrases were used, in SS only 19. Some subjects in SS used their own (student) idiomatic style such as, "What a rotten day," hut in most cases the idiomatic phrases or colloquial markers were added to what was otherwise forma! speech. While subjects in SS felt quite free to mention the status asymmetry explicitly, in RS formulations such as "all paletti", "that would be just wild", '\ve must be able to pull this off," and certain abbreviations and dialect forms were the most characteristic. Typical for the RS was also the amount of such items used by individual subjects: the 37 phrases were produced by only 13 of the 21 subjects. This again confirms the notion of two separate styles of dealing with the situation. Idiomatic phrases, the use of dialect and a certain degree of looseness can be seen as important factors for maintaining a personal relationship and softening the request.

3.3.3 Cooperative strategies Getting someone to do what you want relies essentially on the compliance of the other person. The goal of compliance gaining can be reached by means of several different strategies. Most of our subjects in the reactance situation go to quite some effort to make the compliance as easy as possible for the assistant. The following strategic moves could be isolated: a. Cooperative questions In order to give the assistant the feeling of having influence, some subjects offer her a decisive say in connection with the choice of the new furniture by asking what she would prefer. Anothe.r form of cooperative questions is related to the task itself. In RS more subjects (6 compared to 1 in SS) asked "real" questions as to whether Ms Dom knew the case, whether she had everything she needs and/or whether any necessary information was missing (p<.02). Questions which require an answer 229 from the assistant (rather than rhetorical questions) can be interpreted as attempts to engage in a dialogue. Another way of relating to the assistant is trying to make sure that she has everything she needs. These brief questions are very frequent in both conditions (13 in SS, 15 in RS) and are task-oriented ("Do you have everything you need?") and sometimes personally oriented ("Are you sure you can do it?"). The communicative intention is a reconfirmation by the assistant that everything is ok. b. Offering compensation Many subjects chose to actively mention the apparent conflict (14 in RS cf. 8 in SS). For example they say, "I don't know whether you can finish that within normal working hours," or "I know your normal working day is over." Some of these subjects make offers to the assistant to ease the situation (6 in RS, 3 in SS), thus acknowledging the personal interests of the assistant and demon­ strating respect for her preference to go home. Mostly subjects offer the opportunity to leave earlier or come in later another day: "You can come in later tomorrow then", "You can leave earlier on another.day." Two participants in RS even offer their own time or help: "Can I be of any help?", "l'11 stay here longer, too," are explicit offers to make up for the unpleasant request. Mentioning the conflict openly and offering other forms of compensation can be taken as elements of a cooperative style. The unpleasant demand on the assistant is acknowledged, but it is also made clear that the task must be completed. c. Praise Another, even more obviously cooperative move is praising the abilities of the assistant. In one case the head of department expresses his faith and trust in the capabilities of the assistant to finish the difficult task: "I have faith in you", "l'm sure you can manage that." Other subjects are less direct when mentioning the ability of the assistant: "You know the case, you know what to do." This strategic move was only used in the RS (by 4 subjetcs). The expression of praise has two sides, however. It is a good tactic to demonstrate faith in the abilites of the assistant, because it puts her under pressure and leaves her little opportunity to refuse. Also this kind of praise is a paternalistic way of treating someone, and typically voiced in hierarchical situations (such as mother-child interactions). This is not to say that compli­ menting someone on his or her ablities is not a very important part of work­ place interactions, but depending on the specific situation it can be used to get someone to do something while appearing to be cooperative. 230 d. Making it personal Another way of getting the assistant to cooperate is to show how important the requested task is and demonstrate personal interest. Such utterances can be seen as a kind of self-disclosure, e.g. "I'm showing you how much it means to me", "I would appreciate that," is typical for this strategy. Again, this strategic move does not occur very frequently, hut is used more often in RS than in SS. To give the request such a personal touch isa way of making the whole task so important that a rejection becomes more unlikely. How could an assistant seriously consider not fulfilling a task that is so important to her boss?

3.3.4 Power strategy Very direct utterances in the form of commands, threats or even accusations could be used to make an employee to do what the boss wants. Interestingly, we could not find many cases of such typical power strategy moves. One subject (in SS) employed a power strategy move, putting the assistant under pressure by asking why she had not finished the file already, implying that she was responsible for the delay. Another indicator of power strategy is the use of direct commands. In the SS 12 instances of such direct commands occured, in RS only 8. Without the risk of reactance, the bosses in SS seem to feel more self-confident and more sure of their powerful position. This notion is confirmed by another indicator. Looking more closely at evaluation phrases, we found that more than 50% of the subjects in SS summed up the situation by referring to the work situation, hut only 20% of subjects in RS. They use phrases like "Then we can continue with our regular work", "Is there anything else urgent?" and "Did any problems come up this morning?", thus addressing the work situation itself. It is not surprising that subjects in RS tend to avoid mentioning the sensitive work situation. Being more open about the existing hierarchy is typical of the standard situation.

3.4 Semantic differential The final procedure for the analysis of the role-playing situation is a method common in social psychological research into partner hypotheses. Global ratings of the whole communication situation by externa! observers is quite often used to assess the effectiveness of a partner hypothesis in a given situation. In our study we had 20 students listen to the audio-tapes of the conversations and rate them on a semantic differential (= 12 bipolar adjectives on a 7-point scale). Significant differences (p<.01) between the ratings of SS and RS were found for 7 adjective pairs (see figure 1). 231 -3 -2 -1 0 +1 +2 +3

friendly unfriendly

authoritarian egalitarian

rushed calm

pleasant unpleasant

active passive

secure insecure

clear confused

• standard situation o reactance situation

Figure I. Significant differences in the judgement of standard and reactance situation

We found that speakers in SS are rated as a Iittle less friendly than speakers in RS, also as authoritarian rather than egalitarian, as more rushed, unsympathetic, hut also as more active, sure and clear than speakers in RS. Interpreted in terms of politeness theory this means that subjects in SS focus on being understood rather than being polite.

Discussion The aim of our research project was to find differences in the speech behaviour of subjects who were asked to perform two different role-play situations ending with a request, in a work-place setting. The first role-play condition was a standard situation in which subjects felt that their requests to the subordinate were legitimate and also knew that the subordinate was willing or able to comply. In the second condition, called a reactance-prone situation, the subject again considered his legitimacy to ask for something to be high, hut was not so sure whether the subordinate would comply .or perhaps even refuse to fulfill the task. The reactance prone situation could also be described as a face-threatening situation motivating the requesting subject to be concerned about the hearer's face. 232 Across our three methodological approaches (quantitative analysis of speech markers, analysis of conversational strategies and evaluation of the verbal behaviour by externa! observers) we found that speakers in the standard situation don't go to much trouble to get their way. They use a more direct and more powerful style to verbalise the request. They address work-related issues more clearly and don 't hesitate to point to the status hierarchy. In accordance with these findings externa! observers judged speakers in SS to be more authoritarian, more unsympathetic and more rushed. On the other hand, in the reactance-prone situation, speakers try to mitigate the situation in order to diminish the risk of reactance or resistance from the hearer. Subjects use a less direct and less forceful style, they avoid command­ like requests and prefer more indirect forms of requesting (statements and questions). The subjects in SS mention the task significantly more often than subjects in RS, applying the technique of chunking, i.e. presenting the request in little pieces. Further analysis of interaction strategies demonstrated a tendency in RS to use more vague and impersonal forms for the whole dialogue in order to avoid the responsibility for the situation by referring to externa! sources. On the other hand, speakers in RS quite often engage in cooperative behaviours such as asking questions to ensure that the assistant can fulfil the task, offering compensation to makeup for the unpleasant situation, and/or offering participa­ tion in decision making in other matters (deciding about new furniture for the office). A doser look at interaction strategies in RS reveals two different styles which subjects may adopt when dealing with a situation of potential conflict: some subjects choose to remain forma! and try to conclude the sensitive topic as quickly as possible. Others prefer a more cooperative partner-directed style to secure the socio-emotional relationship with the subordinate. They choose a more personal style, characterised by idiomatic phrases and frequent addressing of the assistant. It is the subjects in RS who ask more cooperative questions and who offer compensation. Although these differences in style cannot be secured statistically, it nevertheless appears useful to follow up the notions of face­ management and compliance-gaining in asymmetrical work situations. To study complex and highly varied communicative speech, a single-method approach is not sufficient. What is called for is an interdisciplinary multimethod approach combining qualitative linguistic measurements, conversational analyti­ cal approaches as well as methods from experimental social psychology.

References Barnicott, E. & Roloff, M., 1978: The situational use of pro- and antisocial compliance gaining strategies by high and low Macchiavellians. Communi­ cation Yearbook, 2: 193-204. 233 Baumgardner & Brownee, 1987: Strategic failure in social interaction: Evidence for expectancy disconfirmation process. Journal oj Personality and Social Psychology, 53(3):525-535. Bradac, J.J. & Wisegarver, 1984: Ascribed status, lexical diversity, and accent: determinants of perceived status, solidarity, and control of speech style. Journal oj language and Social Psychology, J:239-255. Brown, R. & Gilman, A., 1960: The pronouns of power and solidarity. In Th.A. Sebeok (ed.), Style in language. Cambridge, MA, Massachusetts Institute of Technology, pp 253-277. Brown, P. & Levinson, S., 1978: Universals in language usage: Politeness phenomena. In E. Goody (ed.), Questions and politeness. Cambridge, Cambridge University Press, pp 56-324. Cody, M., McLaughlin, M. & Schneider, M., 1981: The impact of relational sequences and intimacy on the selection of interpersonal persuasion tactics: A reanalysis. Communication Quarterly, 29(2):91-106. Dom-Mahler et al., 1989: Intonation bei Aujjorderungen. Teil 2: Eine experimentelle Untersuchung. (Arbeiten aus dem SFB 245 "Sprechen und Sprachverstehen im sozialen Kontext." Heidelberg/Mannheim, Bericht Nr.8.) Mannheim, Universität, Lehrstuhl för Psychologie III. Erickson, B. et al., 1978: Speech style and impression formation in a court setting: The effects of 'powerful' and 'powerless' speech. Journal oj Experi­ mental Social Aychology, 14:266-279. Fine, M.G. et al., 1987: Ethical issues in defining and evaluating women's communication in the workplace. In L. Stewart & S. Teng-Toomey (eds.), Communication, Gender and Sex Rotes in Diverse Interaction Contexts. Norwood, New York, Eilex Publishing Co. Galtung, J., 1959: Expectations and the interaction process. Inquiry, 2:213-234. Goffman, E., 1972: On face-work: An analysis of ritual elements in social interaction. In J. Laver & S. Hutchinson (eds.), Communication in Face to Face Interaction. Harmondsworth. pp 319-346. Harris, M.J. & Rosenthal, R., 1985: Mediation of interpersonal expectancy effects: 31 meta-analyses.Psychological Bulletin, 97:363-386. Herrmann, Th., 1982: Sprechen und Situation. Berlin, Springer. Hilton, J.L. & Darley, J.M., 1985: Constructing other persons: A limit to the effect. Journal oj Experimental Social Psychology, 21: 1-18. Holtgraves, T., Srull, Th.K. & Socall, D., 1989: Conversation memory: The effects of speaker status on memory for the ;assertiveness of conversation remarks. Journal oj Personality and Social Psychology, 54: 149-160. Jussim, L., 1986: Self-fulfilling prophecies: a theoretical and integrative review. Psychological Review, 93:429-445. Linell, P., Gustavsson, L. & Juvonen, P., 1988: Interactional dominance in dyadic communication: a presentation of initiative-response analysis. Linguis­ tics, 26:415-442. 234 Linell, P. & Jönsson, L., 1990: Suspect stories: On perspective setting in an asymmetrical situation. In I. Markova & K. Foppa (eds.), Asymmetries in dialogue. Hemel Hempstead, Harvester Wheatsheaf. Marwell, G. & Schmitt, D., 1967: Dimensions of compliance-gaining behavior. An empirical analysis. Sociometry, 30:350-364. Merton, R.K., 1948: The self-fulfilling prophecy. AntiochReview, 8:193-210. Miller, G. et al., 1977: Compliance-gaining message strategies. Typology and some findings concerning effects of situational differences. Communication Monographs, 44:37-51. · O'Greene, J., Lindsey, E. & Hawn, J., 1990: Social goals and speech production: Effects of multiple goals on pausa! phenomena. Journal oj Language and Social Psychology, 9: 119-134 Penman, R., 1990: Face work and politeness: Multiple goals in courtroom discourse. Journal oj Language and Social Psychology, 9(1-2): 15-38. Rosenthal, R. & Jacobson, L., 1968: Pygmalion in the Classroom. New York, Holt, Rinehard & Winston Scotton, C.M., 1985: What the heck, sir: Style shifting and lexical coloring as features of powerful language. In R.L. Street & J.N. Capella (eds.), Sequence and Pattern in Communicative Behavior. London, Arnold, pp 103-119. Snyder, M., 1981: Seek, and we shall find: Testing hypotheses about other people. In T.T. Higgins, C. Herman & M.P. Zanna (eds.), Social Cognition. (The Ontario Symposium, Vol. 1.) Hillsdale, New York, Erlbaum, pp 277- 304. Snyder, M., 1984: When belief creates reality. Advances in Social Psychology, 18:62-113. Steffen, V. & Eagly, A., 1985: Implicit theories about influence style: The effects of status and sex. Personality and Social Psychology Bulletin, 11(2):191-205. Street, R.L., Brady, R.M. & Putnam, W.B., 1983: The influence of speech rate, stereotype and rate similarity on listeners' evaluations of speakers. Journal oj Language and Social Psychology, 2:37-56. Swann, W.B., 1987: Identity negotiation: Where two roads meet. Journal oj Personality and Social Psychology, 59(6): 1036-1051. Thimm, C., 1990: Dominanz und Sprache-Zur Beschreibung strategischen Handelns im Alltag. Wiesbaden, Deutscher Universitätsverlag. Thimm, C. & Kruse, L. 1991: Dominanz, Macht und Status als Elemente sprach!icher Interaktion. (Arbeiten aus dem Sonderforschungsbereich 245 "Sprechen und Sprachverstehen im sozialen Kontext." Heidelberg/Mannheim, Bericht Nr.39.) Heidelberg. Winterhoff-Spurk, P. & Grabowski-Gellert, J., 1987a: Nonverbale Kommuni­ kation und die Direktheit von Direktiva: Der Ton macht die Musik! Sprache und Kognition, 3: 138-149. 235 Winterhoff-Spurk, P. & Grabowski-Gellert, J., 1987b: Vom Lob der Direktheit: Wirkungen sprachlicher Merkmale des Ftihrungsverhaltens. Zeitschrift fiir Arbeits- und Organisationspsychologie, 31 :55-62.

The Interaction in Conversations Between Aphasic Persons and Speech Therapists 1

Minna Laakso

The study of aphasic persons' actual communication is relatively new. Despite the disordered language ability, functional communication skills are said to be well preserved in aphasia (Holland 1991:199). Variation has been observed in aphasic persons' speech between the clinical setting and their natura! environ­ ment, and aphasics are found to make appropriate linguistic changes in response to nonlinguistic social contextual variables (Glasser et al. 1988: 122). Neverthe­ less, aphasics have obvious problems in conversation due to the difficulties they have in, e.g., finding words or constructing sentences (Glasser & Deser 1990:67). As a result, aphasics' utterances may be erroneous or structurally incomplete, which makes them difficult to understand. In addition to the language problems, conversational problems may derive from conversational­ ists' different goals and expectations concerning the situation or the topic of conversation. Previous studies report that aphasia therapists spend a considerable proportion of the therapy session conversing with their clients (Armstrong 1989:3). In spite of this, conversation is not usually considered as apart of the therapy, and the features of the interaction between a language-disabled person and a language expert are relatively unknown. The few observations that have been made stress the importance of the nonaphasic conversational partner. For example, Feyereisen (1991) argued that the communicative effectiveness of an aphasic person depends on support by an attentive healthy listener, so that the whole conversation with an aphasic relies on active participation of nonaphasic participants. Also Foldi et al. (1983) have observed that the verbal and gestural referents of aphasics rely on the surrounding context and the dialogue of the other conversant, and that the conversations are frequently clarified by non­ aphasic participants. Hence, aphasics' communication cannot be evaluated without taking the interaction into account. This study will first concentrate on some general patterns of conversational interaction between an aphasic person and spccch therapist. After that, the interaction in repair sequences will be discussed in more detail.

1 This study was funded by the Academy of Finland as apart of the project "Communication and Speech Processing in Aphasia", lead by professor Matti Leiwo. 238 1 Material Six posterior fluent aphasics (three female and three male) were videotaped while each were conversing with her or his speech therapist. The background data for the aphasic subjects are presented in Table 1. Aphasic subjects are listed according to the severity of aphasia, which was measured with the Finnish version of the Western Aphasia Battery (WAB), the first person in the table having the most severe aphasia.

Table I. Background information of the aphasic subjects

Subject Sex Age Duration Type of aphasia AQ Comprhn score of illness max HJO max 10

1. N. m 63 2mos Wernicke 19.8 0.6

2.H. m 33 3 mos Wernicke 38.5 5.4

3.S. f 47 2mos Wernicke 42.l 6.8

4. L. f 50 9mos Wernicke 45.2 6.3

5.M. m 61 5mos Transcortical 48.8 4.0 sensory 6.Z. f 55 7mos Conduction 55.8 7.0 (WAB = Western Aphasia Battery; AQ = Aphasia Quotient that reflects the severity of aphasia; Comprhn score = the score in the comprehension subtests of WAB.) The aphasic subjects had all suffered a left hemisphere cerebrovascular accident at least 2 months prior to their participation in the present study. Four of the aphasics had Wernicke's aphasia, and two of them had initially been classified as Wernicke's aphasics, hut following recovery their aphasia type had shifted to transcortical sensory in one case and conduction aphasia in the other. The severity of aphasia varied: subject N. had a very severe aphasia with almost nonexistent skills in the comprehension tasks of the W AB test. The four moder­ ate aphasics (H., S., L. and M.) were quite close to each other in the severity scale. Subject Z. had moderate to mild aphasia and rather well preserved comprehension according to the aphasia test. All aphasics were characterized by fiuent speech. No one had superfluent and unintelligible jargon aphasia, but all expressed many phonemic and semantic approximations in their speech. They also exhibited distinct word-finding problems. The circumlocutory speech of the most severely aphasic subject N. could be described as 'lexically empty speech' (cf. Nicholas et al. 1985:407) with recurring empty idioms, indefinite terms, 239 deictic terms, pronouns without antecedents, and conjunctions, and he also seemed to be somehow unaware of the emptiness of bis speech. In their jobs the aphasics presumably had talked in different quantities; four (N., S., L. and Z.) bad elementary school education, H. bad college education, and M. bad a university degree. All the speech therapists were female, 25-35 years of age with university education. Four of them had therapy experience of between one and five years and two of between six and ten years. The topics discussed in the dialogues were not determined in advance. In practice, however, topics were usually selected by therapists and were contemporary issues or topics familiar to the aphasic. Therapists also possessed previous knowledge on the topics; supposedly the same topics bad been discussed earlier in the course of the therapy.

2 Method From the middle of each conversation a five-minute segment was extracted for quantitative analysis, totalling 30 minutes. Conversational segments were transcribed and divided into conversational tums (for criteria, see Appendix). The speech time and the number of conversational tums, words and other speech units of each participant in the five-minute conversations were measured in order to examine the distribution of interactional space between participants. The tum-exchange characteristics were examined by counting the occurrences of simultaneous talk in the middle and at the end of the overlapping utterances. Speech functions within tums were identified and classified into six general types considered to be the most typical in these conversations on the basis of the raw analysis of the data. The functions were: questions, clarification requests, comments, extended answers, answers, and minimal acknowledgements (for definitions, see Appendix). The trouble sources in speech production were counted and conversational sequences used for dealing with trouble were examined for the means and amount of repair behaviour (see Appendix).

3 Results

3.1 Speech time, the number of conversational turns and words/speech units A rough measure of the use of interactional space was obtained by examining the distributions of speech time and conversational tums (Table 2). Surprisingly, all aphasic-therapist dyads demonstrated rather similar proportionate distri­ butions despite the individual differences between participants. In all conversa­ tions, aphasics had more speech time but fewer tums; therapists had less speech time but more and shorter tums. Conversation 5 was the clearest example: the 240 transcortical sensory aphasic had a significantly greater proportion of the speech time in this conversation, apparently partly because of his slower speech rate. In general, conversations were particularly characterized by aphasics' long, and therapists' short, utterances.

Table 2. The percentage distribution of speech time and conversational tums between aphasic and therapist in each 5-minute conversation

Conversation Speech time Conversational tums a t total a t % % n % % Cl 64.5 35.5 (71) 48 52 C2 63 37 (98) 45 55 C3 62 38 (101) 49 51 C4 63 37 (130) 48 51 C5 82 18 (32) 41 59 C6 62 38 (62) 39 61 (a = aphasic; t = therapist; conversations are presented in the order of aphasia severity as the background data in Table 1, so that conversation Cl is between the most severe aphasic N and his therapist.) In accordance with their bigger proportion of speech time, aphasics also used more words and other speech units (nonwords, segments of words, sound prolongations and interjections) than therapists in the examined five minutes (Table 3).

Table 3. The percentage distribution of words and other speech units between participants and the percentages of other units of all speech units spoken by each participant in 5-minute-conversations

Conversation All speech units Other units/all a t total a t % % n % %

Cl 65 35 (620) 2.5 C2 62 38 (655) 20 5 C3 59 41 (604) 12 1 C4 55 45 (612) 9 2.5 C5 67 33 (299) 24 2 C6 62 38 (6582 5 4.5 (a = aphasic; t = therapist; All speech units = words and other units; Other units/all = percentages of nonwords, parts of word, interjections and sound prolongations of all speech units.) 241 On the whole, aphasics produced more nonwords, segments of words, interjections and sound prolongations than therapists. With two aphasics, in conversations 2 and 5, over 20 percent of their output consisted of unintelligible words and hesitation; whereas in conversation 6 the milder aphasic Z. and her therapist produced other units than words to the same extent.

3.2 Simultaneous talking All aphasics demonstrated awareness of turn-taking and their conversational role in the interaction with their partners. Simultaneous talking (see Table 4) took place in two different positions in the conversations: in the middle of the utterance of the current speaker before a relevant turn transition place, and at the end of the utterance of the current speaker where turn transition already could be anticipated. Part of the simultaneous talking in the middle of the current speaker's speech consisted of acknowledging minimal responses and could be considered cooperative in nature (interactive simultaneous talking, Table 4). Sometimes participants tended to interrupt and start to talk simultaneously although the current speaker had not yet finished her/his speaking turn (interruptive simultaneous talking, Table 4). Simultaneous talk in the middle of the utterance thus occurred in two different forms: interactive and interruptive.

Table 4. The occurrences of simultaneous talking initiated by aphasic and therapist in each 5-minute conversation

Conversation Total middle end internet interrupt all a t a t a a a t

Cl 9 10 2 6 7 4 6 8 2 2 19 C2 14 12 2 4 12 8 14 8 0 4 26 C3 28 10 2 3 26 7 27 9 1 38 C4 20 17 7 8 13 9 20 16 0 37 C5 2 0 2 0 1 2 0 0 3 C6 9 18 2 11 7 7 8 16 2 27

Total 81 69 15 34 66 35 76 59 5 10 (a = aphasie; t = thernpist;total = total amount of oeeurreneesof simultaneoustalking in five minutes per person; middle = oeeurreneesin the middle of partner's tum; end = oceurrenees in the possible end of partner's tum; internet = interaetive simultaneoustalking; interrupt = interruptivesimultaneous talking.) Conversations differed in the total amount of occurrences of simultaneous talking. In four conversations (Cl, C2, C4 and CS) both participants initiated simultaneous talking rather equally, in the two others either the aphasic (S.) or the therapist (T z) overlapped more frequently. For the total amount of overlap, 242 conversation 5 differed from the others: there were only three overlaps, and none of them were interruptive. Therapist T m did not interfere in the aphasic's speech in the middle of the tum although the aphasic exhibited many nonwords and hesitation. In conversation 2 where the aphasic had the same kind of problematic speech, the therapist acted in a different manner: she tended to interrupt the aphasic (four times in five minutes) and take the speaking tum. This resembles the finding of Glindemann (1990) where aphasics were found to lose their speaking tums easily when they had problems with speech. Interrup­ tions reflect the therapist's tendency to control the conversational flow, hut at the same time these interruptions often aimed at helping the aphasic and repairing the conversation. In conversations 3 and 4 (both female-female dyads) simultaneous talking was frequent hut interactive, consisting mostly of acknowl­ edging minimal responses to the current speaker. On the whole, interruptive overlap was not very common, in conversations 1, 2 and 6 the therapist inter­ rupting the aphasic twice or more in five minutes. In addition, the most severe aphasic (N.) interrupted (to request clarification) more than once. When studied in detail, the overlaps of the nonaphasic participants were often completions, suggestions for interpretations, acknowledgement tokens, and clarification requests.

3.3 Speech functions The distributions of six speech functions are presented in Figure 1, where the aphasic subjects and the therapists are compared as groups to each other. In four functions differences between goups were statistically significant according to the analysis of variance (Kruskal-Wallis): Therapists made significantly more questions and requests for clarification, whereas aphasics answered (i.e., had answers and extended answers) significantly more often. As a whole, aphasics frequently made answers, extended answers, comments, and minimal responses. Therapists contributed minimal responses, clarification requests and questions most frequently. Only 3% of aphasics' conversational tums had an initiating and soliciting function (i.e., were requests for information and clarification) com­ pared with 43% of therapists'. On the other hand, 79% of aphasics' and 48% of therapists' speech functions were responsive (extended answers, answers and minimal responses). Moreover, 72% of aphasics' responsive tums were answers, whereas the majority (91 %) of therapists' responsive utterances were only minimal acknowledgements. The distributions of speech functions suggest that conversations are organized as adjacency pairs of therapist's questions and aphasic's answers. This was confirmed when conversational sequences were examined. The whole interaction was sustained largely by questions from the expert which is typical in institutional discourse (Linell et al. 1988). 243

25

n.s. 20

15

10 ...

5

0 1 2 3 4 5 6

0 aphasics LJtherapists ** = p<.01 I = questions; 2 = clarification requests; 3 = comments; 4 = extended answers; 5 = answers; 6 = minimal acknowledgements Figure 1. The mean distributions of aphasics' and therapists' speech functions in 5-minute conversations

Speech functions (Figure 1) can also be viewed in another way: as functions that focus on the regulation of interaction ( clarification requests and minimal responses), or as functions that develop the content of conversation (requests for information, comments and extended answers; cf. Sorjonen 1988). 25% of the speech functions of the aphasics were interactional (mostly minimal responses) and 44% focused on the content; for therapists, the corresponding percentages were 69% and 29%. By frequently giving minimal acknowledgements (mm, yes, yeah) and nodding while the aphasic was talking, therapists obviously indicated their active participation. This displayed the same attentiveness as their usually intensive gaze when the aphasic was talking. Therapists also typically used minimal acknowledgements to close sequences and topics (Example 1). 244 Example I (Conversation 2) t: hmm=m, it (wasn't) = H: ='nglish, yes. t: yes English was also, like that. H: yes t: mmyeah.

(in Finnish) t: hmm+m, ei (ollu) = H: ='!anti, juu. t: joo englanti oli kans, sellane. H: joo t: justjoo. Conversations appear asymmetric and regulated by the therapist, since therapists frequently asked questions to elicit conversation. However, when conversations were examined sequentially, it could be noted that aphasics were active in an indirect way. They extended their answers to Jonger narrations and so developed their own themes in conversation. Severe and moderate aphasics even appeared to go on with their own stories and did not seem to understand and/or react adequately to the questions of therapists. Furthermore, to achieve the elicitation purpose, therapists' first questions on a new topic often had a typical structure; they were declaratives that require either a yes/no-answer, or alternatively a Jonger narration as Raevaara (1991) has suggested. With declaratives, the thera­ pists offered an opportunity for a long speaking tum to the aphasics. Besides using elicitative questions, therapists also modified wh-questions to simpler ones: if aphasics had difficulties in answering wh-questions, therapists reduced the structure to an alternative which the aphasic could either deny or confirm (cf. also Timonen 1989). In this way therapists supplied a mode! of answer (a candidate answer, cf. Pomerantz 1988) for aphasics. According to Pomerantz, model answers like these are typically used when a speaker wants to guide, direct, or assist a respondent. To summarize, the typical sequence structure was as follows: the therapist asks a question that makes a long response possible, the aphasic answers, and the therapist follows this long tum with short utterances (like mm hm). As aphasics often went their separate ways in their longer tums, therapists' short minimal acknowledgements were not always signalling understanding but provided a conversational mechanism to maintain a polite förmal surface consensus. As Jefferson (1984:4) has pointed out, these acknowl­ edgement tokens also exhibit "passive recipiency" indicating that the current speaker may go on talking. 245 3.4 Specific interactional features in repair sequences

3.4.1 Trouble sources All aphasics had several trouble sources that may lead to repair in their speech. The occurrences of trouble sources among aphasics ranged from 91 trouble items of subject H. to 30 of subject Z. in the examined five minutes (Table 5).

Tahle 5. The number of aphasics' trouble sources in 5-minute-conversations

Subject Pho Sem Uni Uns Cir Nonv Total

N. 9 12 27 9 10 68 H. 21 4 47 11 2 6 91 s. 4 6 15 8 5 39 L. 5 7 10 8 6 37 M. 9 0 34 6 0 0 49 z. 4 7 6 3 6 4 30

Total 44 33 124 63 19 31 314 (Pho = phonemic approximation; Sem = semantic approximation; Uni = unidentified expression; Uns = grammaticalform or semantic content unsuitable in sentential context; Cir = circumlocution;Nonv = nonverbalcircumlocution; Total = total frequencyof trouble sources.) The total number of trouble sources of each aphasic

3.4.2 Self-initiation of repair Aphasics used the typical means for self-repair initiation described in everyday conversation (i.e., self-interruptive pauses, sound prolongations, interjections and repetitions). When self-repair initiations (Table 6) were examined, pausing and repetition were found to be the most common indicators of trouble for all subjects. Some subjects also used certain other means to a greater extent: H. made false starts and comments indicating trouble, M. used sound prolongations and Z. produced interjections. 246 Tab!e6. Self-repair initiations of aphasics in 5-minute conversations

Subject Pause Prolong Interj Repelit Comment Gesture Unaccomp Total

N. 43 2 4 20 3 0 7 79 H. 29 3 4 59 20 0 20 135 s. 18 0 4 7 2 0 2 33 L. 26 3 33 0 0 64 M. 39 13 4 53 2 0 116 z. 29 1 23 34 1 0 89

Total 184 20 42 206 28 0 36 516 (Pause = pause over 2 sec.; Prolong = sound prolongation; Interj = interjection; Repetit = repetition; Comment = comment indicating trouble; Gesture = gesture indicating trouble; Unaccomp = unaccomplished, i.e., false start and a syntactic change.) As can be inferred from the great amount of self-repair initiations, aphasics made self-repair efforts when they faced difficulties in conversation. When the total amount of initiations was examined, there were clear individual differences between the aphasics: subjects H. and M. made noticeably more self-repair initiations compared to the others. However, differences between aphasics may not easily be explained by the severity of aphasia, since the total amount of self­ repair initiations of each aphasic did not correlate with the severity of aphasia measured with the Aphasia Quotient (Pearson product moment correlation coefficient r= +0.07).

Example 2 (Conversation 5)

t I think that -,,. M: l2 l 2

(in Finnish) t: musta tuntuu etta tuo -,,. M: 12 l 2 To conclude, aphasics tried the same kind of rapid self-repair as nonaphasics do, but often failed. They did detect trouble and indicated it, although posterior aphasics are described as having difficulties in monitoring their own speech. 247 However, due to their aphasic problems (e.g. in lexical retrieval and syntactic construction of utterances), they had difficulties in completing the repair. This difficulty often resulted in long repair sequences, where both participants tried to resolve the problem (long repair sequences have been described in more detail in Milroy & Perkins 1992). Sometimes aphasics did indirectly appeal for help from the therapist during their self-repair efforts, hut this was not common, as the pattern of gaze aversion suggests.

3.4.3 Other-initiation ofrepair Therapists conversing with aphasics were ready to modify their behaviour and take part in repairing the conversation. They initiated repair using all different means described in the literature on everyday conversations (general requests for clarification, interrogatives, interrogative clauses with repetition of the trouble source, repetition, offers to give an interpretation of the trouble tum; Schegloff 1987, Sorjonen 1990). They also corrected aphasics' speech directly (other-initiated other-correction), which is not usual in everyday conversation (Schegloff et al. 1977). In addition, therapists occasionally gave approving acknowledgements when the word or other item had been repaired correctly (according to their idea of correctness). This is also rather uncommon in everyday talk. Other-initiated other-correction and approving acknowledge­ ments are conversational features typical of classroom interaction. The amount and types of therapists' other-initiations of repair and the number of their direct other-repairs are presented in Table 7.

Table 7. Frequencies of other-inititation of repair and direct other-repair conducted by the therapist in the analyzed 5-minute segment Subject Huh? Wh Wh+Repetit Repetit lnterpr Direct Total 5 1 4 10 7 2 6 12 4 31 3 8 2 4 2 19 I 3 4 8 3 1 5 6 2 8

Total 4 23 4 11 27 12 81 (Subject, T x = therapist conversing with aphasic X; Huh? = other-repair initiations expressed by stating simply huh? or what?; Wh = other-repair initiations expressed by an interrogative; Wh+Repetit = other-repair initiations expressed with an interrogative and a repetition of the trouble source; Repetit = other-repair initiations expressed by repeating the trouble source or a part thereof; Intpr = other-repair initiations expressed by interpreting the trouble; Direct = other-initiated other-repairs.) 248 The total amount of therapists' initiations to repair the conversation

Example 3 (Conversation 3)

s: no but I've been in Tölvee/i looks att l ..... t: wait a mi[riute] s: [that']s like mental hospital's work look at t 2--+ t: wait a min ute, so where, is there a place like that [here] s: [it is] here, () in the corner of Tsvveedi looks att to riteille last spring and now I, I have that, looks att what is that cork [well] over ten kattonen so= t: [is it] s: =it is now only (I stay) looks att 3 ..... t: now I didn't quite unders[tand] s: [whaat] what I, looks att 4 ..... t: no but I just tried to find out where there is one here, ha- have you, been at the clinic of psychi[ atry] s: [no it] is in Mekeliini Street looks at t 5--+ t: Mekeliini Str[ eet?] s: [yes.] and Albrekt's Str[eet] looks at t 6 ..... t: [is it] this Lapinlahti s: no but there's again, but it is, the state's looks att psychiatric hospital 7 ..... t: yes yes. but this belongs to HUCH? s: yes looks att (t = therapist T s; s = aphasic S; HUCH = Helsinki University Central Hospital; Lapinlahti = State Psychiatric Hospital; nonwords are in italics.) In conversations with the more severe aphasics, interpretations were occasion­ ally used to provide a polite way out from a long and troublesome repair sequence. These interpretations were more like general statements, and attempt­ ed no specific understanding of the unclear utterance. On the whole, therapists seemed to think that they should help aphasics in handling the difficulties in conversation. Sometimes the therapist even interrupt­ ed the current self-repair efforts of the aphasic with a "candidate repair", and the 250 aphasic had no chance to finish the repairable turn by her or himself (see Example 4). In those cases the rotes of "the helper" and "someone to be helped" seemed to overcome the turn-taking and repair conventions of everyday conversation.

Example 4 (Conversation 2)

H: it, it it is the one I've mo- , o, most, that I have [most,.] t: [that's] the language you've known very well.

(in Finnish) H: se, se se on semmone eni- enite mitä mä po-, o, enitep p[oon,.] t: [sitä ]sä oot hallinnu oikeen hyvin. On the other hand, self-initiated other-correction may be more plausible in aphasia therapy conversation. As Norrick (1991) has argued, certain settings (e.g. interactions between parents and children, teachers and students, and native and nonnative speakers) may favor second speaker correction. Furthermore, when conversationalists negotiate corrections of one another's speech e.g. in a classroom setting, corrective sequences are based on their respective abilities to complete the correction rather than adhering to the preference for self­ correction. Word searches, and specifically such utterances as and uh and uh are found to be recurrent loci for next-speaker startings and they may function as "interruption invitations" (Jefferson 1984). Also word searches with wh­ questions (e.g. "what was that girl's name?") can have social character when they occur with gaze shift to the conversational partner (Goodwin 1987). However, the conversations studied

4 Discussion The present study of interactional features in aphasia therapy conversations suggests that therapists tended to control the conversational flow when talking with an aphasic person. Interaction was asymmetric although aphasics

1. phonemic approximation of the target item inferred from the immediate sentential or nonverbal context, 2. semantic approximation of target item inferred from immediate context, 3. unidentified expression (nonword or part of the word), 4. grammatical form or semantic content unsuitable in sentential context, 5. circumlocution, e.g. circumlatory description of target, 6. nonverbal circumlocution, e.g. pointing or pantomimic illustration. 255 B. Self-repair initiations

I. pause over 2 seconds, 2. sound prolongation (filled pause, e.g. ee, mm, creaking, coughing), 3. interjection (filled pause, e.g. uh, well, so), 4. repetition of preceding word or utterance, 5. verbal indication of trouble, e.g. comments like What is it called; I don'! rememher, 6. gesture indicating trouble, e.g. scratching one's head, 7. unaccomplished phrase, a false start anda syntactic change. Other-repair initiations

I. general request for clarification, expressions like huh? and what?, 2. interrogative clause, e.g. What did you say?, 3. interrogative clause that contains repetition of the trouble source or a part thereof, 4. repetition of the previous trouble tum or a part of it, 5. offer to give an interpretation of the trouble tum ( ojten Y mean that ...), 6. direct other-initiated other-completed repair of other speaker's utterance. (Classes are based on preinspection of the data and Schegloff, Jefferson & Sacks 1977; Schlenck, Huber & Willmes 1987, Schegloff 1987, Sorjonen 1990.) 256 Transcription notation

1. Overlap and pauses I I beginning and end of overlap (.) micropause (1.1) measured pause (1.1 seconds) () pause over 2 seconds (not measured in more detail) = utterance continues without pause 2. Intonation contour fälling slightly fälling steady ? rising 3. Gestures <> beginning and end of gestures in conversational tums G gaze 4. Other + words pronounced together z z silently pronounced word or utterance (--) unclearly heard word or utterance $ $ phrase spaken with laughter pesramppi neologism (in italics) The Systems Analyst's lnterview

Helen Tebble

The development of a computer system for an oil terminal, an airline, a library, a personnel office, or even a local video hire shop, typically undergoes four major stages: analysis, design, implementation and evaluation. It is the role of the systems analyst to supervise the full project and after the completion of the feasibility study, to engage in the depth study which forms the major part of the analysis stage. It is imperative that the information obtained in the depth study is accurate and complete because the design of the system, the programming, installation and use all depend on it. If the information obtained

The generic structure potential of the systems analyst's interviews Following Hasan (for example, 1978, 1984) and Halliday & Hasan (1985) the generic structure potential is a summary of all the optional and obligatory elements and their sequence for every text that shares the same contextual configuration. The features of field, tenor and mode are drawn from the context of situation to establish the contextual configuration. The field accounts for the subject matter, that is, the content of the context. The tenor accounts for the social roles, the interpersonal relationships of the participants and their social distance in the context. The mode accounts for "the particular part that language is playing in the interactive process" (Halliday & Hasan 1985:24), the channels of communication (e.g. speech, writing, signing), the media of communication (e.g. documents, telephone), and the rhetorical mode of the text. Within the depth study of a systems analyst's project the generic structure potential for fact-finding interviews, those whose main function is to elicit information, will be different from that of walkthrough interviews because although the tenor remains much the same for both, there are differences in the fields and in the modes. The field, refers to the subject matter, which in this case entails finding and checking the fäets in the depth study of the analysis phase of a systems analysis project. This project is the computerization of personnel records of a state personnel branch of a national Australian retail company. Within the fact­ finding interviews the emphasis is on seeking and obtaining information while the emphasis within the walkthrough interviews is on checking that the information obtained previously in the fact-finding interviews is correct. 259 The tenor encompasses the interpersonal relationships of the systems analyst and all her clients who participate in the interviews according to their organi­ zational work role and status. The systems analyst holds most of the power by conducting the interviews and significantly controlling topics, but nevertheless is often in the role of k2 (secondary knower) (Berry 1987) while the clients are usually in the role of kl (primary knower). The social distance between the participants is relatively förmal and ca-operative with the systems analyst being superordinate. The duration of the working relationship is several months. Concerning the mode, language clearly constitutes the systems analyst's interviews. The channels of communication are phonic and graphic. The media are the spaken language, visual contact, and documents such as personnel forms, the systems analyst's notes and activity diagrams. The rhetorical mode is mainly referential and transactional in that aspects of the data processed in the personnel office are identified and referred to, and information is sought and given about those data. In the fact-finding interviews the systems analyst requests information about personnel forms brought to the interview and takes notes. In the walkthrough interviews while checking information she may annotate her prepared list of questions and/or diagrams, or if need be, she may make lengthier notes similar to the style used

an abstract category: [ ...] a description of the totai range of textual structures available within a genre [ ...]. It is designed to highlight the variant and invariant properties of textual structures within the limit of one genre (Hasan 1984: 16). The generic structure potential therefore is not typically expressed as just a list of all the elements of the social activities but rather as a formula that can be used 260 to show how the elements fit together to create the structure. The features of text structure are predicted by the contextual configuration which identifies the obligatory and optional stages as well as the sequence and frequency of the stages of a specified social activity.

Generic structure potential for fact-finding interviews Table 1 shows the major stages of the fact-finding interviews, the speech event which the systems analyst uses to obtain information from a client. These stages are the social activities that occur in this type of interview and are equivalent to the elements of the generic structure of the fact-finding interview. In every actual fact-finding interview these stages do not always follow sequentially from 1-8 because they may be iterated, merged with others, be optional or occur in different orders, but the generic structure can capture these various properties as shown in figure 1. (See end notes for identification of notation conventions.)

Table 1. List of significant stages in the systems analyst's fact-finding interviews

Stage Abbreviation

1. Greeting =G 2. Topic Announcement =TA 3. Eliciting =E 4. Informing =INF 5. Commenting on W ork =CW 6. Summarizing =S 7. Concluding =CON 8. Leave Taking =LT

Figure 1. Generic structure potential for systems analyst's fact-finding interviews

Of the eight elements of the generic structure for fact-finding interviews six are obligatory and two, commenting on work (CW) and leave taking (LT), are optional. The fact-finding interview commences with a greeting (G). The topic(s) of the interview are announced at the outset by the systems analyst. The major part of the interview is spent on the activities of eliciting (E) and informing (INF) in either order; hence the use of the convention of the raised 261 dot and bracketing. The systems analyst initiates most of the items that make up the eliciting activities while the client initiates most of the informing activities. Occasionally the systems analyst makes evaluative comments on the client's work (CW). These remarks serve an interpersonal function. This stage, which is iterative and optional, allows the participants to relax a little after periods of concentrated work. From time to time the systems analyst engages in a substantial activity of summarizing (S) the information that has been provided over a considerable period of the interview, that is, for the duration of several transactions. Summarizing is part of the activity of managing the interview. Leave taking is optional when a fact-finding interview precedes a walkthrough interview or when a clerk leaves at the end of the interview to return later with new documentation for the systems analyst's own use.

Generic structure potential for walkthrough interviews

Table 2 shows the major stages that arise out of the contextual configuration for the systems analyst's walkthrough interviews. These stages are linguistic equivalents to the main social activities that occur in the systems analyst's walkthrough interviews and are thus elements of the generic structure of this type of interview. The generic structure potential for the systems analyst's walkthrough interviews is illustrated in the formula in figure 2. It commences with the short obligatory element of Greeting (G). The systems analyst may then

Table 2. List of significant stages in the systems analyst' s walkthrough interviews

Stage Abbreviation

1. Greeting =G 2. Explaining Activity Diagrams =EAD 3. Requesting Check of Activity Diagrams =RCAD 4. Topic Announcement =TA 5. Checking Previous Data =CPD 6. Concluding ·=CON 7. Commenting on Work =CW 8. Leave Taking =LT show and explain to the client an activity diagram which is the systems analyst's interpretation of information that the client supplied in a prior interview. Thus the substance of the walkthrough interview may start with an explanation of the activity diagrams (EAD) and the accompanying notes that are to be "walked through"

Figure 2. Generic structure potential for systems analyst's walkthrough interviews

The convention of the second raised dot in the formula for the generic structure potential for walkthrough interviews as shown in figure 2 indicates that the matters of dealing with the activity diagram can occur before or after the matters of checking previously collected data. Nevertheless, the interview may start with a more predictable element, the systems analyst's announcement of the topic of the interview (TA), and then move to the next activity of checking previously collected data (CPD). This latter element is iterated and is crucial to the walkthrough interview. When the systems analyst is satisfied with the information obtained through the activity of checking previously collected data (which is usually an outcome of the previous fact-finding interview) she draws the interview toa conclusion (CON). The subsequent optional element of comment on work (CW) is typically initiated by the client who may comment on the size of the project and want to know about the project beyond his/her immediate concerns. Leave taking (LT) is a typical final element in the generic structure but it does not occur at the end of a brief walkthrough interview which may precede a fact­ finding interview

Levels of discourse structure The above approach to the description of the generic structure is top down but for a full analysis of the discourse structure of these lengthy interviews in systems analysis, a bottom up approach is also needed. A typical transcription (using modified conversational analysis conventions) of a systems analyst's hour long interview produces sixty or more A4 sized pages. Faced with this amount of data the linguist needs a comprehensive set of procedures to describe this type of professional discourse. A top down approach should guide the linguist in noting the stages (that is, the main social activities) of the interview which are realizations of the genre elements. But to do this clearly the full structure of the interview needs to be delineated. Sinclair & Coulthard's ( 1975) ranked units within systemic linguistics which start from the speech act and move up the scale to the transaction provide a very useful approach when 263 combined with Hasan's approach to specifying generic structure. By combining these two separate branches of systemic linguistics, Sinclair & Coulthard's bottom up approach with Hasan's top down approach, at the level of the genre element, a full linguistic description of the discourse structure can be achieved for the systems analyst's interviews and potentially for other professional interviews. GENRE t GENRE ELEMENT t TRANSACTION t EXCHANGE t MOVE t ACT

Figure 3. Rank scale of linguistic units of spoken discourse in the systems analyst's interviews

Figure 3 shows the sequence of a rank scale of linguistic discourse units that combine to form the total discourse structure of the systems analyst's interview. Each leve[ on the scale includes a realization of the leve! below it. The basic component of a move isa speech act; an exchange is made up of several moves; a transaction is made up of one or more exchanges; a genre element is made up of one or more transactions of the same type; and the genre comprises the obligatory and optional genre elements.

Dominant elements The main element that typifies the fact-finding interviews is the Eliciting element and the main element that typifies the walkthrough interviews is the element, Checking previous data. Both of these elements are iterative and they reveal the systems analyst's strategies for obtaining information and for checking information. The systems analyst has to obtain very detailed infor­ mation that is often self-evident to the client. Nevertheless every detail of every form in a personnel office whose system is being upgraded has to be asked about and later checked. It is at the leve! of the exchange in these two genre elements that the particular elicitation and checking strategies of the systems analyst are revealed. The relationship between the transaction and the genre element has been discussed e[sewhere (for example, Tebble 1992). 264 The eliciting element is realized by one or more sequences of eliciting transactions. This type of transaction is determined by its first conversational exchange which occurs after the optional initial boundary exchange of the transaction. This first conversational exchange, the eliciting exchange, in turn is determined by its initiating move which is realized by a requestive speech act. This speech act will be uttered typically by the systems analyst.

Eliciting exchanges Within systemic linguistics the structure of a typical exchange comprises an initiating move followed by a responding move. Within conversation analysis they are equivalent to the first pair part and second pair part of adjacency pairs, respectively. The exchange structure has a third move, the follow-up move, which is optional. The category of the exchange is largely determined by the speech act in the initiating move. There are three main types of eliciting ex­ changes in the systems analyst's interviews: the question, the formulation and the observation. The major types of speech acts used by the systems analyst to obtain information are requestives in the form of questions, formulations, and observations (Tebble 1993b).

Questions The question exchange is determined by the indirect speech act of the "request to tell" variety, which is expresssed syntactically by the question in the initiating move of the eliciting exchange. This exchange is made up of three sequential moves each comprising a specific type of speech act: the initiating move containing the question; the responding move containing the answer; and an optional follow up move containing an acknowledgement of the answer, as is illustrated in figure 4. lnitiating --[ Question System of Question Responding -[ Answer Exchange i(Follow Up) -[ Acknowledgement Figure 4. System of the question exchange in the systems analyst's interview

An example of the question exchange is the following (SA= systems analyst, Cl=clerk): 265 Initiating SA: and does this have their address in it as wel12 Responding Cl: NO it just has .hh their employee NUMber Follow Up SA: right (T7SIP28) The above example shows the systems analyst, in terms of speech act theory, making an indirect request for information in the initiating move of a question exchange. The syntactic form of the request is a yes/no question uttered with question intonation (tone 2). In the responding move of the exchange the clerk answers the question with the ellipted negative "NO" and with a clarification of the fäets about that component of the personnel form which the systems analyst was studying. That is, the form just has the employee number. The follow up move of the exchange shows that the systems analyst acknowledges the clerk's answer and clarification. As part of her eliciting techniques for obtaining and checking information the systems analyst uses all of the syntactic forms of interrogative clauses. These interrogative clauses uttered with rising tone 2 (question intonation) and fälling tone I (declarative intonation) (Halliday 1970) are illustrated in the following examples. Tone is indicated with a raised digit. Tone 1 is used when the systems analyst seeks confirmation with some certainty, and tone 2 when seeking information. Wh-questions with fälling tone 1: what is actually USEd to UPDATE this 1 why do you need the p-d-seventy for salaried people 1 Y es/No questions uttered with rising tone 2: do you PUT the house number on there2 do you mean the P-D-forty-three transfers2 Alternative questions with tones 2 and 1: is this purely WAGes a:nd salaried staff2 or JUST salaried staff 1 Question tags: you'd probably have to get that off there wouldn't you 1 THEY HA VE a copy as well do they 2 Multiple questions: how are you advised of that from the storel .hh is it by a phone calJ2 ora 1 Houtkoop-Steenstra (1990) suggests that the use of multiple questions, for example, a wh-question plus an alternative question, is a request for telling. Single or multiple questions as interrogatives perform this speech act. The interrogative clauses that give substance to the initiating moves of the eliciting exchanges function as questions which in speech act theory are requests to tel1. 266 All the questions occur in the first move of the question exchange which is itself one type of eliciting exchange used for obtaining information.

Formulations The second type of eliciting exchange is the formulation. The formulation which is made by the systems analyst typically expresses a summary or inference which isa deduction about prior talk in the interview. It occurs frequently and is used as a checking strategy which enables the systems analyst to have confirmed by the client, what has been said by the client thus far. The formulation and decision together area type of adjacency pair (Heritage & Watson 1979, 1980) like the question and answer. When adapted to Sinclair & Coulthard's conver­ sational exchange structure, the formu!ation and the decision fil! the initiating and responding moves respectively. The decision is usually either a confirmation or a disconfirmation of the systems analyst's formulation. The formulation starts with "so" and is a statement-question with either declarative fälling tone 1 or rising tone 2. Figure 5 shows the three part system of the formulation exchange. lnitiating --[ Formulation System of Formulation Responding -[ Decision Exchange i(Follow Up) -[ Acknowledgement Figure 5. System of the formulation exchange in the systems analyst's interview

The following example of a formulation exchange contains a formulation expressed with tone I. Initiating SA: SO you just ADD that person on to your latest four-two­ three-a 1 Responding Cl: no SEParate form for adds Follow Up SA: OH separate form for adds I see (T9S2P64) The speech act of the formulation occurs in the initiating move; the decision, which in this case isa disconfirmation, occurs in the responding move; and the acknowledgment occurs in the follow up move. The systems analyst sums up (with certainty, using tone I) her understanding of what the clerk has told her: that the clerk adds details about particular employees to the most recent copy of the form known as the "423A". The c!erk elliptically disconfirms the systems analyst's formulation by using the negator, "no", and justifies this contradiction 267 in the responding move. The systems analyst uses the discourse marker "OH" to indicate receipt of new information and acknowledges that her understanding has been corrected.

Observations The third type of eliciting exchange is the observation exchange which is dominated by the observation speech act in the initiating move of the exchange. A series of observation exchanges comprise one type of eliciting transaction, and all these particular transactions together form the genre element checking previous data in the walkthrough interview. The observation is one or a series of statements whose propositions are about the context of situation of the interview. It is used by the systems analyst to make a remark, a comment or a slight criticism, to express an opinion, or to point out something that has been noticed. Like the formulation it is responded to by the client's decision which usually either confirms or disconfirms the observation. Like the other eliciting exchanges, the speech act in the initiating move determines the type of exchange. The responding move by the other interlocutor, typically the system analyst's client, isa decision. The third component of the exchange, the follow up move, to be made by the systems analyst is optional. Figure 6 shows the three part system of the observation exchange and is followed by two examples. Initiating ---[ Observation System of Observation Responding -[ Decision Exchange 1(Follow Up) -[ Acknowledgement Figure 6. System of observation exchange in the systems analyst's interview a. Observation expressed with tone 2: lnitiating SA: NOW that's UM + for all SALaried emPLOYees in vsfd 1 but it's ALSO I GATHer for WAGE people too2 [ Responding Cl: wage people mhm (T7S I :33/34) 268 b. Observation expressed with tone 1:

Initiating 1 SA: for SOME reason I thought there was a manager's= [ Responding 1 Cl: mm Initiating 2 SA: =address book 1 [ Responding 2 Cl: mm (T7S2:2) Both observations point out what the systems analyst has noticed. They are not deductions made during the interview hut observations made when reflecting on the data supplied from the prior interview. The observation serves as a checking strategy hut is intertextually dependent since it is not used to check an item from the immediate interview hut from a prior one. In (a) and (b) an understanding of something the systems analyst noticed in her analysis concerning salaried and wage staff is offered for checking in the form of an observation. The systems analyst has some uncertainty as to whether employees paid by weekly wages as distinct from those on an annual salary are included on the personnel form. In (b) there is some uncertainty about a book referred to in a prior interview. The structure of the initiating move in (a) is realized by three declarative clauses which reveal these observations. The first observation uttered with declarative word order uses the stative verb BE, 's, and tone 1, is a statement of confidence proferred by the systems analyst to the clerk for confirmation. The second observation commences with the contrasting conjunction "hut" and is uttered with tone 2 whose illocutionary function is to express a request for clarification. The third observation is contained in the embedded clause "ALSO I GATHer" which is a suppositive speech act. "People on wages are to be included too" is a notion the systems analyst has deduced when analysing her notes from the previous interview. The deduction does not have the same status as a fäet and thus is proffered to the clerk for clarification. The clerk's decision in the responding move of (a) is a confirmation shown by a repetition of the uncertain item, "wage people" and re-inforced by a non­ verbal confirmation, "mhm". The observation (b) as a remark is confirmed non­ verbally in the second responding move by a short auditor back channel response, "mm". In both examples, (a) and (b), the option of acknowledgment within the observation exchange is not used by the systems analyst. The types of eliciting exchanges, the question and the formulation, are the most frequently occurring exchanges, and series of these form the eliciting transactions, which in tum constitute the eliciting element of the genre of the systems analyst's fact-finding interviews. The observation exchange which is used to remark on information already available prior to the interview is a different type of exchange. Series of these observation exchanges make up the 269 eliciting transactions which themselves comprise that generic element of the walkthrough interview, "checking previous data".

Scale of certainty when seeking confirmation In the depth phase of a systems analysis project it is absolutely critical that the systems analyst not only obtains correct information but also checks to confirm that it is correct. When the systems analyst seeks confirmation of her inferences, summaries and observations this is done so with varying degrees of certainty about their validity. Typically then the systems analyst's interviews contain many formulations, observations and a wide range of yes/no questions with questioning or confirmation seeking intonation (tone 2) and declarative into­ nation (tone 1). It is likely that the choice of confirmation seeking strategy is not due to free variation but rather directly reveals the systems analyst's expressed leve! of confidence about the validity of what she is saying. Thus the systems analyst's leve! of certainty about the validity of the proposition she makes in her formulation, observation, or question can be noted on a scale of certainty. Figure 7 is a scale of certainty of the accuracy of the proposition expressed by the systems analyst at the time of seeking confirmation. The scale ranges from formulations and observations each uttered with tone 1, indicating the greatest degree of certainty, to leve! 8 yes/no list questions with two listed items each uttered with question intonation (tone 2), followed by or creating a further unspecified item for the list, indicating the strongest degree of uncertainty. The formulation and observation are declarations but they also have the function of seeking endorsement. The most confident way of expressing a formulation or observation is with tone 1, which is the normal tone to accom­ pany declarative statements (leve! 1). Yes/no questions uttered with tone 1 (levels 2 and 3) are infrequent in occurrence but they do indicate more certainty on the part of the systems analyst than do her formulations and observations uttered with tone 2. This is because the intonation (tone 1), implying certainty on the part of the speaker, is declarative and over-rides the syntactic structure of the question that requires an answer of either yes or no, for leve! 2, or one alternative as an answer for leve! 3. Question tags do not occur very often in the systems analyst's interview: yes/no questions for example, occur four times as often as question tags. The main patterns of question tags that occur are:

(a) operator+ subject (which is by far the most frequent) is it (b) operator + negator + subject does n't it 270 1. Formulation Tone 1 Observation Tone 1 2. Yes/No Question Tone 1 3. Yes/No Alternative Question Tone 1, Tone 1 Yes/No Alternative Question Tone 1, or ...

right 4a. Formulation Tone 1 + is that correct + Tone 1 tthe same Observation Tone 1 + Question Tag Tone 1

right 4b. Formulation Tone 1 + is that correct + Tone 2 tthe same Formulation Tone 1 + Question Tag Tone 2 Observation Tone 1 + Question Tag Tone 2 5. Formulation Tone 2 Observation Tone 2 6. Yes/No Question Tone 2 7. Yes/No Alternative Question Tone 2, Tone 1 Yes/No Alternative Question Tone 2, or ... 8. Yes/No List Question Tone 2, Tone 2, Tone 1 Yes No List Question Tone 2, Tone 1, or ... Y es/No List Question Tone 2, Tone 2 Y es/No List Question Tone 2, or ... Yes/No List Question Tone 2, Tone 2, or ...

Figure 7. Scale of certainty of the systems analyst's confirmation seeking strategies

They occur after a statement (which may be a formulation or an observation), after a yes/no question, and also as part of a multiple question. The question tag is uttered with fälling tone I (leve! 4a) when the systems analyst seeks confirmation with some certainty and with rising tone 2 when seeking infor­ mation (leve! 4b ). The particular yes/no questions, "is that right", "is that correct" and "is that the same", and question tags are uttered with question intonation (tone 2) and tend to be mid-scale when immediately preceded by a formulation or observation hearing tone 1. The yes/no questions which require a yes/no answer, or the choice of two alternatives as an answer, or one item of a list from two or more options (levels 6, 7, 8) are uttered with rising tone 2. This indicates the systems analyst's 271 uncertainty about the information she is checking. The only small degree of certainty is in the use of tone 1 when the systems analyst at least knows that there can only be two or three specific options. The use of or... implies uncertainty about the content and number of other options that could be offered to the Iistener to check. Examples of some levels of the scale of certainty of the systems analyst's interview follow. Leve) 1. Formulation Tone 1: so that's the ONE-fifty-five card (tone 1) (T9S1Pl 1) Observation Tone 1: .hh that must get fairly lar:ge sometimes with all the people LEAVing (tone 1) (T9S1P26) Leve) 4a. Observation Tone I + Question Tag Tone 1: that's from the ONE-fifty-fi:ve (tone 1) isn't it (tone 1) (T9S1Pl2) Leve) 4b. Formulation Tone I + Question Tag Tone 2: so all that comes back to you (tone 1) does it (tone 2) (T9S1P44) Leve) 5. Formulation Tone 2: so once you get that printout from super [annuation] then you inform the STORE that the person's to go onto super (tone 2) (T9S2P73) Observation Tone 2: .hh and then you OUTput the list and you don't actually FILE a copy (tone 2) (T7S2P58) Leve! 6. Yes/No Question Tone 2: do you PUT the house number on there (tone 2) (T7S1P34)

Leve) 7. Yes/No Alternative Question Tone 2, Tone I: is this purely WAges a:nd salaried staff (tone 2) or JUST salaried staff (tone 1) (TlS1P31) 272 Leve] 8. Yes/No List Question Tone 2, Tone 1, or ... : do they ring you up (Tone 2) or send in a form (Tone 1) or ... The scale is formulated by acknowledging that the intonation is primary in revealing the degree of uncertainty in the systems analyst's confirmation seeking strategies. Tone 1 (fälling intonation) is declarative and therefore is to be taken as indicating certainty about the validity of the content of the utterance. But because it occurs with formulations and observations within the eliciting exchanges the total utterance still has to be confirmed or disconfirmed by the listener's response since it is the listener (kl) who has the knowledge. Tone 2 (rising intonation) has the function of expressing uncertainty, of seeking infor­ mation. Utterances whether they have declarative syntax (as do formulations and observations) or not (questions) when uttered with tone 2 will express greater degrees of uncertainty than utterances expressed with tone 1.

Conclusion The systems analyst's interview is one type of professional discourse that can be studied by using two previously unconnected branches of systemic functional linguistics, Hasan's top clown approach to delineating the generic structure potential, and Sinclair & Coulthard's bottom up approach of identifying levels of discourse structure that extend from the speech act up to the transaction. The connection between both approaches is made at the discourse structure leve! of the genre element. The main features of the systems analyst's interview in the depth phase of a systems analysis project are the sub-genres of the fact-finding and walkthrough interviews; the dominance of the elicitation and checking strategies of questions, formulations and observations; and the scale of degrees of certainty that the systems analyst can use to express levels of understanding of information as she seeks to understand the work of ber clients in the pursuit of improving their data processing systems. The approach outlined here is applicable to the discourse analysis of !arge corpora of transcribed interviews of other professionals at work. For example, it has been applied to interpreted consultations (Tebble 1993a) and used as a prototype mode! for describing the bilingual discourse of English speaking doctors, lawyers and their non-English speaking patients and clients respec­ tively, mediated through professional liaison interpreters. It is also currcntly being applied to the discourse of psychotherapy consultations (Patrick & Tebble forthcoming). Applied linguists need to have full generic descriptions of professional consultations to use in the design and development of syllabuses and materials for the pre- and in-service communication education of professionals. The use of a comprehensive linguistic theory such as systemic 273 functional linguistics for the analysis of !arge texts that constitute transcribed professional consultations can provide a full description of the structure of the discourse from the smallest discourse unit of the speech act to the complete formulaic description of the genre of particular professional consultations or interviews.

Notation conventions " = left to right sequence; ( ) = optionality; [ ] = restraint; curved arrow = iteration; raised dot = more than one option in a sequence; capita! letters serve a mnemonics for stages of the interviews.

References Berry, Margaret, 1987: Is teacher an unanalysed concept? In Michael A.K. Halliday & Robin P. Fawcett (eds.), New Developments in Systemic Linguistics. Vol. 1. London, Frances Pinter, pp 41-63. Chandor, A., 1970: Computers as a Career. London, Rupert Hart-Davies. Halliday, Michael A.K., 1970: A Course in Spaken English: Intonation. Oxford, Oxford University Press. Halliday, Michael A.K. & Hasan, Ruqaiya, 1985: Language, Context, and Text: A.1pects oj Language in a Socia!-Semiotic Perspective. Geelong, Deakin University Press. Hasan, Ruqaiya, 1978: Text in the systemic-functional mode!. In Wolfgang U. Dressler (ed.), Current Trends in Textlinguistics. Berlin & New York, Walter de Gruyter, pp 228-246. Hasan, Ruqaiya, 1984: The nursery tale as a genre. (Mimeo) In M. Berry, M. Stubbs, & R. Carter (eds.), Nottingham Linguistic Circular. Vol. 13. Heritage, J.C. & Watson, D.R., 1979: Formulations as conversational objects. In George Psathas (ed.), Everyday Language. Studies in Ethnomethodology. New York, Irvington, pp 123-162. Heritage, J .C. & Watson, D.R., 1980: Aspects of the properties of formulations in natura! conversations: Some instances analysed. Semiotica, 30:245-262. Houtkoop-Steenstra, Hanneke, 1990: Normative and analytical perspectives on interviewing techniques. In Harm Pinkstcr & Inge Genee (eds.), Unity in Diversity. Dordrecht, Foris, pp 133-149. Patrick, Jon & Tebble, Helen, forthc.: Discourse analysis of a psychotherapy consultation. Paper for the 5th International Conference on Language and Social Psychology, University of Queensland, Brisbane, 1994. Sinclair J. McH. & Coulthard, R.M., 1975: Towards an Analysis oj Discourse. London, Oxford University Press. 274 Tebble, Helen, 1992: The genre element in the systems analyst's interview. Australian Review oj Applied Linguistics, 15(2): 120-136. Tebble, Helen, 1993a: A discourse mode! for dialogue interpreting. In First Practitioners' Seminar. AUSIT [Australian Institute of Interpreters and Translators], 1993. Canberra, National Accreditation Authority for Translators and Interpreters, pp 1-26. Tebble, Helen, 1993b: Formulations and observations in professional interviews: Systems analysis. In Andre Crochetiere, Jean-Claude Boulanger & Conrad Ouellon (eds.), Proceedings oj the XVth International Congress oj Linguistics. Quebec, Presses de I' Universite Laval, pp 265-268. ASLA:s skriftserie

I. Barns tvåspråkighet. Rapport från ASLA:s höstsymposium, Uppsala 6-7 november

1987. B.-L. Gunnarsson, C. Liberg, & A. Nesser (utg). ASLA, Svenska föreningen

för tillämpad språkvetenskap. Uppsala 1988.

2. Skrivande. Rapport från ASLA:s nordiska symposium, Uppsala 10-12 november 1988.

B.-L. Gunnarsson, C. Liberg, & S. Wahlen (utg). ASLA, Svenska föreningen för

tillämpad språkvetenskap. Uppsala 1989.

3. På väg mot ett nytt språk. Rapport från ASLA:s höstsymposium, Linköping 9-10

november 1989. V. Adelswärd & N.F. Davies (utg). ASLA, Svenska föreningen för

tillämpad språkvetenskap. Uppsala 1990. •

4. Barns läsutveckling och läsning. Rapport från ASLA:s höstsymposium, Stockholm

15-16 november 1990. B. Hene & S. Wahlen (utg). ASLA, Svenska föreningen för

tillämpad språkvetenskap. Uppsala 1991.

5. Språk, språkbruk och kön. Rapport från ASLA:s nordiska symposium, Uppsala 7-9

november 1991. B.-L. Gunnarsson & C. Liberg (utg). ASLA, Svenska föreningen för

tillämpad språkvetenskap. Uppsala 1992.

6. Text and Talk in Professional Contexts. Selected Papers from the International

Conference "Discourse and the Professions," Uppsala, 26-29 August, 1992. B.-L.

Gunnarsson, P. Linell & B. Nordberg (eds). ASLA, The Swedish Association of

Applied Linguistics. Uppsala 1994.

Beställning av böckerna sänds till följande adress: ASLA, c/o FUMS, Box 1834,

751 48 Uppsala

ISBN 91-87884-05-4 ISSN 1100-5629

Reprocentralen HSC, 1994