THE PIXI CORPORA: Bookshop encounters in English and Italian Printed version: edited by Laura Gavioli & Gillian Mansfield published by Cooperativa Libraria Universitaria Editrice, Via Marsala 24, 40126 Bologna, Italy copyright Cooperativa Libraria Universitaria Editrice 1990 Computer-readable version: edited by Guy Aston, CISeL, Via Pizzecolli 68, 60121 Ancona, Italy CONTENTS Transcription conventions L. Gavioli & G. Mansfield: Introductory remarks The corpora Lod Bof Paf Pif Rok TRANSCRIPTION CONVENTIONS AL speaker = assistant, where L is the first letter of the assistant's name C(n)w speaker = customer, where n is a consecutive number C(n)m when more than one customer speaks in the sequence, C(n)i w = woman, m = man, i = child C(n)wn speaker = non-native customer C(n)mn [note] comments (paralinguistic and extralinguistic features) (text) tape unclear: tentative transcription (nsyll) tape untranscribable: n = approximate number of syllables spoken (??) tape untranscribable + short pause (less than one second) ++ longer pause (less than one second) (n) long pause (n = length in seconds) = latched to the preceding turn in the transcript == latched to previous-but-one turn in the transcript text spoken in overlap with next text text spoken in overlap with next text "Text" title of book or series text stressed syllable or in loud voice text: lengthening of previous sound or syllable (number of colons indicates extent of lengthening) text- syllable cut short text - tone group interrupted . low fall intonation , fall-rise intonation ? low rise intonation ! rise-fall intonation INTRODUCTORY REMARKS by Laura Gavioli, Gillian Mansfield & Guy Aston These corpora provide a transcription of the data collected and analysed in the course of the PIXI project on the pragmatics of public service encounters in English and Italian. The data consist of 379 naturally-occurring conversations recorded in bookshops in England and in Italy some years ago. While the English data (Lod: 150 encounters) were all recorded in one multi-departmental bookshop in a large university town, the Italian data consist of a main corpus (Bof: 180 encounters) recorded in a similar context to that of the English data, and three minor ones from bookshops of other kinds and in other parts of the country (Paf: 21 encounters; Pif: 19 encounters; Rok: 9 encounters). The transcribed corpora are available in a printed version published by the Cooperativa Libraria Universitaria Editrice, Bologna, and in a computer-readable version deposited in the Oxford Text Archive. We have decided to edit these corpora for publication for three main reasons. First, there is still relatively little naturally-occurring conversational data published in the form of complete encounters: for instance Crystal & Davy (1975) and Svartvik & Quirk (1980) provide mainly extracts from longer episodes. While the importance of examining naturally-occurring data for an understanding of conversation has been widely maintained in the literature, such data are not easily obtainable. Long and/or surreptitious recordings are needed in order to overcome the observer's paradox and elicit maximally spontaneous speech (see Labov 1972, Svartvik & Quirk 1980), while the poor recording quality which is virtually inevitable in real-life settings makes accurate transcription difficult. Second, to the best of our knowledge, there are no published corpora which allow comparison of naturally-occurring conversations in different cultures. Conversational data for contrastive analysis have typically been recorded in experimental settings using role-play techniques (e.g. Edmondson 1981, House 1986), and in those cases where naturally-occurring data have been used, the full corpora are not available; for instance Ventola (1987) includes only the English data from her comparison of service encounters in English and Finnish, and Lindenfeld (1990) provides only French data in her hints at French-American contrasts. We would argue that our two main corpora particularly lend themselves for purposes of cross-cultural comparison since they come from contexts which have similar situational characteristics, and show similar features of basic discourse organisation. Lastly, publication seemed desirable from a methodological perspective. Work from the project which has discussed quantitative tendencies and qualitative characteristics of the data has necessarily cited only sample extracts in illustration. By making available the complete corpora we hope to allow others to better confirm or contest the empirical foundedness of these analyses. The corpora as instrument: aims of the project 1.1 The choice of bookshops Our study of the pragmatics of service encounters has primarily pedagogic objectives. It hopes to contribute to an understanding of the dynamics of a type of conversation which is typically proposed to language learners as a model: encounters between clerks and customers in shops, public offices, banks, stations, hotels and the like have played a central role in second language teaching from Renaissance phrasebooks to the Council of Europe Threshold Level (van Ek 1975). A data-based description and critique of the regularities of such encounters may not only offer more accurate models, but provide insights into the relationship between stereotypical patterns and actual occurrences. In its turn, contrastive analysis may cast light on cross-cultural differences both in terms of stereotypes and of these procedures of instantiation. Insofar as they represent routine institutionalised situations in both cultures, service encounters lend themselves to contrast along these lines. Bookshop encounters seemed particularly suitable objects of study for a number of methodological reasons. In the first place, bookshop encounters are relatively accessible. Most of the shared background which participants draw on is also available to the outsider. Such encounters typically take place between strangers, who do not have a specific shared history to tailor their talk to (Goffman 1983), merely their current context in the shop and their common knowledge of the subject area. This context can be largely known to the analyst. We became familiar with the organisation of the shops while collecting the data, and the main subject areas of the departments concerned (foreign languages and linguistics) were ones where our own backgrounds facilitated understanding. We were also able to turn to catalogues, bibliographical reference works, and indeed the assistants themselves to check names of authors, titles and the like. Secondly, bookshop encounters are fairly homogeneous, and this facilitates the task of observing regularities. While generally going beyond the minimal exchanges typical of booking offices, newsagents or supermarkets (Merritt 1976), they are clearly based on a recurrent sequence: the customer's request and the assistant's response. Requests are furthermore homogeneous as to topic; they usually concern either specific books (e.g. Burchfield's "The English language") or classes of books (architecture, stylistics, etc.). During a pilot stage of our research we also collected data at hotel reception desks and in travel agencies, but we rejected these for the main project for the reason that the wide range of topics and discourse structures made it difficult to compare encounters within each single setting, let alone from two different cultures. Thirdly, bookshops in England and Italy are essentially similar, in the sense of being comparable in ethnographic terms (e.g. those of Hymes' factors in the speech event: 1972) - unlike, say, banks, bars and restaurants where the organisation of the setting, the goods and services available, and/or the roles of participants show clear cross-cultural differences. Finally, it should be stressed that, again for pedagogic reasons, we were interested in examining data where interpersonal conflict was the exception rather than the norm. We aimed to identify regularities of socially `unmarked' collaborative talk in the two cultures, analysing eventual occurrences of conflict in terms of deviation from these regularities. While there are clearly some settings in which encounters regularly assume a conflictual tone, such as service points dealing with customer complaints (George 1990), service encounters are usually collaborative: By a service encounter I mean an instance of face-to-face interaction between a server who is officially posted in some service area and a customer who is present in that service area, that interaction being oriented to the satisfaction of the customer's presumed desire for some service and the server's obligation to provide that service. (Merritt 1976: 321) Our reactions as native speakers to the encounters we recorded in bookshops confirmed that these were generally non-conflictual. 1.2 Analytical approaches It needs to be stressed that while aiming to provide a record of naturally-occurring conversation, the transcripts reproduced here are a far cry from the `real world' which underlies them. The mediating procedures of data collection, selection and transcription are inevitably linked to the transcriber-analyst's interpretations and purposes, and play an unavoidable filtering role (Ochs 1979). A brief account of our analytic approach may therefore help explain some of the procedures we have adopted in creating this written record of the real world of speech in bookshops. Our concern as analysts has been to identify recurrent patterns of strategic action and interaction in the achievement of the prime goals of service encounters, which we see both in the `transactional' terms of getting and providing relevant goods and information, and in the `interactional' ones of establishing and maintaining face and interpersonal rapport (Brown & Yule 1983): following Giddens, we take it that "through recurrent and recursive properties of interaction [...] actors both produce and reproduce social relationships across time and space" (Boden 1990: 246). Our concern with these pragmatic aspects of negotiation has entailed a focus on the text as produced jointly and progressively by participants, stressing the temporal nature of its realisation. In a first phase (see particularly the papers in Aston 1988a), analysis concentrated on identifying the underlying `scripts' (Schank & Abelson 1975) to which service encounters appeared to be oriented in the two cultures: from this perspective we examined the lexicogrammatical and intonational forms and sequential organisation of requests and responses, and of encounter openings and closings (Tucker 1987, 1988a, b; Mansfield 1988; Anderson et al. 1988; Merlini Barbaresi 1988; Anderson 1988a). In looking at deviation from apparently routine patterns, we examined participants' strategies of dominance and control, and interactive procedures of negotiation, stressing the ways in which identities and statuses were established and maintained through such strategies and procedures (Ciliberti 1988; Aston 1987, 1988c; Vincent Marrelli 1988; Anderson 1987, 1988b). In contrast to data which is role-played or where one or more speakers has been instructed how to elicit the conversation (e.g. Scotton Myers & Bernstein 1988), and where in consequence goals and outcomes may fail to reflect real-life concerns, we argued that such analysis of naturally-occurring discourse offered a more reliable basis for the development of teaching materials, in designing both learner activities of discourse production and ones involving metalinguistic observation of verbal behaviour (Aston 1988d, e, f; Anderson 1988c; George 1988). [We do not mean to suggest that the data themselves constitute appropriate materials for direct teaching use, merely that they offer indications for the design of such materials, a process in which pedagogic concerns will be paramount over demands for authenticity tout court. It should also be noted that the tapes cannot directly provide classroom listening material, since background interference and audio quality make comprehension arduous even for advanced students.] In our more recent work, analysis has shifted to comparison of the English and Italian corpora, contrasting basic structural patterns (openings and closings, request and response forms: Zorzi Cal• et al. 1990; Brodine 1990; Zorzi Cal• 1990a, c; Ciliberti 1991c) and the realisation and placement of certain recurrent actions (interruption, laughter, personal disclosure, accounts: Zorzi Cal• 1990a, b; Gavioli 1987, 1990a; Anderson 1990), with an attempt to relate cross-cultural differences in discourse procedure to differences in wider terms of cultural `style'. [The minor corpora, which come from a variety of geographical and organisational settings, have been used to validate generalisations concerning cross-cultural differences derived from the major corpora.] In facing this question we have taken two main approaches, which we see as complementary in their pedagogic implications (Ciliberti & Aston, forthcoming). The first derives from conversational analysis, and sees sociocultural phenomena and personal identities as creatively constituted through the use of culturally-preferred and locally-adhocced discourse procedures (Schegloff 1987; Aston 1989). The second, derived from critical discourse analysis (Fairclough 1989), seeks evidence of predetermined sociocultural norms in discursive conventions, and of individual personality in distinctive patterns of use of these conventions (Ciliberti 1989a, b, 1991a, b). In sensitising learners to culture-specific regularities and their individual use, and in developing their capacity to negotiate discourse goals with reference to such regularities, both approaches would appear pedagogically productive. By describing discourse procedures and relating these, in terms of cause or effect, to features of social structure (personal self-image and positional status, interpersonal rapport and institutional requirement), both approaches provide accounts of interaction which can be employed to design contexts for classroom activities involving analysis or construction of foreign language discourse (Aston 1988d, 1990; Ciliberti 1991a, b; Gavioli 1990b). The corpora as record 2.1 Contexts of situation: the main corpora The main corpora were each collected in large bookshops in the centre of large university towns, in the South of England (Lod) and in the North of Italy (Bof) respectively. Both shops were of a self-service, multi-departmental kind, which while mainly serving university students and staff, also catered for other kinds of readers and tastes. In the departments where we recorded, which each dealt with language and linguistics, a significant number of customers were non-native speakers; 37 encounters where customers are recognisably non-native were collected in the English shop (25%), and 19 in the Italian one (11%). In both shops assistance to customers was provided from service points which were located separately from the cashiers' tills, with two or three assistants (of both sexes) generally assigned to each point. However the vast majority of the encounters involve one assistant and one customer. Two assistants occasionally take part in an encounter, but there is always one primarily responsible for dealing with the customer, while the other only intervenes to provide additional information or comment. Similarly there is almost always one primary customer. While s/he may occasionally be accompanied by a friend or friends, the latter are rarely also potential purchasers. Being self-service bookshops, contact with the service-giver usually occurs when the customer needs help with a problem of some kind. Most of the encounters concern either a specific book which the customer cannot locate, or the whereabouts of a category of books (e.g. dictionaries). A few encounters involve requests for information of other kinds, such as shop rules (where one pays, the possibility of paying by credit card or on account, opening hours) or `outside information' concerning the book trade (e.g. directions to another shop). In all these cases, the customer produces one or more utterances hearable as requests, and, by the adjacency pair rule (Schegloff & Sacks 1973), makes relevant the assistant's reply as the next action. A request-reply sequence is central to both English and Italian encounters. Even when customers produce lists which they show to the assistant, they still generally produce a request, using a deictic this/questo to refer to some item on the list. The main exception to the `customer request - assistant reply' pattern is where the assistant forestalls an eventual request of purchase: "Are you paying cash?" (Lod D-06); "Er: if you want to pay, it's on the + desk in the centre there" (Lod B-31). Some recurrent features of requests and replies in the data relate to the general characteristics of the shops where recordings were made. In particular: a) Their organisation in departments entails that we find frequent redirection of customers to other floors and/or departments. Cm Do you keep (only this copy? Opus:) - mh. + (2syll). AP Well they might have upstairs: in mythology: + section, on the second floor. Cm Mm: mythology section, (pardon)? AP =Yeah. + Or:, failing that the - Penguin department on the ground floor. (Lod B-30) b) Their self-service structure implies that customers usually only ask for help when they cannot find what they are looking for themselves. It is probably for this reason that our data contain a large number of negative replies to requests - saying, for instance that a wanted book is not available - far more than would seem likely for shops where goods can only be obtained through an assistant. c) The uncertain nature of the supply chain in the international book trade arguably explains the vagueness of many such replies: AJ Can I help you. C2w Yeah, can you give me a: - a- a prediction on when: Jacques Lacan's "crits" is likely to come back in? AJ + Frankly no. I can give you a + er + educated guess, of: (4syll). C2w Well. That's hh - better than I can do hh, so: - AJ Hh. (07) AJ [in the distance] About two maybe three weeks. (Lod C-30 /b1) Customers are encouraged to put in an order with a contact address or phone number; in Bof they are often told to try coming back on Wednesday or Thursday, as deliveries normally occur on Tuesdays. d) The assistants are not generally responsible for effecting sales (except, in the case of Lod, for purchases on account: cf. Lod B-07 /a; Lod C-23 /c). This entails that there are no `payment closings' to encounters in either set of data; as the assistants explicitly say, "The cash desk is just behind you" (Lod E-06) or "Si paga sotto alla cassa grazie" (Bof C-04). Following these observations on the general similarities of the contexts of the major corpora, we now outline their settings in more detail. 2.1.1 Lod Lovatts is one of the main bookshops in the town of Nonesuch. [Italics indicate names which have been altered to preserve anonymity: cf. 2.5.6 below.] It has about fifteen departments spread over five floors: basement, ground, mezzanine, first and second floor, with a further floor for offices. The department where recordings were made is on the first floor and deals with philosophy, linguistics and religion. It consists of a rectangular room with open entrances and shelves on each of three walls, and windows on the fourth. The service desk is in front of the window, facing into the room, and is thus immediately visible and identifiable as such from the entrances. The assistants generally stand or sit behind the desk where they also have shelves and drawers for catalogues, a table with a microfiche reader, and a card index of books stocked and on order in the department. Various cash points are provided on the same floor, one of which is visible about ten metres away facing the service desk. There are two main assistants, Jane, who heads the department, and Paul. Both are in their mid-thirties and have worked there for several years. Assistants from other departments - one of whom, Zoe, appears in the corpus - occasionally come to help in peak periods, or to substitute during breaks. In slack periods (for reasons of recording quality, much of our data was collected in these) there is often only one assistant on duty. As well as by customers and other assistants, the desk is occasionally visited by publishers' representatives. No customers in the corpus appear to have been previously known to the assistants (there is a contrast here with Bof, where several customers are known by name and a few are clearly friends of an assistant). A public address system, giving announcements throughout the shop, sometimes renders the recording of the encounters inaudible. 2.1.2 Bof Bosinelli is one of the main bookshops in the town of Menzogna. As an assistant explains (Bof A-01), it belongs to a chain which has branches all over Italy, and references are occasionally made to the Menzogna shop to distinguish it from others. The shop occupies two separate adjacent buildings; in the building in question it occupies two floors, each containing a single department. The data were recorded in the foreign language and literature department, on the upper floor, which is a sort of rectangular balcony overlooking the ground floor, reached by stairs leading up to one end. Shelves cover the side and far walls; other lower shelves are ranked in the centre of the room and along the side of the balcony over the ground floor. The service desk is set at the far end of the balcony, facing the entrance, while the single cash desk is downstairs by the street entrance, visible (and audible) from the balcony when standing up. The rest of the shop (another dozen or so departments) is located next door but one, and is sometimes referred to as the `other' shop ("l'altra libreria": Bof G-07 /b). There are two main assistants: one man, Alberto, and one woman, Franca, who heads the department. Like the assistants in Lod, both are in their mid-thirties and have worked in the department for several years. They do not always sit or stand at the service desk; they are often busy elsewhere tidying the shelves, putting price labels on books or personally finding them for customers. Franca also uses a further desk in a backstage area on the balcony to which customers do not have access; usually only one assistant is near the service desk itself. Assistants from other departments occasionally intervene, often in shouted exchanges with the ground floor: there is a buzzer system which the assistants on one floor can use to attract the attention of those on the other (cf. Bof A-19 /b2). Some of the shelves are sliding ones, which hide further shelves behind them with further books. Their movement accounts for the background noises noted in some encounters (e.g. Bof B-06). With respect to Lod, the atmosphere in Bof is a somewhat closer, more local one. Many of the customers visiting the department are students or teachers from the town's language schools and university, and the assistants are familiar with the books used in the main courses there. Some customers are known by name and in a few cases are friends of an assistant (e.g. Bof F-18 /b). While most of the recordings were made on normal working days, one afternoon (day F in the corpus) a lecture by the poet Yevtushenko had been organised on the ground floor, so that the shop was particularly crowded, a fact to which assistants and customers repeatedly allude: C4w "Madame Bovary" di Flaubert? No:n - c'Š in francese solo? AA =L'ho finito + qua sopra, + forse ce l'ho sotto, + per• non ci vado. ++ Nel senso che non si riesce a entrare nella sala bassa. C5m Com'Š che Š cos? Sempre cos? AA C'Š: - no, c'Š: E- Evtuscenko. C5m + Ah: - AA + ŝ quello. Facciamo la presentazione di: di un suo libro - C5m S? C4w Ah:, (ho capito). (Bof F-04 /b) 2.2 Contexts of situation: the minor corpora The minor corpora consist of selected data from smaller shops in other parts of Italy. Unlike Lod and Bof, which are largely analogous in terms of their organisation, clientele and the roles of the assistants, these corpora derive from contexts which differ markedly in one or more of these respects. 2.2.1 Paf The Paladini bookshop is situated in a small town in Northern Italy. It is, like Bof, one of a national chain, self-service, multi-departmental and multi-storey. The shop is spread over two floors connected by stairs, with a single cash point near the entrance. The ground floor, where most recordings took place (Paf-01 was recorded upstairs), has two main sections: Italian language and literature, and history. As well as the main desk at the cash point, there is a smaller desk at the back of the shop, half hidden by the stairs. The upper floor has sections of foreign language and literature, psychology, arts and cinema, computing, children's and leisure books. There is one service point near the top of the stairs, close to the foreign language and literature section. There are normally two assistants on each floor: two women, Graziella and Fabiola, upstairs, and two men, Paolo and Dario, on the ground floor. All move around the shop, however, and most encounters with customers take place away from the service desks. Paolo, the manager, is older than the others, to whom he often redirects customers: consequently many encounters involve exchanges between assistants. As in Bof, customers are mainly students and teachers, but the shop also caters for general readers. 2.2.2 Pif The Piranesi bookshop, located in a small town in Central Italy, is again one of a chain, and is again self-service. The shop is however much smaller than Bof or Paf, being set on a single floor and run as a single department. There is only one desk, with a cash till, placed near the entrance. There are two assistants: one man, Christian, and one woman, Flavia. Christian typically moves around the shop tidying the shelves, putting price labels on books and occasionally getting books for customers, while Flavia sits at the cash desk ringing up and wrapping up purchases. Like the other shops described so far, Pif serves both students and a more general public. The chain to which Piranesi belongs rewards loyalty by offering a discount after a certain number of purchases have been made. Customers are given a card on which the assistant registers each purchase by punching holes in it: when this tessera is complete, it entitles the customer to a small discount on any book published by the company owning the chain. This mechanism is discussed at some length in the data (cf. Pif-07, Pif-11). 2.2.3 Rok The Rocca bookshop, located in a large town in Central Italy, differs from the others in not catering for a general public, not belonging to a national chain, and not being self-service. It is specialised in one field, serving the needs of students and staff in the nearby psychology faculty of the local university: a sister shop in the same town offers a similar service for another faculty (cf. Rok-06). It is on one floor, with a long central desk behind which all the books are kept, and where both service and payment take place. A further distinctive feature is that the assistants share the social status of their customers, being psychology students who work there on a part-time basis. There are two of them in our data, one female, Francesca, and one male, Daniele: they stand at the desk to take requests and then go behind to fetch books before moving to the till to take payments. Rok carries secondhand as well as new psychology books, and additionally distributes university programmes and calendars (cf. Rok-04). 2.3 The spoken record While audio recording excludes the kinesic aspects that are retrievable from a video, it is easier to manage. A microphone can be hidden on a desk, in the lapel of an assistant's jacket, or in the handbag of a passing researcher, and there is no need for further personnel to operate clearly visible equipment. Editing and transcription time are considerably reduced: the difficulty of transcribing and analysing non-verbal behaviours is avoided. Not least, the ethical problems of data collection are lessened. All the data in the corpora were collected using audio-cassette recorders: a professional Sony with a Levalier button microphone was used for the main corpora, while the minor corpora were gathered using normal portable machines with built-in microphones. Recording of the main corpora was in each case carried out over a fortnight: the tape recorder was left on continuously for an hour or two on as many days in this period as possible. In this way we hoped to get a typical sample of `ordinary' data, and lessen, through familiarity, the influence of recording on the assistants' behaviour. We believe that this strategy was reasonably effective: the tapes captured a range of business and relaxed styles of interaction, including some slightly risqu‚ banter between assistants to which one of them later referred in talk with the researcher, observing that he would have been more discreet had he remembered there was a tape recorder on. Only in one case would it appear that an assistant addresses a remark directly to the tape (Bof F-17). Assistants were told that the purpose of the research was to study different customers' pronunciation patterns: when recording was complete, we obtained assistants' permission to study their own patterns of interaction. While assistants knew they were being recorded, customers were, we believe, in all cases unaware of the tape recorder, and the ethics of recording and publishing these encounters may thus be questioned. We would however stress that: a) the encounters occurred in public places (other customers were almost always present in the department and in a position to eavesdrop); b) customers cannot be identified from the transcripts, in which names, addresses, telephone and account numbers have been altered (the tapes themselves are not being made available); c) talk did not involve disclosure of other private information, or sensitive aspects of behaviour (illegal conduct, drug use, sexual matters: cf. Lamoreaux 1989, note 2; Shuy 1986). Identification of the shops and hence of the assistants is also unlikely: we have altered all names, and both Lod and Bof have changed their layout and organisation since the recordings were made. In Lod, the recorder was placed under the assistants' desk, and the microphone hidden in a typewriter resting on it. Tapes were changed by a researcher who would pass by the department periodically. The first part of the Bof corpus (Bof A-01 - A-20) was collected in a similar manner, but as the assistants tended to move around the department more, accompanying customers to search for books or to communicate with colleagues, the result was that many recordings were incomplete. It was thus decided to transfer the recorder to the jacket pocket of the main assistant on duty, with the button microphone in the lapel. While this inevitably increased the assistant's awareness of recording, we were by this stage convinced (not least by assistants' own accounts) that their behaviour should be little influenced: however, as a precaution, data from the first day of this modality of recording were not included in the corpus. The Bof assistants also kindly offered to carry out tape changing themselves; the fact that they regularly forgot to do this punctually is a further testimony to their limited awareness of the microphone's presence. Overall, the Lod and Bof corpora derive from a total of 9 and 12 hours recording made on 5 and 7 days respectively. The minor corpora were instead collected over shorter periods, with 2-6 hours of recording over one or two days. Again the assistants were aware that recording was or might be taking place. In Pif and Rok, the tape recorder was situated at the service point, being visible on the desk. The Paf data were instead collected by a member of the research group who browsed in the shop with a tape recorder in her bag. This method, unfortunately, gave rise to many incomplete recordings, as whole encounters often could not be taped without obvious eavesdropping. 2.4 The written record: structure While the fact of continuous recording contributed, we hope, to the naturalness of the data, our original tapes clearly included all kinds of material which was irrelevant for our purposes: long silences between different encounters, one side of telephone conversations, chat between assistants during lulls in business, visits from publishers' representatives, or from friends who had just happened to stop by. In selecting encounters for inclusion in the transcribed corpora from the many hours of recording, we have tried to limit ourselves to what we consider service encounters (Aston 1988b): those taking place between assistants and customers face-to-face, normally incorporating a central request-reply sequence, even where not strictly connected to bookshop business. In the main corpora we also excluded all the recordings made on the first day of collection or after a change in mode of collection, hoping to thereby minimise the eventual influence of recording on the assistants' behaviour. The transcripts have been `cleaned' in the sense that noise or talk in the background is not transcribed unless it seems relevant to the encounter in question. Telephone calls and other types of conversations have been included only where they seem necessary to the interpretation of a service encounter which they precede or interrupt: in such cases either some transcription or a brief summary of their gist has been provided (such talk is not always understandable where topics are not explicit, however). We have also discarded those encounters that were largely inaudible or intractably incomprehensible, whether owing to background noise (a loudspeaker announcement or a telephone ringing), to the distance of speakers from the microphone, or to continual overlapping of different speakers. While these selection procedures mean that much of the material recorded has been omitted, the transcriptions of the remaining chunks, and of the encounters within those chunks, are reproduced in chronological order, on a day-by-day basis. These aspects of sequencing are relevant in that: a) Assistants' replies may be influenced by the recurrence of similar requests over a relatively short period. For instance in the Lod bookshop, Burchfield's "The English language" had sold out, and the assistants appeared to have developed a fixed spiel in responding to requests for it (cf. Lod A-04 /a; Lod B-02; Lod B-23; Lod C-26): Cm (5syll), recently published by - Bur:chfield, "The English language", have you got that? AP Yes, it's: um: sold out, it's reprinting already, and it's due to come back again: they're hoping about the eighteenth of: this month. Cm Oh! Cm ==Okay. Thank you. (Lod B-23) b) Encounters are on occasion suspended, as a customer leaves the desk and then returns seconds or minutes later; a further encounter or encounters may be `embedded' in the intervening break. c) Strings of encounters may occur one after the other in quick succession or with partial overlaps. Here the presence of a next customer may affect the closing of the previous encounter and the presence of the previous customer the opening of the next. In the main corpora, a different upper-case letter has been used for each day of recording (A to E in Lod, A to G in Bof). Within each day the various chunks of recording transcribed are numbered consecutively (A-01, A-02, etc.). When more than one encounter occurs in a single chunk, the beginning of each encounter is indicated by a lower-case letter (/a, /b, etc.): where such an encounter is suspended while another `embedded' encounter or encounters takes place, its parts are numbered (/a1, /a2, etc.). When a pause occurs between one encounter and another, its duration is indicated in seconds at the end of the first: Bof G-15 /a1 ... C1m E: - pi— o meno + la cifra? Pensavo, non so - AA Andiamo a vedere. C1m + Perch‚: (02) dipendeva anche dalla - AA (01) Adesso (se me lo rid…) un attimo il biglietto. (01) Sampson. C1m Certo. (01) /b C2w Lo pago alla cassa, questo? AA S, alla cass- si paga gi—. Grazie. C2w =Grazie. Buongiorno. AA Buongiorno. (14) /a2 AA [fra s‚] Sampson. (01) [normale] Sulle:: cinquantamila lire. C1m ++ Mm. (01) Mm. + D'accordo. Allora ripasso per: + eh: abbiamo detto, per il Bromhead. AA S. Dovrebbe gi… esserci per gioved. ... Notice however that if an assistant addresses another customer in an aside or another customer interrupts the current encounter only momentarily (often just to say hello or goodbye), while the main axis remains that between the assistant and the current customer, the interruption is not labelled as a separate encounter or part. Examples are C3's intervention in the next extract, and AA's response to it: C2w E: mm: ce l'avete "L'uomo di Pietroburgo" di Ken Follett in inglese? AA =Credo di averlo finito, un attimo che controllo. C3w Eh - senta per pagare, vado gi—? AA [a C2] ŝ finito "Pietroburgo" + infatti. C2w Non c'Š? AA No. ŝ finito. + Tor:na fra dieci quindici giorni. [a C3] S, lo paga sotto alla cassa, grazie. C3w S. C2w Grazie. Buonasera. (Bof F-11 /b) To locate all talk by a particular customer, therefore, it may be necessary to scan the whole of the chunk in which they appear, not merely the labelled part(s) of that chunk in which an assistant is dealing with them directly. Most previously published work from the project made reference to a different numbering scheme from that employed here. 2.5 The written record: transcription procedure Initial transcriptions of the data were made by various members of the PIXI group, and these transcripts were then independently checked by other members. Further checking and revision were carried out by the editors with the aid of computer-controlled Tandberg tape recorders operated via keyboard macros from within the word processor, permitting synchronised playing of specified portions of the tape with video display of the relevant transcript. Finally the revised transcripts were rechecked against the original tapes by other members of the group. All the transcripts have been checked by at least two native speakers of the language used. Our transcription technique follows the modified `playscript' model developed by Jefferson (1978, 1985, 1988; see also Atkinson & Heritage 1984) - a technique we believe substantially meets the criteria of adequate category definition, accessibility, robustness, economy and adaptability (Du Bois 1991). We have adapted Jefferson's norms in minor respects (generally simplifying them somewhat) in order to deal more readily with data in two different languages and to facilitate computerised elaboration. Our aims required a transcription that would be easily readable but which would provide a reasonably faithful representation of sequencing and timing. We thus excluded both simple playscript transcription (Drew 1982) and the musical score type (Makosh & Spencer 1982), since the former does not provide information on pauses and overlaps, and the latter does not clearly distinguish turns, and makes computer editing difficult. [Ehlich's recently-developed HIAT system (Ehlich, forthcoming) overcomes most of the difficulties of computerised transcription using musical score technique.] For the sake of readability, we also excluded phonetic transcription, which would have involved a complex of signs and symbols over and above the numerous features of a basic transcription, not to mention greater difficulty in obtaining agreement between transcribers (Halliday 1970, Brown et al. 1980, Brazil 1985, Erickson & Shultz 1982, Owen 1978, Crystal & Davy 1975). We have followed Jefferson in using punctuation as a rough indication of stress and intonation, and indicated marked segmental pronunciation features only where these appear relevant for an understanding of the discourse (Sornicola 1982: see 2.5.2 below). A summary of the transcription conventions used is provided prior to these remarks. 2.5.1 Turns and speakers The use of a playscript transcription, where speakers are shown as taking alternate turns, entails deciding when a turn is `complete' - in particular, in what cases further talk by a given speaker following a silence or a `backchannel' contribution from another should be considered a continuation of the current turn. To avoid the subjectivity entailed by pragmatic criteria of turn-completeness, we have arbitrarily transcribed as a single turn any stretch of speech by one speaker in which there are no pauses longer than 2 seconds, and during which any other speakers' talk is fully overlapped. Thus if any speaker pauses for more than two seconds, subsequent talk by them is transcribed as a new turn. Similarly, if a speaker stops talking even briefly, and during that interval talk is produced by another speaker, any subsequent talk by the first is transcribed as a new turn. The speaker of each turn is indicated in bold in the left hand margin by a code showing role (A = assistant, C = customer). Individual assistants are further identified by their initial. Thus in the code AJ, A stands for assistant and J is the initial of the assistant's name. A list of the various assistants is provided at the beginning of each corpus. Customers are coded by their sex/age (m = man, w = woman, i = child). Where two or more customers speak during one recorded chunk, they are also numbered in order of appearance. Thus the code C1w indicates that the speaker is the first customer to appear in the chunk, and is a woman. Where customers are clearly non-native speakers, the code ends with an n. Thus C2mn indicates that the speaker is the second customer to appear in this chunk, and is a male non-native speaker of the language of the corpus. Other speaker codes occasionally used are F (a friend of the assistant's who is not a colleague or customer) and R (a publisher's representative): a following m or w indicates their sex. Where speaker identity is uncertain, an italicised question mark precedes the code. 2.5.2 Pronunciation and pseudo-verbal sounds The orthography we have adopted is closer to conventional written forms than that in Jefferson's transcriptions, which attempt to capture actual pronunciation through the use of `creative' spelling. As phonetic analysis was not one of our objectives, it did not generally seem worthwhile attempting to reproduce pronunciation more precisely. Not only would `creative' orthography have decreased readability (particularly for non-native readers of a foreign language), but it would have also posed particular difficulties given our data. Insofar as the informal conventions for the orthographic representation of pronunciation characteristics are language-specific, we would logically have had to adopt different `creative' transcription principles for different languages, or for different corpora. Either way this would have caused considerable confusion where code-switching takes place: in the Italian corpora, for instance, participants frequently use English (as well as French, German and Spanish) to mention titles, authors and publishers, while in the English corpus we find Italian and French. We have therefore generally limited creative spellings to pseudo-verbal sounds, such as mhm, uhuh and the like. Even these, it should be noted, in some cases entail different transcriptions of the same sounds in the two languages (e.g. erm in English and ehm in Italian). To limit the quantity of such `balloon language', we have also made quite extensive use of comments (e.g. [coughs], [laughs]). Partial phonetic transcriptions have however been provided in comment format (italicised within square brackets) in cases of non-standard or variant pronunciation which are targets of self- or other-repair: C5w ... E poi, e- un libro di Borges [/djz/]. In spagnolo. AA =Di Borges [/ghez/] cosa? (Bof F-10 /d) Here the customer pronounces the author's name as a single syllable with a soft g; the assistant `corrects' her pronunciation, using a two-syllable form with a hard g. Foreign words, which occur frequently in the discussion of foreign-language books, are transcribed using their conventional spellings and without specific markings. Transcription of laughter poses particular problems in Jefferson's format (1985, Jefferson et al. 1976, 1987). In our data, while laughter is present in many encounters, participants do not laugh extensively and they seldom laugh together. We have therefore only made a broad distinction between intensive laughter ([laughs / ride]) and less intensive laughter (transcribed as "hh" in both languages). 2.5.3 Use of capitals, question marks and hyphens Capital letters are used according to the orthographic conventions of the language of the corpus: for the first word of a turn and for the first word following a full stop, question or exclamation mark within a turn; for the first letter of proper names (including publishing houses and authors' initials); and, in English, for proper adjectives, days and months. Titles of books and of series (placed in quotation marks, see below) have their first letter capitalised: AJ There's: a book called "Essays in modern linguistics", by D C Freeman, (Lod B-07 /c2) Capital letters are also used for acronyms ("EFL", "livello A"), and in English, spaced capitals are used to transcribe spelling aloud: AJ + Don't know it. How do you spell the name (or something)? Cm D Y, C H, T W A L D. It's er Wildwood: House. AJ D Y, C H? Cm + Yeah. (Lod C-08) For spellings aloud in the Italian data we have instead followed the Italian convention of reproducing the pronunciation, i.e. "enne" for the letter n, "esse" for s, etc.: AA E: a nome, signore? Scusi che non mi ricordo mai i nomi hh. (01) Ha detto? Cmn E:hm Birley. Cmn Birley. Bi i erre, AA Ah. AA (Ah s). Cmn Elle e ipsilon. Birley. (Bof G-03) Quotation marks are used to indicate titles of books and series. In some cases, particularly ones of widely-used foreign language courses, titles are abbreviated by participants: in Bof, The Cambridge English course is regularly referred to as "Cambridge", Life styles as "Elle esse". Such courses are generally divided into a series of levels (primary/intermediate/advanced or one/two etc.), with a number of components at each level (coursebook, workbook, practice book, tests etc.): where the level is stated along with the name of the course we have treated it as part of the title, placing it within quotation marks (e.g. "Meaning into words intermediate", "Cambridge uno"). References to the various components, which are made in a wide range of manners, have instead been treated as normal text. AA "English today tre" con il workbook - vediamo. (Bof C-06) Cmn E: - e poi del "Cambridge due", ehm + due practice. (Bof G-03) Hyphens (en dashes) are not used in compound words, but instead indicate features of timing and intonation (see 2.5.7 below). In a few cases this has entailed the printing as two words of compound names (e.g. "Selby Bigge": Lod C-23 /b), and also of English numbers (e.g. "twenty five": Lod C-21). Words are never broken at line ends. 2.5.4 Comments and notes Comments (italicised within square brackets) and notes (in the computer-readable version, references to notes are italicised within square brackets, and listed in the notes file) are provided in the language of the corpus. Comments are used to provide information about characteristics of the following portion of talk and about `noises' heard on the tape, based on our understanding of the ongoing conversation and of the context. They principally describe: - paralinguistic features: [laughs / ride], [clears throat / si schiarisce la voce], [sighs / sospira], [hums / canticchia], [whistles / fischietta], etc. - speech rhythm and loudness: [slowly / lentamente], [sottovoce / fra s‚], [voice raised / ad alta voce], [normal / normale], [in the distance / in lontananza], etc. - changes in speech direction, leaving or returning to the main customer-assistant axis: [to another C / ad altro C], [to C4 / a C4], [al telefono], etc. - background noises: [scaffale scorrevole], [passi], [sound of paper / rumore di carta], [PA announcement], [cicalino interno], etc. - participant actions inferred from background noises: [they move away / si allontanano], [scende al piano inferiore], [mostra una lista] - characteristics of untranscribed talk: [unclear / confuso], [talk between As / conversazione tra commessi], [talk between C1 and C2 in background / conversazione tra C1 e C2 in sottofondo], etc. - role-relationships of participants: [C1 and C2 are together / C1 e C2 sono insieme], [C is a friend of A / C Š amico di A], [same C as in / stesso C di ], etc. - incomplete recordings: [in medias res], [interrupted / interrotta], etc. An initial question mark indicates uncertainty as to speech direction or as to a non-verbal action, e.g. [?to another C], [?consults catalogue]. When the length of a noise or action is significant, its duration is indicated in seconds, prior to the comment, but within the square brackets, e.g.: [(10) conversazione in sottofondo] Footnotes explain speaker references to local facts and institutions which are shared knowledge for participants, or provide translations of slang or dialect expressions. 2.5.5 Dubious transcriptions Like Jefferson, we have placed dubious or tentative transcriptions within round brackets. In the face of uncertainty, the transcriber is torn between the alternative of providing some hypothesis, however unintelligible the tape is, so as to give the reader the benefit of her/his closehand knowledge, and that of avoiding interpretation where unsure, so as not to influence the reader unjustifiably. We have transcribed our guesses where they seem acoustically congruous and they improve the readability of the data; when no plausible and meaningful transcription could be arrived at, we have indicated the approximate length of the speech segment in syllables, italicised within round brackets: e.g. (3syll). Where not even the rough number of syllables could be established, two question marks have been used, again italicised within round brackets: (??). Following punctuation indicates the intonation pattern where this is clear (cf. 2.5.7 below). Doubts in identifying names of authors, titles and publishers have been resolved, as far as possible, by consulting catalogues and the assistants themselves. 2.5.6 Altered transcriptions All names of assistants and customers, their addresses, telephone and account numbers, the names of specific shops and of local institutions, towns, streets and districts have been altered. The number of syllables and stress pattern of the original forms have been maintained as far as possible in their replacements, which are italicised in the text. Names of authors, books and publishers have not been changed. 2.5.7 Timing, stress and intonation Pauses of one second or more were measured on a seconds tape counter, and their length in seconds is indicated between round brackets: (02), (15), (120), etc. Pauses less than one second long have instead been subjectively classified as shorter (+) or longer (++). Measured pauses three or more seconds long, or separating different encounters, are placed on a separate line in the transcript and followed by a new turn, irrespective of whether there is a change of speaker. All other pauses are indicated within the turn of the current speaker. Jefferson's conventions for transcribing overlaps and latching have been modified to facilitate computer search for recurrent patterns and to avoid the need to vertically align overlapping sections of different turns. Overlaps and simultaneous start-ups are marked by  at the beginning of the first overlapping portion and by  at the beginning of the portion of the subsequent turn in the transcript which is simultaneously produced. The end of the overlapping portion is indicated by  in both cases. Where more than one part of a turn overlaps with subsequent ones in the transcription, the various overlaps are distinguished by doubling or tripling these symbols. [In early papers from the project we used the symbols $ and % (and multiples of these) to mark overlap beginning and overlap end respectively, in both the affected turns.] An equals sign (=) at the beginning of a turn indicates that it is latched to the end of the previous turn in the transcript, beginning immediately upon its completion; a multiple symbol (==, === or ====) indicates that the turn in question is latched to the end of a turn that respectively appears two, three or four turns above in the transcript. The following example may clarify these conventions: AP Ah. Right. + Ah, perhaps you're looking for the: poetry thing. Yes hh, I've thought about that too. Cw I am. Cw Yes. Cw ===Should I go hh? (Lod A-10) Here A's "thing" is overlapped by C's "I am", his "I've thought" by C's "Yes", while C's "Should I go hh?" follows immediately on completion of A's "about that too". On the few occasions where three speakers' talk overlaps, this is noted in a comment prior to the first overlapping section, and the `overlapped' symbols  and  are repeated in both the subsequent turns affected: AA Dica? C3m + [parzialmente sovrapposto ai due turni successivi] No no si- AA Ah, siete insieme? C2w (Noi siamo con:) (2syll). (Bof F-01 /b) Here, C3's "no si-" overlaps with both A's "Ah sie-" and C2's "Noi si-", while the continuations of A's and C2's turns overlap each other. As in Jefferson, sound or syllable lengthening is indicated by a colon (:) placed immediately after the letter(s) representing the lengthened sound(s). Double and triple colons are used for particularly extensive lengthening. In the case of occlusives, it should be noted, the colon actually represents lengthening of an indefinite following vowel. When a syllable is instead cut short an en dash (-) is attached to the letter representing the final sound heard. Emphasis, either through relatively loud speech and/or marked stress (other than that predictable for the word in the context of the tone group), is indicated by underlining of the stressed syllable(s). [In earlier papers from the project stressed syllables were printed in capitals.] E.g.: C1m Thank you. I'm looking for a book that- that is: either just been published or about to be published, called I think "The secret cult". (Lod B-01) Intonation is roughly shown by punctuation, as follows: low fall . low rise ? fall-rise , rise-fall ! When a tone group is incomplete, the point of interruption is marked by an en dash (-) preceded by a space. References Anderson, L. 1987. "Intonation, turn-taking and dysfluency: non-natives conversing". In Morley & Partington. 93-102. Revised version in M. Hewings (ed.) 1990, Studies in discourse intonation. Birmingham: University of Birmingham English Language Research. 102-113. Anderson, L. 1988a. "Access routines in service encounters". In Aston 1988a. 99-134. Anderson, L. 1988b. "Issues in native-nonnative interaction". In Aston 1988a. 267-285. Anderson, L. 1988c. "Learning to serve: developing oral skills for ESP". In S. Holden (ed.), Language and literature. Oxford: Modern English. 95-99, 133. Anderson, L. 1990. "Epistemic qualification in conversational accounts". In De Stasio et al. 331-340. Anderson, L., G. Aston & G. Tucker 1988. "The joint production of requests in service encounters". In Aston 1988a. 135-151. Aston, G. 1987. "Ah: a corpus-based exercise in conversational analysis". In Morley & Partington. 123-137. Aston, G. (ed.) 1988a. Negotiating service: studies in the discourse of bookshop encounters. Bologna: Cooperativa Libraria Universitaria Editrice. Aston, G. 1988b. "What's a public service encounter anyway?" In Aston 1988a. 25-42. Aston, G. 1988c. "Interactional speech in service encounters". In Aston 1988a. 73-97. Aston, G. 1988d. "From discourse to pedagogy: some implications". In Aston 1988a. 287-301. Aston, G. 1988e. "An applied linguistic approach to the analysis of naturally-occurring conversation". In E. Rigotti & C. Cipolli (eds.) Dal senso al testo. Brescia: La Scuola. 237-260. Aston, G. 1988f. Learning comity: an approach to the description and pedagogy of interactional speech. Bologna: CLUEB. Aston, G. 1989. "Solidarity and conflict as metaphors of the discourse process". In L. Curti, L. Di Michele, T. Frank & M. Vitale (eds.), Il muro del linguaggio: conflitto e tragedia. Naples: Istituto Universitario Orientale. 413-426. Aston, G. 1990. "Cultural norms of conversational procedure in a pedagogic perspective". Paper presented at the 3rd international congress on dialogue, Bologna. (Forthcoming in the proceedings, Tbingen: Niemeyer) Atkinson, J.M. & J. Heritage 1984. "Transcript notation". In J.M. Atkinson & J. Heritage (eds.), Structures of social action. Cambridge: Cambridge University Press. ix-xvi. Boden, D. 1990. "People are talking: conversation analysis and symbolic interaction". In H.S. Becker & M.M. McCall (eds.), Symbolic interaction and cultural studies. Chicago: University of Chicago Press. 244-274. Brazil, D. 1985. The communicative value of intonation in English. Birmingham: Bleak House/English Language Research. Brodine, R. 1990. "Requesting and responding in Italian and English service encounters". Paper presented at the 3rd international congress on dialogue, Bologna. (Forthcoming in the proceedings, Tbingen: Niemeyer) Brown, G., K. Currie & J. Kenworthy 1980. Questions of intonation. London: Croom Helm. Brown, G. & G. Yule 1983. Teaching the spoken language. Cambridge: Cambridge University Press. Ciliberti, A. 1988. "Strategies in service encounters in Italian bookshops". In Aston 1988a. 43-71. Ciliberti, A. 1989a. "Lo studio dell'uso linguistico". Lingua e nuova didattica, 3. 32-43. Ciliberti, A. 1989b. "I luoghi del discorso deputati ad esprimere realt… sociale". Paper given at Convegno nazionale LEND, Rapporto lingua/cultura, Urbino. (Forthcoming in the proceedings) Ciliberti, A. 1991a. Grammatica, pedagogia, discorso. Firenze: La Nuova Italia. Ciliberti, A. 1991b. "The relevance of contrastive studies for the description and the teaching of English". In M.T. Zagrebelsky (ed.), The study of English in Italian universities. Alessandria: Dell'Orso. In press. Ciliberti, A. 1991c. "Request compliance/non-compliance in English and Italian public service encounters". Issues and developments in English and applied linguistics, 6. In press. Ciliberti, A. & G. Aston (forthcoming). "Research methodology and affect". Atti del XIII convegno nazionale dell'AIA. Pisa. Crystal, D. & D. Davy 1975. Advanced conversational English. London: Longman. De Stasio, C., M. Gotti & R. Bonadei (eds.) 1990. La rappresentazione verbale e iconica: valori estetici e funzionali. Milano: Guerini. Drew, P. 1982. "Defence sequences in cross examination". University of York. Unpublished paper. Du Bois, J.W. 1991. "Transcription design principles for spoken discourse research". Pragmatics, 1. 71-106. Edmondson, W. 1981. Spoken discourse: a model for analysis. London: Longman. Ehlich, K. (forthcoming). "HIAT - a transcription system for discourse data". In J.A. Edwards & M.D. Lampert (eds.), Talking data. Hillsdale NJ: Erlbaum. Erickson, F. & J. Shultz 1982. The counselor as gatekeeper. New York: Academic Press. Fairclough, N. 1989. Language and power. London: Longman. Gavioli, L. 1987. "Ridere e deridere con simpatia: analisi delle risate nei PSE". Paper presented at 2nd international colloquium of contrastive pragmatics, Bagni di Lucca. Gavioli, L. 1990a. "The role of personal and impersonal information in service encounters: a contrastive study". Paper presented at 7th AILA conference, Thessaloniki. Gavioli, L. 1990b. "Il testo parlato: spunti di analisi e osservazione". Laboratorio degli studi linguistici, 1990. 13-26. George, S.E. 1988. "Postscript: discourse descriptions and pedagogic proposals". In Aston 1988a. 303-331. George, S.E. 1990. Getting things done in Naples: action, language and context in discourse description. Bologna: CLUEB. Goffman, E. 1983. "Felicity's condition". American journal of sociology, 89. 1-53. Halliday, M.A.K. 1970. A course in spoken English: intonation. London: Oxford University Press. House, J. 1986. "Learning to talk: talking to learn. An investigation of learner performance in two types of discourse". In G. Kasper (ed.), Learning, teaching and communication in the foreign language classroom. Aarhus: Aarhus University Press. 43-58. Hymes, D. 1972. "Models of the interaction of language and social life". In J.J. Gumperz & D. Hymes (eds.), Directions in sociolinguistics. New York: Holt Rinehart. 35-71. Jefferson, G. 1978. "Explanation of transcript notation". In J. Schenkein (ed.), Studies in the organization of conversational interaction. New York: Academic Press. xii-xvi. Jefferson, G. 1985. "An exercise in the transcription and analysis of laughter". In T. van Dijk (ed.), Handbook of discourse analysis: discourse and dialogue. New York: Academic Press. Jefferson, G. 1988. "Preliminary notes on a possible metric which provides for a `standard maximum' silence of approximately one second in conversation". In D. Roger & P. Bull (eds.), Conversation. Clevedon: Multilingual Matters. 166-196. Jefferson, G., H. Sacks & E. Schegloff 1976. "Some notes on laughing together". Pragmatics microfiche, 1, 8. A3-D9. Jefferson, G., H. Sacks & E. Schegloff 1987. "Notes on laughter in the pursuit of intimacy". In G. Button & J.R.E. Lee (eds.), Talk and social organisation. Clevedon: Multilingual matters. 152-205. Labov, W. 1972. Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press. Lamoreaux, E.L. 1989. "Rhetoric and conversation in service encounters". Research on language and social interaction, 22. 93-114. Lindenfeld, J. 1990. Speech and sociability at French urban marketplaces. Amsterdam: Benjamins. Makosh, M. & S. Spencer 1982. "Working transcript of BBC `Skinhead' data". University of Lancaster. Unpublished paper. Mansfield, G. 1988. "The supportive role of intonation in service encounters". In Aston 1988a. 205-231. Merlini Barbaresi, L. 1988. "Parameters of discourse markedness in public service encounters". In Aston 1988a. 167-204. Merritt, M. 1976. "On questions following questions in service encounters". Language in society, 5. 315-357. Morley, J. & A. Partington (eds.) 1988. Spoken discourse. Camerino: Universit… di Camerino. Ochs, E. 1979. "Transcription as theory". In E. Ochs & B. Schieffelin (eds.), Developmental pragmatics. New York: Academic Press. 43-72. Owen, M. 1978. "Units of natural conversation". Cambridge University. Unpublished paper. Schank, R. & P. Abelson 1975. "Scripts, plans and knowledge". In P. Johnson-Laird & P. Wason (eds.) 1977, Thinking: readings in cognitive science. Cambridge: Cambridge University Press. 421-432. Schegloff, E. 1987. "Between micro and macro: contexts and other connections". In J.C. Alexander, B. Giesen, R. Mnch & N.J. Smeiser (eds.), The micro-macro link. Berkeley: University of California Press. 207-236. Schegloff, E. & H. Sacks 1973. "Opening up closings". Semiotica, 7. 289-327. Scotton Myers, C. & J. Bernstein 1988. "Natural conversations as a model for textbook dialogue". Applied linguistics, 9. 372-384. Shuy, R. 1986. "Ethical issues in analysing surreptitious tapes". International journal of the sociology of language, 62. 119-128. Sornicola, R. 1982. Sul parlato. Bologna: Il Mulino. Svartvik, J. & Quirk, R. 1980. A corpus of English conversation. Lund: Gleerup. Tucker, G. 1987. "Syntactic and prosodic mapping in service encounter requests". In Morley & Partington. 139-150. Tucker, G. 1988a. "The role of lexicogrammar in the realisation of requests". In Aston 1988a. 153-165. Tucker, G. 1988b. "From illocution to syntactic and prosodic realisation in making requests". In E. Steiner & R. Veltman (eds.), Pragmatics and discourse: some systemically inspired approaches. London: Pinter. 120-131. van Ek, J. 1975. "The threshold level in a European unit/credit system for modern language learning by adults". Strasbourg: Council of Europe. Ventola, E. 1987. The structure of social interaction. London: Pinter. Vincent Marrelli, J. 1988. "Notes on power and solidarity in bookshops: goals, roles and the functions of laughing". In Aston 1988a. 233-265. Zorzi Cal•, D. 1990a. Parlare insieme: la co-produzione dell'ordine conversazionale in italiano e in inglese. Bologna: CLUEB. Zorzi Cal•, D. 1990b. "Non-conflictual interruption in English and Italian". Paper presented at 7th AILA congress, Thessaloniki. Zorzi Cal•, D. 1990c. "Le procedure per l'accessibilit… e per l'accettabilit…: in confronto fra l'italiano e l'inglese". Paper presented at the 3rd international congress on dialogue, Bologna. (Forthcoming in the proceedings, Tbingen: Niemeyer) Zorzi Cal•, D., R. Brodine, L. Gavioli & G. Aston 1990. "Opening and closing service encounters: some differences between English and Italian". In De Stasio et al. 303-313.